In Vitro Treatment of Human Monocytes/Macrophages with Myristoylated Recombinant Nef of Human Immunodeficiency Virus Type 1 Leads to the Activation of Mitogen-Activated Protein Kinases, IκB Kinases, and Interferon Regulatory Factor 3 and to the Release of Beta Interferon

Abstract

The viral protein Nef is a virulence factor that plays multiple roles during the early and late phases of human immunodeficiency virus (HIV) replication. Nef regulates the cell surface expression of critical proteins (including down-regulation of CD4 and major histocompatibility complex class I), T-cell receptor signaling, and apoptosis, inducing proapoptotic effects in uninfected bystander cells and antiapoptotic effects in infected cells. It has been proposed that Nef intersects the CD40 ligand signaling pathway in macrophages, leading to modification in the pattern of secreted factors that appear able to recruit and activate T lymphocytes, rendering them susceptible to HIV infection. There is also increasing evidence that in vitro cell treatment with Nef induces signaling effects. Exogenous Nef treatment is able to induce apoptosis in uninfected T cells, maturation in dendritic cells, and suppression of CD40-dependent immunoglobulin class switching in B cells. Previously, we reported that Nef treatment of primary human monocyte-derived macrophages (MDMs) induces a cycloheximide-independent activation of NF-κB and the synthesis and secretion of a set of chemokines/cytokines that activate STAT1 and STAT3. Here, we show that Nef treatment is capable of hijacking cellular signaling pathways, inducing a very rapid regulatory response in MDMs that is characterized by the rapid and transient phosphorylation of the α and β subunits of the IκB kinase complex and of JNK, ERK1/2, and p38 mitogen-activated protein kinase family members. In addition, we have observed the activation of interferon regulatory factor 3, leading to the synthesis of beta interferon mRNA and protein, which in turn induces STAT2 phosphorylation. All of these effects require Nef myristoylation.

The 27- to 34-kDa Nef protein is an important virulence factor of primate lentiviruses; it is the regulatory protein expressed earliest and most abundantly in the infection cycle. Studies of animal models and seropositive patients showed that Nef-defective viruses led to an attenuated clinical phenotype with a reduced viral load (29, 30, 50, 59, 60). In addition, nef transgenic mice develop an AIDS-like disease (51) characterized by failure to thrive/weight loss, diarrhea, wasting, premature death, thymus atrophy, loss of CD4⁺ T cells, interstitial pneumonitis, and tubulointerstitial nephritis. Inside the cell, Nef induces effects that are genetically distinguishable yet highly conserved and that appear to be mediated via specific protein-protein interaction domains (7, 33, 44). Nef is cotranslationally modified by an N-terminal myristoylation site whose lipidation is required for membrane association. However, cellular-fractionation assays from transient transfections showed that less than 50% of the protein was localized at membranes, while the remaining portion was found to be cytosolic (25, 34, 58, 68, 98). The protein adopts a two-domain structure that is characterized by a flexible N-terminal arm of about 60 amino acids, followed by a well-conserved and folded core domain of about 120 residues that comprises a flexible loop of 30 amino acids projecting out from the core domain (44). Protein structures of Nef have been determined for the core domain and the flexible anchor domain independently, but not yet for the full-length protein due to the low stability and solubility and the high degree of intrinsic flexibility (5, 6, 10, 46, 49, 64).

Many cellular Nef targets have been reported, most of them representing functionally related or connected sets of proteins that can be divided mainly into two classes: proteins involved in the trafficking of cell surface receptors and those acting as signaling molecules. Two main functions of Nef have been well characterized: (i) the acceleration of endocytosis and lysosomal degradation of the transmembrane glycoprotein CD4, thereby preventing superinfection of infected cells and the interaction of budding virions on infected cells with CD4, and (ii) the down-regulation of the A and B alleles of major histocompatibility complex class I, thereby protecting infected cells from destruction by cytotoxic T lymphocytes (26). To recruit CD4 into the endocytotic pathway, Nef acts as an adaptor between CD4 and components of the clathrin-coated pits, disrupting the CD4-Lck complex and coupling CD4 with adaptor protein complexes or the regulatory subunit H of the vacuolar-membrane V-ATPase (47, 66, 75). Nef was also suggested to bind to the β subunit of COP-I coatomers to direct CD4 to a degradation pathway (76). Surprisingly, Nef uses different domains and mechanisms to down-regulate major histocompatibility complex class I molecules (16, 33).

In addition to these effects, Nef interacts with several molecules involved in the T-cell receptor signal transduction pathway (i.e., CD3 ζ chain, Lck, and Vav) (13, 32, 35), promoting their recruitment into glycolipid-enriched microdomains and inducing a state of preactivation in T cells (90) that could favor viral replication. Nef is able to promote apoptosis in uninfected bystander cells through the induction of FasL expression on the infected CD4⁺ T cell (101, 102), meanwhile protecting the infected cells by apoptotic stimuli through more than one mechanism (40, 41, 48, 100). Nef effects have been mainly studied in a CD4⁺ T-cell context, whereas those induced in cells of the monocyte/macrophage lineage are less well characterized. Nevertheless, monocytes/macrophages are one of the main cellular targets of human immunodeficiency virus (HIV) infection and represent an important reservoir of the virus; they show long-term survival after HIV infection and are at the same time particularly resistant to the current antiretroviral therapies (4). It has been reported that macrophages expressing Nef, or that are stimulated through the CD40 receptor, release paracrine factors that are able to recruit T lymphocytes, making them susceptible to HIV replication (93, 94). Recently, we have reported that HIV-1 Nef specifically activates STAT1 and STAT3 by inducing the synthesis and secretion of soluble activating factor(s) in 7-day-old cultures of human monocyte-derived macrophages (MDMs) purified from peripheral blood mononuclear cells (PBMC) of healthy donors (36, 73). STATs (for signal transducer and activator of transcription) are involved in the responses of a wide number of cytokines, growth factors, and hormones. We observed Nef-induced STAT1 and STAT3 tyrosine phosporylation 8 h after the infection of MDMs with nef-expressing HIV type 1 (HIV-1) pseudotyped with glycoprotein G of vesicular stomatitis virus (VSV-G), but also by treating MDMs for 2 h with the recombinant Nef protein (recNef) (36, 73). Indeed, there is increasing evidence that Nef induces signaling effects when added to cell cultures (2, 18, 39, 54, 56, 80-82, 96, 97). In particular, it has recently been reported that Nef induces apoptosis in CD4⁺ T lymphocytes interacting with the CXCR4 receptor, suggesting that exogenous Nef could contribute to the CD4⁺ lymphocyte depletion that occurs prior to and during the onset of AIDS (54, 56). recNef is internalized by primary human MDMs and dendritic cells in culture (2, 81-83) and mimics the effects of the virus-encoded proteins inducing CD4 down-regulation (2). The profile of mRNA expression of cytokines, chemokines, growth factors, and their receptors in Nef-treated MDMs, performed using cDNA expression macroarray technology, showed a marked increase in CCL2/MIP-1α and CCL4/MIP-1β mRNA expression, as well as regulation of the expression of several transcripts, reinforcing the idea that Nef treatment regulates the gene expression program in monocytes/macrophages (70). Enzyme-linked immunosorbent assay detected the release of MIP-1α, MIP-1β, interleukin 6 (IL-6), tumor necrosis factor alpha (TNF-α), and IL-1β in the culture medium, which has been correlated with the cycloheximide-independent activation of NF-κB inside cells (70).

The results described here indicate that Nef is able to induce a very early activation signal in treated MDM cultures, leading to the rapid phosphorylation of mitogen-activated protein kinase (MAPK) family members (i.e., ERK1/2, p38, and JNK), interferon regulatory factor 3 (IRF-3), and both the α and β subunits of the IκB kinase (IKK) complex, required for the activation of the NF-κB pathway. Activation of all these pathways required N-terminal myristoylation of Nef. We also show that the IKK activation is essential for the subsequent production of the STAT-activating factors. Finally, we show that the activation of IRF-3 is followed by the induction of beta interferon (IFN-β) mRNA and its production.

MATERIALS AND METHODS

Cells, recombinant Nef preparations, and reagents.

PBMC were isolated from buffy coats obtained from healthy donors. Monocytes were isolated by positive selection using CD14 microbeads and LS columns, all purchased from Miltenyi Biotech (Auburn, CA), following the manufacturer's recommendations. The purity of the recovered cell populations was assayed by fluorescence-activated cell sorter (FACS) analysis by means of phycoerythrin-conjugated anti-CD14 monoclonal antibody (Chemicon-Cymbus, Temecula, CA) labeling. Cell preparations staining below 95% positive for CD14 (a cell surface marker specific for monocyte/macrophage cell populations) were discarded. Seven-day-old MDMs were obtained by culturing monocytes for the first 3 days in RPMI 1640 (Cambrex, Milan, Italy) supplemented with 20% heat-inactivated fetal calf serum (FCS) (Eurobio, Courtaboeuf, France) and 50 ng/ml of granulocyte-macrophage colony-stimulating factor (a kind gift from Schering-Plough, Milan, Italy) to promote macrophage differentiation and for the last 4 days in the same medium without granulocyte-macrophage colony-stimulating factor.

The A549 cell line, derived from a human lung carcinoma, was grown in Dulbecco's modified minimal essential medium supplemented with 10% FCS.

Recombinant Nef proteins were obtained as hexahistidine-tagged fusion proteins as previously described (2, 17, 31). In particular, Nef proteins were generated from the different alleles NL4-3, BH10 (a generous gift from Mark Harris, Institute of Molecular and Cellular Biology and Astbury Centre for Structural Molecular Biology, University of Leeds), and SF2. Nef preparations were analyzed for the presence of endotoxin using the chromogenic Limulus amebocyte lysate endpoint assay QCL-1000 and, if required, purified using the EndoTrap endotoxin removal system (both from Cambrex). To avoid possible signaling effects due to residual lipopolysaccharide (LPS) traces in Nef preparations, all of the experiments were performed in the presence of 1 μg/ml of polymyxin B (Sigma-Aldrich), a cationic antibiotic that binds to the lipid A portion of bacterial LPS. In our hands, this polymyxin B treatment did not interfere with the signaling events analyzed and blocked the signaling activity of up to 50 endotoxin units (EU)/ml LPS.

The MEK/ERK pathway inhibitor PD98059 (2′-amino-3′-methoxyflavone) and the p38 inhibitor SB203580 [4-(4-fluorophenyl)-2-(4-methylsulfinylphenyl)-5-(4-pyridyl)1H-imidazole] were purchased from Merck Biosciences-Calbiochem (Nottingham, United Kingdom). The highly specific IKKα/β inhibitor BMS-345541 {4-(2′-aminoethyl) amino-1,8-dimethylimidazo[1,2-a]quinoxaline} (19) was a kind gift from James R. Burke, Department of Immunology, Inflammation and Pulmonary Drug Discovery, Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton, NJ.

Virus preparations, infections, and detection.

Preparations of VSV-G-pseudotyped HIV-1_NL4-3, produced as previously described (36, 73), were obtained as supernatants of 293 cells 48 h after the cotransfection of different derivatives of the pUc19/NL4-3 molecular clone with a plasmid expressing VSV-G under the control of the immediate-early cytomegalovirus promoter (molar ratio, 5:1), performed by the calcium phosphate method (99). The supernatants were clarified and concentrated by ultracentrifugation as described previously (24). The Δenv HIV-1 construct was obtained by inserting the SalI/BamHI fragment from the Δenv HXB2 HIV-1 molecular clone (85) in the SalI and BamHI sites of the pNL4-3 plasmid (1). To obtain the Δenv/Δnef pNL4-3 double mutant, this fragment was also inserted in the SalI and BamHI sites of the Δnef pNL4-3 molecular clone (91). The HIV-1 genome expressing the G2A nef mutant was produced as previously reported (73).

Virus preparations were titrated by measuring HIV-1 p24 contents by quantitative enzyme-linked immunosorbent assay (Abbott, Abbott Park, IL) and through a reverse transcriptase assay as described previously (84). Pseudotyped HIV-1 (10 ng [corresponding to approximately 5 × 10⁵ cpm]/10⁶ cells) was used to infect 7-day-old MDMs. The virus adsorption was performed in 48-well plates by incubating the cells for 1 h at 37°C with the viral inoculum diluted in 100 μl of complete medium. Afterwards, the viral inoculum was removed, and the cells were washed and refed with 300 μl of complete medium.

The percentages of cells expressing intracytoplasmic HIV-1 Gag-related products were evaluated by FACS analyses after they were treated with Permafix (Ortho Diagnostic, Raritan, NJ) for 30 min at room temperature (RT) and labeled for 1 h at RT with a l:50 dilution of KC57-RD1 phycoerythrin-conjugated anti-HIV-1 Gag monoclonal antibody (Coulter Corp., Hialeah, FL).

Immunodepletion of recNef.

To ensure total and specific depletion of recNef, complete medium supplemented with 100 ng/ml recNef was incubated for 8 h at +4°C with a 1:50 dilution of a cocktail containing six different monoclonal and polyclonal anti-Nef antibodies (all obtained from the National Institutes of Health AIDS Research and Reference Program). As a control, recNef-complemented medium was incubated with equal amounts of irrelevant isotype- and species-matched antibodies. Then, immunocomplexes were reacted with detergent-free protein A-G agarose beads (Pierce, Rockford, IL) overnight at +4°C. Afterwards, the immunocomplexes bound to protein A-G agarose were discarded through centrifugation; the supernatants were filtered (0.22-μm pore diameter) and added to MDM cultures.

Western blot assay.

MDMs were washed twice with phosphate-buffered saline (PBS), pH 7.4, and lysed in 20 mM HEPES, pH 7.9, 50 mM NaCl, 10 mM EDTA, 2 mM EGTA, 0.5% nonionic detergent IGEPAL CA-630 (Sigma), 0.5 mM dithiothreitol (DTT), 20 mM sodium molybdate, 10 mM sodium orthovanadate, 100 mM sodium fluoride, 10 μg/ml leupeptin, 0.5 mM phenylmethylsulfonyl fluoride for 20 min in ice. Whole-cell lysates were centrifuged at 6,000 × g for 10 min at 4°C, and the supernatants were frozen at −80°C. The protein concentrations of cell extracts were determined by the Lowry protein assay. Aliquots of cell extracts containing 20 to 50 μg of total proteins were resolved on 7 to 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred by electroblotting them on nitrocellulose membranes (Sartorius, Göttingen, Germany) for 60 min at 100 V with a Bio-Rad Trans-Blot. For the immunoassay, membranes were blocked in 3% bovine serum albumin (BSA) fraction V (Sigma-Aldrich, Milan, Italy) in TTBS/EDTA (10 mM Tris, pH 7.4, 100 mM NaCl, 1 mM EDTA, 0.1% Tween 20) for 1 h at RT and then incubated for 1 h at RT or overnight at +4°C with specific antibodies diluted in 1% BSA/TTBS-EDTA. Antibodies used in the different immunoblottings were as follows: rabbit polyclonal antibodies anti-phospho-IKKα(Ser180)/IKKβ(Ser181), anti-IKKβ, anti-IKKα, anti-phospho-Stat1(Tyr701), anti-phospho-Stat3(Tyr705), anti-phospho-SAPK/JNK(Thr183/Tyr185), anti-SAPK/JNK, anti-phospho-p38(Thr180/Tyr182), antip38 MAP kinase, and anti-phospho-p44/42 MAP kinase(Thr202/Tyr204), all from Cell Signaling Technology (Beverly, MA); rabbit polyclonal antibody anti-ERK1/2 from Promega (Madison, WI); rabbit polyclonal antibodies anti-STAT1 (E-23), anti-STAT2 (C-20), anti-STAT3 (C-20), and anti-IRF-3 (FL-425), all from Santa Cruz Biotechnology (Santa Cruz, CA); mouse monoclonal anti β-tubulin from ICN Biomedicals (Costa Mesa, CA); and rabbit polyclonal anti-JAK2 and anti-phospho-STAT2(Tyr689) from Upstate Biotechnology/Millipore (Billerica, MA). Immune complexes were detected with horseradish peroxidase-conjugated goat anti-rabbit or goat anti-mouse antiserum (Calbiochem), followed by enhanced chemiluminescence reaction (ECL; Amersham Pharmacia Biotech, Milan, Italy). To reprobe membranes with antibodies having different specificities, nitrocellulose membranes were stripped for 5 min at RT with Restore Western Blot Stripping Buffer (Pierce, Rockford, IL) and then extensively washed with TTBS/EDTA.

EMSA.

MDMs were washed twice in ice-cold PBS and lysed for 20 min at +4°C in electrophoretic mobility shift assay (EMSA) lysis buffer (20 mM HEPES, pH 7.9, 50 mM NaCl, 10 mM EDTA, 2 mM EGTA, 0.5% IGEPAL CA-630 [Sigma], 0.5 mM phenylmethylsulfonyl fluoride, 20 mM sodium molybdate, 10 mM sodium orthovanadate, 100 mM NaF, 10 μg/ml leupeptin, and 0.5 mM DTT). Three picomoles of the double-stranded oligonucleotide containing the NF-κB binding site of the IRF-1 promoter (5′-GGG CCG GCC AGG GCT GGG GAA TCC CGC TAA GTG TTT GGA T-3′) was end labeled with [γ-³²P]ATP (74 MBq/ml; 220 TBq/mmol; Amersham Bioscience Europe GmbH) by T4 polynucleotide kinase (New England BioLabs, Beverly, MA). Total cell extracts (20 μg of protein) were incubated in the presence of the labeled oligonucleotide probe (30,000 cpm) at +4°C for 30 min and at RT for 40 min in 20 μl of binding buffer [20 mM Tris, pH 7.5, 75 mM KCl, 13% glycerol, 1 mM DTT, 1 μg of bovine serum albumin, and 2 μg of poly(dI)·poly(dC)]. Cold competitor was added in a 100-fold molar excess of the radiolabeled probe. Identification of the NF-κB subunits contained in the DNA-protein complex was performed by adding 2 μg of anti-NF-κB p50 (NLS; Santa Cruz Biotechnology, Santa Cruz, CA), which allowed supershifting of the complex, or 2 μg of anti-NF-κB p65 (A; Santa Cruz Biotechnology), which abrogated the formation of the DNA-protein complex. DNA-protein complexes were resolved on 5% polyacrylamide gels in 25 mM Tris-borate buffer, pH 8.2, 0.5 mM EDTA and visualized by autoradiography.

RNA isolation and real-time PCR.

Real-time PCR assays were performed on total RNA isolated from MDMs treated or not for 2 h with 100 ng/ml of recombinant Nef or its mutants, using the RNeasy mini kit (QIAGEN, Milan, Italy) according to the manufacturer's instructions. Five hundred nanograms of total RNA was reverse transcribed using oligo(dT)_12-18 (Pharmacia-Biotech) as a primer and 50 units of Moloney murine leukemia virus reverse transcriptase enzyme (Gibco-BRL). Quantitative real-time PCR was then performed on reverse-transcribed IFN-β mRNA. The expression of GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was used to normalize the IFN-β mRNA level. IFN-β expression observed in control cells was chosen as a calibrator.

The following forward (F) and reverse (R) oligonucleotides were designed on the basis of GenBank sequences and used for the amplification: IFN-β (F), 5′-CAG CAG TTC CAG AAG GAG GA-3′, and IFN-β (R), 5′-AGT CTC ATT CCA GCC AGT GC-3′; GAPDH (F), 5′-GGG AAG GTG AAG GTC GGA GT-3′, and GAPDH (R), 5′-TCA TTG ATG GCA ACA ATA TCC ACT-3′.

Antiviral assay and detection of IFN-β.

The antiviral activities of supernatants collected from Nef-treated MDMs were tested on A549 cells. A 96-well microtiter plate (Greiner bio-one, Frickenhausen, Germany) was seeded with 4 × 10⁴ A549 cells/well in Dulbecco's modified minimal essential medium, 2% FCS. After 24 h, the culture medium was replaced with twofold dilutions of Nef-treated MDM conditioned medium (100 μl). Twenty-four hours later, the cell cultures were infected with 100 μl of a 1:1,000 dilution of murine encephalomyocarditis virus (viral titer, 2.8 × 10⁸ PFU/ml) and incubated for an additional 24 to 48 h. The cytopathic effect was visualized under inverted microscopy, and antiviral activity was calculated using a preparation of recombinant IFN-β (Rebif; 9 × 10⁷ IU/ml, 3 × 10⁸ IU/mg of protein; Ares-Serono) as a standard. Neutralization of the antiviral activity present in the supernatants was performed using anti-IFN-β polyclonal antibodies (neutralizing titer, 1:200,000 IU/ml; a generous gift from M. Capobianchi, IRCCS Lazzaro Spallanzani, Rome, Italy).

Immunofluorescence cell staining.

To detect IRF-3 nuclear translocation, MDMs were treated with a recombinant myristoylated Nef protein of the SF2 (17) strain of HIV-1 (myr⁺ recNef_SF2) for 2 h and then washed with 1× PBS without Ca²⁺ and Mg²⁺, fixed with 2% paraformaldehyde for 15 min on ice, washed four times with PBS, and permeabilized with 0.2% Triton X-100 in PBS for 5 min on ice. Afterwards, the specimens were incubated for 1 h in 1% BSA in PBS to decrease nonspecific binding to immunoglobulin G (IgG) and then incubated for 1.5 h at RT with anti-IRF-3 antibody (FL-425; Santa Cruz Biotechnology), washed four times with 1% BSA in PBS, incubated with goat anti-rabbit IgG-fluorescein-conjugated antibody (sc-2012; Santa Cruz Biotechnology), and washed again four times with 1% BSA in PBS. Nuclear DNA was counterstained with 5 μg/ml DAPI (4′,6′-diamidino-2-phenylindole; Sigma-Aldrich, Inc.) in PBS. Coverslips were mounted using antifade solution (Vectashield H-1000; Vector Laboratories Inc., Burlingame, CA) and analyzed using an Axiophot fluorescence microscope (Zeiss, Gottingen, Germany). Images were captured with a charge-coupled-device camera and reelaborated with Adobe Photoshop 6.0 software (Adobe Systems Inc., Mountain View, CA).

RESULTS

Nef treatment of MDMs induces rapid activation of NF-κB and IKK complexes.

We have previously shown that a 2-h treatment of human monocytes/macrophages with recNef is able to activate the NF-κB pathway, leading to the formation of p50/p50 homodimer- and p50/p65 heterodimer-containing DNA-protein complexes in EMSA (70). NF-κB activation correlated with the synthesis and the release of proinflammatory cytokines and chemokines (i.e., IL-1β, IL-6, TNF-α, CCL2/MIP-1α, and CCL4/MIP1-β) (70). To gain more insight into this phenomenon, we studied the kinetics of NF-κB activation by treating purified MDMs with myr⁺ recNef. The modified protein is produced in Escherichia coli by coexpression of vectors coding for Nef and N-myristoyl transferase (NMT) and the addition of myristic acid before induction (17, 31). The DNA binding activities of NF-κB family members to a specific DNA probe by EMSA were evaluated using total cellular extracts. As shown in Fig. 1A, a 30-min treatment with myr⁺ recNef_BH10 induces the appearance of a DNA-protein complex containing the p50/p65 heterodimer. The p50/p50 homodimer-containing complex, already present in untreated cells, was increased only at later time points (2 and 6 h), when the signal corresponding to the heterodimer decreased. NF-κB activation was also evident after 15 min (data not shown). To verify the specificity of NF-κB activation, cells were stimulated for 15 min with Nef-complemented medium after immunodepletion with anti-Nef specific or isotype-matched nonspecific antibodies. As shown in Fig. 1B, Nef-immunodepleted medium was unable to efficiently activate NF-κB, ruling out the possibility that the effect was due to contaminants present in the recombinant-protein preparation (Fig. 1B).

FIG. 1.

Nef treatment of MDMs activates NF-κB. (A) EMSA performed using total extract of MDMs treated for the indicated time with 100 ng/ml myr⁺ recNef_BH10. The extracts were incubated in the presence of a γ-³²P-labeled probe containing the NF-κB sequence of the irf-1 gene promoter. Identification of NF-κB subunits contained in the complex was obtained using a polyclonal antibody, anti-p50 or anti-p65 (see Materials and Methods). As a competitor, a molar excess (200×) of the same unlabeled probe was used. (B) Medium supplemented with 100 ng/ml of myr⁺ BH10 Nef was immunodepleted as described in Materials and Methods and used to stimulate the MDMs for 15 min. EMSA was performed as for panel A. The results shown in panels A and B were obtained with two different donors.

The IKK complex is the main regulator of the NF-κB pathway (for a review, see references 55 and 105). The two catalytic subunits of the complex, named IKKα and IKKβ, are activated through serine phosphorylation and can, in turn, phosphorylate members of the IκB family, leading to their degradation via the ubiquitin-proteasome pathway and to NF-κB nuclear translocation. Therefore, we thought that Nef-induced NF-κB activation could be the consequence of IKK activation. The phosphorylation of IKKα and -β was analyzed in cell extracts using anti-phospho specific antibodies. As shown in Fig. 2, the treatment of MDMs with myr⁺ recNef induced the rapid phosphorylation of both IKKα and -β, which was already evident after 15 min, peaked at 30 min to 45 min, and remained above the control levels after 2 h. The experiment was performed using both the myristoylated recombinant proteins of the BH10 (31) (Fig. 2A) and SF2 (17) (Fig. 2B) strains of HIV-1. Unlike BH10, which is highly laboratory adapted, the SF2 strain was cloned directly from clinical material and is closer to a primary isolate of HIV-1. The SF2 Nef protein is four residues longer than BH10, because it contains an insert at positions 23 to 26, and the primary amino acid sequence is significantly divergent (12%) from that of BH10. Taken together, these results suggest that IKK activation is an allele-independent event and that this feature of Nef is not related to the adaptation of HIV-1 strains in the laboratory.

FIG. 2.

Nef treatment of MDMs induces the rapid phosphorylation of IKKα and IKKβ. MDM cultures were treated as shown with 100 ng/ml of myr⁺ BH10 (A) or 100 ng/ml myr⁺ SF2 (B) Nef. Cellular extracts (50 μg) were analyzed by Western blotting using specific anti-phospho-IKKα/β, -IKKα, and -IKKβ antibodies. As a control (Ctr), the membranes were blotted with anti-β-tubulin (A) or anti-JAK-2 (B) antibodies. The results shown in panels A and B were obtained with two different donors.

In addition to the myr⁺ recNef_BH10 and recNef_SF2 proteins, we also used a recombinant protein derived from the T-tropic strain NL4-3 that was expressed in E. coli without the vector coding for the NMT (36) (NMT⁻ NL4-3 recNef). This protein is not myristoylated, but it could be lipidated inside the cells following its internalization because of its intact myristoylation consensus sequence. In agreement with this hypothesis, the NMT⁻ NL4-3 wild-type protein partially colocalized with the membrane marker CD14 in confocal analysis performed in MDM-treated cells, whereas the G2A NL4-3 mutant did not (2). Previous results showed that NMT⁻ NL4-3 recNef was less effective than the myr⁺ recNef_BH10 in the activation of NF-κB binding activity to DNA (70), suggesting that Nef myristoylation plays an important role in the activation of the NF-κB pathway. To address this point, we treated MDMs with NMT-coexpressed recNef_SF2 mutated in the myristoylation sequence (G2A) or with the N-terminal 44 amino acids deleted (ΔN-term). Both mutants were unable to activate IKKs at all time points tested (Fig. 3), demonstrating that myristoylation of the protein is indeed a fundamental prerequisite to induce efficient IKK phosphorylation.

FIG. 3.

IKK activation depends on Nef myristoylation. MDMs were treated for the indicated times with 100 ng/ml of myr⁺ SF2 Nef (wild type [WT]), a mutant in the myristoylation site (G2A), or a Nef protein with the first 44 amino acids deleted (ΔN-term). The cells were lysed as reported in Materials and Methods, and total cell extracts (50 μg) were analyzed by Western blotting. In the upper blot, the phosphorylation levels of IKKs; in the middle blots, the expression of IKKβ and IKKα, respectively; and in the bottom blot, the expression of β-tubulin used as an internal control are shown. Ctr, untreated cells.

Nef treatment of MDMs induces rapid activation of ERK1/2, JNK, and p38.

Recently, it has been reported that in monocyte/macrophage cell lineages, Nef is able to intersect the CD40/CD40L signaling pathway when expressed via adenovirus-based vectors or in the context of HIV-1 infection (93). In several cell types, CD40 triggering allows the activation of both the NF-κB and MAPK (i.e., ERK, JNK, and p38) pathways (27, 88, 92). To verify if exogenously added recNef could activate MAPKs besides NF-κB, we treated MDMs with myr⁺ recNef_BH10, testing the levels of phosphorylation of the three MAPKs using specific anti-phospho antibodies. The results in Fig. 4 show that Nef treatment is able to induce phosphorylation of ERK1/2, JNK, and p38. The kinetics of activation of both JNK and p38 parallel those of NF-κB (compare Fig. 4A and B with 2A); conversely, ERK1/2 activation seems to precede NF-κB activation, reaching a peak at 15 min (Fig. 4C). Similar results were obtained using the myr⁺ SF2 Nef allele (data not shown).

FIG. 4.

Nef treatment of MDMs induces the rapid phosphorylation of ERK1/2, JNK, and p38 MAP kinase. MDMs were treated for the indicated times with 100 ng/ml of myr⁺ BH10 Nef and lysed in lysis buffer as reported in Materials and Methods. (Top) Cellular extracts (30 μg) were analyzed by Western blotting using specific antibodies, anti-phospho-JNK (A), -p38 (B), and -ERK1/2 (C). (Middle) Membranes were reblotted using antibodies recognizing total JNK (A), p38 (B), and ERK1/2 (C). (Bottom) β-Tubulin expression used as a control. Ctr, untreated cells.

Nef-induced IKK activation is required for the production of STAT1- and -3-activating factors.

We previously correlated the production of STAT-activating factors with the cycloheximide-independent activation of the NF-κB pathway (36, 70, 73). To demonstrate a direct link between the activation of STAT1 and -3 and those of IKKs and/or MAPKs, we treated MDMs with myr⁺ recNef_SF2 in the presence of specific inhibitors of IKKs (i.e., BMS-345541) and ERK1/2 and p38 (PD98059 and SB203580, respectively). As shown in Fig. 5, Nef did not induce STAT1 and -3 phosphorylation in the presence of the IKK-specific inhibitor used alone or in combination with PD98059 and/or SB203580. The p38 inhibitor, SB203580, showed only a slight effect that was more pronounced in the combined treatment with PD98059; on the other hand, PD98059 alone had no effect on Nef-mediated STAT activation. These results indicate that IKK activation is absolutely required for the subsequent production of STAT1- and -3-activating factors, whereas p38 and ERK are not, even if they could be involved in a fine-tuning of the phenomenon.

FIG. 5.

Tyrosine phosphorylation of STAT1 and STAT3 depends on Nef-dependent IKK activation. MDMs were pretreated for 1 h with BMS-345541 (25 μM), PD98059 (10 μM), and SB203580 (10 μM) alone or in combined treatment and then incubated for an additional 2 h in the presence of myr⁺ SF2 Nef (100 ng/ml). The cells were lysed, and cellular extracts (30 μg) were analyzed by Western blotting for the phosphorylation of STAT1 and STAT3, using anti-phospho-tyrosine specific antibodies. Protein levels were monitored using antibodies recognizing total STAT1 and STAT3, whereas β-tubulin was used as an internal control.

Nef treatment of MDMs induces IRF-3 phosphorylation, production of IFN-β, and activation of STAT2.

Using a gene chip microarray procedure, we observed that treatment of MDMs with recNef regulates the expression of many genes and, interestingly, up-regulates the level of IFN-β mRNA (G. Fiorucci, unpublished data). To confirm this result, real-time PCR was performed on total mRNA isolated from recNef-treated MDMs. As shown in Table 1, recNefs were all able to up-regulate the IFN-β mRNA. Again, myristoylated BH10 and SF2 proteins induced a much stronger accumulation of the mRNA than the NMT⁻ NL4-3 protein. We also tested whether the increase in the expression of the IFN-β mRNA is followed by the production and release of this cytokine. For this purpose, supernatants of recNef-treated MDM cultures were used to test the induction of the antiviral state in the A549 human cell line. The antiviral effect was quantified by comparison with that induced by a standard preparation of human recIFN-β and expressed as units/ml. As shown in Fig. 6A, supernatants collected from MDMs treated with NMT⁻ NL4-3 or with myr⁺ recNef_BH10 and recNef_SF2 induced the antiviral state in A549 cells. As in the real-time PCR analysis, the supernatants of the cells treated with the myr⁺ SF2 and BH10 proteins induced a stronger antiviral response than the one collected from cells treated with NMT⁻ NL4-3 recNef (Fig. 6A and B). The units of antiviral activity contained in the supernatants were also compared with those present in the supernatant of MDMs treated with 100 EU/ml of LPS (Fig. 6B). It is well known that LPS is able to induce the production of IFN-β in MDMs via the signal transduction events triggered by its engagement with TLR4 (37). The amount of antiviral activity detected in the supernatant of MDMs treated for 4 h with 100 ng/ml of myr⁺ SF2 Nef equaled the amount detected in the supernatant of MDMs treated for 8 h with 100 EU/ml of LPS. The Nef-induced antiviral state in A549 cells is entirely due to the presence of IFN-β in the supernatants of Nef-treated MDMs, because it appears to be completely neutralized by anti-IFN-β specific antibodies (Fig. 6C). Among the STAT transcription factors, STAT2 is activated only in response to type I IFNs and epidermal growth factor, even if only the former activates STAT2 through phosphorylation on tyrosine (57, 65). To verify if Nef treatment induces STAT2 tyrosine phosphorylation, cellular extracts of MDMs, treated for different durations with myr⁺ recNef_SF2, were analyzed by Western blotting using anti-phosphotyrosine-STAT2 specific antibodies. STAT2 tyrosine phosphorylation became evident after 1.5 h of treatment and persisted at least until 4.5 h, in agreement with the kinetics of IFN-β release (compare Fig. 6A to D). Using cell extracts of MDMs treated with nonlipidated NL4-3 recNef, we observed a delayed tyrosine phosphorylation of STAT2 (clearly evident only after 3.5 h of treatment) compared with those induced by both myr⁺ BH10 and SF2 Nef. This result is in agreement with previous observations showing the lack of ISGF3(STAT1-STAT2-IRF-9)/interferon-stimulated response element (ISRE) complex formation in EMSA performed using cell extract of MDMs treated for 2 h with NMT⁻ NL4-3 recNef (36).

TABLE 1.

IFN-β mRNA induction in MDMs^a

Expt	Treatment	Induction (n-fold)
I	Control	1
	recNefNL4-3	415
	recNefBH10	24,312
II	Control	1
	recNefNL4-3	784
	recNefSF2	18,363

^aMDMs were treated with 100 ng/ml recNef for 2 h. Real-time PCR was performed on total RNA isolated from MDM-treated cells as described in Materials and Methods.

FIG. 6.

Nef induces the release of IFN-β. (A) Supernatants collected from MDMs treated for the indicated times with 100 ng/ml of myr⁺ BH10 Nef (open circles), 100 ng/ml of myr⁺ SF2 Nef (black circles), or 100 ng/ml of NL4-3 Nef (open squares) were used to treat A549 cells. The antiviral state induced by these supernatants on A549 cells was evaluated in triplicate as reported in Materials and Methods. The error bars indicate standard deviations. (B) Supernatants collected from MDMs treated for 4 h with 100 ng/ml of NL4-3 Nef (light-gray column) or 100 ng/ml of myr⁺ SF2 Nef (gray column) or for 4 and 8 h with LPS at 100 EU/ml (stippled columns) were used to treat A549 cells. The supernatants collected from those cultures were tested for the induction of the antiviral state in the A549 cells as for panel A. (C) The supernatants collected from MDMs treated for 4 h with 100 ng/ml of myr⁺ SF2 Nef (gray columns) or with 100 EU/ml of LPS (white columns) were incubated with anti-IFN-β specific antibodies (hatched columns) and tested for the induction of the antiviral state in A549 cells as for panel A. In both panels B and C, the black columns represent the control (Ctr; untreated cells). (D) MDM cultures were treated with 100 ng/ml myr⁺ SF2 Nef (left) or with 100 ng/ml of myr⁺ BH10, NMT⁻ NL4-3, and myr⁺ SF2 Nef, respectively (right), for the indicated times. Cellular extracts (30 μg) were analyzed by Western blotting using specific anti-phospho-tyrosine-STAT2, anti-STAT2, and anti-β-tubulin antibodies. (E) MDM cultures were infected with VSV-G-pseudotyped Δenv HIV-1 expressing wild-type Nef (Δenv), with nef deleted (Δenv/Δnef), or expressing a Nef mutant in the myristoylation site (Δenv/G2A). Sixteen hours after infection, the cells were lysed, and the cell lysates were analyzed for the activation of STAT1 and STAT2 using anti-phosphotyrosine specific antibodies. The efficiencies of infections were monitored by evaluating the percentages of p24_gag-positive cells by FACS analysis (right).

Next, we assessed whether STAT2 tyrosine phosphorylation is achieved in the context of the viral infection by performing a high-efficiency single-cycle infection procedure with VSV-G-pseudotyped HIV-1_NL4-3. We previously showed that this procedure is able to activate STAT1 in MDMs and that this activation correlates with the expression of both nef and env genes (36). It has also been shown that treatment of monocytes/macrophages with gp120 is able to induce the production of IFN-β (42). For both these reasons, we decided to infect MDMs with VSV-G-pseudotyped HIV-1 lacking env but able to express wild-type Nef (Δenv) or a Nef mutant in the myristoylation site (Δenv/G2A) or lacking both env and nef (Δenv/Δnef). We independently analyzed the responses of MDMs purified from three different donors. Figure 6E shows the Western blot analysis of the cell extracts from one donor after 16 h of infection. STAT1 and STAT2 appear tyrosine phosphorylated in cells infected with the HIV-1 pseudotypes coding for wild-type Nef, but not with the virus coding for the G2A mutant or Δenv/Δnef virus. Depending on the donor, the phosphorylation signals were also detected as early as 8 h after infection (data not shown), in agreement with previous data on STAT1 and STAT3 tyrosine phosphorylation (36, 73). Taken together, these results strongly suggest that type I IFN is produced during the early phases of infection, since, as mentioned above, STAT2 tyrosine phosphorylation is a good marker of type I IFN production.

It is well established that IFN-β gene expression requires the activation of IRF-3 (53, 74). IRF-3 is a constitutively expressed member of the IRF family, a group of related transcription factors first identified as regulators of the IFN-α and -β gene promoters, as well as of promoters of some IFN-stimulated genes via their interaction with the ISRE. To test whether Nef is able to activate IRF-3, leading to the transcription of the IFN-β gene, MDMs were treated with myr⁺ recNef_BH10 and cellular extracts were analyzed by Western blotting using anti-IRF-3 specific antibodies. Activation of IRF-3 was detected as a shift in the electrophoretic mobility of the protein, due to the increase in its phosphorylation state (103). As shown in Fig. 7A, the accumulation of the slower-migrating form of IRF-3 after myr⁺ recNef_BH10 treatment of MDMs started at 15 min, became more evident at 30 min to 45 min, and was still present after 2 h. Analysis performed using fluorescence microscopy on MDMs treated for 2 h with myr⁺ recNef_SF2 and stained with anti-IRF-3 antibodies showed that the phosphorylation of IRF-3 was followed by its nuclear translocation (Fig. 7B). Once again, the phosphorylation of IRF-3 was achieved using both myr⁺ BH10 and myr⁺ SF2 Nef alleles (Fig. 7A and 8B) but was not observed when MDMs were treated with G2A or ΔN-term SF2 Nef mutants (Fig. 8B), indicating that the myristoylation of the protein is required to obtain this effect.

FIG. 7.

Nef activates IRF-3. (A) MDMs were treated for the indicated times with 100 ng/ml of myr⁺ BH10 Nef, and the cellular extracts (30 μg), separated on 9% SDS-PAGE, were analyzed by Western blotting using a specific anti-IRF3 antibody. The phosphorylation of IRF-3 was visualized as the increase of the more slowly migrating form corresponding to the hyperphosphorylated form of IRF-3. β-Tubulin expression was used as an internal control. (B) MDMs were treated for 2 h with myr⁺ recNef_SF2 (panels II, IV, and VI) or left untreated (panels I, III, and V) and then fixed and stained using anti IRF-3 antibodies as reported in Materials and Methods. (I and II) Anti-IRF-3 staining. (III and IV) DAPI staining of nuclear DNA. (V and VI) Merge.

FIG. 8.

Effects of Nef mutants on IKKs and IRF-3 activation, IFN-β production, and STAT2 phosphorylation. MDMs were treated for 30 min with 100 ng/ml of myr⁺ SF2 Nef or mutants thereof. The cells were lysed as described in Materials and Methods, and the cellular lysates (30 μg) were analyzed by Western blotting for the phosphorylation of IKKα and IKKβ on 7% SDS-PAGE (A) or for the activation of IRF-3 on 9% SDS-PAGE (B). (C) MDM cultures were treated for the indicated times with 100 ng/ml of myr⁺ SF2 Nef or mutants thereof, and then the supernatants were collected and analyzed in triplicate for the induction of the antiviral state in the A549 cells as described in Materials and Methods. The error bars indicate standard deviations. (D) Cells were treated for 2 h with 100 ng/ml of myr⁺ SF2 Nef or mutants thereof, and the cellular lysates were analyzed by Western blotting on 7% SDS-PAGE using anti-phospho-STAT2 and anti-phospho-STAT1 specific antibodies. Anti-β-tubulin was used as a gel-loading control. Ctr, untreated cells; WT, wild type.

Analysis of Nef mutants.

Since we had identified Nef myristoylation as being of fundamental importance for the activation of NF-κB, MAPKs, and IRF-3, we tried to dissect the roles of other sequence motifs of the viral protein in those functions. For this purpose, besides G2A and ΔN-term mutants, we used different myristoylated SF2 proteins mutated in functional motifs involved in the interaction and down-regulation of CD4 (mutant C⁵⁹AWL⁶²-AAAA) or in the interaction with SH3-containing proteins (the proline-rich mutant P⁷⁶XXP⁷⁹XR⁸¹-AXXAXA) or with cellular proteins involved in endocytic pathways, such as the V1H subunit of the vacuolar-membrane ATPase (mutant E¹⁷⁸D¹⁷⁹-AA, called DDAA) (45) or the adaptor protein complex (mutant L¹⁶⁸L¹⁶⁹-AA). We also used a Δ-Loop mutant (Δ158-178) lacking the C-terminal flexible loop, projecting out from the protein core domain, that contains both the ED and the LL motifs described above and a motif involved in the interaction with the β subunit of the COP-I coatomer (E¹⁵⁸E¹⁵⁹). All these mutants, except G2A and ΔN-term, induce phosphorylation of IKKs (Fig. 8A), activation of IRF-3 (Fig. 8B), the production of IFN-β (Fig. 8C), and tyrosine phosphorylation of STAT2 and STAT1 (Fig. 8D), indicating that those interaction motifs are dispensable for the Nef-induced effects described here.

DISCUSSION

There is increasing evidence of the ability of extracellular Nef to activate signaling pathways in uninfected cells of different lineages (18, 39, 80, 96, 97), although it is described as an intracellular viral protein expressed inside the HIV-infected cell by the transcription of the proviral genome. Indeed, Nef is internalized by MDMs and dendritic cells, but not by T cells (2), when added to cell cultures (2, 81, 82) and induces apoptosis in uninfected T cells by interacting with CXCR4 (54, 56). Recently, Qiao et al. (80) reported that Nef is internalized in B cells in vitro, thereby suppressing CD40-dependent immunoglobulin class switching. In vivo analysis performed on infected follicles of lymphoid tissue from symptomatic HIV-1 patients showed that a variable portion of IgD⁺ B cells at the edge of the germinal center and in the interfollicular area contained Nef. These IgD⁺ B cells were unlikely to be infected by HIV-1, because they lacked p24 and p17 viral proteins (80). The presence of Nef in the sera of HIV-infected patients at concentrations ranging from 1 to 10 ng/ml has also been described (39). This concentration may be higher in the lymphonodal germinal centers, where virion-trapping dendritic cells and virion-infected CD4⁺ T cells and macrophages are densely packed (63, 71). Infected cells may release Nef through a nonclassical secretory pathway or after lysis, and then bystander cells could internalize Nef via endocytosis, pinocytosis, or other yet-unknown mechanisms. Regarding intracellular signaling induced by Nef treatment of MDMs, we have previously reported that this protein modulates the expression of a significant number of genes as early as 2 h after treatment (70; Fiorucci, unpublished). This was suggestive of a prompt transcriptional cell reprogramming induced by Nef that leads to the synthesis and the release of proinflammatory cytokines/chemokines that in turn activate the signal transducers and activators of transcription STAT1 and STAT3 (36, 73). In agreement with these results, we have shown that Nef treatment of MDMs induces a rapid activation of IKK/NF-κB, MAPK, and IRF-3 signaling pathways. Regarding MAPK activation, Nef induces prompt phosphorylation of the three MAP kinases ERK1/2, JNK, and p38 (Fig. 4). Nef-induced ERK and JNK activation has been also observed previously by other authors (87, 97). To our knowledge, only one paper has previously reported that Nef is able to activate p38 in a Jurkat T-cell line inducing FasL on infected cells and apoptosis in bystander cells (67). In that case, the p38 phosphorylation was detected 24 h after HIV-1 infection and was sensitive to the p38 inhibitor RWJ67657, suggesting that this phosphorylation was an effect mediated by p38-activating factors rather than a primary Nef-mediated effect. Here, we showed that a Nef treatment as short as 15 min was able to induce p38 phosphorylation (Fig. 4) that, being insensitive to the p38 inhibitor SB203580 (data not shown), appears to be due to rapid recruitment and activation of signaling intermediates upstream of p38.

Using the IKK inhibitor BMS-345541, we also obtained evidence that the synthesis of STAT-activating factors requires Nef-induced IKKα and -β activation, whereas ERK and p38 activation is not strictly required (Fig. 5). From this point of view, it is interesting that many of the transcripts induced in MDMs by Nef treatment are encoded by genes regulated by κB-like responsive elements (70; Fiorucci, unpublished). Our observation points to a capacity of Nef to “hijack” this basic signaling pathway, as already observed after endogenous expression in MDMs (93). NF-κB activation promotes HIV-1 replication via both direct and cytokine-mediated effects. Indeed, it has been shown that NF-κB activation is required for optimal long-terminal-repeat-driven expression (8). It has also been observed that the Nef-inducible cytokines TNF-α and IL-1β (70) are known to increase the replication of HIV-1, especially in cells of monocyte/macrophage lineage (52, 78, 79).

In addition to MAPK and IKK phosphorylation, we observed that exogenously added Nef also induces the rapid phosphorylation of the transcription factor IRF-3, the main regulator of IFN-β gene expression (53, 86, 104). In agreement with this observation, we also detected the induction of IFN-β mRNA and protein and the tyrosine phosphorylation of STAT2, which is well known to be induced by type I IFN signaling. The possibility that these signaling events could be induced by triggering of TLR4 by LPS traces present in the recNef preparations was carefully evaluated and excluded (see Materials and Methods). The evidence that the G2A and ΔN-term mutants do not induce signaling effects (Fig. 8A to D) is genuine proof that they are specifically induced by Nef. We have also checked whether LPS and Nef could act synergistically to produce IFN-β. In particular, combined treatment of MDMs with suboptimal amounts of LPS and Nef was unable to induce any synergistic effect on IFN-β production, whether the cells were simultaneously treated with both molecules or pretreated with one of them for 30 min (data not shown).

It has been shown that other HIV products, i.e., gp120, are able to induce the production of type I IFNs in MDMs, as well as in PBMC (3, 20, 21, 42). Induction of both IFN-α and -β production as a consequence of the in vitro infection of MDM cultures has been reported (43, 95). In particular, IFN-α production was detected over 7 to 21 days following infection (95), whereas the presence of anti-IFN-β antibodies during the infection process increased the release of p24 in the supernatants of MDMs infected with HIV-1_Ba-L (43). Using a highly efficient single-cycle infection of MDM with VSV-G Δenv HIV-1_NL4-3 pseudotypes, we detected the induction of STAT2 tyrosine phosphorylation in cells infected with the pseudotypes coding for wild-type Nef but not in those coding for the G2A Nef mutant or the mutant with nef deleted. STAT2 tyrosine phosphorylation was revealed at an early stage of infection (8 to 16 h, depending on the donor). These results, together with our previous results on STAT1 tyrosine phosphorylation in MDM-infected cells (36), further support the idea that the expression of both nef and env genes triggers the induction of type I IFN production in MDMs. To date, the reason for such IFN-β production in MDMs, a cell population very sensitive to the antiviral effects of IFNs, is unclear and underlines the complex physiological function of these cells in maintaining normal homeostasis in vivo in response to virus infection. A possible explanation could be found in the capability of type I IFNs to protect the macrophages from a productive infection, as reported in early publications (9, 61, 77), leading to the establishment of a persistent infection and to the creation of a viral reservoir in a cell type that is particularly resistant to antiretroviral therapies (28). In fact, it is well known that type I IFNs inhibit the release of viral particles in cells chronically infected by retroviruses in the absence of a significant reduction of viral macromolecular synthesis (9, 15, 22, 23, 38, 69, 72, 77, 89) and that antiserum to type I IFNs can increase retrovirus production by a factor of 10 to 50 (12). In this respect, it is also interesting to remember that in the early 1980s, one of the first clear-cut HIV isolates was obtained from cultured T lymphocytes derived from a lymphonode biopsy specimen from a patient with lymphadenopathy, with the help of IL-2 and anti-IFN-α serum (11). It has also been reported that primary HIV-1 isolates display a broad range of sensitivities to IFN-α2. In particular, the prevalence of IFN-α2 resistance was low in the absence of AIDS but dramatically increased once HIV infection progressed to AIDS (62).

The rapid phosphorylation of MAPKs, IRF-3, and both IKKα and -β induced by Nef treatment of MDMs is compatible with two different hypotheses: (i) the interaction between Nef and a putative receptor present on the cell membranes of MDMs or (ii) the internalization of the protein, followed by interaction between Nef and upstream activators of these signaling cascades.

In an attempt to identify the Nef structural motifs required for the activation of these signaling pathways, we treated MDMs with different myr⁺ recNef_SF2 proteins mutated in conserved functional regions. In particular, we used mutants in the polyproline-rich region (PXXPXR) or in domains involved in the interaction with CD4 (CAWL) or in interactions with elements of the endocytic machinery (LL and DD) (see Results). We also used two deletion mutants lacking the first 44 amino acids (ΔN-term) or the C-terminal flexible loop (Δ-Loop) and a mutant in the myristoylation site (G2A). The data reported here indicate that myristoylation of the protein is required for the activation of the signaling cascades, as the G2A Nef mutant is unable to activate IKK/NF-κB and IRF-3, as well as the synthesis and the release of IFN-β and STAT2 tyrosine phosphorylation. These data are in agreement with a lack of induction of both STAT1 and STAT3 tyrosine phosphorylation in MDMs 8 to 16 h after infection with VSV-G Δenv HIV-1 pseudotypes expressing the G2A Nef mutant (Fig. 6E) (73). It has recently been shown that myristoylation is only a weak membrane-targeting signal and that N-terminal basic residues, especially an arginine-rich cluster (R17 to R22), are needed for the stable association of the viral protein with cellular membranes (14, 98). We cannot formally exclude here the hypothesis that these residues are important for the activation of NF-κB and IRF-3, as the arginine cluster is already present in the G2A mutant, but the data obtained (Fig. 8) indicate that Nef myristoylation is an absolute requirement for the activation of those pathways.

Regarding the Nef alleles tested, we have observed that both of the myristoylated recombinant proteins used (SF2 and BH10) are more effective than the NMT⁻ NL4-3 recNef that was expressed in E. coli without the vector coding for the human N-myristoyl transferase (Table 1; Fig. 6A, B, and D; and data not shown). The differences in the amplitude of the responses evoked by the fully myristoylated recombinant proteins and the G2A mutant or the NMT⁻ NL4-3 recNef might be explained by conformational changes in the protein upon lipidation. A recent study has found that myristoylated Nef appears more compactly folded than its nonmodified variant, suggesting that the cytosolic form of Nef presents different accessible surfaces than the membrane-bound form of Nef when the myristate and parts of the anchor domain are hidden in the lipid bilayer (17). Another explanation might lie in the different oligomerization properties of myr⁺ versus myr⁻ Nef proteins in solution. Indeed, myr⁺ Nef is normally present as a monomer (17, 31) and, in this form, might be internalized faster than the nonlipidated protein, or it could be targeted more efficiently to a subcellular localization in a signaling-active compartment, such as glycolipid-enriched microdomains. Conversely, nonmyristoylated proteins form dimers, trimers (17), or even oligomers of greater magnitude (31), suggesting that such oligomers could be less active than monomers when added to tissue culture or might be targeted less efficiently to the signaling compartment. As described in Results, the nonmyristoylated NL4-3 recombinant protein could be myristoylated, at least in part, inside the cells. The reduced efficiency of NMT⁻ NL4-3 recNef protein to activate the signals, described here, could also explain the apparent contradiction in the data reported here, showing that both the full myr⁺ SF2 DDAA and ΔLoop mutants activate IKKs and IRF-3 (Fig. 8A and B), and those data previously obtained using nonlipidated DDAA and EEQQ NL4-3 mutants that were unable to activate NF-κB after 2 h of treatment (70). This suggests that these motifs are required to help the signaling activity only when nonlipidated proteins are used to treat the cells.

The data reported here extend our previous model to the activation of MAPKs, IκB kinases, and type I IFN regulation (Fig. 9), strengthening the concept that Nef, as a polyvalent viral protein, is able to “hijack” different cellular signaling pathways in MDMs, inducing a reprogramming of gene expression that could favor both the establishment of a persistent infection in monocytes/macrophages and the recruitment of other HIV-sensitive cellular targets.

FIG. 9.

Model for Nef-induced signaling in monocytes/macrophages. Exogenous added myr⁺ Nef activates the IKK/NF-κB pathway, the MAPKs, and IRF-3. The activation of the NF-κB and MAPK pathways might increase the expression of the proviral genome in infected cells and also take part in the regulation of cellular-gene expression that results in the synthesis and release of inflammatory factors, such as IL-1, IL-6, TNF-α, CCL2/MIP1α, and CCL4/MIP1β, and in the production of IFN-β. Upon binding to their specific receptors on the same or neighboring cells, those factors lead to the activation of STAT1, -2, and -3, further regulating cellular functions.