|Home | About | Journals | Submit | Contact Us | Français|
Many enveloped viruses use the ESCRT proteins of the cellular vacuolar protein sorting pathway for efficient egress from the cell. Recruitment of the ESCRT proteins by human immunodeficiency virus type 1 (HIV-1) Gag is required for HIV-1 particle budding and egress. ESCRT proteins normally function at endosomal membranes, where they facilitate the downregulation of mitogen-activated receptors such as EGF receptor (EGFR) through multivesicular body biogenesis. It is not known whether the Gag-mediated recruitment of ESCRT proteins functionally depletes the pool of these molecules that is available for the downregulation of EGFR. Here we show that the expression of HIV-1 Gag decreases the rate of EGFR downregulation, as assessed by decreases in the rates of 125I-EGF and EGFR degradation. The effect of Gag was dependent on the presence of the TSG101 binding motif (PTAP) within the Gag C-terminal p6 domain. Cells expressing HIV-1 Gag retained more EGFR in late endosomes. This effect occurred when Gag was expressed alone from a heterologous promoter and when Gag expression was driven by the HIV-1 long terminal repeat within pHXB2ΔBalD25S, a noninfectious lentiviral vector. Gag-expressing cells exhibited higher levels of activated mitogen-activated protein kinase for longer times after EGF addition than did cells that did not express HIV-1 Gag. These results indicate that HIV-1 Gag can impinge upon the functioning of the cellular vacuolar protein sorting pathway and reveal yet another facet of the intricate effects of HIV-1 infection on host cell physiology.
The Gag polyprotein precursor of human immunodeficiency virus type 1 (HIV-1) drives viral assembly and is sufficient for the assembly and budding of virus-like particles (6, 8, 10, 12). After their synthesis on cytosolic ribosomes, Gag molecules are targeted to cellular membranes, where Gag-Gag multimerization results in membrane deformation and budding (27). Particle release is dependent on the function of a conserved P(T/S)AP motif, located in the p6 domain of Gag (13), that recruits cellular proteins to assist in the viral budding and membrane fission process (5, 7, 9, 20, 29). Recently, the entire cellular protein network that participates in HIV-1 budding was mapped, with TSG101 and AIP1 identified as direct interaction partners of the Gag p6 domain (19, 25, 30). All of these proteins, termed class E proteins or ESCRTs I, II, and III, are normally involved in the sorting of ubiquitinated cargo, such as ligand-activated cell surface receptors, for delivery into the lumens of multivesicular bodies (MVBs) (1, 3, 15).
It has been postulated that HIV-1 Gag may have evolved to mimic Hrs, a cellular protein that uses a PSAP motif to recruit TSG101 to endosomal membranes (2, 17). Gag has a sevenfold-higher affinity for TSG101 than Hrs does and can effectively compete with Hrs for TSG101 binding in vitro (22). Given these observations, we asked whether HIV-1 Gag and Hrs compete for TSG101/ESCRT proteins in vivo. Specifically, does the overexpression of HIV-1 Gag, which may be physiologically relevant during a productive HIV-1 infection, antagonize TSG101 recruitment by the endosomal machinery, resulting in inefficient cell surface receptor downregulation?
To address this question, we monitored the ligand-induced downregulation of the EGF receptor (EGFR), a cell surface receptor tyrosine kinase whose trafficking through the endocytic pathway has been well characterized (24, 34). The binding of EGF to EGFR results in receptor dimerization and phosphorylation, followed by the activation of a signaling cascade involving multiple mitogen-activated protein (MAP) kinases. The timely downregulation of this signaling is essential to avoid perturbations in cell physiology. EGFR downregulation is achieved through the internalization of activated receptors by clathrin-dependent endocytosis, the delivery of receptors into the lumens of MVBs, and their subsequent lysosomal degradation. The overexpression or depletion of Hrs or TSG101 results in an attenuated degradation of EGFR, indicating that this process, like HIV-1 egress, is dependent on the function of intact ESCRT complexes (1, 3, 11, 28). EGF-induced EGFR downregulation is therefore a good model system for studying the potential antagonistic effect of HIV-1 Gag expression on the cellular vacuolar sorting pathway. Here we show that the expression of HIV-1 Gag decreases the rate of EGFR downregulation. The increased intracellular retention of EGFR results in prolonged EGFR-mediated signaling, as evidenced by the hyperactivation of ERK/MAP kinase.
A rabbit anti-p24 CA antiserum was used to detect Gag. A human anti-HIV-1 serum was obtained from the NIH AIDS Research and Reference Reagent Program. Anti-TSG101 and anti-phosphoERK monoclonal antibodies and anti-ERK and anti-phosphoEGFR polyclonal antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, Calif.). Anti-EEA1 and anti-CD63 monoclonal antibodies were obtained from BD Transduction Laboratories (San Diego, Calif.). Na-125I was purchased from Perkin-Elmer Life and Analytical Sciences (Boston, Mass.). EGF was obtained from Oncogene Research Products (San Diego, Calif.). 125I-EGF was prepared by using the Iodo-Beads iodination reagent (Pierce Biotechnology, Rockford, Ill.) according to the manufacturer's instructions.
The Rev-independent Gag expression vectors pCMV5 Gag and pGag-GFP have been discussed elsewhere (14, 27). pHXB2ΔBalD25S is a lentiviral vector expressing HIV-1 Gag, Env, Rev, Tat, Vif, Vpu, Vpr, and truncated Nef (35). pEGFP-TSG101 (full length) was a kind gift from Stanley Cohen, Stanford University. EGFR-CFP was a kind gift from Larry Samelson, National Institutes of Helath. Point mutations within the Gag sequence were generated by PCR and were verified by DNA sequencing.
COS-1 cells were maintained in Dulbecco's modified Eagle's medium (DMEM)-10% fetal bovine serum. For transfections, COS-1 cells were seeded to an approximately 50% density and transfected the next day with 6 to 12 μg of plasmid DNA by the use of Lipofectamine 2000 (Gibco BRL Life Technologies, Carlsbad, Calif.). HeLa cells were maintained in DMEM-10% fetal bovine serum and transfected via a calcium phosphate-based protocol.
An assay of the internalization and degradation of 125I-EGF was performed as described by Tebar et al. (26). Briefly, COS-1 cells cultured in 35-mm-diameter dishes were incubated with 125I-EGF (~20 ng/ml) in DMEM-3% bovine serum albumin (BSA) at 37°C for 0 to 15 min. At each time point, the medium was aspirated and the cells were washed two times in cold DMEM-3% BSA followed by a 5-min cold acid wash (0.2 M acetic acid, 0.5 M NaCl, pH 2.8) to remove unbound and surface-bound EGF, respectively. The cells were lysed with 1 N NaOH to measure the internalized 125I-EGF. An aliquot of each lysate was used to determine the total amount of protein. The internalized radioactivity (minus nonspecific binding in the presence of a 100-fold molar excess of cold EGF) per unit of total protein was plotted against time. For measurements of 125I-EGF degradation, cells cultured in 35-mm-diameter dishes were incubated with 20 ng of 125I-EGF/ml for 10 min at 37°C and then washed in cold DMEM-3% BSA. Surface-bound EGF was removed by a 2.5-min acid wash (0.2 M sodium acetate, 0.5 M NaCl, pH 4.5) at 0°C. The cells were then incubated in fresh medium containing a 100-fold molar excess of cold EGF (to prevent reinternalization of the recycled EGF) at 37°C for 5 to 120 min. At each time point, the medium was collected, and intact 125I-EGF was precipitated with trichloroacetic acid (TCA). The cells were solubilized with 1 N NaOH. The degraded fraction was calculated as the ratio of TCA-soluble counts per minute (cpm) at each time point to the total cpm internalized in 10 min.
Transfected COS-1 cells grown on coverslips were starved of serum for 16 h and then processed at 48 h posttransfection as follows (14). The cells were treated with 20 ng of EGF/ml for 10 min at 37°C, washed with acid (pH 4.5) for 2.5 min at 4°C, washed with cold PBS (twice, for 5 min each time), and reincubated in serum-free DMEM for 0 to 3 h at 37°C. After fixation and permeabilization, the cells were incubated with a 1:250 dilution of human anti-HIV-1 serum and/or a 1:500 dilution of an anti-phosphoEGFR, anti-EEA1, or anti-CD63 antibody in phosphate-buffered saline (PBS) for 90 min. After four 5-min washes with PBS, the cells were incubated with a Cy5-conjugated anti-human antibody, an Alexa Fluor 594-conjugated donkey anti-goat antibody and/or an Alexa Fluor 633-conjugated goat anti-mouse antibody (Molecular Probes, Eugene, Oreg.) for 45 min. The cells were then washed four times for 5 min each with PBS and were mounted on microscope slides. For nuclear staining, Hoechst dye was added to the cells during the first PBS wash after the secondary antibody incubation. Enhanced green fluorescent protein (GFP) fluorescence was visualized directly. Laser scanning confocal microscopy was performed with a Zeiss LSM510 confocal microscope equipped with an Axiovert 100 M inverted microscope using a 63×, 1.2-numerical-aperture (NA) water immersion lens for imaging, as previously described (14). Colocalization and absolute fluorescence intensities were measured on a pixel-by-pixel basis with MetaMorph software (Universal Imaging Corp., Downingtown, Pa.).
Transfected COS-1 or HeLa cells grown on 60-mm-diameter dishes were starved of serum for 16 h. The cells were then treated with 20 ng of EGF/ml for 10 min at 37°C, followed by a 2.5-min cold acid wash (pH 4.5) and two 5-min PBS washes. The cells were reincubated in serum-free DMEM for 0 to 190 min. At each time point, cells were lysed with RIPA buffer (150 mM NaCl, 1% NP-40, 0.5% deoxycholic acid, 0.1% sodium dodecyl sulfate [SDS], 50 mM Tris [pH 8.0]) containing leupeptin (10 μg/ml), aprotinin (10 μg/ml), 4-(2-aminoethyl)benzene sulfonyl fluoride (0.25 mg/ml), NaF (1 mM), and Na3VO4 (0.5 mM). The lysates were clarified at 14,000 rpm in an Eppendorf centrifuge for 10 min at 4°C. Western blotting was performed with the indicated antibodies. Proteins were detected by the use of horseradish peroxidase-conjugated secondary antibodies and ECL Western blotting detection reagents according to the manufacturer's instructions. Western blots were quantified with MacBas, version 2.0, software.
The first set of experiments analyzed the effect of HIV-1 Gag expression on EGF internalization and degradation rates. COS-1 cells transfected with plasmids encoding HIV-1 Gag or with an empty vector were incubated with 125I-EGF for various times to allow binding to EGFR and internalization. After an acid wash to remove bound ligands, the cells were lysed and the amounts of intracellular (internalized) radioactivity (in cpm) were determined. The rate of internalization of the ligand was unaffected by the expression of HIV-1 Gag (Fig. (Fig.1A).1A). The rate of degradation of the internalized ligand was then monitored by determining the amount of free 125I (TCA-soluble fraction) that accumulated in the medium at various times after the acid wash. As shown in Fig. Fig.1B,1B, a 30% reduction in the rate of EGF degradation in the Gag-transfected population was detected relative to that for the control. However, a fluorescence-activated cell sorting analysis of the transiently transfected cells revealed that transfection of the HIV-1 Gag construct occurred with an efficiency of only 20 to 30%. Thus, the effect on EGF degradation in the subpopulation of Gag-expressing cells may have been masked by the presence of an excess of untransfected cells within each plate. We therefore designed a single-cell analysis approach to determine the effects of Gag expression within individual cells.
The binding of EGF to EGFR induces tyrosine phosphorylation and activation of the receptor. The movement of tyrosine-phosphorylated, activated EGFR (pEGFR) was monitored in cells expressing either Gag or TSG101 as GFP fusion proteins. Previous studies have shown that HIV-1 Gag-GFP localizes and produces virus-like particles in an identical manner to Gag (14, 16). We observed that TSG101-GFP displays a punctate perinuclear expression pattern similar to that of untagged TSG101. Moreover, TSG101-GFP inhibits both HIV-1 virus-like particle production and endogenous TSG/MVB function when it is overexpressed, in a manner similar to that reported for untagged TSG101 (11). These results indicate that GFP-tagged Gag and TSG101 are functional counterparts of their respective untagged versions. COS-1 cells were transfected with a plasmid encoding GFP, Gag-GFP, or TSG101-GFP, starved of serum, and then allowed to internalize 20 ng of EGF/ml for 10 min. The cells were washed in acid, reincubated in fresh serum-free medium for various times, and then fixed and immunolabeled with an anti-phosphoEGFR antibody (pEGFR) to detect the activated form of the EGF receptor (Fig. (Fig.2A).2A). The movement of pEGFR from the cell surface (chase t = 0) (Fig. (Fig.2A)2A) to internal membranes (chase t = 60 or 120 min after EGF addition) (not shown) was clearly observed. By 180 min, nearly all of the pEGFR signal had disappeared from the cells expressing GFP (Fig. (Fig.2A,2A, left panels), indicating that the activated receptor had been degraded. In contrast, distinct punctae remained in the TSG101-GFP-expressing cells at the 180-min time point (Fig. (Fig.2A,2A, right panels).
In order to perform quantitative analyses on individual cells, we established a scoring system. Fifty to 80 green cells were imaged and grouped into four classes based on the number of pEGFR punctae remaining at the 180-min time point, as follows: class 1, no pEGFR punctae; class 2, 1 to 5 punctae; class 3, 6 to 10 punctae; and class 4, >10 punctae (Fig. (Fig.2B).2B). Almost 70% of the GFP-expressing cells (negative control) were assigned to class 1. In contrast, only 20% of the cells expressing TSG101-GFP were assigned to class 1; instead, most of these cells were assigned to classes 2 and 3 (Fig. (Fig.2A).2A). This result confirmed previous observations that the overexpression of TSG101 induces defects in endosomal sorting (11). Cells expressing wild-type (wt) Gag-GFP exhibited a class distribution similar to that of TSG101-GFP-expressing cells. Only 30% of the cells were in class 1, and nearly 60% were in class 2 (Fig. 3A and B). Mutation of the PTAP motif within Gag reverted this distribution such that the pEGFR staining pattern resembled that of control cells expressing GFP (Fig. (Fig.3A).3A). Interestingly, the expression of G2A Gag-GFP, a nonmyristoylated Gag mutant, also decreased the number of cells in class 1. Again, the expression of a PTAP mutant form of G2A Gag-GFP reverted the distribution to a pattern similar to that in control GFP-expressing cells (Fig. (Fig.3A).3A). These observations indicate that at the single-cell level, Gag expression attenuates EGFR degradation in a late-domain-dependent manner and that membrane binding is not required for TSG101/ESCRT sequestration by Gag.
In order to determine the correlation between Gag expression levels and the amount of pEGFR retained at the single-cell level, we measured the integrated green and red fluorescence intensities in individual Gag-GFP-expressing cells at the 180-min time point. As shown in Fig. Fig.3C,3C, a positive correlation between Gag expression and the amount of pEGFR retained at 180 min was clearly observed over a seven- to eightfold range of Gag levels.
The subcellular localization of the pEGFR punctae observed at 180 min was determined for TSG101-GFP- and Gag-GFP-expressing cells. Cells were treated as described above for Fig. Fig.22 and then were fixed and costained for pEGFR and either EEA1, a marker for early endosomes, or CD63, an MVB-late endosome marker protein. pEGFR colocalized partially with EEA1 and mostly with CD63 in both TSG101- and Gag-expressing cells (Fig. (Fig.4),4), implying that the receptor was primarily present in late endosomal compartments (MVBs) by this time. Notably, there was no marked clustering of EEA1- or CD63-positive compartments in Gag-expressing cells. This was in contrast to the prominent clustering of endosomal compartments observed when cellular levels of Hrs and TSG101 are perturbed (2, 11).
The activation of the MAP kinases ERK1 and ERK2 is a key signaling event that is mediated by the activated EGFR. The subsequent decay of MAP kinase signaling emanating from the activated EGFR is a clear indication of efficient MVB biogenesis. The retention of pEGFR in intracellular compartments predicts that Gag-expressing cells should exhibit enhanced and/or prolonged EGFR signaling. To test this hypothesis, we stimulated serum-starved COS-1 cells that were cotransfected with either Gag, TSG101, or an empty vector and EGFR-CFP with EGF for 10 min, washed them, and reincubated them at 37°C in serum-free medium for various times. The cells were lysed and the amounts of activated ERK1/2 were determined by Western blotting. Both Gag- and TSG101-expressing cells exhibited higher ERK activation peaks than the control and, subsequently, prolonged signaling (Fig. 5A and B). This result was consistent with the presence of pEGFR in endosomal compartments in these cells. A similar result was obtained with HeLa cells (Fig. 5C and D). HeLa cells expressing wt Gag exhibited higher pERK levels than control cells transfected with an empty vector or cells expressing the PTAP Gag mutant (LTAL Gag). We observed, however, that 200 min after the addition of EGF, pERK levels in the Gag-transfected population were indistinguishable from those in the empty vector-transfected population. In contrast, pERK levels in the TSG101-transfected population were still significantly higher than those in control cells at this time. These results indicate that HIV-1 Gag expression slows down EGF-induced EGFR downregulation, resulting in a transient hyperactivation of MAP kinase signaling.
In all of the experiments described above, Gag expression was driven by a cytomegalovirus (CMV) promoter and occurred in the absence of other viral proteins that are normally present in the cell during HIV-1 infection. In order to determine whether the efficiency of Gag expression from a CMV promoter was comparable to that from the HIV-1 long terminal repeat (LTR), we performed an SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting analysis of lysates obtained from cells transfected with molar equivalent amounts of either pCMV5-Gag or a pHXB2-derived plasmid DNA. pHXB2ΔBalD25S is a lentiviral vector that encodes HIV-1 Gag, Env, Tat, Rev, Vif, Vpu, Vpr, and truncated Nef and that uses the HIV-1 LTR as a promoter to drive the expression of these viral proteins. As shown in Fig. Fig.3D,3D, equivalent amounts of HIV-1 Gag protein were expressed from the two constructs.
In order to test whether the HIV-1 Gag-mediated inhibition of EGFR degradation also occurs in the presence of other viral proteins, we transfected COS-1 cells with pHXB2ΔBalD25S and repeated the experimental procedures described above for Fig. Fig.2.2. Transfected cells were identified by staining with a human anti-HIV antiserum and Cy5-conjugated anti-human antibodies. The transfected cells were grouped into classes based on the number of pEGFR punctae remaining after a 180-min chase. As shown in Fig. Fig.3E,3E, only 35% of the pHXB2-expressing cells were assigned to class 1. The general distribution of the pHXB2-expressing cells among the four classes was reminiscent of the distribution of wt Gag-GFP- and TSG101-GFP-expressing cells (Fig. (Fig.3B).3B). This observation indicates that in a system that resembles HIV-1-infected cells, EGFR degradation is attenuated.
There is abundant evidence in the literature that a perturbation of normal cellular levels of the vacuolar protein sorting (VPS) machinery results in defective endosomal sorting of receptors (3, 11, 28). Likewise, a perturbation of the levels of TSG101 and other VPS components clearly results in inefficient HIV-1 particle release (9, 11, 19). What happens when both of these processes need to take place simultaneously in the same cell? Do both processes utilize and/or compete for the same cellular pool of VPS machinery? If so, is any one of the two processes compromised in efficiency? In this study, we addressed these questions directly by monitoring ligand-induced downregulation of EGFR in cells expressing HIV-1 Gag alone or along with other viral proteins. We demonstrated that the rate of receptor downregulation is decreased when HIV-1 Gag is present in the cell.
COS-1 cells were used in this study as a model system for the demonstration of proof of principle. Although COS-1 cells are not authentic targets of HIV-1 infection, they exhibit three important features that were relevant for this study. First, these cells are permissive for HIV-1 assembly and particle production. Second, COS-1 cells express abundant levels of endogenous EGFR. Third, the envelope-mediated cell death and cell fusion that are induced by HIV-1 infection of target cells do not occur in Gag-expressing COS-1 cells. We were therefore able to demonstrate a correlation between the amount of HIV-1 Gag expressed and the amount of endogenous EGFR remaining after EGF stimulation on a single-cell level. As depicted in Fig. Fig.3C,3C, there was a positive correlation between Gag and EGFR fluorescence intensities over a seven- to eightfold range of Gag expression levels.
Although it is now known that approximately 5,000 Gag molecules are packaged per immature HIV-1 virion (4), the amount of cellular Gag that is expressed on a per-cell basis in an HIV-1-infected cell has not been reported. It is likely that Gag expression levels vary considerably during the course of the infection. However, two lines of evidence suggest that our findings with Gag-transfected COS-1 cells may be applicable to the situation that occurs in HIV-1-infected cells. First, equivalent amounts of Gag protein were produced when expression was driven from a heterologous promoter (CMV) or from the HIV-1 LTR (pHXB2ΔBalD25S) (Fig. (Fig.3D).3D). Second, the pattern of pEGFR staining in cells transfected with pHXB2ΔBalD25S was similar to that observed for cells transfected with pCMV5 Gag-GFP (Fig. (Fig.3E).3E). Thus, we hypothesize that it is possible for the HIV-1 LTR to produce the levels of Gag expression that are needed to attenuate EGFR downregulation.
The scenario during an actual HIV-1 infection, however, may be much more complicated. The presence of other viral proteins may serve to either further enhance or mask Gag-mediated effects on receptor trafficking and signaling (23). It would therefore be interesting to further investigate whether the ESCRT-dependent endosomal sorting of receptors is perturbed in HIV-1-infected target cells, e.g., T cells. A potential candidate receptor in the T cell is CXCR4, which has been shown to be dependent on ESCRT components for downregulation (18).
Interestingly, the effect of HIV-1 Gag overexpression on receptor downregulation was not as strong as one would expect if the virus completely usurped the cellular machinery and redirected it towards sites of viral egress. This is emphasized by the lack of clustering of endosomal compartments that are normally observed when the cellular levels of VPS components are perturbed. There are three potential explanations for this observation. (i) HIV-1 Gag induces an increased expression of ESCRT proteins to accommodate its needs. This possibility is unlikely, since no differences in endogenous TSG101 protein levels were observed for HIV-1 Gag-transfected versus untransfected cells (not shown). (ii) The cellular TSG101/ESCRT pool is too large to be completely depleted by the overexpression of Gag/PTAP-containing proteins. (iii) The stoichiometry of the Gag-ESCRT interaction is not 1:1. The number of ESCRT complexes that are recruited per Gag assembly unit in vivo is not known. The multimerization of Gag into assembly complexes, which consist of ~5,000 Gag molecules, could physically restrict its ability to sequester large, 350-kDa ESCRT complexes.
An obvious consequence of altered receptor trafficking is altered intracellular signaling. As depicted in Fig. Fig.5,5, cells expressing HIV-1 Gag exhibited increased EGFR-mediated signaling, as evidenced by hyperactivation of the ERK/MAP kinase pathway. This finding is consistent with the presence of increased amounts of pEGFR in endosomal compartments 180 min after the addition of EGF. The ability of endosomal EGFR to induce intracellular signaling has been documented (31). Changes in intracellular signaling are a well-established characteristic of HIV-1 pathogenesis. The binding of HIV-1 gp120 to CD4 and a chemokine receptor (CXCR4 or CCR5) during virus entry triggers a plethora of signaling pathways (21). The upregulation of cellular signaling regulates the ability of the virus to replicate efficiently within the target cell (21). After provirus integration into the host genome, either a productive or a latent infection can be established. The switch from a latent to productive infection has been shown to be triggered and maintained by extracellular mitogen/agonist-induced MAP kinase activation through the Ras/Raf/Mek signaling pathway (32). MAP kinase stimulates AP-1, which, along with NF-κB, transactivates the HIV-1 LTR (32). The activation of MAP kinase also increases the infectivity of the HIV-1 virions that are subsequently produced (33). The observations presented in this study strongly suggest that Gag-mediated budding and egress during the late stages of the viral life cycle may indirectly contribute to enhanced HIV-1 replication and/or infectivity by slowing down mitogen-induced cell surface receptor downregulation.
We thank Wolf Lindwasser and Mira Perlman for fruitful discussions and critical analysis of the data and Raisa Louft-Nisenbaum for expert technical assistance.
This work was supported by NIH grant CA 72309.