Antibody-Induced Internalization of HIV-1 Env Proteins Limits Surface Expression of the Closed Conformation of Env

HIV-1 has evolved to acquire several strategies to limit the exposure of its envelope glycoproteins (Env) on the surface of infected cells. In this study, we show that antibody-induced Env internalization is conformation specific and reduces the susceptibility of infected cells to antibody-dependent cellular cytotoxicity (ADCC). Thus, a better understanding of this mechanism might help develop antibodies with improved capacities to mediate ADCC. ABSTRACT To minimize immune responses against infected cells, HIV-1 limits the surface expression of its envelope glycoprotein (Env). Here, we demonstrate that this mechanism is specific for the Env conformation and affects the efficiency of antibody-dependent cellular cytotoxicity (ADCC). Using flow cytometry and confocal microscopy, we show that broadly neutralizing antibodies (bNAbs) targeting the “closed” conformation of Env induce its internalization from the surface. In contrast, non-neutralizing antibodies (nNAbs) are displayed on the cell surface for prolonged period of times. The bNAb-induced Env internalization can be decreased by blocking dynamin function, which translates into higher susceptibilities of infected cells to ADCC. Our results suggest that antibody-mediated Env internalization is a mechanism used by HIV-1 to evade immune responses against the “closed” conformation of Env expressed on HIV-1-infected cells. IMPORTANCE HIV-1 has evolved to acquire several strategies to limit the exposure of its envelope glycoproteins (Env) on the surface of infected cells. In this study, we show that antibody-induced Env internalization is conformation specific and reduces the susceptibility of infected cells to antibody-dependent cellular cytotoxicity (ADCC). Thus, a better understanding of this mechanism might help develop antibodies with improved capacities to mediate ADCC.

BST-2 (9,(12)(13)(14) through its Vpu accessory protein. In addition to these evasion mechanisms, the virus minimizes Env accumulation at the cell surface by internalizing Env (15). HIV-1 Env has a long C-terminal cytoplasmic domain, which is involved in the regulation of Env trafficking and contains several endocytic signals, including membrane-proximal tyrosine-based sorting motifs and dileucine motifs (16)(17)(18). Mutations of these motifs have been shown to result in increased cell surface expression of Env, which correlates with increased ADCC responses (15,19,20). Internalizing envelope glycoproteins from the cell surface to avoid humoral immune responses does not appear to be restricted to HIV-1; equine herpesvirus 1 (21) and pseudorabies virus (22) also downregulate their envelope glycoproteins in order to avoid antibody-dependent complement-mediated lysis. Thus, internalization of viral envelope glycoproteins from the surface of infected host cells appears to be a mechanism employed by many viruses to minimize recognition by the immune system (23)(24)(25)(26).
The ability of antigen-Ab complexes to remain on the surfaces of infected cells might be another factor modulating humoral responses against different viruses. It has been reported previously that the interaction of some Abs with Env glycoproteins expressed on the surfaces of respiratory syncytial virus (RSV) (27)and feline coronavirus (28)-infected cells accelerates their internalization and decreases their exposure to the immune system. We evaluated here the fate of HIV-1 Env expression at the surfaces of cells upon antibody binding and made the surprising observation that the binding of broadly neutralizing antibodies (bNAbs) but not that of nonneutralizing antibodies (nNAbs) induced Env-Ab complex internalization and reduced ADCC responses mediated by bNAbs. These data indicate that HIV-1 specifically evades immune responses against the closed conformation of Env.
Primary CD4 ϩ T cells were infected with a previously reported wild-type (wt) HIV-1 NL4.3 strain encoding a gfp reporter gene and an R5-tropic (ADA) envelope (9). Infected cells were identified based on GFP expression (GFP ϩ ). Cells were first incubated with monoclonal Abs (MAbs) at 4°C for 30 min, washed to remove excess antibody, and then incubated at 37°C for internalization to occur. After incubation at 37°C for different time intervals, Env-Ab complexes remaining at the cell surface were visualized with a fluorescent secondary anti-human antibody by flow cytometry. Under these experimental settings we observed that the binding of bNAbs (PG126, PGT151, PG9, and 2G12) significantly declined over time indicating Env internalization (Fig. 1). In contrast, binding of the nNAbs (17b, N12-i2, A32, and N5-i5) was fairly steady. The reduction in the cell surface levels of Env was ϳ60% after 6 h of incubation with PGT126, PG9, PGT151, and 2G12 at 37°C. Conversely, binding to surface Env was only reduced by ϳ20% upon incubation for the same time period at 37°C with N12-i2, N5-i5, A32, and 17b ( Fig. 1B and C).
To evaluate whether this phenotype was restricted to the infectious molecular clone used, we infected primary CD4 ϩ T cells with the transmitted/founder (T/F) virus CH58 (CH58 T/F) and performed the same experiments described above with the exception that infected cells were identified by intracellular Gag staining. Similar to what we obtained with NL4.3 ADA-infected cells (Fig. 1), bNAbs reduced the cell surface levels of CH58T/F Env by 80% after 6 h of incubation at 37°C, whereas incubation with nNAbs only reduced Env by ϳ20% during the same time period (Fig. 2).
It is now well established that the conformation of Env at the cell surface influences antibody binding and ADCC responses (38)(39)(40). Nonneutralizing antibodies and bNAbs target different Env conformations, with most bNAbs preferentially recognizing Env in its "closed" conformation (state 1) (36), whereas the nNAbs used here target epitopes that are only exposed upon Env interaction with CD4 (9,33,37,42). To verify whether the differences in cell surface reduction of Env between bNAbs and nNAbs was due to their differential recognition of infected cells, we infected primary CD4 ϩ T cells with NL4.3 ADA defective for Nef and Vpu accessory proteins (Nef-Vpu-) that fails to downregulate CD4. This virus was used to present Env in its CD4-bound "open" conformation at the surfaces of infected cells and enhance the recognition of infected cells by the nNAbs used in this study (8,9) (Fig. 3). As expected, deletion of Vpu enhances the overall levels of Env at the cell surface as measured by 2G12, known to bind to both "open" and "closed" conformations of Env (7,9,43) (Fig. 3A). This phenomenon has been well established and is due to the accumulation of BST-2trapped viral particles at the cell surface, which results in Env accumulation (9,12,14). In agreement with their CD4-induced nature, the recognition of infected cells by the nNAbs tested (A32, 17b, N12-i2, and N5-i5) was dramatically enhanced by deletion of Nef and Vpu. Deletion of these accessory genes impair the ability of HIV-1 to downregulate CD4 from the cell surface, thus resulting in increased Env-CD4 interactions and the subsequent exposure of CD4i epitopes (Fig. 3B) (9). Despite their improved capacity to recognize Env at the surfaces of Nef-Vpu-infected cells, nNAbs only reduced Env from the cell surface by ϳ20% after 6 h, while bNAbs reduced Env levels by ϳ60% within the same time period (Fig. 4). Altogether, these results indicate that the conformation specificity of antibody-mediated internalization of Env is not affected by changes in surface expression and that Nef and Vpu accessory proteins are not involved in the bNAb-mediated reduction of Env from the cell surface.
Antibody-induced internalization of cell surface-expressed HIV-1 Env. To confirm that Env-bNAb complexes were internalized upon incubation at 37°C, we performed confocal microscopy studies in parallel to flow cytometry. Since primary CD4 ϩ Error bars indicate means Ϯ the SEM. Statistical significance was tested using an unpaired t test or a Mann-Whitney U test based on statistical normality (****, P Ͻ 0.0001).
T cells are small and poorly adherent, we used HEK 293T cells for these studies. 293T cells were transfected with a primary tier 2 Env (JRFL) and incubated 48 h posttransfection with Alexa Fluor-conjugated bNAbs (2G12 and PGT151) or nNAbs (17b and A32) for various time points at 37°C before fixing and analysis by flow cytometry and confocal imaging. Since A32 and 17b do not recognize the unliganded Env JRFL , Env was cotransfected, together with the CD4 receptor, to trigger the exposure of the CD4i epitopes recognized by these nNAbs (9,43). Flow cytometry confirmed that the bNAbs 2G12 and PGT151 were internalized much faster than nNAbs 17b and A32, thus phenocopying the results from T cells (Fig. 5A). Parallel confocal imaging demonstrated that, upon the addition of bNAbs PGT151 and 2G12, Env moved from the plasma membrane into intracellular compartments after 60 min of incubation at 37°C. In contrast, the binding of A32 and 17b did not change the cell surface localization of Env ( Fig. 5B and C). These results were consistent with those obtained using infected primary CD4 ϩ T cells ( Fig. 1, 2, and 4). Interestingly, the quantification of internalization measured by microscopy correlated well with the quantification of cell surface Env measured by flow cytometry (Fig. 5D), suggesting that both assays measure Env internalization. Further supporting this possibility, we observed that bNAb-induced Env reduction from the cell surface was blocked if cells were incubated at 4°C, a temperature known to block molecular trafficking events, for the entirety of the internalization assay (Fig. 6).
It has been previously suggested that Ab-mediated cross-linking facilitates internalization of RSV fusion proteins. In this study, RSV Env endocytosis was significantly reduced when Fab fragments were used instead of their full MAb counterpart (27). To verify whether this was the case for HIV-1 Env, we performed side-by-side comparisons on the ability of full-length Abs versus their Fab fragments to reduce Env levels at the cell surface upon incubation at 37°C, as measured by flow cytometry. A32 and 17b Fab fragments behaved similar to the full-length Abs. However, PG9, PGT151 and 2G12 Fab fragments were significantly less efficient than their IgG counterparts at inducing internalization of Env (Fig. 7). These data suggest that the cross-linking of Env trimers at the cell surface stimulates bNAb-mediated internalization.
A dynamin inhibitor attenuates bNAb-induced Env internalization and increases the susceptibility of infected cells to ADCC. To further confirm that internalization was involved in the observed bNAb-induced reduction of Env from the cell surface, we decided to block the function of dynamin, a GTPase implicated in endocytic membrane fission events (44). Primary CD4 ϩ T cells were infected with NL4.3 ADA green fluorescent protein (GFP) or CH58 T/F and dynamin was inhibited with the dynamin inhibitory peptide (DIP). This peptide blocks the binding of dynamin to amphiphysin and has been shown to reduce endocytic events (27,45,46). The addition of DIP significantly reduced bNAb-induced Env internalization in infected cells as measured by flow cytometry (Fig. 8A and B) and confocal microscopy using Env JRFLtransfected 293T cells ( Fig. 8C and D). No statistical differences in the internalization rates of Env bound by nNAbs, with or without added DIP, were observed. Thus, this reiterates the observation that the binding of these Abs does not promote Env internalization. Correlation of the quantification of Env internalization by confocal microscopy with internalization by flow cytometry using a Pearson correlation test. Error bars indicate means Ϯ the SEM. Statistical significance was tested using an unpaired t test or a Mann-Whitney U test based on statistical normality (***, P Ͻ 0.001; ****, P Ͻ 0.0001).
Since DIP treatment resulted in an accumulation of bNAb/Env complexes at the surface of HIV-1-infected cells, we evaluated whether it had any impact on the ability of bNAbs to mediate ADCC. Primary CD4 ϩ T cells were infected with the NL4.3 ADA GFP wt virus, and their susceptibility to ADCC was measured using a previously described fluorescence-activated cell sorting (FACS)-based assay (9,47). Interestingly, we observed a significant increase in ADCC responses mediated by PGT126, PGT151, PG9, and 2G12 upon DIP treatment ( Fig. 9) but not of direct killing (i.e., in the absence of Abs [not shown]). Thus, this suggests that under normal conditions bNAb-mediated Env internalization reduces their capacity to trigger ADCC.

DISCUSSION
Here we report that antibody-bound Env proteins are internalized from the surface of HIV-1-infected cells in a conformation-specific manner. bNAbs that preferentially recognize "closed" Env conformations trigger rapid Env internalization. When this process was decreased using a dynamin inhibitor, the susceptibility of infected cells to ADCC mediated by bNAbs was enhanced, suggesting that bNAb-triggered Env internalization impairs their ability to mediate robust ADCC responses. In contrast, nNAb specific for "open" Env conformation remain exposed on the surfaces of infected cells for prolonged periods of time. As such, differential internalization of Env-antibody complexes is likely an immune evasion mechanism that HIV-1 evolved to limit the surface exposure of Env in "closed" conformations while distracting the immune system Error bars indicate means Ϯ the SEM. Statistical significance was tested using an unpaired t test or a Mann-Whitney U test based on statistical normality (*, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.001; ****, P Ͻ 0.0001).

Env Internalization and ADCC
Journal of Virology with the display of Env proteins in "open" conformations that ultimately result in nonneutralizing immune responses. Specifically, the native Env trimer mainly exists in an untriggered "closed" conformation (state 1). The interaction with the CD4 receptor lowers the energy barrier to reach the "open" states 2/3 (34,36). Natural HIV-1 infection elicits mainly nNAbs, which poorly recognize the "closed" Env and are only able to recognize highly conserved epitopes exposed upon the "opening" of Env (8,48). Despite poor neutralizing activity, these nNAbs have been shown to exert a constant selection pressure and alter the course of HIV-1-infection in vivo (49)(50)(51) and can mediate potent ADCC activity against cells presenting Env in its "open" CD4-bound conformation (7,9,38,52). It is interesting to note that the nNAbs used in this study can form a stable complex with Env on the cell surface for a prolonged amount of time. Conversely, the bNAbs tested induced faster Env internalization rates. Since these bNAbs recognize the "closed" conformation of Env, these results suggest that Env sampling its State 1 conformation might be located in discrete membrane microdomains that could be prone to antibodymediated internalization. Additional studies are required to explore this interesting possibility. Of note, bNAb-Env complexes do not appear to follow the degradative pathway since the intracellular compartments where they accumulate over time are negative for the lysosomal marker Lamp1 (Fig. 10). Rather, Env accumulated in endosomes positive for early endosome marker EEA1 (Fig. 10). It is intriguing to speculate that this endocytic pathway could be related to the observed role for recycling endosomes in the incorporation of Env into budding virions (53)(54)(55).
Previous reports have shown that humoral immune responses, such as antibodydependent complement-mediated lysis, are decreased by internalization of surfaceexpressed viral glycoproteins upon the binding of antibodies (22,23). We show here that this also applies to ADCC. The surprising differences observed between bNAbs and nNAbs studied here suggest that different Env populations sampling different conformations coexist at the surfaces of infected cells. Envs sampling the "closed" conformation could potentially facilitate the elicitation of bNAbs. Therefore, it is tempting to speculate that the virus minimizes the exposure of Env sampling the "closed" conformation while tolerating the exposure of limited amounts of Env in the "open" conformation. The "open" Env could act as a decoy, since it is well established that this conformation fails to elicit bNAbs. The nNAbs that are elicited instead fail to neutralize viral particles or mediate ADCC against wild type-infected cells (9-11, 43, 56, 57).
It is becoming increasingly clear that several factors contribute to ADCC responses Env Internalization and ADCC Journal of Virology against HIV-1-infected cells, i.e., Env conformation (38), CD4 (8,9,11) and BST-2 (12)(13)(14) downregulation, gp120 shedding (37,59), and the stability of Env-Ab complexes at the cell surface (60), which are driven by the affinity of Abs for Env (61), Env internalization (15), cell surface expression of stress ligands (56,62), and now antibody-induced Env internalization. Additional work will be required to tease apart the differential contribution of each of these factors in ADCC. Further dissecting the mechanisms underlying antibody-induced Env internalization might help in the development of new generations of bNAbs that are able to efficiently eliminate HIV-1-infected cells.

MATERIALS AND METHODS
Plasmids and cell lines. 293T human embryonic kidney cells (American Type Culture Collection) were maintained at 37°C under 5% CO 2 in Dulbecco modified Eagle medium (Invitrogen) containing 5% fetal bovine serum (Sigma) and 100 g/ml of penicillin-streptomycin (Mediatech). The E168K mutation was introduced into the previously described pcDNA3.1 expressing codon-optimized HIV-1 JRFL enve- Autologous peripheral blood mononuclear cells were used as effector cells in a FACS-based ADCC assay (9,47). The percentages of ADCC-mediated killing obtained with A32, 17b, PGT126, PGT151, PG9, and 2G12 MAbs in the presence of 50 M DIP or the vehicle control are shown. Error bars indicate means Ϯ the SEM. Statistical significance was tested using an unpaired t test or a Mann-Whitney U test based on statistical normality (*, P Ͻ 0.05; **, P Ͻ 0.01). lope glycoproteins (34) using a QuikChange II XL site-directed mutagenesis protocol (Stratagene). Other plasmids used to transfect 293T cells include pcDNA3.1 human CD4 expressor and pIRES-GFP vector (9,43). The pCAGGS codon-optimized JR-FL gp160 plasmid was kindly provided by Joseph Sodroski (63). The pECFP lamin B receptor plasmid was kindly provided by Melissa Rolls and Tom Rapoport (64).
Isolation of primary cells, viral production, and infections. CD4 ϩ T lymphocytes were purified from resting PBMCs by negative selection and activated as previously described (8,9). Briefly, PBMC were obtained by leukapheresis from at least five different healthy HIV-uninfected individuals. CD4 ϩ T lymphocytes were purified by negative selection using immunomagnetic beads as per the manufacturer's instructions (StemCell Technologies). CD4 ϩ T lymphocytes were activated with phytohemagglutinin-L (10 g/ml) for 48 h and then maintained in RPMI 1640 complete medium supplemented with rIL-2 (100 U/ml).
To ensure similar levels of infection between viruses, vesicular stomatitis virus G-pseudotyped viruses were produced and titrated as described previously (8). Viruses were used to infect activated primary CD4 T cells from healthy HIV-1-negative donors by spin infection at 800 ϫ g for 1 h in 96-well plates at 25°C.
Antibodies. Anti-HIV-1 gp120 MAbs recognizing CD4-induced epitopes (A32 and 17b; obtained from the NIH AIDS Reagent Program), the outer domain (2G12; obtained from the NIH AIDS Reagent Program), and the gp120-gp41 interface (PGT151; obtained from International AIDS Vaccine Initiative [IAVI]) were conjugated with Alexa Fluor 488 or Alexa Fluor 594 (Thermo Fisher Scientific) according to the manufacturer's protocol and used for confocal microscopy analyses of cell surface staining of 293Ttransfected cells. In addition, the following anti-Env MAbs were also used for cell surface staining: PGT126, PGT121 (IAVI), PG9 (Polymun), N5-i5, and N12-i2. Fab fragments were generated by papain digest of the corresponding antibody. Briefly, purified IgGs were incubated at 37°C overnight with immobilized papain (G Biosciences) and then filtered to remove the papain. Fabs were separated from undigested IgG and Fc by passage over a HiTrap protein A column (GE Healthcare) equilibrated with phosphate-buffered saline (PBS; pH 7.2). Flowthrough fractions were concentrated and further purified by gel filtration chromatography on a Superdex 200 gel filtration column (GE Healthcare) equilibrated in 25 mM Tris-HCl (pH 8.5) and 150 mM sodium chloride. Fab fractions with elution times roughly correlating to 50 kDa were combined and concentrated prior to use. Goat anti-human Alexa Fluor 647conjugated secondary Ab (Thermo Fisher Scientific) was used to determine overall antibody binding and AquaVivid (Thermo Fisher Scientific) as a viability dye. Mouse monoclonal EEA1 (14/EEA1) and LAMP1 (H4A3) antibodies were obtained from BD Transduction Laboratories. Alexa Fluor 568-conjugated goat anti-mouse was obtained from Thermo Fisher Scientific.
Flow cytometry analysis of cell surface staining. Cell surface staining was performed as previously described (8,9). Cells infected with HIV-1 primary isolates were identified by intracellular staining of HIV-1 Gag using the Cytofix/Cytoperm fixation/permeabilization kit (BD Biosciences) and the PE-conjugated anti-Gag MAb, clone KC57 (Beckman Coulter). The percentage of infected cells (Gag ϩ or GFP ϩ cells) was determined by gating the living cell population based on viability dye staining with AquaVivid (Thermo Fisher Scientific). For the cell surface staining of transfected 293T cells, 3 ϫ 10 5 293T cells were transfected by the calcium phosphate method with the Env-expressing plasmids, along with a pIRES-GFP vector, at a ratio of 2 g of pcDNA3.1 or JRFL Env to 0.5 g of GFP. At 16 h posttransfection, cells were washed with fresh medium, and cell surface staining was carried out 24 h later. Samples were analyzed on an LSRII cytometer (BD Biosciences), and data analysis was performed using FlowJo vX.0.7 (Tree Star).
Antibody-induced internalization assay. 48 h postinfection or posttransfection, HIV-1-infected primary CD4 ϩ T cells or Env-transfected 293T cells, respectively, were incubated with 5 g/ml anti-Env antibodies for 30 min at 4°C. Excess antibodies were washed three times with cold PBS. This was followed by incubation at 37°C in complete media to start the internalization process. After different time points, cells were fixed with 2% paraformaldehyde. For flow cytometry analyses, to visualize remaining antigenantibody complexes on the cell surface, cells were stained with a goat anti-human conjugated with Alexa Fluor-647 secondary Ab (Thermo Fisher Scientific). As a control, cells were fixed after 4°C incubation (time point, 0 min). Dead cells were excluded by staining the cells with live/dead fixable AquaVivid stain (Thermo Fisher Scientific). The reduction in surface expression for a given time point was normalized by using the following equation: [(mean fluorescence intensity at x min)/(mean fluorescence intensity at 0 min)] ϫ 100.
For confocal microscopy analyses, 293T cells were plated in 10-mm MatTek dishes with #0 coverslip bottoms. 293T cells were transfected with 1 g of pCAGGS JR-FL Env (codon-optimized) plasmid and 100 ng of lamin B receptor-CFP plasmid (to locate the nuclear envelope) with or without 1 g of human CD4 plasmid using Fugene 6 reagent according to the manufacturer's instructions. At 48 h posttransfection, the cells were incubated with prelabeled antibodies diluted 1:250 in fresh medium for the indicated times, washed once with PBS plus 0.5% bovine serum albumin (BSA), and fixed with PBS plus 4% paraformaldehyde for 10 min. Cells were then placed in PBS plus 0.01% sodium azide prior to imaging. For the 0-min control, cells were fixed prior to antibody staining for 1 h in PBS plus 0.5% BSA. For endosomal staining, cells were first incubated with prelabeled 2G12 antibody for 180 min, followed by fixation, permeabilization (5 min with 0.1% Triton X-100), and staining for endogenous EEA1 or LAMP1.
Internalization inhibition assay. To inhibit antibody-induced internalization, a dynamin inhibitor was used before and during the internalization assay. Myristoylated dynamin inhibitory peptide (DIP; Tocris Bioscience) was diluted in water prior to dilution in serum-free cell culture medium. After 1 h of pretreatment of infected or transfected cells with 50 and 40 M DIP, respectively, at 37°C, fresh inhibitor was added to the cells, together with antibodies, to allow for their attachment and then during incubations at 37°C to induce their internalization. As a control, 20 g/ml of biotinylated transferrin was Env Internalization and ADCC Journal of Virology used to confirm the effectiveness of DIP, and this was visualized with streptavidin-Alexa Fluor 647 (data not shown). Antibody-dependent cellular cytotoxicity assay. Measurement of ADCC using the FACS-based assay was performed at 48 h postinfection as previously described (9,47). Briefly, infected primary CD4 ϩ T cells were stained with AquaVivid viability dye and cell proliferation dye (eFluor670; eBioscience) and used as target cells. Next, the target cells were treated with and without 50 M DIP for 1 h at 37°C. Autologous PBMC effector cells, stained with another cellular marker (cell proliferation dye eFluor450; eBioscience), were added at an effector/target ratio of 10:1 in 96-well V-bottom plates (Corning). DIP was washed out before incubating the infected cells with effector cells to guarantee that the drug affected only the target cells and not the effector cells. Then, 5 g/ml concentrations of MAbs were added to appropriate wells, and cells were incubated for 15 min at room temperature. Subsequently the plates were centrifuged for 1 min at 300 ϫ g and incubated at 37°C and 5% CO 2 for 5 h before being fixed in a 2% PBS-formaldehyde solution. Samples were acquired on an LSRII cytometer (BD Biosciences), and data analysis was performed using FlowJo vX.0.7 (Tree Star). The percentage of cytotoxicity was calculated using the following formula: (% of GFP ϩ cells in targets plus effectors) Ϫ (% of GFP ϩ cells in targets plus effectors plus MAbs)/(% of GFP ϩ cells in targets) by gating on infected live target cells.
Confocal microscopy. Confocal imaging was performed using a Nikon Eclipse TE2000-E microscope equipped with 444-, 488-, and 561-nm lasers, a Yokogawa CSU 10 spinning disc confocal laser scanning unit, and an Andor Zyla sCMOS camera. Images were analyzed manually using ImageJ (65). Briefly, regions of interest were drawn around entire cells and cytoplasmic regions. Total fluorescence for regions of interest were calculated as the area ϫ (mean -minimum). Surface fluorescence was calculated as the total fluorescence minus the cytoplasmic fluorescence. Values are represented as surface/total and cytoplasmic/total ratios, normalized to the 0-min time point. The average intensity and area were measured and used to calculate the total fluorescence and the cytoplasmic fluorescence. The minimum intensity was subtracted from the mean intensity to correct for cytoplasmic background fluorescence. Colocalization between Env and endosomal markers was quantified with the Pearson correlation function using the JACoP plugin (66) for ImageJ.
Statistical analyses. Statistics were analyzed using Prism v6.01 (GraphPad, San Diego, CA). Every data set was tested for statistical normality, and this information was used to apply the appropriate (parametric or nonparametric) statistical test. P values of Ͻ0.05 were considered significant; significance values are indicated by asterisks in the figures (*, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.001; ****, P Ͻ 0.0001).