Involvement of Clathrin-Mediated Endocytosis in Human Immunodeficiency Virus Type 1 Entry

Plasmids.

Plasmids Eps15GFP (originally named DIIIΔ2) and dnEps15GFP (originally named EΔ95/295) have been described previously (3, 4) and were provided by A. Dautry-Varsat. The caveolin-1 expression plasmid cavGFP (originally named caveolin1-GFP) has been described (48). Plasmid dncavGFP (originally named pINDEGFPVIP) has also been described (32). Mutant dynamin^K44A was described by van der Bliek et al. (60). Dyn^wtGFP (originally named Dyn2-GFP) was described previously (8). CavGFP, dncavGFP, dyn^wtGFP, and dyn^K44AGFP were provided by A. Helenius. The expression vector for the fusion protein of β-lactamase and Vpr (pMM310) (59) was a gift from N. Landau. The expression plasmid for the envelope glycoprotein of VSV (VSV-G) under control of the cytomegalovirus promoter (pM3) was provided by D. von Laer. An HIV-1 proviral clone with a deletion of the env open reading frame (pNL4-3env⁻GFP) was provided by D. Gabuzda (25).

Cells and tissue culture.

HeLa cells, CD4-positive HeLaP4 cells (10), and 293T cells were grown in Dulbecco modified Eagle medium (Gibco) supplemented with 10% heat-inactivated fetal calf serum (FCS), penicillin (100 U/ml), streptomycin (100 μg/ml), and 4 mM glutamine. MT-4 cells were maintained in RPMI 1640 medium supplemented with 10% FCS, penicillin (100 U/ml), streptomycin (100 μg/ml), 4 mM glutamine, and 5 mM HEPES-free acid. All cells were cultivated at 37°C and 5% CO₂.

The stably transformed tTA-HeLa cells with tightly regulated expression of dyn^ts were kindly provided by S. Schmid (13). This cell line expresses a dynamin variant with a point mutation from G to D in position 273 under control of a tetracycline-regulated promoter. The G273D mutation was first discovered in the Drosophila mutant shibire, which fails to recycle neurotransmitters at the synapse at the nonpermissive temperature (28). The packaging cell line PA317/CD4 releasing amphotropic retrovirus vector particles transducing the human CD4 gene has been described (11) and was provided by P. Clapham. To produce a CD4-positive derivative of dyn^ts cells, retrovirus vector particles were harvested from PA317/CD4 cells, filtered, and used to transduce the stably transformed tTA HeLa cells with regulated dyn^ts expression in the presence of 8 μl of Polybrene/ml. Cells were cultured in Dulbecco modified Eagle medium with the described supplements and 200 ng of puromycin (Sigma)/ml, 400 μg of Geneticin (Gibco)/ml, 1 μg of tetracycline (tet) (Sigma)/ml, and 1 mM sodium pyruvate. Transduced cells were incubated with CD4 antibodies conjugated to superparamagnetic MicroBeads (Miltenyi Biotec) and CD4-positive cells were preselected by magnetic activated cell sorting. Clonal cell lines were selected by staining with an antibody against CD4 conjugated with fluorescein isothiocyanate (FITC) and sorting of single CD4-positive cells by using a FACSsorter. Cell lines were analyzed for expression of CD4 and tet-inducible expression of dyn^ts by fluorescence-activated cell sorting (FACS) and immunofluorescence, respectively, resulting in the cell line HeLa4D9.

For single-round entry assays, 3 × 10⁵ HeLa4D9 cells in 10-cm plates were incubated overnight in medium supplemented with 1 μg of tet/ml at 37°C, followed by incubation in the presence or absence of tet for 24 h at 32°C. Subsequently, cells were treated with trypsin and plated according to the type of the experiment (1.1 × 10⁴ cells per 48 wells for analysis of viral capsid [CA] expression by immunostaining, 2.5 × 10⁴ cells per 24 wells for real-time PCR, and 1.5 × 10⁵ cells per 6-cm plate for analysis of HIV-1 infection by fluorescence-activated cell sorting [FACS], for Blam assay, and to control uptake of transferrin) in the absence or presence of tet, respectively. After 48 h at 32°C, cells were incubated at 37°C for 1.5 to 2 h to establish the endocytosis-deficient phenotype or kept at 30°C. Virus was prewarmed to 37°C and added at the indicated multiplicity of infection (MOI). To avoid influences by tet on infection, cells were not exposed to tet during virus incubation. To control the efficiency of endocytosis inhibition, parallel cultures on coverslips were analyzed for transferrin uptake.

Virus production and titration.

Stocks of HIV-1_NL4-3, HIV-1_SF2, and HIV-1_MVP8161 were produced by cocultivation of infected and uninfected MT-4 cells (61). Uninfected cells (5 × 10⁵ cells/ml) were mixed with infected cells at a ratio of 5:1 and virus was harvested before pronounced cytopathic effects were observed (24 to 36 h postinfection). Virus-containing supernatants were cleared by low-speed centrifugation, filtered through 0.45-μm-pore-size filters (Schleicher & Schuell), and adjusted to 10 mM HEPES (pH 7.4). Aliquots were frozen at −80°C or in liquid nitrogen.

HIV-1 particles carrying the β-lactamase (Blam)-Vpr fusion protein were produced by cotransfection of 293T cells with the HIV-1 proviral plasmid pNLC-43 (6) and pMM310. HIV-1 particles pseudotyped with the VSV-G glycoprotein were produced by cotransfection of 293T cells with pNL4-3env⁻GFP and pM3. Transfections were performed by using Lipofectamine 2000 (Invitrogen) according to the description of the manufacturer. After 48 h at 37°C, virus-containing culture media were harvested, cleared as described above, and concentrated 10-fold by centrifugation through an Amicon Ultra filter (30-kDa exclusion limit).

Virus titrations were performed in quadruplicate by using serial 10-fold dilutions of the respective virus. After incubation for 2 days at 37°C, HIV-1-infected cells were detected by staining with a monoclonal antibody to the CA protein and a secondary antibody conjugated with β-galactosidase (BIOZOL) as described previously (12).

Immunofluorescence and FACS analysis.

The expression of dyn^ts was detected by using a FITC-conjugated monoclonal antibody to the epitope tag from influenza virus hemagglutinin (HA; Roche). Cells were fixed in 3.7% paraformaldehyde in phosphate-buffered saline (PBS) for 20 min at room temperature, followed by incubation in 50 mM ammonium chloride for 15 min and immunostaining. To analyze the inhibition of clathrin-dependent endocytosis by dominant-negative dynamin, cells were incubated with Alexa Fluor 568-labeled transferrin (Molecular Probes) 20 min prior to fixation as described above. Fluorescence was visualized on an IX70 epifluorescence microscope (Olympus), and images were captured and processed with Soft Imaging System and Adobe Photoshop.

For FACS analysis of HIV receptor expression, cells were fixed (Fix & Perm kit; Becton Dickinson), incubated with FITC-conjugated antibodies to CD4 and phycoerythrin (PE)-conjugated antibodies against CXCR4 (Becton Dickinson), respectively, and analyzed by using a FACScan (Becton Dickinson). FACS analysis of HIV-1-infected HeLa4D9 cells was performed 48 h after infection. Cells were fixed for 1 h at room temperature in 3.7% paraformaldehyde in PBS and stained for 30 min at room temperature in the dark with a monoclonal antibody to CA conjugated with PE (KC57-RD1; Coulter) diluted 1:20 in 0.1% Triton X-100 in PBS containing 3% FCS.

Detection of virus entry using the β-lactamase assay.

Cytosolic entry of HIV-1 particles carrying the Blam-Vpr fusion protein was analyzed as described previously (59). HeLa4D9 cells or transfected HeLa cells grown on coverslips were infected with HIV-1_NL4-3 particles carrying the Blam-Vpr protein at the indicated MOI for 5 h. Fluorescently labeled transferrin (Molecular Probes) was added to transfected HeLa cells 20 min prior to washing. Cells were washed with CO₂-independent medium (Invitrogen) and loaded with the fluorescent Blam substrate CCF2/AM as described by the manufacturer (Pan Vera). Cells were incubated for 17 h at room temperature to allow cleavage of the substrate and subsequently washed and fixed with 3.7% paraformaldehyde in PBS. The relative number of cells fluorescent at 447 nm among GFP-expressing (detected after short bleaching of the uncleaved substrate at 520 nm) and GFP-negative cells was determined by microscopic analysis of at least 400 cells each.

Quantitation of RT products.

To remove DNA from the virus preparation, HIV-1 particles were treated with 300 U of DNase (Roche)/ml for 30 min at 37°C. Cells were infected with DNase-treated virus and total DNA was isolated at different time points by using a QIAamp Blood Minikit (Qiagen) or DNAzol (Invitrogen) according to the recommendations of the manufacturers. Quantification of RT products from the U3 region was performed by real-time PCR in the LightCycler (Roche) with the primers 48U3 (sense, TGGATCTACCACACACAAGGCTA) and 118U3 (antisense, AGCACCATCCAAAGGTCAGTG). Three independent analyses were performed for each sample, and serial dilutions of a pNL4-3 derivative in unspecific genomic DNA were used as a standard. For normalization, total DNA amounts of each sample were measured by photometry at 260 nm or the cellular single-copy gene encoding gapdh was quantified in parallel by real-time PCR.

Establishment of an HIV-infectible HeLa cell line expressing a dominant-negative variant of dynamin.

To investigate the role of the endocytic pathway in the entry of HIV-1, we transduced a HeLa cell line containing a tightly regulated variant of dynamin (15) with a retrovirus vector carrying the human CD4 gene. The parental cells express a trans dominant-negative, temperature-sensitive mutant of dynamin (G273D; dyn^ts) (29) under control of a tetracycline-responsive promoter. In the absence of tet, expression of the dynamin mutant is induced, and efficient inhibition of clathrin-dependent endocytosis is observed at the nonpermissive temperature of 37 or 38°C (15). No inhibition occurs at the permissive temperature of 30°C. To score for inhibition, dyn^ts is first produced at the permissive temperature (30 to 32°C), allowing oligomerization with endogenous dynamin, and the temperature is then shifted to 37°C to induce the trans-dominant phenotype.

CD4-positive cells were initially isolated by magnetic separation of transduced cell populations and were shown to retain inducible expression of dyn^ts (data not shown). To establish a clonal cell line, single CD4-positive cells were isolated by FACS. Individual cell clones were analyzed for stable CD4 expression (Fig. 1B) and for inducible expression of dyn^ts, and the cell line HeLa4D9 was selected for further experiments. To test the inducible inhibition of clathrin-dependent endocytosis, HeLa4D9 cells were cultivated in the presence or absence of tet at 32°C for 3 days. Subsequently, cells were either shifted to the nonpermissive temperature for 1.5 h or were incubated for 1 h at 30°C after the 37°C shift. Twenty minutes before fixation, fluorescently labeled transferrin was added to monitor clathrin-mediated internalization. Expression of dyn^ts was analyzed with an antibody directed against the HA epitope tag on dyn^ts. No HA signal was detected when cells were repressed by cultivation in the presence of tet (Fig. 1Ac), and normal transferrin endocytosis was observed as indicated by the punctate staining in the perinuclear region (Fig. 1Ac′). Cultivation for 3 days in the absence of tet, on the other hand, yielded a strong signal for dyn^ts in HeLa4D9 cells, which was mainly observed in membrane ruffles and lamellipodia (Fig. 1Aa and b). When cells were kept at the permissive temperature, the uptake of fluorescently labeled transferrin was largely unaltered (Fig. 1Ab′), whereas clathrin-mediated endocytosis of transferrin was efficiently blocked after incubation at the nonpermissive temperature (Fig. 1Aa′). Weak staining at the plasma membrane, but no intracellular punctate staining, was observed in 80 to 95% of HeLa4D9 cells cultivated at 37°C. Thus, this cell line retains the inducible inhibition of clathrin-dependent endocytosis.

To determine whether inhibition of endocytosis interferes with the cell surface expression of the HIV-1 entry receptors, flow cytometric analyses were performed (Fig. 1B). Expression of dyn^ts was induced or suppressed in HeLa4D9 cells as described above, and cells cultivated for an additional 1.5 or 5 h at either 30 or 37°C were fixed and stained with antibodies against CD4 or CXCR4, respectively. As shown in Fig. 1B (left panel), there was no difference in CD4 surface expression and >98% of the cells were CD4 positive in every case, with a similar mean fluorescence intensity. The same result was observed for the HIV-1 coreceptor CXCR4 (Fig. 1B, right panel). CD4 and CXCR4 surface expression were also unchanged after incubation for 5 h at 37°C (data not shown). Thus, HeLa4D9 cells represent a CD4-positive HeLa cell line that is suitable for the investigation of the influence of endocytosis on HIV-1 entry.

Dominant-negative dynamin leads to a strong reduction of HIV-1 infection in HeLa4D9 cells.

To examine the influence of dynamin-dependent endocytosis on HIV-1 entry in a single-round infection assay, the following experimental protocol was established (Fig. 2A). The expression of dyn^ts was induced or suppressed in HeLa4D9 cells as described above, and cells were cultivated for an additional 1.5 h at 37°C to induce the dominant-negative phenotype. The effect of dyn^ts was confirmed by analyzing the internalization of transferrin in parallel samples (data not shown). Subsequently, HIV-1 particles were added at different MOIs, and incubation was continued for a further 5 to 7 h at 37°C. Medium was replaced, and tet was added to repress expression of dyn^ts. The half time of dyn^ts at 37°C was shown to be about 8 h, and no inhibition of endocytosis was observed 20 h after the readdition of tet (data not shown). Thus, a potential influence of residual dyn^ts on HIV-1 gene expression and virion formation could be excluded. Upon removal of the virus inoculum, the reverse transcriptase inhibitor azidothymidine was added to avoid subsequent infection by remaining particles. HIV-1 protein expression was detected after incubation for 2 days at 37°C by immunostaining for the HIV-1 CA protein.

Figure 2B shows the results of representative experiments. All assays were performed in 48 wells in triplicates, and mean values of the total number of infected cells are depicted. Infection of repressed HeLa4D9 cells cultivated in the presence of tet with HIV-1_NL4-3 at an MOI of 0.06 yielded a mean number of 224 infected cells per well (Fig. 2B, bar 2). Infection of dyn^ts-expressing cells cultivated without tet, on the other hand, produced only 108 infected cells per well (Fig. 2B, bar 1), corresponding to a reduction by 52%. In an independent experiment, HIV-1_NL4-3 infection in dyn^ts-expressing cells was reduced by 74% (Fig. 2B, bars 9 and 10). In total, more than 10 independent experiments were performed with an inhibitory range between 40 and 80% in dyn^ts-expressing cells. The results were similar when cells were shifted to 30°C after 5 h of virus infection, which led to an immediate reversal of the dominant-negative effect (data not shown). Furthermore, expression of dominant-negative dynamin also strongly reduced infection of HeLa4D9 cells with the isolates HIV-1_SF2 (data not shown) and HIV-1_MVP8161, which belongs to the divergent group O of HIV-1. In the latter case, the number of infected cells was reduced by 85% (Fig. 2B, bars 11 and 12). There was no reduction in the number of infected cells, however, when dyn^ts-expressing HeLa4D9 cells were infected with HIV-1 particles lacking the viral glycoproteins and pseudotyped with the G glycoprotein of the pH-dependent VSV (VSV-G) (Fig. 2B, bars 5 and 6). This result was unexpected since VSV is known to enter target cells through the endosomal pathway (36).

To test whether acidification of the endosomal compartment plays a role in virus entry, the effect of ammonium chloride (NH₄Cl) on HIV-1 entry was analyzed. Infection of HeLa4D9 cells not expressing dyn^ts in the presence of 10 mM NH₄Cl led to an increase in the number of HIV-1-positive cells by ca. 50% (Fig. 2B, bars 2 and 4), a finding similar to previously published results (1, 21). A similar increase in infectivity was also observed when dyn^ts-expressing cells were infected in the presence of NH₄Cl, and the negative effect of dyn^ts on HIV-1 entry was maintained in this case (Fig. 2B, bars 3 and 4). HIV-1 particles pseudotyped with VSV-G were used as a control, and infection with these particles was completely blocked by NH₄Cl (Fig. 2B, bars 7 and 8) as expected.

To analyze whether the inhibition of HIV-1 infection by dyn^ts is dependent on the MOI, the experimental protocol was adapted to FACS analysis. HeLa4D9 cells were suppressed or induced to express dyn^ts, infected with HIV-1_NL4-3, and stained for intracellular CA antigen 2 days after infection. Figure 3A shows the result of infection at an MOI of 0.08, again confirming a reduction of infectivity in dyn^ts-expressing cells by ca. 50%. Very similar results were obtained when infections were performed with an MOI of 0.24 or 0.72 (Fig. 3B). Taken together, these results indicate that a block in endocytosis strongly reduces HIV-1 entry in HeLa4D9 cells independent of the MOI, whereas endosomal acidification is not required.

The negative effect of dyn^ts on HIV-1 infection occurs prior to reverse transcription.

To determine whether dyn^ts affects a step in early HIV-1 entry, the accumulation of RT products was examined in newly infected cells. A quantitative real-time DNA-amplification protocol was developed by detecting the U3 region of the HIV-1 genome, which is reverse transcribed at an intermediate stage of genome replication (Fig. 4A). To validate the protocol, HeLa cells lacking the CD4 receptor and HeLaP4 cells containing CD4 were infected in parallel with HIV-1_NL4-3 at an MOI of 0.5. Total DNA was extracted at different times after infection and analyzed by real-time PCR (Fig. 4B). No U3 signal was observed immediately after infection (time zero). Subsequently, increasing amounts of RT products were detected in HeLaP4 cells, starting at 2 h postinfection (Fig. 4B). No RT products were detected in HeLa cells lacking CD4, indicating that endosomal uptake of HIV-1 in the absence of CD4, which has been shown to occur in HeLa cells (33), does not lead to reverse transcription (Fig. 4B).

To analyze the effect of dyn^ts-mediated inhibition of endocytosis on the generation of HIV-1 RT products, HeLa4D9 cells were induced or suppressed to express dyn^ts as described above, infected with HIV-1_NL4-3 at an MOI of 0.29, and extracted at different time points after infection. Subsequently, U3 copy numbers were determined by real-time PCR and normalized for total DNA in the sample. Three independent PCRs were performed for each sample, and infections were done in triplicate (time zero) or quadruplicate (time 7 h). No U3 products were detected immediately after infection (Fig. 4C). After 7 h, ca. 240 copies of the U3 product per ng of DNA were detected in cells cultivated in the presence of tet. This number was reduced by 60% in cells that had been induced to express dyn^ts. Thus, the effect of dyn^ts on HIV-1 infection occurs at an early time point prior to the completion of reverse transcription.

Dominant-negative dynamin reduces cytosolic HIV-1 entry.

To further define the step affected by dyn^ts, the cytosolic entry of HIV-1 cores was analyzed in HeLa4D9 cells by using the previously described β-lactamase (Blam) assay (59). This method involves infection of target cells with HIV-1 particles, carrying a fusion protein of Blam and the HIV-1 protein Vpr. Target cells are then loaded with a Blam substrate (CCF2/AM) that changes its fluorescence emission spectrum from green to blue after cleavage by the enzyme. Cleavage depends on receptor-mediated entry of HIV-1 releasing the Blam-Vpr fusion protein into the cytosol of the newly infected cell (9, 40) and can thus be used as a direct measure of viral entry. To analyze the effect of dyn^ts-mediated inhibition of endocytosis on cytosolic entry of HIV-1, HeLa4D9 cells were induced or suppressed to express dyn^ts as before and infected for 5 h with HIV-1_NL4-3 particles carrying the Blam-Vpr fusion protein at MOIs of 0.074 and 0.185, respectively. Cells were subsequently loaded with the Blam substrate, and the accumulation of the blue fluorescent product was analyzed by fluorescence microscopy. Uptake of the Blam substrate was detected in the green channel (Fig. 5Aa′ and b′), whereas the cleaved product was observed in the blue channel (Fig. 5Aa and b). HIV-1 entry led to a strong blue fluorescence (Fig. 5A, arrows). The number of blue fluorescent cells was significantly reduced in HeLa4D9 cells expressing dyn^ts compared to control cells (Fig. 5A, compare panels a and b, and B).

Figure 5B shows a quantitative analysis of HIV-1 entry in HeLa4D9 cells induced or repressed for dyn^ts expression. Approximately 50% of control cells but only 20% of cells expressing dyn^ts exhibited blue fluorescence at an MOI of 0.185, a finding corresponding to a reduction in cytosolic HIV-1 entry by 60% (Fig. 5B). A similar reduction in HIV-1 entry was observed at an MOI of 0.074 (28% versus 11%; Fig. 5B). One should also note that the percentage of blue cells is significantly higher than the respective MOI. Virus titers had been determined on HeLa4D9 cells by analysis of CA expression, and the differences indicate that cytosolic entry may not necessarily lead to productive infection.

Blocking endocytosis in transiently transfected cells also reduces HIV-1 entry.

The described results showed that the dominant-negative dynamin variant G273D strongly reduces entry of various HIV-1 isolates in the HeLa4D9 cell line. To determine whether this effect is dependent on the specific cell line and mutant, we performed infection experiments on transiently transfected HeLaP4 cells. Cells were transfected with expression constructs for fusion proteins of GFP with either wild-type (wt) dynamin or mutant dyn^K44A. This mutant also has a trans dominant-negative effect on the release of endocytic vesicles from the plasma membrane but differs from dyn^G273D in the ability to affect actin at the plasma membrane and is not temperature sensitive (42). At 42 h after transient transfection, cells were infected with HIV-1_NL4-3 particles carrying the Blam-Vpr protein at an MOI of 0.2 for 4 to 5 h at 37°C. At 20 min before the end of the infection period, fluorescently labeled transferrin was added to monitor the inhibition of endocytosis. Subsequently, cells were loaded with the Blam substrate as described above and investigated under the fluorescence microscope. The upper two panels in Fig. 6 show the results for cells transfected with dyn^K44AGFP (Fig. 6a) and dyn^wtGFP (Fig. 6b). Detection of the respective GFP fusion protein in the green channel was partly obscured by the uncleaved Blam substrate and was only possible after rapid bleaching of this substrate (Fig. 6a" and b"). Thus, only cells showing a strong expression of the GFP fusion protein were identified as productively transfected cells and are marked in Fig. 6. Cells transfected with dyn^K44AGFP generally exhibited a strong block in transferrin endocytosis (Fig. 6a and a", arrows), whereas cells transfected with dyn^wtGFP revealed normal transferrin uptake (Fig. 6b and b", arrowheads) as expected. HIV-1 entry was analyzed in the blue channel, revealing virtually no blue fluorescent cells among the strongly GFP-positive population in the case of dyn^K44AGFP transfection (Fig. 6a′, arrows), whereas blue fluorescent cells were readily detected in the GFP-positive population of dyn^wtGFP transfection (Fig. 6b′, arrowheads). Importantly, relative HIV-1 entry as measured by blue fluorescence in neighboring cells not expressing detectable amounts of the respective dynamin variant was similar in both cultures (Fig. 7A, black bars), indicating that the entry block was due to expression of the dominant-negative dynamin variant.

Quantitative determination of HIV-1 entry was performed for three different cell populations in the case of dyn^K44AGFP as follows: (i) GFP-negative cells not expressing the respective dynamin variant (Fig. 7, solid bars), (ii) GFP-positive cells not inhibited in transferrin uptake (i.e., functionally wild-type; Fig. 7A, open bars), and (iii) GFP-positive cells blocked in endocytosis (e.g., arrows in Fig. 6a; Fig. 7A, shaded bars). In the case of dyn^wtGFP transfection, HIV-1 entry was determined for GFP-positive and GFP-negative cells (Fig. 7A). Evaluation of at least 400 cells per population in triplicate samples revealed blue fluorescence in 50 to 60% of the GFP-negative population in both transfections after infection at an MOI of 0.2 (Fig. 7A, black bars). HIV-1 entry was reduced 10-fold in HeLa cells productively transfected with dyn^K44AGFP and blocked in endocytosis (Fig. 7A, gray bar). Little inhibition was observed in dyn^K44AGFP-transfected cells that were not blocked in endocytosis and in cells transfected with dyn^wtGFP (Fig. 7A, open bars). Thus, blocking endocytosis by expression of the dyn^K44A variant in HeLaP4 cells strongly reduced HIV-1 entry, confirming the results obtained for the HeLa4D9 cell line.

HIV-1 entry is inhibited by expression of dominant-negative Eps15 but not by expression of dominant-negative caveolin.

Dynamin plays a role in several endocytic pathways and the observed effect of dominant-negative dynamin expression does not directly identify the pathway important for HIV-1 entry. To further delineate the molecular mechanism of entry inhibition, we performed experiments with trans dominant-negative versions of Eps15 and caveolin, respectively. Dominant-negative Eps15 specifically affects the internalization of clathrin-coated vesicles by interacting with AP2 (3), whereas caveolin-1 fused to GFP at its N terminus (dncavGFP) has been shown to inhibit the entry of the caveola-dependent SV40. In contrast, C-terminal fusion to GFP (cavGFP) did not interfere with the entry of SV40 (49).

HeLaP4 cells were transiently transfected with constructs expressing GFP fused with wild-type or dominant-negative Eps15 and with wild-type or dominant-negative caveolin-1. At 42 h after transient transfection, cells were infected with HIV-1_NL4-3, carrying the Blam-Vpr protein at an MOI of 0.2 for 4 to 5 h at 37°C. Fluorescently labeled transferrin was added to monitor the inhibition of endocytosis, and cells were loaded with the Blam substrate as described above. The lower four panels in Fig. 6 show the results for cells transfected with dnEps15GFP (Fig. 6c), Eps15GFP (wild-type, Fig. 6d), dncavGFP (Fig. 6e), and cavGFP (Fig. 6f). Expression of the respective GFP fusion protein was analyzed in the green channel after rapid bleaching of the uncleaved Blam substrate as described above. This was clearly detectable in the case of the caveolin-1 GFP fusion proteins (Fig. 6e" and f", arrowheads) and of wild-type Eps15GFP (Fig. 6d"), whereas no GFP signal was detectable in the case of the dnEps15 fusion due to low expression levels (Fig. 6c"). Cells expressing dnEps15GFP were therefore identified by their strong block in transferrin endocytosis (Fig. 6c, arrows). HIV-1 entry was analyzed in the blue channel, revealing a strong reduction in the number of blue fluorescent cells when dnEps15GFP-transfected cells that were blocked in transferrin endocytosis (Fig. 6c, arrows) were compared to neighboring cells exhibiting normal transferrin endocytosis. Quantitative analysis of more than 400 cells revealed a 20-fold decrease in HIV-1 entry from ca. 60% in the case of cells with normal transferrin endocytosis to ca. 3% in the case of cells blocked in transferrin endocytosis (Fig. 7B, left panel). In contrast, there was no significant difference in the number of blue fluorescent cells when the strongly GFP-positive and -negative populations were compared after transfection of wild-type Eps15GFP (Fig. 6d and 7B, right panel). Analysis of cells transfected with dominant-negative dncavGFP or wild-type cavGFP also revealed no significant difference in the relative entry of HIV-1. Similar numbers of blue cells were observed in the GFP-positive and GFP-negative cell populations in both cases (Fig. 6e and f and 7C), indicating that caveolae do not play a role in HIV-1 entry.