Human Rhinovirus 14 Enters Rhabdomyosarcoma Cells Expressing ICAM-1 by a Clathrin-, Caveolin-, and Flotillin-Independent Pathway ^▿

Cells and viruses.

Rhabdomyosarcoma cells stably transfected with human ICAM-1 (RD-ICAM), a kind gift of Darren Shafren (University of Newcastle, Australia), were cultured as previously described (25, 36). HeLa-H1 cells were grown in minimal essential medium (MEM) supplemented with 10% heat-inactivated fetal calf serum (FCS), 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM l-glutamine. For infection, Dulbecco's modified Eagle's medium (DMEM) supplemented with 30 mM MgCl₂, 2% FCS, and antibiotics (infection medium) was used. HRVs were originally obtained from ATCC. All experiments were conducted with plaque-purified virus.

Antibodies and reagents.

Mouse anti-HRV14, rabbit anti-HRV14, and monoclonal mouse anti-HRV2 (8F5) (19) were produced by following standard protocols. Rabbit anti-caveolin and rabbit anti-flotillin-1 antibodies were purchased from Santa Cruz Biotechnology. Secondary antibodies labeled with Texas Red or with Alexa Fluor 488 or Alexa Fluor 591 were from Jackson Laboratories and Molecular Probes, respectively. Alexa Fluor 568-transferrin, Alexa Fluor 488-cholera toxin B (CtxB), and lysine-fixable fluorescein isothiocyanate (FITC)-dextran (70 kDa) were purchased from Molecular Probes. The following drugs were purchased from Sigma if not stated otherwise: chlorpromazine dissolved in phosphate-buffered saline (PBS; 5 μg/ml), Filipin dissolved in methanol (10 μg/ml), and methyl-β-cyclodextrin (MßCD) dissolved in PBS (5 mM). Amiloride, cytochalasin D, and dynasore were dissolved in dimethyl sulfoxide (DMSO) and used at final concentrations of 5 mM, 10 μg/ml, and 80 μM, respectively.

Drug treatment.

RD-ICAM cells were seeded as two parallel sets either in 24-well plates or on coverslips. Cells were preincubated in the presence of the selected drug for 30 min and infected with HRV2 or HRV14 at 300 50% tissue culture infective doses (TCID₅₀)/cell in serum-free infection medium. After an additional incubation for 30 min at 34°C, one set of samples was fixed and prepared for indirect fluorescent microscopy. The second set was harvested in 50 μl of sample buffer at 3 h postinfection (p.i.) for assessment of eukaryotic translation initiation factor 4GI (eIF4GI) cleavage via Western blotting. Drugs were present throughout the incubation.

Western blot analysis.

Proteins from whole-cell lysates from each drug treatment were separated on an SDS-6% polyacrylamide gel and electrotransferred onto a polyvinylidene difluoride (PVDF) membrane. The membrane was blocked with 5% milk powder in PBS-T (PBS containing 0.1% Tween 20), incubated with rabbit anti-eIF4GI (1:8,000) kindly donated by R. Rhoads, Louisiana State University, followed by horseradish peroxidase (HRP)-labeled anti-rabbit antibody (1:20,000; Jackson Laboratories), and eIF4GI and its cleavage products were visualized by using an ECL chemiluminescence detection system (Thermoscientific). Cleavage of eIF4GI was quantified with ImageJ (http://rsbweb.nih.gov/ij/) software by using scans of the X-ray films.

Uptake of FITC-dextran.

RD-ICAM cells were grown in 24-well plates up to 90% confluence and incubated with serum-free medium for 4 h at 37°C. Lysine-fixable FITC-dextran at 1 mg/ml was added in the presence or absence of 5 μg/ml purified HRV14 and 100 ng/ml epidermal growth factor (EGF) (Sigma), respectively, and incubated for 10 min. Cells were washed with cold PBS, detached with trypsin at 4°C, collected, and resuspended in PBS containing 7% FCS. In each experiment, the mean fluorescence of 10,000 cells was determined in a BD FACSCalibur.

Confocal fluorescence microscopy.

Cells were grown on glass coverslips until 40 to 50% confluent. After cells were washed with serum-free medium, they were challenged with HRV2 or HRV14 at 300 TCID₅₀/cell as specified in the figure legends. Unbound virus was removed by three washings with PBS⁺⁺ (PBS with 1 mM MgCl₂ and 1 mM CaCl₂). Cells were fixed in 4% paraformaldehyde and permeabilized in 0.2% Triton X-100 for 5 min. After samples were quenched with 50 mM NH₄Cl for 10 min, unspecific binding sites were blocked in 1% goat serum either for 1 h at room temperature or overnight at 4°C. Cells were incubated for 1 h at room temperature with primary antibody (1:500) and washed three times with PBS⁺⁺, followed by secondary antibody (1:400), all diluted in 1% goat serum. Nuclei were stained with 0.1 μg/ml 4′,6′-diamidino-2-phenylindole (DAPI). For assessment of colocalization, cells were infected in the presence of 10 μg/ml Alexa Fluor 568-transferrin, 1 μg/ml Alexa Fluor 488-CtxB, and 500 μg/ml lysine-fixable FITC-dextran (70 kDa). After the times specified on the figures, unbound material was removed, and cells were fixed. For better distinction between internal and external virus, differential staining before and after permeabilization of the cells was carried out with antibodies conjugated to different dyes. Coverslips were mounted in Mowiol and viewed in a Carl Zeiss LSM510-Meta confocal laser scanning microscope (LSM). Stacks were spaced by 0.5 μm. Contrast of the images was adjusted using the LSM image browser, and figures were labeled with Adobe Photoshop CS3 software.

Transfection.

The plasmids pEGFP-C3, encoding the clathrin light chain fused to green fluorescent protein (GFP-clathrin), myc-tagged pCMV-amphi-SH3 (where amphi-SH3 is the Src homology 3 domain of amphiphysin), originally donated by E. Greene (Laboratory of Cell Biology, Bethesda, MD) and H. T. McMahon (Laboratory of Molecular Biology, Cambridge, United Kingdom), respectively, and pCMV-AP180 (44) were kind gifts from Luc Snyers. Cells were seeded on coverslips in 24-well plates and grown to 30% confluence. One microgram of plasmid DNA was mixed with 50 μl of OptiMem (Gibco). Separately, 3 μl of Lipofectamine 2000 (Invitrogen) was mixed with 50 μl of OptiMem and incubated for 5 min at room temperature; the two solutions were mixed gently and incubated for 20 min at room temperature. Cells were washed twice with antibiotic-free growth medium and laid down with 500 μl of the same medium. The mixture of plasmid DNA and Lipofectamine 2000 was applied onto the cells drop by drop with gentle rotation to avoid cell toxicity. Cells were incubated for 3 h with the mixture; then, the medium was replaced with fresh antibiotic-free growth medium. At 24 h posttransfection cells were infected with HRV14 and processed for immunofluorescence microscopy. For amphi-SH3 and AP180-C (the C-terminal domain of AP180), anti-myc antibody was used to verify expression. Viral entry and replication were monitored after 30 min and 7 h by using HRV14 at 300 and 5 TCID₅₀/cell, respectively.

Electron microscopy.

RD-ICAM cells were seeded on Aclar coverslips in 24-well plates and grown to almost 100% confluence. Cells were washed and infected with HRV2 or HRV14 at 10,000 TCID₅₀/cell. After incubation at 34°C for the times given on Fig. 7, unbound virus was removed, and cells were fixed with 2% freshly prepared glutaraldehyde for 1 h at room temperature. Cells were washed three times with 0.1 M Sorensen buffer (pH 7.3) and incubated for 1 h in 2% osmium tetroxide in the same buffer. After cells were washed, they were dehydrated with ascending concentrations of ethanol (40, 60, 80, 95, and 100%; kept over a molecular sieve). Cells were incubated with Epon-100% ethanol (EtOH; 1:1) for 30 min and twice with Epon only, embedded in resin-filled beam capsule lids, and incubated at 60°C to polymerize for 2 days. Sections of 70 nm were prepared, stained with 2% uranyl acetate and lead citrate for 10 and 5 min, respectively, and viewed under an FEI Morgagni 100 kV electron microscope equipped with an 11-megapixel charge-coupled-device (CCD) camera.

Virus entry via ICAM-1 is clathrin independent.

Although inhibition of HRV14 replication in HeLa cells expressing a DN mutant of dynamin-2 has been investigated (9), the clathrin dependence of uptake per se has not been analyzed. We thus first cointernalized HRV14 with human transferrin, a marker for CME. RD-ICAM cells were incubated with virus at 4°C in the presence of Alexa Fluor 568-transferrin for 1 h; cells were washed, shifted to 34°C, and fixed at 0, 5, and 15 min of incubation. Confocal microscopy revealed no colocalization at the plasma membrane (Fig. 1, 0 min). At 5 min after the temperature shift, most of the transferrin was internalized and accumulated in early endosomes and in the perinuclear region while HRV14 largely remained at the cell periphery. At 15 min virtually all transferrin had accumulated in the perinuclear region, whereas HRV14 was evenly distributed throughout the cell body, with a substantial fraction remaining at or close to the cell surface. The uptake kinetics of HRV14 is slow (26), and we consistently observed this plasma membrane accumulation even in HeLa cells that express less ICAM-1 than RD-ICAM cells. Nevertheless, HRV14 was unequivocally internalized, as also seen by differential staining prior to and after permeabilization (data not shown). These data clearly demonstrate that transferrin and HRV14 localize to distinct plasma membrane subdomains and enter with different kinetics, making clathrin-mediated endocytosis for HRV14 highly unlikely. On the other hand, HRV2, used as a control, completely overlapped with transferrin at the plasma membrane and in early endosomes (44; also data not shown) although their intracellular pathways separated thereafter (6).

A more direct proof for clathrin-independent entry via ICAM-1 was obtained by using a GFP-tagged clathrin light chain. RD-ICAM cells expressing GFP-clathrin were challenged with virus and fixed at 5 and 15 min p.i. In accordance with the results shown above, GFP-clathrin did not colocalize with HRV14 at any time (data not shown).

DN inhibitors of clathrin-mediated endocytosis do not modify HRV14 entry via ICAM-1.

DN mutants of key proteins of the clathrin-dependent pathway have been shown to potently interfere with the normal functions of the respective wild-type proteins and hence specifically inhibit clathrin-mediated endocytosis. The SH3 domain of amphiphysin (amphi-SH3) interferes with recruitment of dynamin to coated pits and has been shown to inhibit receptor-mediated endocytosis of HRV2, transferrin, and epidermal growth factor, among many other ligands (44, 49). Thus, we made use of this DN inhibitor to confirm our data of the colocalization experiment and rule out the possible involvement of clathrin in HRV14 entry. Accordingly, RD-ICAM cells were transfected with myc-tagged amphi-SH3, followed by infection with HRV14. Confocal microscopy revealed no difference in entry in the presence or absence of the expression of amphi-SH3 (Fig. 2A, top row). amphi-SH3 was shown to prevent self-assembly of dynamin rings, which impedes constriction of clathrin-coated pits but does not affect clathrin assembly by itself (38). Therefore, to rather more directly assess clathrin function, we also used the C-terminal domain of AP180 (AP180-C), which affects clathrin-mediated endocytosis without interfering with dynamin function (11). Again, overexpression of DN AP180-C failed to modify entry of HRV14 (Fig. 2A, bottom row); in contrast, transferrin uptake was prevented by both DN inhibitors. In order to exclude the improbable possibility that the virus observed followed an abortive pathway, we also monitored viral replication after infection at a much lower multiplicity of infection ([MOI] 5 TCID₅₀/cell). As shown in Fig. 2B, cells strongly overexpressing the DN protein mutants replicated HRV14 to a degree indistinguishable from untransfected cells. This not only proves that the clathrin-dependent endocytosis machinery is dispensable for HRV14 entry but also excludes any effect of the DN mutants on productive viral entry and replication.

HRV14 does not colocalize with CtxB.

The entry pathway involving caveolin is the second well-characterized endocytic route. Thus, the eventual contribution of caveolin was evaluated by assessing a possible colocalization of virus with cholera toxin subunit B (CtxB). CtxB is a widely accepted marker of this pathway although it has also been shown to enter via other routes, such as by flotillin- (14) and clathrin-dependent endocytosis, in various cell types (16, 46). RD-ICAM cells were incubated with HRV14 in the presence of Alexa Fluor 488-CtxB for 60 min at 4°C, washed, and shifted to 37°C for the times indicated in Fig. 3. Confocal microscopy showed some overlap of HRV14 with CtxB at the cell surface at 0 min, but after being shifted to 34°C for 5 min, HRV14 and CtxB separated from each other. CtxB moved to the Golgi apparatus and concentrated at the perinuclear region at 15 min after the temperature shift while the HRV14 signal remained spread throughout the cytosol. These results indicate that HRV14 and CtxB enter with different kinetics and use different pathways to reach different destinations within the cell.

Further confirmation by more direct means was obtained by staining caveolin-1 and virus with appropriate antibodies. HRV14 did not colocalize with caveolin-1 (data not shown). These results definitely exclude the involvement of caveolin-1 in viral entry and intracellular trafficking via ICAM-1.

Entry via ICAM-1 is flotillin-1 independent.

Flotillin-1 and -2 have been described as markers of a clathrin- and caveolin-independent pathway (14). Having excluded entry of HRV14 via clathrin- and caveolin-dependent routes, we wanted to investigate whether flotillin-1 was involved. Again, by using fluorescence microscopy, no colocalization of flotillin-1 and HRV14 was seen (Fig. 4). Therefore, it is very unlikely that this virus follows a flotillin-1-dependent route when accessing the cell via ICAM-1.

Effect of pharmacologic inhibitors on entry of HRV2 and HRV14.

Because of the overexpression of ICAM-1 in RD-ICAM cells, large amounts of virus are taken up by the cell. However, only a fraction might undergo productive uncoating. Therefore, using a functional assay (by measuring translation of the uncoated viral RNA), we studied the effect of various drugs that have been frequently employed to specifically inhibit selected endocytic pathways. In order to avoid exposure of the cells to the chemicals for extended times, as is required to detect virus production, we made use of cleavage of the translation initiation factor eIF4GI as a reporter assay. Upon virus uncoating, the viral RNA is released into the cytoplasm and translated into a polyprotein; autocatalytic processing sets free the viral proteinase 2A, which efficiently cleaves eIF4GI into fragments detectable even at 3 h p.i. RD-ICAM cells were preincubated with the selected drugs for 30 min, as indicated in Fig. 5A, challenged with HRV14 and HRV2, and collected after 3 h. As previously described for poliovirus (21), processing of eIF4GI was assessed via Western blot analysis. The extent of cleavage was quantified from three gels obtained from three independent experiments (Fig. 5B).

Chlorpromazine causes the loss of coated pits from the plasma membrane (47) and has been extensively used to demonstrate clathrin-mediated endocytosis of viruses. Although chlorpromazine is known to have a number of side effects (22), it is still a good choice when entry of different ligands into the same cell type is studied. In agreement with the results above, eIF4GI cleavage upon HRV2 infection was prevented by the drug but only marginally reduced upon infection with HRV14.

Caveolin-mediated endocytosis is associated with lipid rafts. Removal or sequestering of cholesterol profoundly inhibits entry of CtxB and other ligands internalized via this route (42). Filipin destroys the rafts by intercalating between cholesterol molecules while methyl-ß-cyclodextrin (MßCD) sequesters cholesterol. Both lead to impairment of caveolin/lipid raft-dependent endocytosis. Neither Filipin nor MßCD significantly inhibited entry of HRV14 and HRV2. This confirms that caveolin/lipid rafts have no major importance in endocytosis of HRV2 and HRV14.

Amiloride, an inhibitor of the Na⁺/H⁺ exchanger, and cytochalasin D, which disrupts actin filaments, have both been shown to inhibit macropinocytosis. Although the mechanism of action of amiloride on macropinocytosis is still unclear, it has been employed at mM concentrations to study entry of various viruses and other ligands (32, 41, 48); cytochalasin, on the other hand, inhibits membrane ruffling and actin reorganization, which are required for plasma membrane engulfment of liquid and ligands. Amiloride completely inhibited cleavage of eIF4GI by both virus types. While these results suggest a macropinocytic pathway of HRV14, they disagree with the established clathrin dependence of HRV2 internalization (44). However, immunofluorescence microscopy unequivocally demonstrated that the drug did not prevent cell entry of HRV2 even at 5 mM (Fig. 5D), whereas eIF4GI cleavage was abrogated already at 200 μM amiloride (data not shown). This points to steps different from endocytosis, such as RNA release and/or polyprotein processing/translation, as targets of amiloride in HRV2 infection. Cytochalasin D exhibited a moderate effect (about 40% inhibition) on HRV14- and HRV2-mediated eIF4GI cleavage. Similar to amiloride, cytochalasin led to accumulation of HRV14 at the plasma membrane (Fig. 5C), whereas it apparently affected uptake of HRV2 only marginally (Fig. 5D). Note that plasma membrane staining of HRV2 was virtually absent because of the low steady-state concentration of the receptors on the membrane. Even though HRV14 entry in the presence of cytochalasin is below the detection limit of immunofluorescence microscopy, the cleavage assay demonstrated that enough virus entered the cells to exert a notable effect. Therefore, this functional assay is clearly more sensitive.

Finally, dynasore, an inhibitor of the GTPase activity of dynamin (28), virtually prevented eIF4GI cleavage when cells were infected with HRV2 but showed only about 40% inhibition in the case of HRV14 infection. This correlates with clathrin-dependent entry of HRV2. However, it disagrees with earlier findings of a strictly dynamin-dependent entry of HRV14 into HeLa-H1 cells (9) and indicates that the degree to which uptake of ICAM-1-bound ligands is dynamin dependent might be cell type specific.

The cleavage assay monitors viral entry via translation of the viral RNA without interference with later processes such as viral replication. Nevertheless, in order to lend more support to the above results, we also attempted to measure the number of cells that became infected in the presence of the drugs via staining of de novo synthesized viral protein and quantification by use of a TissueFAXS instrument. As pointed out above, this required maintaining the cells in the presence of the drugs for at least 7 h. Examination of the images taken by the instrument revealed that some of the compounds led to a more diffuse staining pattern than that of the controls (infected in the absence of drugs). As this strongly interfered with recognition of infected cells, these experiments did not produce meaningful results.

HRV14 uptake shows some, but not all, characteristics of macropinocytosis.

Based on the previous results, we concentrated on macropinocytosis as a candidate pathway for viral entry. RD-ICAM cells were challenged with HRV14 in the presence of 500 μg/ml of the fluid-phase marker FITC-dextran for 15 and 30 min at 34°C. Cells were fixed and examined by confocal fluorescence microscopy. Figure 6A shows substantial colocalization of HRV14 with FITC-dextran, which is considered a relatively specific marker of macropinocytosis (10). In most cell lines macropinocytosis is not constitutive but is induced by signaling events (23). We thus asked whether uptake of FITC-dextran was increased in RD-ICAM cells upon challenge with HRV14. Serum-starved cells were incubated with HRV14 and, as a control, with epidermal growth factor (EGF), together with FITC-dextran for 10 min. Fluorescence-activated cell sorter (FACS) analysis demonstrated a substantial stimulation of fluid-phase uptake by HRV14 and EGF (Fig. 6B).

Electron microscopy of virus entry.

The tubovesicular compartments involved in the diverse entry pathways exhibit characteristic morphological features, and clathrin-coated pits, caveolae, and macropinocytic membrane invaginations are distinguishable by electron microscopy. In addition, structures morphologically different from the three mentioned above have also been recently implicated in other uptake mechanisms (27). We thus studied the early events of HRV14 entry by electron microscopy. RD-ICAM cells were challenged with HRV14 for the times indicated in Fig. 7, and thin sections were prepared. As expected, HRV2, used as a control for clathrin-mediated endocytosis, accumulated in coated pits and vesicles at 5 min p.i. Conversely, many HRV14 virions appeared aligned in tubular structures like pearls on a string. This is very similar to the structures observed previously in HRV14-infected BHK cells transfected to express human ICAM-1 (15) and in adenovirus 3 (Ad3)-infected HeLa cells (1). Although apparently exploiting macropinocytosis, HRV14 thus enters via membrane invaginations very different from “conventional” macropinosomes.

HRV14 entry into HeLa cells.

The experiments described above were all carried out with RD cells stably transfected to express human ICAM-1. We chose this cell line because of its use in the elucidation of HRV8v uptake via heparan sulfate (A.G. Khan et al., unpublished data). However, in most of the previous work with rhinoviruses, HeLa-H1, a particularly susceptible substrain of HeLa cells (2), has been employed. Furthermore, this same cell line has also been used in experiments that suggest the clathrin-dependent endocytosis of HRV14 mentioned above (9). Therefore, we repeated some of the most significant experiments with this cell type. Virus uptake into HeLa cells was unaffected by the DN inhibitors amphi-SH3 and AP180-C. In electron microscopy, the virions did not appear as pearls on a string but, rather, were located to small, uncoated invaginations and short tubules. Similar virus-containing vesicles have been previously detected, but some coated structures were also seen (15). Therefore, although highly unlikely, clathrin-dependent endocytosis of HRV14 in HeLa cells cannot be excluded with absolute certainty. Unfortunately, the cleavage assay could not be employed with HeLa cells because proteolysis of eIF4GI, even at 4 h p.i., was only about 10% under experimental control conditions otherwise identical to those used for RD-ICAM cells. Therefore, partial inhibition could not be evaluated. Nevertheless, cleavage of eIF4GI in the presence of chlorpromazine was identical to that in the control in the absence of the drug. This is again evidence that clathrin does not play any role in HRV14 uptake into HeLa cells.

Entry of HRV2.

Although HRV2 was used merely as a control for clathrin-dependent endocytosis, we made some new findings with the pharmacological inhibitors. The endocytosis-blocking effect of chlorpromazine and dynasore was expected. However, we were surprised to find complete inhibition of eIF4GI cleavage by amiloride. The possibility that this drug affects entry can be excluded because immunofluorescence microscopy clearly showed that virus entered the cells in its presence. For coxsackievirus B3 and for HRV2, inhibition of RNA replication and of virus release, respectively, by amiloride has been suggested (12, 18). Our data, rather, point to inhibition of RNA release into the cytosol, RNA translation, or even inhibition of polyprotein processing. It will be interesting to find out which one of these is the target of amiloride. We also found only marginal inhibition by 5 mM MßCD in the cleavage assay, which contrasts with earlier findings indicating reduction of entry and accumulation of virus at the plasma membrane when the drug was present at 10 mM (44). In the absence of additional experimentation, we tentatively explain this by the difference in cell type used (i.e., HeLa-H1 cells in the previous study and RD-ICAM cells in the present work) and a strong concentration-dependent effect of drug action. In the case of HRV2, the presence of cytochalasin reduced viral entry and led to viral replication in small dots rather than in large areas (data not shown). Actin is also known to be involved in CME, and inhibition of its normal function might reduce HRV2 entry (8). The underlying mechanisms deserve further study.

Entry of the closely related poliovirus into HeLa cells occurs via a caveolin- and dynamin-independent pathway, whereas entry into brain microvascular endothelial cells depends on both components (7). We have now demonstrated a similar situation for HRV14. Entry of this virus into different cell types depends on dynamin to different degrees. Furthermore, by using HRV2 and HRV14, which bind different receptors and belong to species HRV-A and HRV-B, respectively, we show the existence of distinct entry pathways with different final destinations in identical cell types. Furthermore, the demonstration of the tubules involved in uptake of two different virus types (HRV14 and Ad3) raises the question of whether these structures are involved in a new type of viral cell entry mechanism. It will be interesting to study the mechanism of their induction.