|Home | About | Journals | Submit | Contact Us | Français|
The early steps of human parvovirus B19 (B19V) infection were investigated in UT7/Epo cells. B19V and its receptor globoside (Gb4Cer) associate with lipid rafts, predominantly of the noncaveolar type. Pharmacological disruption of the lipid rafts inhibited infection when the drug was added prior to virus attachment but not after virus uptake. B19V is internalized by clathrin-dependent endocytosis and spreads rapidly throughout the endocytic pathway, reaching the lysosomal compartment within minutes, where a substantial proportion is degraded. B19V did not permeabilize the endocytic vesicles, indicating a mechanism of endosomal escape without apparent membrane damage. Bafilomycin A1 (BafA1) and NH4Cl, which raise endosomal pH, blocked the infection by preventing endosomal escape, resulting in a massive accumulation of capsids in the lysosomes. In contrast, in the presence of chloroquine (CQ), the transfer of incoming viruses from late endosomes to lysosomes was prevented; the viral DNA was not degraded; and the infection was boosted. In contrast to the findings for untreated or BafA1-treated cells, the viral DNA was progressively associated with the nucleus in CQ-treated cells, reaching a plateau by 3 h postinternalization, a time coinciding with the initiation of viral transcription. At this time, more than half of the total intracellular viral DNA was associated with the nucleus; however, the capsids remained extranuclear. Our studies provide the first insight into the early steps of B19V infection and reveal mechanisms involved in virus uptake, endocytic trafficking, and nuclear penetration.
Human parvovirus B19 (B19 virus [B19V]) was discovered in 1975 (16) and has been classified within the Erythrovirus genus of the Parvoviridae family. Transmitted mainly via the respiratory route, B19V is generally associated with a mild, self-limiting childhood disease named erythema infectiosum, or fifth disease. However, during pregnancy or in individuals with underlying immune or hematologic disorders, B19V can cause more-severe syndromes, such as acute and chronic arthropathies, severe cytopenias, hydrops fetalis, and fetal death (49). The single-stranded DNA genome of B19V is packaged into a small, nonenveloped, icosahedral capsid consisting of 60 structural subunits, of which approximately 95% are VP2 (58 kDa) and 5% are VP1 (83 kDa) (17). Two other open reading frames coding for small proteins of unclear functions have also been identified (51, 64).
Parvoviruses have similar strategies for the delivery of the viral genome into the nucleus for replication; however, there are significant differences depending on the specific virus and cell type (18, 25, 42). While some parvoviruses are internalized following interaction with a single receptor, others require complex interactions with receptors and coreceptors before they can be internalized. In addition to clathrin-mediated internalization, caveola-dependent internalization and macropinocytosis have also been described (2, 5). A common feature of the endocytic trafficking of parvoviruses is the requirement for endosomal acidification. However, the timing of virus trafficking through the endosomal pathway, the cellular elements involved, and the sites of escape into the cytosol differ considerably among virus species and cells. Moreover, the mechanisms by which parvoviruses enter the nucleus remain unclear. Since they are small (22 to 25 nm), they can theoretically pass through the nuclear pore complex (NPC) without disassembly. However, an alternative mechanism based on local disruption of the outer nuclear envelope has been proposed (14, 15). For most parvoviruses, the majority of capsids accumulate in a perinuclear location from which the viral DNA is imported into the nucleus either as an intact viral particle (24, 50, 61) or after capsid disassembly (33). Three cell receptors/coreceptors have been identified for B19V: the glycosphingolipid globoside (globotetraosylceramide [Gb4Cer]) (12), the α5β1 integrin (59), and the Ku80 autoantigen (40). Gb4Cer is expressed mainly in the human erythroid progenitor cells in the bone marrow, which are also the main target cells for the virus, and it seems to be the primary attachment receptor. Ku80 might also function in virus attachment in certain cells (40), while the α5β1 integrin is thought to act as a coreceptor required for internalization (58, 59).
The mechanisms of B19V uptake and intracellular trafficking have remained elusive (48). These studies are limited because a well-established cell line system for B19V infection is lacking; thus, it is not possible to propagate the virus to high titers. Therefore, highly viremic plasma from naturally infected individuals without virus-specific antibodies remains the main source of infectious native virus. UT7/Epo cells are commonly used to study B19V infection (30). Although the entry and intracellular trafficking of B19V are not restricted in UT7/Epo cells, the infection is limited to a small number of cells due to intracellular factors induced by erythroid differentiation (23). An improved B19V infection has been described in ex vivo-expanded primary human erythroid progenitor cells (EPCs) (21, 60), where B19V infects a larger number of cells. The improvement operates at the level of replication/transcription and not at the level of entry and trafficking (13); however, the production and egress of infectious progeny is still restricted. Considering that the susceptibility of EPCs to B19V infection is limited to a narrow time frame due to large variations in virus receptor expression (60), UT7/Epo cells provide more stable conditions for the study of entry/trafficking events.
Using complementary approaches, we have studied the early events of B19V infection in UT7/Epo cells. Although B19V and its receptor globoside associate with lipid rafts, the internalization mechanism occurs through clathrin-mediated endocytosis. Viral particles are routed to the lysosome for degradation. Escape into the cytosol depends on low endosomal pH and occurs without apparent membrane damage. Nuclear entry is highly inefficient but can be boosted by chloroquine (CQ), which prevents the degradation of incoming viruses by blocking their transfer to lysosomes. The viral DNA is then efficiently imported into the nucleus, while the capsids remain extranuclear.
UT7/Epo cells were cultured in RPMI medium with 10% fetal calf serum (FCS) and were complemented with 2 U/ml of recombinant human erythropoietin (Epo; Janssen-Cilag, Midrand, South Africa) at 37°C under 7.5% CO2. A B19V-infected plasma sample (genotype 1; CSL Behring AG, Charlotte, NC), without detectable levels of B19V-specific IgM or IgG antibodies, was used as a source of native infectious virus. The virus was pelleted by ultracentrifugation through 20% (wt/vol) sucrose, and the concentration of virions was determined using quantitative PCR.
For the detection of B19V by immunoprecipitation and immunofluorescence, a monoclonal antibody (MAb) against B19V (MAb 860-55D) was kindly provided by S. Modrow (Regensburg, Germany). MAb 860-55D targets a conformational epitope in VP2 and recognizes intact capsids exclusively (22). A rabbit polyclonal antibody (α-VP1u) against VP1u (amino acids [aa] 142 to 163) was obtained as described previously (7). Early endosomes were detected using immunofluorescence with a rabbit polyclonal antibody against EEA1 (Abcam, Cambridge, MA). A mouse MAb against the mannose-6-phosphate receptor and a mouse MAb against Lamp1 (Abcam) were used to detect late endosomes and lysosomes, respectively. Flotillin-1 was obtained from BD Biosciences (San Jose, CA), and caveolin-1 was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Fluorescein isothiocyanate (FITC)-labeled cholera toxin was purchased from Sigma, and FITC-labeled human holotransferrin was purchased from Exbio (Vestec, Czech Republic). Rhodamine-conjugated, lysine-fixable dextran (Mr, 3,000) was obtained from Invitrogen (Carlsbad, CA). All drugs were purchased from Sigma (St. Louis, MO). Ammonium chloride (NH4Cl) was dissolved in water; nystatin and bafilomycin were dissolved in dimethyl sulfoxide (DMSO); and filipin was dissolved in ethanol.
To test drugs for cytotoxicity and specificity, B19V replication was quantified in the presence of increasing doses of drugs added a few hours after internalization. Only doses that showed no effect on B19V replication were used. In addition to ensuring that none of the doses used have an effect on the cellular S phase, which is required for parvovirus replication, the method allows one to rule out possible pleiotropic effects on later stages of the infection not related to virus binding/internalization and trafficking.
The insolubility of lipid rafts in cold nonionic detergents provides the most common method for lipid raft isolation. However, the presence of detergents can produce artifacts or interfere with sphingolipid distribution. Therefore, we used a novel detergent-free method in which rafts are purified by shearing cells in an isotonic buffer with cations, followed by separation along an OptiPrep gradient (34). A total of 5 × 107 UT7/Epo cells were infected or not at 4°C for 1 h and were intensively washed to remove unbound virus. Cells were lysed in 500 μl of detergent-free lysis buffer (1× Tris-buffered saline [TBS], pH 8), in the presence of 1 mM CaCl2 and 1 mM MgCl2 to stabilize the rafts and were supplemented with protease inhibitor cocktail (Complete Mini; Roche). The homogenate was sheared through a 22G by 3-in needle and was centrifuged at 1,000 × g for 10 min at 4°C. The resulting postnuclear supernatant was collected and maintained on ice. The procedure was repeated on the pellet, and the supernatant obtained was combined with the first. The pooled supernatant (225 μl) was mixed with an equal volume of 50% OptiPrep (Sigma) in SW60 ultracentrifuge tubes. The mixture was carefully overlaid with 3 ml of 20% OptiPrep and 675 μl of 10% OptiPrep. The tubes were centrifuged at 38,500 rpm (200,000 × g) for 18 h at 4°C. Acceleration and deceleration rates were set to zero. Fractions of 520 μl each were collected from the top of the tube. An aliquot (20 μl) from each fraction was analyzed by Western blotting with an antibody against the lipid raft component flotillin-1.
UT7/Epo cells were infected with B19V at 2 × 104 viral particles per cell in RPMI medium for 1 to 2 h at 4°C to allow binding, followed by several washing steps to remove unbound virus. After different incubation times at 37°C, the cells were fixed with acetone-methanol (1/1 [vol/vol]) for 5 min at −20°C. Following staining with specific antibodies, the cells were washed and mounted with Mowiol (Calbiochem, La Jolla, CA) containing 30 mg/ml of 1,4-diazabicyclo(2,2,2)octane (DABCO; Sigma) as an antifade agent. Confocal laser scanning microscopy was performed using an LSM 510 Meta laser scanning microscope with an inverted Zeiss microscope (Axiovert 200M; Carl Zeiss AG, Feldbach, Switzerland).
UT7/Epo cells were infected with B19V at 3 × 104 viral particles per cell in RPMI medium without serum for 1.5 h at 4°C. Following a washing step to remove unbound virus, the cells were incubated at 37°C. After 0, 1, 5, 10, or 30 min, the cells were fixed in 2.5% glutaraldehyde. Cells were washed in phosphate-buffered saline (PBS) and were embedded in 3% agarose. The cubes were fixed for 1 h in osmium tetroxide (1% in PBS) on ice, washed, and further incubated for 1 h in a tannin solution (0.05 M HEPES plus 0.01 g tannin) at room temperature (RT). The samples were washed with a sodium sulfate solution (1% in 0.05 M HEPES) and water. A filtrated 2% uranyl solution was added to the samples for 1 h at RT; then the samples were washed with water and were dehydrated in increasing ethanol concentrations. Samples were incubated in Epon, embedded in gelatin capsules, and polymerized for 48 h at 60°C. Capsules were dissolved at 70°C, and 70-nm ultrathin sections were cut, stained with 1% lead citrate and 2% uranyl acetate, and finally analyzed in a Zeiss EM 902 electron microscope equipped with a TRS digital camera.
Cells (3 × 105) in RPMI medium containing 10% fetal calf serum and 2 U/ml of Epo were seeded into 12-well plates and were preincubated with drugs for 30 min at 37°C before infection. In some cases, cells were pulse-treated with drugs for specific times. As controls, drugs were added 3 to 4 h postinfection. Cells were infected with B19V at 5 × 103 genome equivalents per cell and were further incubated at 37°C for 24 h. The cells were transferred to RNase-free tubes (Eppendorf Biopur, Hamburg, Germany) and were pelleted. The pellets were washed twice with PBS and were stored at −20°C until use. Total poly(A) mRNA was isolated using the Dynabeads mRNA direct kit (Invitrogen) according to the manufacturer's instructions. The isolated viral NS1 mRNA was reverse transcribed, and cDNA was quantified. Cells collected at 30 min postinfection served as input controls to define the background signal.
Amplification of B19V DNA or cDNA and real-time detection of PCR products were performed using a LightCycler system (Roche Diagnostics, Rotkreuz, Switzerland) with SYBR green (Roche). PCR was performed using a FastStart DNA SYBR green kit (Roche) according to the manufacturer's instructions. Plasmids containing the genome of B19V were used at 10-fold dilutions as external standards. The number of cells used for each experiment was determined by the quantification of the cellular β-actin gene, as described previously (44).
UT7/Epo cells were infected with B19V, as described above. Cells were lysed in NP-40 lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1% NP-40, and 5 mM EDTA) supplemented with a protease inhibitor cocktail (Complete Mini; Roche). Viral particles were immunoprecipitated with a human MAb against intact capsids (MAb 860-55D). After overnight incubation with 20 μl protein G agarose beads at 4°C, the beads were washed four times (three times with PBS–1% bovine serum albumin and once with PBS) and were resuspended either in protein loading buffer, for analysis of the immunoprecipitated capsids by Western blotting, or in PBS, for quantification of the viral DNA by PCR. Total DNA was extracted using the DNeasy tissue kit (Qiagen) and was quantified as specified above.
To examine whether B19V can induce detectable endosome membrane permeabilization or damage during the escape process, rhodamine-labeled dextrans (Mr, 3,000) were coendocytosed with B19V for 2 h at 37°C. The cells were washed and were further incubated at 37°C in the presence of B19V to maintain a constant flow of incoming viruses in endocytic vesicles. Uninfected cells were used as controls. The capacity of the endosomal vesicles to retain the endocytosed dextrans was monitored by fluorescence microscopy at increasing incubation times.
UT7/Epo cells were infected as specified above. At increasing times postinfection, the cells were washed twice with ice-cold PBS, and the pellets were resuspended in 100 μl EZ buffer before the addition of an additional 900 μl EZ buffer (Sigma). The samples were vortexed and were kept on ice for 5 min; then they were pelleted at 2,400 rpm for 5 min at 4°C. This step was repeated. Pellets were then resuspended in 500 μl EZ buffer containing 0.25 M sucrose and were layered on top of 500 μl EZ buffer with 0.5 M sucrose. The purified nuclei were collected by centrifugation at 5,000 rpm for 10 min at 4°C. The purity and integrity of the isolated nuclei were assessed via light microscopy after trypan blue staining. Nuclear pellets were resuspended in nuclear lysis buffer (10 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, and 1 mM EDTA [pH 7.4]) supplemented with protease inhibitor cocktail (Complete Mini; Roche) and were maintained on ice for 10 min. The homogenate was sheared through a 26G by 1-in needle and was centrifuged at 8,000 × g for 10 min at 4°C. The supernatant was used to quantify the viral DNA or to immunoprecipitate capsids, as specified above.
Lipid rafts are specialized plasma membrane microdomains that are enriched in cholesterol and glycosphingolipids and play roles in membrane receptor dynamics, signal transduction, intracellular trafficking, cell polarization, and cell migration (31). The glycosphingolipid globoside (Gb4Cer) is the receptor of B19V (12). The presence of Gb4Cer in lipid rafts was examined using immunofluorescence staining with FITC-labeled cholera toxin B (CTxB), which binds to the ganglioside GM1, a common component of lipid rafts (62). The expression of GM1 in UT7 cells was irregular; some cells showed little or no detectable expression. Gb4Cer colocalized partially with CTxB in cells expressing GM1 abundantly (Fig. 1A). Like Gb4Cer, B19V partially colocalized with the GM1-CTxB complex (Fig. 1A). Additionally, the distribution of the lipid raft marker flotillin-1 (Flot-1) along OptiPrep density gradients was examined. As expected, in uninfected cells, Flot-1 was found in the light, buoyant membrane fractions, predominantly in fractions 1 and 2. In contrast, when lipid rafts were isolated from infected cells, Flot-1 shifted significantly to denser fractions, suggesting a virus-mediated modification in the structure and/or lipid composition of the rafts (Fig. 1B).
Depending on the presence of the membrane protein caveolin-1 (Cav-1), lipid rafts can be divided into caveolar and noncaveolar rafts (38, 47), which are enriched in caveolin and flotillin proteins, respectively. Gb4Cer had variable colocalization with Cav-1, from extensive colocalization in some cells to modest or no colocalization in others (Fig. 1C). However, B19V rarely colocalized with Cav-1 (Fig. 1C), indicating that the large majority of B19V associates preferentially with noncaveolar lipid rafts.
The infectivity of B19V was analyzed in the presence of nystatin, an antifungal reagent that disrupts cholesterol-enriched microdomains (6, 46). When nystatin was added after virus binding/internalization, none of the doses tested caused significant inhibition. However, when added before virus infection, nystatin caused dose-dependent inhibition of B19V infectivity (Fig. 1D). B19V infectivity was also sensitive to filipin, a sterol-binding agent that binds to cholesterol and disrupts lipid raft formation (1) (Fig. 1D). Like nystatin, filipin had no effect on the infection when added after virus binding/internalization, suggesting that the plasma membrane microdomains play a critical role during the process of B19V entry.
Electron micrographs from B19V-infected cells during the first minutes of incubation at 37°C showed the uptake of B19V capsids into plasma membrane invaginations and vesicles with the characteristic clathrin electrodense coating (Fig. 2A). One single virus particle was sufficient to accomplish the process of internalization into clathrin-coated vesicles. No virus internalization in the typical small flask-shaped caveolar invaginations or in other structures was observed at any time. Furthermore, B19V did not colocalize with the endoplasmic reticulum (ER), which is typically involved in caveola-dependent internalization (29) (data not shown).
The uptake mechanism of B19V was followed using FITC-labeled human holotransferrin, which is typically internalized through clathrin (26). B19V and transferrin did not colocalize during the binding step at 4°C. However, when the temperature was shifted to 37°C, B19V and transferrin colocalized intensively, indicating a common internalization pathway (Fig. 2B). These results, taken together, indicate that clathrin-mediated endocytosis is the primary internalization mechanism used by B19V.
The progression of intracellular capsids was examined using immunofluorescence confocal microscopy at increasing time points after internalization. Cells were washed, fixed, and stained with an antibody against intact capsids (MAb 860-55D) and with antibodies against early endosomes (EE) (EEA1), late endosomes (LE) (mannose-6-phosphate receptor), lysosomes (Lys) (Lamp1), recycling vesicles (RV) (Rab 11), the Golgi apparatus (58K Golgi protein), or the ER (GRP78 BIP). By 5 min postinternalization, B19V capsids were detectable in early endosomes, and by 30 min, capsids were observed throughout the endocytic pathway (Fig. 3A). The colocalization signals from an average of 40 cells were quantified with BioImage XD software (27). Maximal colocalization with LE occurred by 10 min postinternalization, and maximal colocalization with lysosomes occurred by 30 min (Fig. 3B). At later times postinfection, colocalization with both LE and Lys decreased progressively. By 1 h, capsids were detected in lysosomes, but only a few were detected in the LE. By 2 to 3 h postinternalization, capsids were clearly less abundant and appeared scattered in the cytoplasm, and with the exception of a few capsids detectable in lysosomes, colocalization with EE or LE was no longer observed. No significant colocalization of incoming particles with recycling vesicles, the Golgi apparatus, or the ER was observed at any time point (data not shown). The lack of colocalization with any relevant organelle marker suggests that following endocytic trafficking, a proportion of incoming viruses escapes into the cytosol.
Acidification inside the endosomes is required by certain viruses as part of their infectious entry route. All parvoviruses analyzed to date require endosomal acidification for the infection process (18, 25, 42). To investigate the requirement of B19V for a low endosomal pH, cells were treated with either bafilomycin A1 (BafA1), a selective inhibitor of vacuolar ATPases, or the lysosomotropic weak base ammonium chloride (NH4Cl). These compounds raise the pHs of intracellular compartments, which can be restored upon removal of the inhibitory substances (11, 36). UT7/Epo cells were pulse-treated for different times with BafA1 or NH4Cl. Although their mechanisms of action are different, BafA1 and NH4Cl treatments resulted in remarkably similar reductions in B19V infectivity when applied at the beginning of the infection (Fig. 4A). However, BafA1 and NH4Cl had no significant effect on infectivity when added 1 h postinternalization, suggesting that most of the infectious particles had escaped from the endocytic vesicles.
The effect of BafA1 was further investigated using immunofluorescence. By 5 h postinfection (p.i.) in untreated cells, intact capsids did not colocalize with endocytic markers (Fig. 4B). Furthermore, no colocalization was observed with markers of recycling endosomes (Rab 11 or transferrin), caveolin-containing vesicles, the Golgi apparatus, or the ER (data not shown), indicating that these capsids may be in the cytosol. However, in the presence of BafA1, the capsids were retained inside late endocytic elements, notably in the lysosomes (Fig. 4B). These observations suggest that endosomal acidification is not essential for the progression of the virus along the endocytic pathway or, in particular, for the transfer of the incoming viruses from late endosomes to lysosomes but is required by B19V for escape into the cytosol.
During the process of entry, B19V is routed to the lysosomes, reaching maximal colocalization by 30 min to 1 h after internalization (Fig. 3). To determine whether particles become degraded in this compartment, the integrity of the incoming viral DNA was examined using quantitative PCR. As shown in Fig. 4C, from 1 to 5 h postinfection, significant degradation of the viral DNA was observed. The reduction was not due to virus detachment from the receptor or recycling, since no increase in the level of viral DNA was observed in the supernatant of the infected cells from 1 to 5 h postinfection (Fig. 4D). These observations indicate that during the process of viral entry, a proportion of incoming capsids are routed to the lysosomes for degradation.
At later times postinfection, intracellular capsids did not colocalize with markers of the endocytic pathway, including recycling endosomes (Rab 11 or transferrin), and no significant colocalization with caveolin-containing vesicles, the Golgi apparatus, or the endoplasmic reticulum was observed at any time (data not shown). This observation would suggest that the capsids had escaped into the cytosol. The mechanism by which these capsids reach the cytosol is unknown. Increasing evidence indicates that the phospholipase A2 (PLA2) activity of VP1u plays a role by disrupting the endosomal membrane (63). We have examined whether B19V can induce detectable permeabilization or damage in the endocytic membranes by monitoring the capacity of the endosomal vesicles to retain small (Mr, 3,000) coendocytosed dextran molecules. Dextrans were retained in the endocytic vesicles, without detectable cytosolic or nuclear staining (Fig. 5A and andB).B). The lack of detectable dextran leakage would suggest that endosomal escape occurs either without membrane permeabilization/disruption through a yet unidentified mechanism or from only a few vesicles, resulting in undetectable leakage.
Parvoviruses deliver their genomes into the nucleus for replication. In general, this step has been found to be inefficient in all parvoviruses analyzed (25). We have shown previously that incoming minute virus of mice (MVM) accumulates and persists in the lysosomal compartment without noticeable cytosolic capsids and that incoming nuclear viral DNA or proteins remain undetectable even after the onset of viral DNA replication and RNA transcription in the nucleus (35). To examine the presence of incoming B19V DNA in the nucleus, total DNA was extracted from purified nuclei at increasing postinternalization times in untreated and BafA1-treated cells. A sample taken 10 min postinternalization was used to define the background signal. As shown in Fig. 6A, no increase over the background was observed in untreated or BafA1-treated cells, even after the onset of viral transcription in the nucleus (Fig. 6B).
The presence of viral particles in the nucleus was examined by confocal microscopy. Viral capsids appeared scattered throughout the cytoplasm, and although some particles were in close contact with the nuclear envelope and nuclear invaginations, they were not observed inside the nucleus (Fig. 6C). Immunoprecipitation of capsids from total cells or from isolated nuclei further confirmed that viral capsids did not enter the nucleus in detectable quantities (Fig. 6D).
In previous studies, we have shown that chloroquine (CQ) enhances B19V infection in UT7, HepG2, and primary bone marrow mononuclear cells (8). To understand the mechanism underlying the boosting effect of CQ, UT7/Epo cells were pulse-treated for different times with CQ. As shown in Fig. 7A, when added at the time of the infection for 1 h, CQ had a modest inhibitory effect on the infection. However, when added 1 h after the infection, CQ was beneficial for the infection, and this effect was more significant with increasing time in the presence of the drug. This observation would suggest that immediately after internalization, endosomal acidification is beneficial for the infection, and that the effect of CQ can be attributed to a subsequent process.
We next analyzed the efficiency of B19V nuclear entry in the presence of CQ. In sharp contrast to no treatment or BafA1 treatment, CQ boosted nuclear import. The viral DNA accumulated in the nuclei and reached a plateau by 3 h postinfection (Fig. 7B). Comparison of the amount of viral DNA from complete cells and from isolated nuclei at 3 h postinfection revealed that the viral DNA associated with the nuclei represents a significant fraction of the total intracellular viral DNA (Fig. 7C). Although more than half of the incoming viral DNA was imported into the nuclei, viral capsids were not detectable inside the nucleus (Fig. 7D). Immunoprecipitation of viral capsids from total cells or from isolated nuclei confirmed that viral capsids remained extranuclear (Fig. 7E and andF).F). These results would suggest either that viral DNA is imported into the nucleus from capsids that are immediately uncoated or that uncoating takes place prior to the import of the DNA into the nucleus.
The increase of virus nuclear import by CQ could be explained through the particular effects of CQ on endosomal vesicles. Although CQ, BafA1, and NH4Cl have similar effects on endosomal pH, CQ has the ability to destabilize endocytic membranes, leading to vesicle swelling. The intense lysosomal dysfunction caused by CQ results in a deficient transfer of cargo to the degradative lysosomes, promoting escape from a prelysosomal vesicle and avoiding degradation in lysosomes (28, 37). This ability has been routinely used to improve the efficiency of transfection experiments (32). Using confocal microscopy, we confirmed that in CQ-treated cells, but not in untreated or BafA1-treated cells, lysosomes appeared severely enlarged (Fig. 8A). The extensive vacuolization of endocytic vesicles did not induce detectable leakage of endocytosed dextrans (Fig. 8B). Confocal microscopy images taken 5 h p.i. confirmed the swelling of the endocytic vesicles, the accumulation of intact capsids in late endosomes, but not in lysosomes, and the lack of viral DNA degradation (Fig. 8C and andD).D). Although the viral DNA was not degraded, the number of intact capsids decreased significantly during the entry process (Fig. 8E). Therefore, this result cannot be attributed to degradation, as is the case for untreated cells; the more likely cause is the uncoating of the incoming virus.
Lipid rafts are small, heterogeneous, cholesterol- and glycosphingolipid-enriched domains that play important roles in cellular processes such as membrane signaling and trafficking (31). They are also involved in multiple stages of the virus life cycle, such as attachment, internalization, uncoating, protein transport, assembly, and budding (53, 55). Several pieces of evidence indicate that B19V exploits lipid rafts during the process of entry: (i) colocalization of B19V and its receptor globoside (Gb4Cer) with the lipid raft marker GM1 (Fig. 1A), (ii) a shift of flotillin migration toward nonraft fractions in infected cells (Fig. 1B), and (iii) inhibition of B19V infection following lipid raft disruption (Fig. 1D). The fact that lipid raft disruption had no effect a few hours after infection indicates that plasma membrane rafts are important for the infectious entry/trafficking of B19V and not for later steps. The exact mechanism by which these membrane microdomains contribute to virus entry is not known. Preliminary results indicate that lipid raft disruption does not inhibit virus attachment (data not shown). Gb4Cer is required but is not sufficient for B19V infection. Other receptors have been shown to be important in the process of viral entry (12, 40, 58, 59). Thus, lipid rafts may act as platforms for the concentration of virus receptors/coreceptors required for B19V infection, as has been shown for other viruses (55).
Clathrin-dependent endocytosis is the default uptake mechanism for parvoviruses (3, 5, 20, 41, 54). To date, only adeno-associated virus 5 (AAV-5) and porcine parvovirus use additional routes based on caveolae and macropinocytosis, respectively (2, 5). Since caveolae are integral parts of some lipid rafts and B19V associates with rafts, a possibility of caveola-mediated internalization of B19V into UT7/Epo cells could be envisioned. Although Gb4Cer was associated with both caveolar and noncaveolar rafts, B19V associated preferentially with noncaveolar rafts (Fig. 1C). By means of electron microscopy, B19V was recurrently observed in clathrin-coated invaginations and vesicles (Fig. 2A); however, viruses in flask-shaped invaginations characteristic of caveolae were not observed. While at the attachment step B19V did not colocalize with transferrin, which is internalized through clathrin-mediated endocytosis (26), extensive colocalization was observed during the internalization step (Fig. 2B). In contrast to the slow internalization by caveolae, cargo internalized by clathrin-mediated endocytosis is quickly delivered to early endosomes (39). Consistent with a rapid internalization mechanism, immunofluorescence pictures taken 5 min p.i. confirmed the presence of B19V capsids in early endosomes (Fig. 3A). Therefore, although B19V interacts with plasma membrane rafts, internalization occurs by clathrin-mediated endocytosis and does not involve caveolae. This mechanism of internalization based on lipid rafts and clathrin has been observed in other viruses (19).
We have found that BafA1 and NH4Cl inhibited viral infection by blocking viral escape, resulting in the accumulation of viral particles in the degradative lysosomes. However, CQ, which also alkalinizes the endosomes, enhances the infection. We have shown previously that CQ enhances B19V infection in UT7, HepG2, and primary bone marrow mononuclear cells (8). The case of HepG2 cells is particularly striking. These cells are considered nonpermissive for B19V infection (10); however, in the presence of CQ, HepG2 cells support B19V infection (8). The reason for this enhancement can be explained, at least in part, by the particular effects of CQ on endosomal vesicles. CQ causes vesicle swelling and hampers the fusion of endosomes and lysosomes, preventing the transfer of endocytosed material to the degradative lysosomal compartment (28, 37). Because of these particular properties, CQ is frequently used in transfection experiments to increase transduction efficiency (32). We confirmed that CQ, but not BafA1, induced the vacuolization of endocytic vesicles (Fig. 8A) and prevented the transfer of the incoming capsids to the degradative lysosomes (Fig. 8C). We could also confirm that CQ, but not BafA1, prevented the degradation of incoming B19V DNA (Fig. 8D). The capsids retained in vacuolated prelysosomal vesicles would profit by progressively escaping into the cytosol.
Although the effect of CQ on endosomal vesicles is independent of the cell type, only B19V, and no other parvovirus, can benefit (4, 20, 44, 50). All parvoviruses studied to date exploit endosomal acidification for capsid modifications required for subsequent steps, primarily the externalization of N-VP1 and its constitutive PLA2 domain, which is thought to be required for endosomal escape (35, 50, 56). In addition, nuclear localization signal (NLS) sequences have been identified in the N-VP1 proteins from some parvoviruses, which might assist in the transport of capsids toward the nucleus (57). We have recently shown that B19V is unique among parvoviruses in that N-VP1 becomes externalized upon receptor binding (7, 9, 45). Therefore, in contrast to other parvoviruses, B19V would not depend on low endosomal pH for this critical conformational rearrangement. However, B19V would require acidification to facilitate the process of endosomal escape, for example, by promoting interactions between the PLA2 domain of N-VP1 and the endocytic membranes.
At increasing postinternalization times, incoming capsids appeared dispersed throughout the cytoplasm and did not colocalize with endocytic markers, the ER, caveolin-containing vesicles, the Golgi apparatus, or recycling vesicles, suggesting that they had escaped into the cytosol. The PLA2 activity of N-VP1 is thought to play a role in endosomal escape; however, the mechanism is not known. Adenovirus is able to release coendocytosed dextrans of different sizes, implying a mechanism of escape based on endosomal disruption. Human rhinovirus 2 is able to release only small dextrans, suggesting a less invasive mechanism of escape based on pore formation (43). The escape mechanism of canine parvovirus (CPV) has been studied by cointernalization of alpha-sarcin, where no disruption of endosomes was found (41). In other studies, however, CPV was able to release rhodamine-labeled dextrans of 3 kDa but not of 10 kDa (52). However, the effect was evident only 20 h postinfection, while endosomal escape of CPV occurs earlier. In addition, a parvovirus capsid would not be able to escape through a pore that selectively allows the escape of dextrans of 3 kDa but not of 10 kDa. In the case of B19V, there was no detectable leakage of dextrans of 3 kDa at any time or in the presence of CQ (Fig. 5 and and8B).8B). Two possibilities can be envisaged: either capsids permeabilize only a minor amount of endocytic vesicles or capsids escape through a yet unidentified mechanism, which does not involve membrane damage.
Viral capsids (Fig. 6C and andD)D) or viral DNA (Fig. 6A) was not detected inside the nucleus, even after the onset of viral transcription (Fig. 6B). Thus, B19V is inefficiently imported into the nucleus, making it difficult to study this step of the infection. In clear contrast, in CQ-treated cells, B19V DNA was efficiently imported into the nucleus, reaching a plateau by 3 h postinfection (Fig. 7B). At this time, more than half of the total viral DNA was found associated with the nuclear fraction (Fig. 7C). Examination of the incoming nuclear DNA confirmed that the viral DNA was not associated with capsids, which remained extranuclear (Fig. 7D, ,E,E, and andF).F). In CQ-treated cells, the number of intact capsids decreased significantly (Fig. 8E). Since the viral DNA remained stable (Fig. 8D) and a large proportion was found in the nucleus (Fig. 7B and andC),C), the decrease in the number of intact capsids cannot be attributed to degradation, as was the case for untreated cells. These results would suggest either that the viral DNA is imported into the nucleus from intact capsids that are immediately uncoated or that uncoating takes place prior to the import of the DNA into the nucleus. The second possibility seems more plausible, since no capsids were found in the nucleus at any time point.
Our studies provide the first insight into the early steps of B19V infection and reveal mechanisms involved in virus uptake, endocytic trafficking, and nuclear import. This study also outlines novel questions that warrant further investigation, such as the precise role of lipid rafts in the process of virus entry, the mechanism by which B19V escapes from endosomes without detectable permeabilization/damage, and the pathway involved in the nuclear import of viral DNA.
We are grateful to S. Modrow and J. Lindner (Regensburg, Germany) for kindly providing MAb 860-55D. The skillful technical assistance of R. Eberle (Paul-Ehrlich-Institute, Langen, Germany) in EM sample preparation is gratefully acknowledged.
Published ahead of print 20 June 2012