![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article 
Journal of Virology, November 2006, p. 11040-11054, Vol. 80, No. 22
0022-538X/06/$08.00+0 doi:10.1128/JVI.01056-06
Copyright © 2006, American Society for Microbiology. All Rights Reserved.
German Cancer Research Center, Infection and Cancer Research Program,1 Cell and Tumor Biology Research Program, Im Neuenheimer Feld 242, D-69120 Heidelberg, Germany2
Received 23 May 2006/ Accepted 28 August 2006
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Infection of cultured cells by adeno-associated virus type 2 (AAV2) is initiated by binding to heparan sulfate proteoglycan as an attachment receptor followed by interaction with either human fibroblast growth factor receptor 1, integrin 
Vß5, or hepatocyte growth factor receptor (35, 50, 62, 63). After receptor binding, AAV2 enters the host cell via clathrin-coated vesicles in a dynamin-dependent process (2, 15) and is transported to a perinuclear vesicle compartment (1, 2, 16, 27, 31, 55, 73). Acidification of the endosomal compartment appears to be important for AAV2 infection, since treatment of cells with bafilomycin A1 significantly, albeit cell type dependently, decreased infection by wild-type (wt) virions or AAV gene delivery vectors (2, 14, 28, 47). It is generally believed that AAV2 exits from early or late endosomes before it delivers its genome to the nucleus. The catalytic activity of a phospholipase A2 (PLA) domain located on the minor capsid protein VP1 is required for infection with AAV2 at a postentry step (21), and based on recent findings with autonomous parvoviruses, canine parvovirus (CPV) and minute virus of mice (MVM), this activity is needed for endosome release (20, 61). The PLA domain is buried inside the capsid structure of AAV2 (37, 38) and has to become exposed during infection to develop its catalytic activity. Proteasome inhibitors significantly increase gene transduction with AAV2 (12, 14, 16, 23, 34, 75, 76); however, the mechanism of inhibitor action remains unclear, although ubiquitylation and ubiquitin-dependent degradation of AAV2 and AAV5 capsids have been demonstrated (16, 75). Several reports claim that intact AAV particles enter the cell nucleus (2, 29, 55, 57, 73), possibly by a nuclear pore-independent route (29, 73). Detailed analysis of nuclear and perinuclear localization of viruses by optical sectioning, however, suggested that this is a rather inefficient process and genome delivery might not require whole capsid entry into the nucleus (41).
The AAV capsid has a diameter of approximately 25 nm and is composed of 60 subunits of the viral proteins VP1, VP2, and VP3 in an approximate stoichiometric ratio of 1:1:10. The three coat proteins are derived from the same open reading frame but differ in their N-terminal extensions due to different expression start sites (4, 8, 64). The capsid structure of AAV2 has been resolved by electron cyromicroscopy and X-ray crystallography (38, 74). The outer surface is characterized by 3 elongated spikes surrounding the threefold axes and pores at the 12 fivefold axes. The N termini of VP1 and VP2 which are not resolved in the X-ray crystal structure are not accessible to antibodies or to trypsin digestion on intact capsids but can become exposed in vitro by limited heat treatment (5, 37). Genetic and structural data suggest that the N termini of the three capsid proteins—including the VP1 and VP2 N termini—are located at the twofold symmetry axes inside the capsid and traverse through the pores at the fivefold symmetry axes upon heat treatment (5, 37). Similar observations and interpretations were made for autonomous parvoviruses (7, 11, 19, 51, 67). Accessibility of VP1 N termini for antibodies during viral entry has been shown for autonomous parvoviruses (42, 53, 61, 67).
The VP1 N termini not only harbor the PLA catalytic domain (21, 77) but also three short basic amino acid sequence elements (basic clusters [BCs]) with the character of nuclear localization signals (NLS). Two of these elements (BC2 and BC3) are also located on VP2 N termini. These sequences are conserved among all AAV serotypes analyzed so far. It has been suggested that BC3 is able to act as an NLS (33). Homologous sequences present in VP1 N termini of CPV and MVM are required for virus infection (40, 67), possibly by targeting the viral genome to the cell nucleus. Although a role of the BCs in nuclear targeting of parvoviral genomes is very suggestive, direct evidence that they are necessary and even sufficient for nuclear transport of virions is lacking.
In this study we intended to analyze the activation and function of the PLA and the three BC elements located on the VP1 and VP2 N termini during AAV2 infection. We show that the BC elements are required for infection and become exposed on the capsid surface, probably during passage of the virion through the endosomes. The capsids are released with exposed N termini into the cytoplasm and transfer the genomes into the nucleus while retaining these structural characteristics until the final uncoating reaction occurs. BC1 and BC2 showed nuclear transport activity on a heterologous protein, while BC3 failed to do so.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Plasmids and site-directed mutagenesis.
The pTAV2.0 construct (32) contains the entire AAV2 genome from pAV-2 (39), including both inverted terminal repeats cloned into the BamHI site of Bluescript II. The N335A, L336A, and V221Y constructs are based on pTAV2.0 with the designated mutations in the cap open reading frame (5). The construct HD/AN, with two mutated residues in the catalytic center of the PLA2 domain, has been described previously (21). Plasmid pDG
VP is based on plasmid pDG (26) and expresses the AAV2 Rep and all essential adenovirus helper proteins but does not express the AAV2 VPs (17). In plasmid pDM, the rep-cap open reading frame of plasmid pDG has been replaced by the full-length AAV2 genome (37). Plasmid pJ407 (36), containing the BamHI-NotI fragment of pTAV2.0 cloned into pUC131, was used as a template for site-directed mutagenesis. Mutagenesis was performed using the QuickChange site-directed mutagenesis kit (Stratagene, Amsterdam, The Netherlands) according to the manufacturer's manual. For each mutation, two complementary PCR primers were designed, coding for the indicated substitutions (see Fig. 1C), flanked on each side by 20 to 23 homologous base pairs. Mutant plasmids were verified by DNA sequencing. Fragments containing the mutations were then subcloned into the pTAV2.0 backbone.
|
Preparation of full AAV2 virions.
293T cells (2.3 x 108) were seeded in a 6,360-cm2 Cell Stack culture chamber (Corning, Schiphol-Rijk, The Netherlands) 24 h prior to transfection of 1,525 µg pDM for wt AAV2 production. Three hundred ninety-two micrograms of plasmid BC1–2–3–, carrying a triple BC mutation, or plasmid HD/AN and 1,525 µg of helper plasmid pDGVP were transfected for production of the different mutants. Cells were incubated for 48 to 72 h at 37°C, harvested, washed once with phosphate-buffered saline (PBS) (18.4 mM Na2HPO4, 10.9 mM KH2PO4, 125 mM NaCl), and lysed by three rounds of freezing-thawing in 20 ml lysis buffer (150 mM NaCl, 50 mM Tris, pH 8.5). Lysates were digested with Benzonase (50 U/ml) (Sigma, Munich, Germany) for 30 min at 37°C. Cell debris was removed by centrifugation at 3,700 x g for 20 min, and the supernatant was loaded onto an iodixanol step gradient (7 ml of 15% iodixanol in PBS-MKN [1 mM MgCl2, 2.5 mM KCl, 0.75 M NaCl] and 5 ml of 25% iodixanol, 4 ml of 40% iodixanol, and 4 ml of 60% iodixanol in PBS-MK [1 mM MgCl2, 2.5 mM KCl]) in Beckman Quickseal tubes (25 by 89 mm) (78). Samples were centrifuged for 2 h at 50,000 rpm (50.2 Ti rotor) at 4°C. Subsequently, the 40% iodixanol fraction containing the virus was collected and mixed with 1 volume of PBS-MK. Then, virus was further purified by heparin affinity chromatography (10 ml heparin-agarose) (78). Full virions were enriched on a continuous iodixanol gradient (48). Therefore, the elution fraction of the heparin affinity chromatography (in PBS-MK, 1 M NaCl) was mixed with 1 volume of 60% iodixanol, 1.8 M MgCl2, and centrifuged for 8 h at 70,000 rpm and 15°C (70.1 Ti rotor) in Beckman Quickseal tubes (16 by 76 mm). Fractions (0.5 ml each) were collected from the bottom of the tube, and fractions containing predominantly full virions were pooled based on their refraction index. Buffer exchange to PBS and concentration of the sample were performed in a Vivaspin 30 concentrator (Vivascience, Goettingen, Germany). Electron microscopy analysis of negatively stained virus stocks proved that approximately 95% of the particles were full virions. Virus preparations were stored in small aliquots at –80°C.
Viral protein synthesis and virus titration. For analysis of viral protein synthesis, identical portions of harvested cells were processed for Western blot analysis. Monoclonal antibodies 303.9 and B1, specific for the Rep and VP proteins, respectively, were used as described previously (70). Capsid titers, genomic titers, and infectious viral titers were determined as described previously (25).
Heparin-binding assay. Binding of virions to heparin was determined as described previously (36).
PLA2 assay. PLA2 activity of heat-treated virions (iodixanol step gradient purified) was determined as described previously (5).
Detection of exposed VP N termini. For detection of exposed N termini by fluorescence microscopy, 6 x 104 HeLa cells/well were seeded in 24-well plates containing coverslides. After 48 h, the cells were incubated for 30 min at 4°C with 25,000 particles (purified by iodixanol step gradient)/cell in serum-free DMEM, 20 mM HEPES. After washing of the cells three times in cold medium, temperature was shifted to 37°C. The cells were fixed in 100% methanol for 10 min (–20°C) at the designated time points and subsequently washed with PBS. For signal enhancement, a tyramide signal amplification kit (Molecular Probes, Leiden, The Netherlands) was used according to the manufacturer's manual. First, endogenous peroxidase activity was quenched by incubation with 1.5% H2O2 for 15 min. Then the slides were blocked in tyramide signal amplification blocking reagent for 30 min prior to incubation with antibody A69 or B1 (69) overnight at 4°C. Following several washes with PBS, goat antimouse peroxidase antibody (1:200) was applied for 45 min and the coverslides were washed again. After that, tyramide working solution (1:150) was incubated for 5 min. Confocal sections of 0.5 µm were obtained with a Leica DM IRBE laser scanning microscope. Images were processed using Adobe Photoshop CS software.
For detection of N termini by a native immuno-dot blot assay, 1.7 x 106 HeLa cells were seeded in 10-cm dishes. After 24 h, the cells were incubated with 10,000 wild-type (wt) or mutant AAV2 particles (purified by iodixanol step gradient)/cell for 30 min at 4°C in serum-free medium. The cells were washed once and incubated at 37°C for the designated time periods. Then, the cells were harvested in PBS, pelleted at 2,000 x g for 4 min, and resuspended in 500 µl PBS-MK. After four freeze-thaw cycles, cell debris was removed by centrifugation at 2,000 x g for 5 min. The virus-containing supernatant was transferred to Protran nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany) using a vacuum blotter. Membranes were blocked for 1 h in PBS containing 10% skim milk powder (blocking solution) and then incubated for 1 h with monoclonal antibodies A1, A69, B1, or A20 (69-71) diluted 1:10 in blocking solution. Membranes were washed several times with PBS and incubated for 1 h with a peroxidase-coupled goat antimouse antibody (Dianova, Hamburg, Germany). After several washes with PBS, the antibody reaction was visualized using an enhanced chemiluminescence detection kit (Amersham, Braunschweig, Germany). Signal intensity was quantified with Image Quant TL V2003.02 software. To correct for antibody background reaction, the signal intensity of noninfected cells was subtracted.
For detection of N termini after in vitro heat treatment, a total of 5 x 109 iodixanol step gradient-purified capsids in PBS were heated at various temperatures for 3 min. Then, samples were cooled to 4°C, transferred to nitrocellulose membranes, and processed further as described above.
Lysosomotropic drugs. HeLa cells were preincubated for 30 min in medium containing 50 nM bafilomycin A1 (Sigma) or 50 µM chloroquine (Sigma). During infection and further incubation, the cells were maintained permanently in the presence of the drugs.
Early viral gene expression after infection. HeLa cells (7 x 104/well) were seeded in 24-well plates containing coverslides 24 h before infection. The cells were infected with full wt AAV2, the HD/AN mutant, or theBC1–2–3– mutant (17 genome-containing particles/cell). In some experiments, wt AAV2 was heat treated at 65°C for 3 min and then cooled to 4°C directly before infection. After superinfection with Ad5 (MOI = 4) and incubation for 20 h at 37°C, cells were fixed with 100% methanol (7 min, –20°C) and 100% acetone (3 min, –20°C), air dried, and incubated for 1 h with monoclonal antibody 76.3 directed against the Rep proteins (69). Then, secondary chicken antimouse antibody coupled to Alexa 488 (Molecular Probes) was applied for 1 h. Cell nuclei were visualized with 4',6'-diamidino-2-phenylindole (DAPI), and the specimens were embedded in Permafluor mounting medium (Beckman Coulter, Marseille, France). The percentage of Rep-positive cells was quantified using a Leica DM RBE fluorescence microscope, evaluating at least 200 cells, and particle/infectivity ratios were calculated.
Neutralization assay. wt AAV2 particles (2.5 x 108) were incubated with 100 µg/ml purified antibody A1, A69, B1, or A20 (71) in a total volume of 5 µl (PBS) for 30 min at room temperature. Then, the samples were diluted in 1 ml medium and added in combination with Ad5 (MOI = 4) to 1.4 x 105 HeLa cells grown overnight on coverslides. Rep expression was analyzed after 20 h as described above.
Subcellular localization of virus particles. HeLa cells (6 x 104/well) were seeded in 24-well plates containing coverslides. After 24 h, the cells were incubated for 45 min at 4°C with 500,000 particles (iodixanol step gradient purified)/cell in serum-free medium. Then, the cells were washed three times in cold medium and the temperature was shifted to 37°C. The cells were fixed at the designated time points with 2% paraformaldehyde for 15 min. Quenching was carried out twice with 50 mM NH4Cl for 5 min, followed by a permeabilization step with 0.2% Triton X-100 for 10 min. Primary antibodies directed against capsids (A20; mouse) and Lamin B (goat) (Santa Cruz Biotechnology, Santa Cruz, Calif.) were applied overnight at 4°C. Then cells were incubated for 1 h with chicken antimouse Alexa 594 and chicken antigoat Alexa 488 secondary antibodies (Molecular Probes). Confocal images were taken by H. Spring with a Zeiss LSM 510 META laser scanning microscope (0.3-µm sections). Several images were further examined according to the nuclear uptake of particles, using the deconvolution module of the Zeiss software.
Microinjection settings. Microinjection experiments were performed with a Zeiss AIS automated injection system equipped with an Eppendorf 5242 injector system. Capillaries were prepared from GC120TF-10 glass tubing (Harvard Bioscience, Edenbridge, United Kingdom) on a P87 capillary puller (Sutter Instruments, Novato, Calif.). Throughout this study, capillaries with an inner diameter of 0.4 to 0.5 µm were used. HeLa cells (1.25 x 105) were seeded in 35-mm dishes containing coverslides 48 h before injection in DMEM-20 mM HEPES. The average injection volume was approximately 200 fl for the cytoplasm and 50 fl for the nucleus. A minimum of 100 cells was injected per experiment.
Microinjection of peptide-protein conjugates. The fluorescently labeled peptide-protein conjugates were obtained from Peptide Specialty Laboratories (Heidelberg, Germany). Briefly, the peptides were synthesized containing a C-terminal cysteine residue for the chemical coupling and purified by reverse-phase high-performance liquid chromatography. Bovine serum albumin (BSA) or ovalbumin was labeled with fluorescein by incubation with carboxyfluorescein succinimidyl ester prior to the coupling of the peptides in order to prevent a reaction of the dye with the basic amino acids of the peptides. Then, the peptides were conjugated to BSA or ovalbumin using the cross-linking agent sulfosuccinimidyl 4-(N-maleimidolmethyl)cyclohexane-1-carboxylate. Free maleimide groups were saturated with cysteine, followed by extensive dialysis against PBS. The labeled peptide-protein conjugates were injected at a concentration of 1.0 to 1.5 mg/ml into the cytoplasm of HeLa cells. Two hours after injection, the cells were fixed in 2% paraformaldehyde for 15 min and washed two times with PBS. Confocal images (0.5-µm sections) were obtained with a Leica DM IRBE laser scanning microscope and further processed using Adobe Photoshop CS software.
Microinjection of antibodies and detection of early gene expression. Control goat immunoglobulin G (IgG) antibody (Dianova) or purified monoclonal antibody A1, A69, B1, A20, C24-B, or C37-B (71) was injected into the cytoplasm or the nucleus of HeLa cells at a concentration of 5 mg/ml (in PBS). Monoclonal antibody (MAb) A20 belongs to the IgG3 subclass and is weakly detectable with standard secondary antibodies. Therefore, A20 was injected in combination with goat IgG (3 mg/ml) as an injection marker. Nuclear injection was performed using low pressure (maximum, 100 hPa) and a long injection time (approximately 1 s) in order to reduce the proportion of antibody which was lost in the cytoplasm when traversing this compartment. We have estimated that this leakage was less than 10% of the total amount injected into the nucleus. In a control experiment, the antibodies were injected at a 1:10 (0.5 mg/ml) or 1:5 (1.0 mg/ml) dilution into the cytoplasm. After injection, the cells were washed, infected with AAV2 (MOI = 20) and Ad5 (MOI = 4), and incubated for 20 h. Then, the cells were fixed with paraformaldeyde and permeabilized as described above, followed by incubation with a rabbit anti-Rep polyclonal serum and tetramethyl rhodamine isocyanate-labeled antirabbit secondary antibody (Dianova) to detect infected cells. Injected cells were visualized with Alexa 488-labeled secondary antibody directed against mouse or goat IgG (Molecular Probes). The percentage of Rep-positive injected cells and also the percentage of Rep-positive noninjected cells in the same culture were quantified using a Leica DM RBE fluorescence microscope.
Microinjection of virus and detection of early gene expression. Antibody C37-B (200 µg/ml) was added to the medium to neutralize infection of the cells with virus leaked into the medium during injection. Full wt and mutant AAV2 virions (3.5 x 1011 genome-containing particles/ml) (in PBS) were injected in combination with goat IgG antibody (3 mg/ml) as an injection marker into the cytoplasm or nucleus of HeLa cells. In some experiments, virus was heat treated at 65°C or 70°C for 3 min and then cooled to 4°C directly before injection. The cells were extensively washed and infected with Ad5 (MOI = 4). After 20 h, the cells were fixed with paraformaldeyde and permeabilized as described above, followed by incubation with a rabbit anti-Rep polyclonal serum and tetramethyl rhodamine isocyanate-labeled antirabbit secondary antibody (Dianova) to visualize infected cells. Injected cells were stained with an Alexa 488-labeled secondary antibody (Molecular Probes) directed against the goat IgG marker. The percentage of Rep-positive injected cells was quantified using a Leica DM RBE fluorescence microscope, and the particle/infectivity ratios were calculated.
Subcellular localization of microinjected virus. Cells were preincubated for 1 h with monoclonal antibody C37-B (50 µg/ml) to neutralize infection with virus leaked into the medium during injection and with the proteasome inhibitor MG-132 (30 µM) (Merck, Bad Soden, Germany) to prevent intracellular degradation by the proteasome. Then, purified AAV2 (iodixanol step gradient plus heparin agarose) with a titer of 1 x 1013 particles/ml (in PBS) was injected in combination with goat IgG antibody (3 mg/ml) as an injection marker into the cytoplasm or nucleus of HeLa cells. In one experiment, the virus was heat treated at 60°C for 3 min to externalize the VP N termini before injection. The cells were washed, infected with Ad5 (MOI = 5), and incubated for 8 h in the permanent presence of MG-132. After fixation in paraformaldehyde and permeabilization as described above, antibody A20 and Lamin B antibody (Santa Cruz Biotechnology) were applied overnight at 4°C. Then, the samples were incubated for 1 h with chicken antimouse Alexa 488 (directed against A20) and chicken antigoat Alexa 594 (directed against both Lamin B and the goat IgG injection marker) (Molecular Probes). Confocal images (0.16-µm sections) were obtained with a Leica DM IRBE laser scanning microscope and further processed using Adobe Photoshop CS software.
DNA release. HeLa cells (1.5 x 106) were seeded in 10-cm dishes. After 24 h, the cells were infected for 1 h at 37°C with 5,000 particles (purified by iodixanol step gradient)/cell in serum-free medium. The cells were harvested in PBS at the indicated time points, resuspended in 1 ml RYMO buffer (10 mM Tris, pH 7.5, 200 mM sucrose, 2 mM magnesium acetate, 0.1 mM EDTA, 1% NP-40) containing protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany), left on ice for 20 min, and centrifuged at 1,000 x g for 10 min at 4°C. The supernatant was designated as the cytoplasmic fraction. The pellet was washed once with RYMO buffer and resuspended in 1 ml RIPA buffer (150 mM NaCl, 10 mM Tris, pH 7.5, 0.1% sodium dodecyl sulfate, 1% NP-40, 1% sodium deoxycholate, 1 mM EDTA) containing protease inhibitor cocktail, which was designated as the nuclear fraction. Samples were incubated at 4°C for 2 h with gentle inversion and cleared by centrifugation at 12,000 x g for 5 min at 4°C. Immunoprecipitation of capsids was performed with protein A-Sepharose CL-4B (Amersham Pharmacia) in NET-N buffer (20 mM Tris, 100 mM NaCl, 1 mM EDTA, 0.5% NP-40, pH 7.5). Three milligrams/sample of protein A-Sepharose was added to 600 µl of antibody A20, and the mixture was incubated with gentle inversion for 2 h at 4°C. The protein A-Sepharose-antibody mixture was washed twice with NET-N buffer to remove unbound antibody. Aliquots in NET-N buffer were added to 300 µl of cytoplasmic or nuclear fractions, and the mixtures were incubated for 2 h at 4°C with gentle inversion. The samples were washed three times with NET-N buffer and resuspended in 200 µl MNase buffer (10 mM Tris, pH 8.0, 1 mM CaCl2), of which 100 µl was either incubated or not with MNase (300 U/ml) (Roche Diagnostics) for 3 h at 37°C. Subsequently, 100 µl of 2x proteinase K digestion buffer (20 mM Tris, 20 mM EDTA, 1% sodium dodecyl sulfate, pH 7.5) and proteinase K (320 µg/ml) (Roche Diagnostics) were added, and the mixtures were incubated for 3 h at 37°C. The DNA was extracted with phenol-chloroform and concentrated by ethanol precipitation. The DNA pellet was resuspended in 0.4 M NaOH-10 mM EDTA and then transferred to GeneScreen Plus nylon membranes using a vacuum blotter. DNA samples were hybridized with a 32P-labeled rep-specific probe, generated from a rep fragment (SalI) of plasmid pTAV2.0. Autoradiography was carried out with Kodak Biomax MS X-ray film.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Exposure of VP1/VP2 N termini during infection. Because the PLA and BC domains are buried inside the capsid structure of the virion (37, 38) and they are obviously required for infection, they should become externalized during infection to develop their PLA and potential nuclear targeting activities. In HeLa cells infected with wt AAV2, VP1/VP2 N termini were detectable already 1 h postinfection (p.i.) by indirect immunofluorescence using a VP1/VP2 N-terminus-specific antibody (A69) (71) (Fig. 2A). The fluorescence signals were first visible as a few patches in the cytoplasm, distant from the nucleus, which increased with time and accumulated in a perinuclear compartment, confirming the observations made with CPV and MVM (42, 53, 61). Quantification of externalization of N termini by native immuno-dot blot analysis of freeze-thaw lysates obtained at different time points p.i. showed that externalization was already detectable with MAb A1 (detecting only the VP1 unique N terminus [71]) and A69 after 30 min and reached a plateau between 4 and 6 h p.i. (Fig. 2B). The quantity of capsids detected with antibody A20 (69) remained stable during this time, and the VP protein C terminus (analyzed with MAb B1) became not detectable at all. The increase of the A1/A20 and A69/A20 signal ratios suggest that the reaction with N-terminus-specific antibodies occurred on capsids and not on disassembled capsid proteins.
|
Since AAV2 enters the host cell by receptor-mediated endocytosis into endosomal vesicles (2, 15), the time course of VP1/VP2 N-terminal externalization suggests that the conformational change occurs in the endosomes. Therefore, we asked whether the low endosomal pH is required for the structural rearrangement of the capsid leading to the exposure of the N termini of VP1 and VP2. We treated the cells with chloroquine or bafilomycin A1, which are thought to raise the endosomal pH (3, 6, 56). Both drugs efficiently inhibited infection of HeLa cells with AAV2, as tested by the appearance of Rep-positive cells 20 h p.i. (Fig. 3A) in accordance with previous publications (14, 28, 47). Exposure of VP1/VP2 N termini was measured by the native immuno-dot blot assay using freeze-thaw lysates harvested at different time points after infection. The ratio of the reactions with MAbs A1 and A20 or A69 and A20 in the presence and absence of the drugs was calculated to quantify the exposure of VP1/VP2 N termini throughout the course of infection. While bafilomycin A1 treatment inhibited exposure of VP1/VP2 N termini by 40 to 60% (Fig. 3E and F), treatment with chloroquine transiently increased the presentation of N termini on the capsid up to 2 h p.i. and then reduced it up to 6 h p.i. (Fig. 3C and D). This effect was more pronounced when analyzed with MAb A1, whose epitope is more N-terminally located than the epitope of MAb A69, which reacts with both capsid proteins VP1 and VP2 (71). At higher chloroquine concentrations (100 µM), this transient increase in N-terminal exposure was not observed. Capsids remained stable during the time of analysis; only treatment with bafilomycin A1 generated some capsid destruction after 3 h of incubation (Fig. 3B).
|
VP1/VP2 infection-relevant domains are exposed in the cytoplasm. It is assumed that AAV2 enters the cytoplasm before it delivers its genome to the nucleus. However, this step has never been directly documented, and the question remains of whether the VP1/VP2 BC elements are exposed in the cytoplasm to possibly function there as nuclear targeting signals. To investigate this question, we microinjected HeLa cells with our panel of N- and C-terminus-specific VP protein-recognizing antibodies (A1, A69, and B1) and with the capsid-detecting antibody A20 and infected the cells with AAV2 (MOI = 20) and Ad5 (MOI = 4). Cells were fixed 20 h later, and successful infection was monitored by immunostaining for Rep proteins. The N-terminus-specific antibodies A1 and A69 injected into the cytoplasm inhibited infection nearly completely (Fig. 4B). Also, A20 caused a similar dramatic reduction of Rep-positive cells when injected into the cytoplasm, showing that intact virions pass through the cytoplasm with exposed VP1/VP2 N termini before they deliver the genome to the nucleus. Goat IgG and MAb B1 had no effect in the cytoplasm, indicating that antibody injection per se did not inhibit AAV2 infection. For a control, we also tested the capacities of the different antibodies to neutralize infection when applied extracellularly. To this end, we preincubated them with wt AAV2 virus before infection and counted the Rep-positive cells 20 h p.i. (Fig. 4A). As expected, A1, A69, and B1 did not neutralize infection after preincubation, because N and C termini of the capsid proteins are not accessible on intact AAV2 virions, whereas A20 preincubation blocked infection as shown previously (71). These results show that passage of the virus through the cytoplasm is an obligatory step in AAV2 infection and the BC elements are presented at the capsid surface in this compartment.
|
|
|
|
Final uncoating of AAV2 occurs in the nucleus. As initially mentioned, it remains an open question whether the viruses detectable in the cell nucleus (2, 41, 55, 57, 73) represent the infectious entity and bring about AAV2 infection. In particular, in light of the recently described observation that a substantial amount of MVM genomes is already released in a cytoplasmic compartment, alternative mechanisms of nuclear transport of the parvovirus genomes should be considered (42). To test whether the genome of AAV enters the cell nucleus protected within the capsid, we injected capsid-specific antibodies (5 mg/ml) into the nucleus and infected the cells subsequently with wt AAV2 and Ad5. Successful infection was monitored by immunofluorescence detection of Rep protein synthesis. Nuclear injection of three different capsid-specific antibodies (A20, C24-B, and C37-B) nearly completely prevented AAV infection, whereas injection of unrelated goat IgG or MAb B1 did not inhibit infection (Fig. 8A). To our surprise, nuclear injection of the N-terminus-specific antibodies A1 and A69 also prevented rep gene expression. Leakage of antibody into the cytoplasm during nuclear injection cannot be excluded, although low pressure and a long injection time were applied to reduce this amount to a minimum. We have tested that this leakage would be less than 10% of the amount injected into the nucleus (Fig. 8C). After injection of antibody into the nucleus at concentrations used for inhibition of infection (5 mg/ml), leakage into the cytoplasm was barely detectable, while antibody injected into the cytoplasm at 5- or 10-fold-lower concentrations could easily be seen by immunofluorescence detection. To control for the effect of a small amount of antibody in the cytoplasm, we injected a 1:10 dilution of each antibody into the cytoplasm (for A20 a 1:5 dilution also was tested; data not shown) to see whether it could inhibit infection. As shown in Fig. 8B, this amount of the different MAbs did not inhibit infection at all. These experiments strongly support the conclusion that the infectious AAV genome is transported into the nucleus within the capsid.
|
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Endosomal processing and release. The VP1 N terminus of parvoviruses harbors infection-relevant protein domains (21, 40, 67, 77). Several studies have shown that these VP1 N termini—and in the case of AAV, also the VP2 N termini—are located inside the capsid, with the exception of the N termini of B19 (11, 38, 53, 54, 61, 67, 77). In vitro studies using limited heat treatment or incubation with urea demonstrated that the parvovirus virion exists in a metastable state and can expose the VP1 N termini on the capsid surface to present protein domains involved in viral trafficking (11, 37, 53, 61, 67). Combined genetic and structural data indicate that these sequences become exposed through the channels at the fivefold axes of symmetry (5, 18, 19, 51). Here we demonstrate that the exposure of VP1/VP2 N termini of AAV2 also occurs during cell infection. Analysis of capsid pore mutants further shows that in vivo several mutants with exchanges of channel-forming amino acids are impaired in the presentation of the N termini, in accordance with the in vitro data. The time course and localization of exposure of the N termini during infection suggest that this conformational rearrangement occurs in the endosome, which is in agreement with observations made with the autonomous parvovirus MVM (42, 53). Accessibility of CPV N termini to antibodies, however, was also observed after cytoplasmic injection (67), suggesting that low pH is not necessary for externalization. Treatment of cells with bafilomycin A1, a drug that prevents endosomal acidification, partially inhibited externalization of capsid protein N termini. This was also observed with MVM (42) but not with CPV (61), arguing for the interpretation that acidification influences the presentation of N termini in vivo but that it is not necessary. In addition, one should keep in mind that bafilomycin A1 not only prevents lowering of the endosomal pH but also influences Ca2+ influx, ligand-receptor dissociation, or formation of endosomal carrier vesicles and fusion of late endosomes with lysosomes (9, 10, 30, 49, 65). In vitro low-pH treatment of AAV2 did not lead to exposure of N termini up to pH 5. At pH 3 and to some extent at pH 4, VP1/VP2 N termini could be detected on capsids; however, the C termini also were exposed under these conditions (data not shown), which was not the case in vivo. These data together support the conclusion that low pH achieved under physiological conditions also is not sufficient to induce the conformational change in the AAV capsid. Treatment of cells with chloroquine, a drug that buffers intracellular endosome acidification (56), did not prevent exposure of N termini. Even a transient stimulation of presentation of N termini at early time points p.i. could be observed, followed by a reduction of reaction with antibody A1, while reaction with antibody A69 remained largely preserved. This could be due to a dynamic retraction of exposed sequences (18) or to degradation of the N termini, leading to a loss of antibody epitopes. The latter explanation is supported by the fact that the more N-terminally located epitope of A1 disappeared with time, whereas the A69 epitope located more downstream of the N terminus remained more or less completely accessible. If the N termini were retracted, one would expect the opposite result. Degradation of VP1 N termini at a neutral pH and preservation of VP1 N termini at a low pH is in line with in vitro observations with MVM (18). This means that detection of N termini in acidic compartments might also be a consequence of protection of a VP1 domain against degradation and not necessarily the manifestation of low-pH-triggered externalization. Taken together, the physiological trigger which leads to the conformational rearrangement of the AAV capsid remains unknown.
Detection of VP1 N termini during infection with N-terminus-specific antibodies has previously been shown on the basis of immunofluorescence (42, 53, 61). Colocalization of N termini with capsids, however, does not allow distinguishing between colocalization of capsids with disassembled capsid proteins (which would present the N termini) and capsids on which the N termini are exposed. Our biochemical assay using a dot blot under nondenaturing conditions showed quantitatively that VP1 N termini became accessible to antibodies while the reaction with the capsid-specific antibody remained constant, suggesting that the N termini were exposed on the capsid.
Nuclear targeting of AAV2 capsids. Although nuclear localization of AAV2 capsids has been described several times (2, 29, 55, 57, 73), there exists some controversy as to the efficiency of this process. A recent report showed that after detailed optical sectioning, viral capsids could be detected in the nucleus only rarely (41), leading to the conclusion that nuclear targeting of AAV2 capsids is a very inefficient process. We were interested in whether the basic clusters of the VP1/VP2 N termini act as NLSs and permit nuclear uptake of the virus and infection. Three prerequisites for such a role of the BC elements were given: (i) the BC elements were exposed when the virus entered the cytoplasm (Fig. 4); (ii) some BC elements targeted a nonnucleophilic protein (BSA) to the nucleus (Fig. 5); and (iii) mutation of the BC elements reduced viral infectivity (Fig. 1). These observations support the interpretation that the BC elements mediate nuclear uptake of AAV virions. Very recently, further evidence has been presented that BC3 acts in nuclear transport of AAV genomes, based on a difference in nuclear accumulation of recombinant AAV genomes packaged into wt or BC3 mutant capsids shown by in situ hybridization (24). However, our attempts to directly show the role of the BC elements in nuclear transport of the virus failed in three experiments, probably due to the low efficiency of the nuclear transport process. We could not observe incoming virus with a capsid-specific antibody in the nucleus after deconvolution of fluorescence signals above and below the nuclei. Also, cytoplasmically microinjected virus—with or without exposed N termini—remained exclusively in the cytoplasm. This was also observed with Alexa 568-labeled recombinant AAV2 (13). Even amplification of the nuclear signal by replication and rep gene expression was negative when the virus was injected into the cytoplasm and the cells were infected with Ad5. These results may be explained by the low number of viral particles which were able to interact with the nuclear transport machinery due to either limited release from the endosomes or the accessibility for degradation by the proteasome in the cytoplasm in the infection experiment. Also, the large size of the capsids might reduce transport through the nuclear pore complexes. Additionally, the punctuate distribution of microinjected virus in the cytoplasm might indicate AAV particles sticking to cytoplasmic structures, leading to cytoplasmic retention. Such cytoplasmic retention could explain why microinjection of rather high concentrations of 104 to 105 viral particles/cell (which might saturate such retention sites) led to strong nuclear accumulation of CPV and NS1 expression (60, 67). Failure to induce infection with CPV after cytoplasmic injection in earlier experiments is in line with our observation and may be explained by lower concentrations of virus used for injection in these studies (66). Navigation of AAV in vesicles to the nuclear periphery could facilitate nuclear uptake of virus by shortening the ways through the cytoplasm where nonspecific binding and/or proteasomal degradation might interfere with successful infection.
Uncoating of AAV2 in the nucleus. So far it has not been shown that the AAV particles detected in the nuclei are the infectious entities. Moreover, the observation that VP1 externalization and viral DNA release of MVM already occurred simultaneously in the endosomes suggested that the endosomal pathway may be the site of uncoating (42). Similarly, Lux et al. (41) concluded that uncoating probably occurs before or during nuclear entry of the genome. This conclusion was based on the observation that AAV2 DNA was detectable in the nucleus, while almost no capsids were detectable, and that not all perinuclear DNA signals coincided with capsid signals. In contrast to these interpretations, our studies provide evidence that viral capsids transfer the genome into the nucleus, where final uncoating occurs. This evidence is based on inhibition of AAV infection by injection of two types of monoclonal antibodies into the nucleus. The first type of antibodies consists of three different capsid-specific antibodies, for which it has been shown that they react with assembled capsids and not with denatured or dissociated capsid proteins (71). We reexamined the capsid specificities of these antibodies by sucrose density gradient fractionation of cytoplasmic extracts of cells infected with AAV2 and Ad5. The cytoplasm of such infected cells is enriched for nonassembled capsid proteins which sediment in different oligomeric states (70). We could not see a reaction of these antibodies with any of these fractions, while they clearly detected assembled capsids run on parallel gradients (data not shown). This means that the injected antibodies are specific for capsids and do not react with completely or partially disassembled capsid proteins. Also, the inhibition of infection by injection of the second type of antibodies, which react only with the VP protein N termini (A1 and A69), supports this interpretation. Interaction with free or oligomeric capsid proteins in complex with the genome would be less likely to lead to a complete inhibition of AAV2 rep gene expression than to precipitation of virions on which the N termini are exposed. Further support for the interpretation that uncoating can occur in the cell nucleus comes from the detectable gene expression from virions microinjected directly into the nucleus. These virus preparations were nuclease treated and were highly purified by iodixanol step gradient, heparin affinity chromatography, and a continuous iodixanol gradient. That means they should not contain free virus DNA or damaged virus. We consider the possibility that a small amount of viral DNA is released due to sheering of the virions during injection to be very unlikely because of the high stability of the virus and the relatively large diameter (approximately 500 nm) of the injection capillary. Finally, the observation that BrdU-labeled single-stranded DNA could be detected in infected cells only after denaturation can be interpreted by the assumption that the DNA is protected within the capsid (31).
It is an intriguing question how disassembly and assembly of the parvovirus capsid occur in the same compartment, namely, the nucleus. A solution to this paradox could provide the observation that genome release does not necessarily require disassembly. In vitro heat-treated parvoviral capsids are able to release their DNA from the intact capsid (5, 11, 52). In the case of AAV2, this is accompanied by the appearance of the antibody B1 epitope on the capsid surface (5). However, a signal with the B1 antibody could not be detected during the first 6 h of AAV2 infection. A reaction with B1 was observed only late in infection (16 to 20 h p.i.) in a cytoplasmic compartment, presumably the lysosome (data not shown). Nevertheless, it is possible that the DNA exits through the same pore at the fivefold symmetry axis where it entered the capsid, and the number of capsids which release their DNA is below the detection level of our B1 antibody reaction. This is in line with the DNA release experiments showing no vast increase in the DNase sensitivity of the genome during infection. Processing of the capsid in endosomal, cytoplasmic, and nuclear compartments might prepare the capsid for genome release.
AAV infection pathway. The data presented in this paper modify and supplement the picture of the AAV2 entry pathway. After uptake of AAV into endosomal vesicles, the virus undergoes a conformational change which leads to the exposure of VP1/VP2 N termini. This process is not simply triggered by a lowered pH, although lysosomotropic drugs have an influence on release or presentation of the N termini. Thereafter, the virus obligatorily enters the cytoplasm to proceed with the infection process. Transport of the genome into the nucleus is a major limiting step for AAV2 infection and requires probably additional modifications of the AAV capsid during the endosomal route or transport to perinuclear sites. Virus injected into the cytoplasm could not lead to productive infection even in the presence of Ad5. Only a few viruses, not directly visible with the techniques we applied, enter the nucleus and release their genome possibly without capsid disassembly. To our knowledge, AAV is the first virus for which functional evidence exists that uncoating of the genome might occur in the nucleus. The factor(s) that lead to exposure of N termini as well as to genome release have still to be investigated.
|
ACKNOWLEDGMENTS |
|---|
F.S. was supported by a grant from the Deutsche Krebshilfe, 10-1912-Kl I.
|
FOOTNOTES |
|---|
Published ahead of print on 6 September 2006.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
