JOURNAL OF VIROLOGY, July 2008, p. 6492–6500 0022-538X/08/$08.00⫹0 doi:10.1128/JVI.00557-08 Copyright © 2008, American Society for Microbiology. All Rights Reserved.
Vol. 82, No. 13
Neutralizing Antibody Blocks Adenovirus Infection by Arresting Microtubule-Dependent Cytoplasmic Transport䌤 Jason G. Smith,1 Aurelia Cassany,2† Larry Gerace,2 Robert Ralston,3 and Glen R. Nemerow1* Department of Immunology and Microbial Science, The Scripps Research Institute, 10550 N. Torrey Pines Road, La Jolla, California 920371; Department of Cell Biology, The Scripps Research Institute, 10550 N. Torrey Pines Road, La Jolla, California 920372; and Department of Virology, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, New Jersey 070333 Received 12 March 2008/Accepted 17 April 2008
Neutralizing antibodies are commonly elicited by viral infection. Most antibodies that have been characterized block early stages of virus entry that occur before membrane penetration, whereas inhibition of late stages in entry that occurs after membrane penetration has been poorly characterized. Here we provide evidence that the neutralizing antihexon monoclonal antibody 9C12 inhibits adenovirus infection by blocking microtubuledependent translocation of the virus to the microtubule-organizing center following endosome penetration. These studies identify a previously undescribed mechanism by which neutralizing antibodies block virus infection, a situation that may be relevant for other nonenveloped viruses that use microtubule-dependent transport during cell entry. the penton base (41). The first step in virus disassembly, dissociation of the fiber, occurs at or near the plasma membrane (25). Upon acidification in early endosomes, additional uncoating events occur, including dissociation of the vertices from the rest of the capsid and release of protein VI, which has been recently identified as a capsid component with membrane-lytic activity (12, 42). Protein VI likely plays a central role in endosome disruption mediated by the partially uncoated capsid. Upon escape from the endosome, the remaining capsid, containing the nucleoprotein core and viral genome, undergoes microtubule-dependent transport toward the nucleus (11). In a poorly understood process, the capsid then docks with the nuclear pore complex, allowing the viral genome to be imported into the nucleus (10, 30, 43). As has been shown for a number of viruses, the host immune response is known to interfere with several steps of the HAdV entry pathway. Recently, we described the mechanism of action of defensins, antimicrobial peptide effectors of the innate immune system that block HAdV uncoating, including protein VI release, and escape from the endosome (31). HAdV also elicits a potent humoral immune response, and considerable effort has been devoted to characterizing this response as well as mechanisms of antibody-mediated neutralization. Antihexon neutralizing antibodies (NAbs) constitute the majority of the neutralizing activity of antisera, although antifiber and anti-penton base NAbs likely also play a role (8, 14, 34, 45). While the ability of antibodies to block early steps in virus entry has been well-described, their role in blocking late steps in entry is largely unknown. For HAdV, a recently described antihexon monoclonal NAb (MNAb) appears to block a postentry step in infection (39). This MNAb, designated 9C12, has been shown by cryo-electron microscopy (cryo-EM) studies to enmesh the virus in a partially ordered lattice of molecules that cross-links the majority of the hexon trimers. The 9C12 MNAb does not aggregate virus particles at neutralizing concentrations or induce measurable conformational changes in hexon, nor does the hexon cross-linking appear to occlude the
Human adenovirus (HAdV) is a nonenveloped, doublestranded DNA virus with a genome of approximately 36 kDa. The virion is comprised of an icosahedral capsid surrounding the viral nucleoprotein core. The capsid contains 240 hexon capsomers, and each of the 12 vertices is anchored by a penton capsomer comprised of five copies of the penton base protein that are associated noncovalently with a homotrimeric fiber protein. In addition to these major constituents, the capsid contains minor proteins IIIa, VI, VIII, and IX, which likely stabilize the capsid (29), and the genome is associated with the nonstructural proteins V, VII, mu, and terminal protein. There are 51 recognized serotypes of HAdV divided into six species (HAdV-A to -F). HAdVs are common pathogens that invade the respiratory, gastrointestinal, ocular, and excretory systems and cause generally nonfatal, self-limiting infections in immunocompetent patients; however, these viruses can cause severe and sometimes fatal diseases in immunocompromised patients, including AIDS and transplant patients (9, 16). In addition, they are the most common cause of acute febrile respiratory disease among military recruits (17, 24). Due to their ability to efficiently infect a broad range of cell types, several HAdV serotypes have been studied extensively as gene transfer vectors. In addition, HAdV vectors elicit potent innate and adaptive immune responses, making them attractive vaccine candidates. HAdV infection is initiated by fiber binding to a primary attachment receptor (e.g., coxsackievirus and adenovirus receptor or CD46) on the target cell membrane (49). Internalization via clathrin-coated vesicles is promoted by the interaction between cellular integrins and a conserved RGD motif in
* Corresponding author. Mailing address: 10550 N. Torrey Pines Rd. IMM19, La Jolla, CA 92037. Phone: (858) 784-8072. Fax: (858) 784-8472. E-mail:
[email protected]. † Present address: CNRS MCMP UMR 5234, Universite´ Bordeaux 2, 146 rue Le´o Saignat, 33076 Bordeaux Cedex, France. 䌤 Published ahead of print on 30 April 2008. 6492
VOL. 82, 2008
POSTENTRY NEUTRALIZATION OF ADENOVIRUS INFECTION
integrin binding sites on the penton base. Consistent with these observations, immunofluorescence studies demonstrated that the MNAb/virus complex was internalized. Confocal microscopy studies also showed that at least some of the 9C12/virus complexes could be detected adjacent to the nuclear margin. However, the previous studies did not provide direct evidence that 9C12/virus complexes were able to penetrate the endosome or to be transported to the nuclear pore. The new studies presented here demonstrate that the MNAb 9C12 does not block the vertex dissociation step or inhibit capsid escape from the endosome. Rather, 9C12 MNAb blocks infection at the stage of microtubule-dependent transport. To our knowledge, this is the first demonstration of this type of antibody-mediated virus neutralization. MATERIALS AND METHODS Cell lines and reagents. Tissue culture reagents were obtained from Invitrogen (Carlsbad, CA). Human A549, HeLa, and 293 cells and normal rat kidney (NRK) cells were propagated in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (Omega Scientific, Tarzana, CA), 10 mM HEPES, pH 7.55, 4 mM L-glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, and 0.1 mM nonessential amino acids. Antibodies. Isolation and production of the 9C12 and 27F11 anti-HAdV-5 hexon MAbs were described previously (39). The antihexon MAb 8C4 was obtained from Fitzgerald Industries International, Inc. (Concord, MA). Fluorescein isothiocyanate (FITC)-conjugated antitubulin MAb (DM1A) was from Abcam, Inc. (Cambridge, MA). 9C12 MAb was conjugated to Cy3 monoreactive dye (GE Healthcare, Piscataway, NJ), following the manufacturer’s instructions. Labeling of 9C12 did not affect its ability to inhibit HAdV-5 infection (data not shown). Ad5.eGFP production and fluorescent labeling. The replication-defective HAdV-5 vector used in these studies had E1/E3 deleted and contained a cytomegalovirus promoter-driven enhanced green fluorescent protein (eGFP) reporter gene cassette. The temperature-sensitive mutant of HAdV-2, HAdV-2ts1 (ts1), was obtained from Joseph Weber (University of Sherbrooke, Quebec, Canada). HAdVs were propagated in 293 cells and purified by CsCl density gradient as previously described (47). The ts1 mutant was propagated at the nonpermissive temperature of 39.5°C as previously described (42). Viral protein concentration was determined as previously described (31). Purified virus was labeled with Cy3 monoreactive dye (GE Healthcare) following the manufacturer’s instructions. Free dye was removed by dialysis against 40 mM Tris, 150 mM NaCl, 2 mM MgCl2, 10% glycerol, pH 8.1. Aliquots of labeled virus were snap-frozen in liquid nitrogen and stored at ⫺80°C. Dye incorporation was determined by UV spectroscopy. Typical dye/capsomer ratios were between 1.5 and 3. The titer of labeled virus on infection of A549 cells was compared to that of mock-labeled virus, and only labeled virus with a titer within 95% of control was used for subsequent experiments. HAdV thermostability assay. Ad5.eGFP, 9C12 MAb, and 27F11 MAb were dialyzed against 10 mM HEPES, 100 mM NaCl, pH 7.4. Ad5.eGFP (25 g/ sample) was incubated with 9C12 or 27F11 (600 nM, or a threefold molar excess of antibody) for 1 h in 50 mM HEPES, 100 mM NaCl, pH 7.4, plus complete, non-EDTA protease inhibitors (Roche) at 4°C. Parallel samples (200 l) were incubated for 10 min at the indicated temperatures, cooled to room temperature (RT), and loaded onto discontinuous gradients consisting of 250 l of 40% nycodenz and 200 l of 80% nycodenz in 130 mM NaCl, 50 mM HEPES pH 7.4. Gradients were centrifuged at 209,000 ⫻ g (average) for 1.5 h at 4°C using an SW55ti rotor with adaptors (Beckman) in ultraclear centrifuge tubes. The visible virus band was collected, reduced with dithiothreitol, boiled, and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The gel was stained with Simply Blue (Invitrogen). HAdV-mediated endosomalysis. The ␣-sarcin ribotoxin coentry assay was performed as previously described (31) with the following modifications. Ad5.F5 or ts1 (100 ng/ml) was incubated with or without 10 g/ml 9C12 or 27F11 MAbs and 0.1 mg/ml ␣-sarcin (EMD Chemicals, Inc., San Diego, CA) in Dulbecco’s modified Eagle’s medium for 1 h at RT. This concentration of 9C12 is approximately a 40-fold molar excess of antibody at the highest concentration of virus tested and approximately 200-fold higher than the concentration required for maximal inhibition of Ad5.eGFP infection (39) (data not shown). Before being added to
6493
cells, the virus-antibody-sarcin mixtures were serially diluted into medium containing the same concentrations of ␣-sarcin and antibody such that only the virus concentrations were diluted. Fluorescence microscopy. Ad5.eGFP (1 ⫻ 109 particles/sample) was incubated with or without 1 g/ml Cy3-labeled 9C12 in medium for 45 min on ice. A549 cells plated on glass coverslips were washed twice with cold medium and placed on ice. Virus or virus/antibody was added (50 l/coverslip), and samples were incubated at 4°C. After 45 min, samples were washed twice with cold medium. Immediately after washing or following a 90-min incubation at 37°C, cells were fixed in PHEMO buffer [68 mM piperazine-N,N⬘-bis(2-ethanesulfonic acid) (PIPES), 25 mM HEPES, 15 mM EGTA, 3 mM MgCl2, 10% dimethyl sulfoxide] containing 3.7% paraformaldehyde, 0.05% glutaraldehyde, and 0.5% Triton X-100 for 10 min at RT, washed once with PHEMO buffer, quenched with 50 mM NH4Cl in phosphate-buffered saline for 10 min at RT, and blocked in 2% bovine serum albumin (BSA) in Tris-buffered saline. Cells incubated with virus alone were stained with a mixture of Cy3-9C12 and DM1A-FITC (␣-tubulin). Cells incubated with Cy3-9C12-bound virus were stained with DM1A alone. Samples were incubated with primary antibodies in 2% BSA in Tris-buffered saline for 45 min at RT, stained with 1 M 4⬘,6⬘-diamidino-2-phenylindole (DAPI; Sigma) in water, and mounted with Prolong Gold antifade (Invitrogen). z-series images were acquired with a DeltaVision optical sectioning microscope (model 283; Applied Precision, LLC, Issaquah, WA). The acquired images were deconvolved using DeltaVision softWoRx version 2.5. The z-profiles were created with Image J (NIH Imaging), and these images were analyzed by first determining the threshold for signal over background and then for colocalization above this threshold using LSM Examiner software (Zeiss, Inc., Peabody, MA). To determine the effect of CRM1 inhibition on HAdV infection, cells were preincubated in medium containing 20 nM leptomycin B (LMB; Sigma) for 30 min at 37°C and infected, washed, and incubated with medium containing 20 nM LMB. Samples were fixed and stained as above except that all samples were stained with both Cy3-9C12 and DM1A. HAdV association with microtubules. Cy3-Ad5.eGFP (6 ⫻ 109 particles/sample) was incubated with or without a 20-fold molar excess of 9C12 or control 8C4 or 27F11 MAbs (both control antibodies gave equivalent results) for 1 h at 4°C in the presence of complete EDTA-free protease inhibitors (Roche) in general tubulin buffer (GTB; 80 mM PIPES, 1 mM MgCl2, 1 mM EGTA, pH 7). Microtubules were polymerized from a 5:4 ratio of purified bovine brain tubulin to biotinylated, purified bovine brain tubulin and stabilized with 20 M taxol following the manufacturer’s instructions (all reagents from Cytoskeleton, Inc., Denver, CO). Virus/antibody complexes were incubated with or without taxolstabilized microtubules (10 g/sample) and a purified microtubule-associated protein fraction (MAPF; 2 g/sample; Cytoskeleton) in the presence or absence of 500 mM NaCl in GTB for 45 min at RT. A small aliquot of each sample was removed to measure input hexon signal. Immobilized streptavidin beads (Pierce, Rockford, IL) were blocked with 20 mg/ml BSA in GTB. BSA (10 mg/ml) and BSA-blocked beads were added, and samples were incubated for 45 min at RT. Beads were washed four times in GTB, and bound proteins were eluted by boiling in gel loading buffer containing dithiothreitol. Samples were separated by SDS-PAGE, and the gel was imaged for Cy3 fluorescence on a Typhoon 9410 imager (GE Healthcare). Hexon fluorescence was quantified with ImageQuant TL software (GE Healthcare) and normalized to input hexon. The value for hexon bound to beads alone was subtracted from each sample, and the net value was compared to virus incubated with microtubules and MAPF but without antibodies or NaCl (positive control). Importin  expression and purification. Human importin  (5) with N-terminal His and S tags and a C-terminal myc tag was cloned into the pET30 expression plasmid. His-S-importin  was expressed in Escherichia coli strain BL21. Protein expression was induced with 1 mM isopropyl -D-thiogalactopyranoside. After incubation for 4 h at 30°C, cells were collected and sonicated in lysis buffer (50 mM Tris pH 8, 250 mM NaCl, 2 mM MgCl2, 1 mg/ml lysozyme, and 1 mg/ml each of aprotinin, leupeptin, and pepstatin). The lysate was cleared by centrifugation in a Beckman Ti45 rotor (100,000 ⫻ g for 30 min at 4°C). His-S-importin  was affinity purified with Talon beads (Clontech, Mountain View, CA) and eluted with imidazole. Purified protein was precipitated with 65% ammonium sulfate overnight at 4°C. The precipitate was dissolved in transport buffer (20 mM HEPES pH 7.4, 110 mM potassium acetate, 2 mM magnesium acetate, 2 mM dithiothreitol, and 1 mg/ml each of aprotinin, leupeptin, and pepstatin), and the cleared supernatant was purified on a MonoQ column (Pfizer, New York, NY) and eluted with a 0 to 1 M NaCl linear gradient in transport buffer. Fractions containing His-S-importin  were pooled and desalted using a PD10 column (GE Healthcare) equilibrated with transport buffer. The purified protein was frozen in liquid nitrogen and stored at ⫺80°C.
6494
SMITH ET AL.
J. VIROL.
FIG. 1. 9C12 binding does not inhibit HAdV-mediated endosomalysis. (A) 9C12 binding does not stabilize the virus capsid for vertex dissociation. Capsid-associated proteins from Ad5.eGFP samples incubated at the indicated temperatures in the presence of the neutralizing 9C12 antihexon antibody or a control, nonneutralizing antihexon antibody were analyzed by SDS-PAGE and Simply Blue staining. Bands corresponding to hexon (H), penton base (PB), fiber (F), and proteins IIIa, V, VI, and VII are indicated. Note the equivalent loss of vertex proteins (bold). A labeled molecular mass standard is in the left lane. Data are representative of three independent experiments. (B) [35S]Met incorporation into cellular proteins was measured as a function of Ad5.eGFP concentration upon infection of A549 cells in the presence or absence of the neutralizing 9C12 antihexon antibody or a nonneutralizing antihexon antibody (control) and the ribotoxin ␣-sarcin. The ts1 mutant of HAdV-2 (ts1) was included as a control. Data are the mean percentages of [35S]Met incorporation compared to control uninfected cells for triplicate samples ⫾ standard deviations and are representative of three experiments.
Hexon binding to the nuclear envelope. Hexon from 293 cells infected with Ad5.eGFP was purified by fast-performance liquid chromatography using a MonoQ column (Pfizer) as previously described (44), desalted on a PD10 column (GE Healthcare) equilibrated with transport buffer, and concentrated with a Microcon device (Millipore, Billerica, MA). Hexon was conjugated to Alexa Fluor 488 dye (Invitrogen) following the manufacturer’s instructions. Labeled hexon was separated from free dye by size exclusion chromatography. Labeled hexon (1.3 g) was incubated with 9C12 or 8C4 MAb (fivefold molar excess) or purified His-S-importin  (fivefold molar excess) in transport buffer for 1 h on ice. NRK cells on glass coverslips were washed with transport buffer and permeabilized with 0.005% digitonin in transport buffer for 5 min at RT. Permeabilized cells were washed with cold transport buffer, incubated for 15 min at 30°C, and washed extensively with cold transport buffer to remove cytosol. Hexon samples were incubated with permeabilized cells for 30 min at 30°C. Cells were washed in transport buffer, fixed with 3.7% formaldehyde in phosphate-buffered saline, stained with 1 M DAPI, and mounted with SlowFade antifade (Invitrogen). Images were acquired using a Bio-Rad 1024 laser scanning confocal microscope and Bio-Rad Lasersharp 2000 software.
RESULTS 9C12 MNAb binding does not block HAdV-mediated endosome penetration. The initial report describing the mechanism of 9C12 MNAb inhibition of HAdV infection suggested that MAb binding did not alter virus internalization; however, the ability of the antibody to impact virus escape from the endosome was not determined (39). In previous studies, we established a close correlation between capsid thermostability, as determined in an assay that mimics virus uncoating in the endosome, and the ability of the virus to penetrate the endosome (31, 42). In this assay, virus uncoating is indicated by dissociation of the vertex proteins, including fiber, penton base, and peripentonal hexons, as well as the internal capsid protein VI from the virus capsid. Using this assay, we asked whether the 9C12 MNAb stabilized the virus capsid such that uncoating was prevented. An HAdV-5-based vector encoding enhanced green fluorescent protein (Ad5.eGFP) was incubated with the MNAb 9C12 or with a control, isotype-matched, but nonneutralizing antihexon MAb to allow binding. The vi-
rus-antibody mixtures were then heated, and soluble proteins were separated from capsid-associated proteins by density gradient centrifugation. As observed previously (31, 42), the virus capsid remained intact upon incubation at 43°C; however, upon incubation at 46°C or 50°C, the majority of penton base, fiber, and protein VI dissociated from the capsid (Fig. 1A). The presence of the 9C12 antibody did not alter this profile, suggesting that the antibody does not block vertex dissociation and, thus, allows the release of the membrane-lytic protein VI. We next asked whether the 9C12 MNAb blocked HAdVmediated endosome disruption using a ribotoxin coentry assay. In this assay, endosomalysis is indicated by a reduction in cellular protein synthesis as measured by [35S]methionine incorporation into host proteins upon HAdV-mediated endosomal escape of the ribotoxin ␣-sarcin, which on its own is unable to penetrate the host cell membrane. Infection of A549 cells by increasing concentrations of Ad5.eGFP in the presence of ␣-sarcin resulted in a dose-dependent loss of [35S]methionine incorporation into cellular proteins with a 90% inhibitory concentration (IC90) of approximately 100 ng/ml, similar to our previous results (Fig. 1B) (31). In contrast, a temperature-sensitive mutant of the closely related HAdV-2 (ts1), with a mutation in the protease gene and a known defect in endosome penetration (26), was significantly impaired in its ability to mediate ␣-sarcin entry, with an IC90 of 1,000 ng/ml. [35S]methionine incorporation upon infection with Ad5.eGFP in the presence of either the 9C12 MNAb or a control antihexon MAb was equivalent to infection in the absence of antibody, strongly suggesting that the 9C12 antibody does not block virus escape from the endosome. These findings are also consistent with previous studies showing that 9C12 does not block early steps in infection, including virus attachment or internalization, and indicate that a later step in entry (i.e., postendosomal) is altered by the 9C12 MNAb.
VOL. 82, 2008
POSTENTRY NEUTRALIZATION OF ADENOVIRUS INFECTION
6495
FIG. 2. 9C12 binding blocks accumulation of HAdV-5 at the nuclear envelope. (A) Representative A549 cells at the indicated times postinfection with Ad5.eGFP in the presence or absence of the neutralizing 9C12 antibody stained for tubulin (FITC, green), Ad5.eGFP (Cy3, red), and the nucleus (DAPI, blue) are depicted as z-profiles of sequential deconvolved z-series. Bars, 10 m. Percentages are the colocalization values for each cell. (B) Colocalization of Ad5.eGFP with the nucleus (DAPI) was quantified. Data are the mean percent colocalization ⫾ the standard error of the mean for an average of 60 cells per data point.
9C12 MNAb blocks HAdV-5 nuclear localization. To further delineate the step in infection blocked by 9C12, we infected A549 cells with Ad5.eGFP unbound or bound to 9C12 and quantified the colocalization of the virus with the nucleus (stained with DAPI) immediately after a 1-h binding step at 4°C or following a 90-min internalization step at 37°C. In z-profiles of z-series from an average of 60 cells per condition, we observed that in the absence of 9C12 the nuclear colocalization of Ad5.eGFP increased from approximately 25% to 55% upon incubation for 90 min at 37°C (Fig. 2). This is in agreement with our previous studies (31). The colocalization observed after the 4°C binding step is due to random colocalization of virus with the nucleus in z-profile. In contrast, Ad5.eGFP was evenly distributed throughout the cytoplasm at 90 min postinfection (p.i.) in the presence of 9C12, and no increase in nuclear colocalization was observed. These studies suggest that the 9C12 antibody blocks HAdV-5 infection at a step between endosome disruption and docking at the nuclear pore complex. 9C12 MNAb binding enhances MAP-dependent and MAPindependent microtubule association. According to current models, HAdV associates with microtubules following endo-
some penetration. This association is enhanced by the presence of microtubule-associated proteins (MAPs), specifically the microtubule-dependent, minus-end-directed dynein motor (15). We hypothesized that 9C12 binding could inhibit nuclear localization of HAdV capsids by blocking their association with MAPs and/or microtubules. To address this hypothesis, Cy3Ad.eGFP alone or bound to the 9C12 MNAb or a control antihexon MAb was incubated with biotinylated microtubules in the presence or absence of a bovine brain MAP fraction, and microtubule-associated virus was isolated and quantified by measuring the fluorescent hexon signal (Fig. 3). As has been reported previously (15), a significant increase in Cy3-Ad.eGFP association with microtubules was observed in the presence of MAPs. Remarkably, we found that rather than blocking this association, 9C12 significantly enhanced both MAP-dependent and MAP-independent association between Cy3-Ad5.eGFP and microtubules. Control, nonneutralizing antihexon antibodies did not have this effect, suggesting that hexon binding alone was not sufficient to promote enhanced MT binding. Incubation in high salt (500 mM NaCl), which is known to disrupt most MAPdependent binding (15), reversed HAdV association with microtubules in the absence of 9C12 but failed to completely abrogate
6496
SMITH ET AL.
FIG. 3. 9C12 binding increases the association of HAdV-5 with microtubules in the presence and absence of microtubule-associated proteins. Ad5.eGFP binding to microtubules (MTs) was measured in the presence or absence of the neutralizing 9C12 antihexon antibody or a control, nonneutralizing antihexon antibody. Virus was incubated with MTs alone or in the presence of a MAPF with or without 500 mM NaCl. Data are the fold increase of bound hexon for each sample compared to the no-antibody sample incubated with MTs and MAPF without NaCl (*, positive control) ⫾ the standard deviation for three independent experiments. Bands are fluorescent scans of bound and 8% of input Cy3-labeled hexon from a representative experiment.
microtubule association in the presence of 9C12. This confirms that 9C12 enhances both MAP-dependent and -independent association of HAdV with microtubules. These studies support a model in which 9C12-bound HAdV-5 escapes from the endosome and, upon encountering microtubules in the cytoplasm, associates with microtubules in a manner that precludes virion translocation to the nuclear periphery. 9C12 binding prevents HAdV-5 accumulation at the microtubule-organizing center. We reasoned that if this model of HAdV neutralization were valid, then we could observe an altered transport of 9C12-HAdV-5 to the microtubule-organizing center (MTOC) due to this change in microtubule association. Recent biochemical and confocal studies have provided additional details for the steps in HAdV entry between endosomal escape, dynein-dependent translocation through the cytoplasm along microtubules, and docking at the nuclear pore complex (1, 33). These studies have shown that upon reaching the minus end of microtubules, HAdV remains bound at the MTOC and must be actively rescued by a process dependent either directly or indirectly upon the functions of the nuclear export factor CRM1. Thus, in the presence of CRM1 inhibitors such as LMB or ratjadone A or in enucleated cells, HAdV accumulates at the MTOC. To test the model that 9C12 binding blocks virus translocation by altering microtubule-dependent trafficking, we determined the relative position of the 9C12-mediated block compared to that imposed by CRM1 inhibitors. A549 cells were infected with Ad5.eGFP in the presence or absence of an inhibitory concentration of the 9C12 MNAb and 20 nM LMB. This concentration of LMB was determined to be approximately sixfold higher than the IC99 for luciferase activity in
J. VIROL.
A549 cells upon transduction by a HAdV-5-based vector expressing luciferase (data not shown). The colocalization of the virus with the nucleus (stained with DAPI) was quantified after a 1-h binding step at 4°C followed by a 90-min internalization step at 37°C. In z-profiles of z-series from an average of 80 cells per condition, we observed that in the absence of 9C12 neutralization, the nuclear colocalization of Ad5.eGFP was approximately 73% (Fig. 4). The colocalization for virus bound to 9C12 was 25%, equivalent to cells incubated with virus at 4°C but without the internalization step (Fig. 2). In cells incubated with LMB, nuclear colocalization of Ad5.eGFP was 32% and 41% in the presence and absence of 9C12, respectively. Although nuclear localization was similarly restricted for unbound virus and virus bound to 9C12 in cells exposed to LMB, the distribution of the virus was distinct. LMB alone caused a striking accumulation of virus at the MTOC, revealed by staining for tubulin, as has been observed previously (33). In contrast, virus in a complex with 9C12 was evenly distributed throughout the cytoplasm in both LMB-treated and untreated cells and appeared to be associated with microtubules; however, the resolution of the microscopy was insufficient to quantify the precise juxtaposition of virus with individual microtubules. This experiment indicates that the entry block imposed by 9C12 is upstream of the block imposed by CRM1 inhibitors and supports the model that the 9C12 MNAb blocks HAdV-5 infection by arresting HAdV movement along microtubules, thus preventing translocation to the MTOC. 9C12 does not directly block hexon binding to the nuclear pore complex. One hypothesis proposed in the previous study on the mechanism of 9C12 inhibition was that this antibody could inhibit infection by blocking virus binding to the nuclear pore complex (39). To explore this hypothesis, we examined the ability of 9C12 to block HAdV-5 hexon labeled with the fluorescent dye Alexa Fluor 488 from binding to the nuclear envelope in detergent-permeabilized cells. Fluorescent hexon was incubated with a fivefold molar ratio of 9C12 MNAb or a control antihexon MAb. Labeling with fluorescent dyes did not affect the ability of either antibody to bind to hexon (data not shown). The antibody-hexon complex or hexon alone was then added to NRK cells that had been permeabilized with digitonin and washed to remove cytosol. Hexon incubated with cells in the presence of an excess of the nuclear import factor importin , which competes for hexon binding to the nuclear envelope (30, 38; A. Cassany and L. Gerace, unpublished data), was included as a control. Cells were then incubated at 30°C to allow for diffusion and binding to the nuclear envelope, fixed for microscopy, and stained with DAPI to identify the nucleus. Under these conditions, hexon alone bound efficiently to the nucleus with bright signal intensity around the nuclear envelope in cross-sectional images (Fig. 5). In contrast, no hexon signal could be detected at the nuclear envelope in cells exposed to an excess of importin . Neither the 9C12 MNAb nor the control antibody had any effect on the binding of hexon to the nuclear envelope. Identical results were observed in HeLa cells (data not shown). Although we cannot rule out the possibility that the interaction of hexon trimers with the nuclear envelope is different than that of the uncoated capsid, these studies suggest that if virus-9C12 complexes were able to reach the nuclear envelope, their ability to dock with the nuclear pore complex would not be impaired.
VOL. 82, 2008
POSTENTRY NEUTRALIZATION OF ADENOVIRUS INFECTION
6497
FIG. 4. 9C12 binding prevents HAdV-5 from reaching the microtubule-organizing center. (A) Representative A549 cells at 90 min p.i. with Ad5.eGFP in the presence or absence of the neutralizing 9C12 antibody and the CRM1 inhibitor LMB stained for tubulin (FITC, green), Ad5.eGFP (Cy3, red), and the nucleus (DAPI, blue) are depicted as z-profiles of sequential deconvolved z-series. Bars, 10 m. Percentages are the colocalization values for each cell. (B) Colocalization of Ad5.eGFP with the nucleus (DAPI) was quantified. Data are the mean percent colocalization ⫾ standard error of the mean for an average of 80 cells per data point.
DISCUSSION Taken together, our studies support a model in which the 9C12 MNAb blocks HAdV-5 infection by altering the association of the capsid with microtubules or microtubule motors, resulting in a failure of the capsid to translocate through the cytoplasm from the site of endosome penetration to the nuclear envelope. The observation that a NAb could block the interaction of HAdV with microtubules at such a late step in the entry pathway was unexpected. A few examples of NAbs that block late steps in enveloped virus entry have been described, but their modes of action remain poorly characterized (27). During entry, the HAdV capsid mediates a series of interactions with the host cell to effect delivery of the viral chromosome into the nucleus. As the HAdV capsid, like that of all nonenveloped viruses, is not shielded by a lipid membrane, much of the capsid machinery that mediates these functions is exposed and can be targeted by the immune system. Antibody-dependent neutralization of virus infection is complex. HAdV in particular elicits a potent humoral response, and NAbs can be readily detected. Using several approaches, it has been shown convincingly that hexon is the major target for neutralization (8, 14, 28, 34, 45). Antifiber and anti-penton base NAbs that block primary receptor interactions and integrin-me-
diated internalization, respectively, have been identified (7, 13, 19, 32, 45); however, they likely represent only a small portion of the neutralizing activity of serum. Studies of the mechanisms of antiHAdV NAbs have identified modes of action that are common to antibody neutralization of other viruses. These include aggregation, which reduces the effective concentration of virus, particle destabilization, which leads to premature uncoating and particle destruction, failure to internalize, likely due to steric hindrance with integrin interactions, and failure to uncoat due to crosslinking of the capsid (2, 6, 37, 46). Collectively, the previous analysis of 9C12 and our studies have ruled out each of these mechanisms as explanations for the neutralizing activity of 9C12. Rather, these studies support the model that 9C12 represents a previously unidentified mode of action for antibody-mediated neutralization of HAdV. This mechanism may be applicable to other viruses, as microtubule-dependent transport is a common feature of the entry pathway of multiple pathogens (e.g., human papillomavirus and adeno-associated virus); however, the lipid membrane of enveloped viruses (e.g., herpes simplex virus) that use microtubules during entry would likely shield the critical capsid components from neutralization (11). One major unresolved question that had not been answered
6498
SMITH ET AL.
J. VIROL.
FIG. 5. 9C12 does not directly block hexon binding to the nuclear envelope. Accumulation of Alexa Fluor 488-labeled HAdV-5 hexon at the nuclear envelope of digitonin-permeabilized NRK cells was visualized in the presence or absence of the neutralizing 9C12 antihexon antibody or a control, nonneutralizing antihexon antibody. Cells incubated with the nuclear import factor importin  (Imp) were included as a positive control for blocking hexon binding to the nuclear envelope. Individual z-sections of cells stained with DAPI to identify the nucleus are shown and are representative of at least three experiments.
in previous studies was whether 9C12-bound virus was able to penetrate the endosome. Quantitative fluorescence microscopy indicated that virus internalization was not significantly affected by 9C12 binding; however, it was unclear if the internalized virus visible in the cytoplasm was free or in endosomes (39). Our biochemical studies support the conclusion that HAdV-mediated endosomalysis is unaffected by 9C12 binding. In previous studies, we established a close correlation between endosome escape measured by the ␣-sarcin coentry assay and by fluorescence colocalization with endosomal markers (31). Moreover, the failure of the HAdV-2 ts1 mutant to mediate ␣-sarcin entry also supports the validity of this assay to measure endosomalysis, as this mutant virus has been shown in microscopy studies to be arrested in the endosome (31, 36, 42). In addition, the uniform distribution of 9C12-bound virus in the cytoplasm is distinct from virus blocked at the stage of endosome penetration that accumulates perinuclearly in the endosomal/lysosomal pathway (4, 31). Because this biochemical assay is dependent upon virus binding to a primary receptor and internalization into endosomes, this study also supports the previous conclusions that the 9C12-dependent block is downstream of these early events. A second outstanding issue from the previous study was the role of the antibody in stabilizing the capsid. Cryo-EM image reconstruction of the 9C12-virus complex revealed extensive hexon-hexon cross-linking by the antibody (39). The cryo-EM reconstructions also showed that one FAb arm of 9C12 bound to a specific tower on peripentonal hexons. Electron density of the other FAb arm was consistent with minor binding sites on adjacent hexons, suggesting the possibility that antibody bridging of peripentonal hexons might prevent dissociation of the capsid vertices. Our thermostability analysis suggests that any bridging interactions at the peripentonal hexons are insufficient to block vertex removal. Moreover, these results are consistent with the close correlation that we established in previous studies between capsid stabilization, endosome penetration, and the requirement for protein VI release from the capsid to mediate endosome disruption (31, 42).
One discrepancy with the previous study was the effect of 9C12 on HAdV-5 nuclear localization. Both studies documented a statistically significant reduction in the amount of 9C12-bound virus at the nuclear envelope compared to control; however, an increase in the perinuclear localization of 9C12-bound virus was observed in cells at 1 h p.i. compared to cells at 0 h p.i. in the first study that was not observed in our experiments (39). This discrepancy may be due to differences in fluorescence image analysis. Based on observations in the previous study, Varghese et al. favored the hypothesis that 9C12 blocked capsid binding to the nuclear pore complex, whereas our confocal studies suggest that the 9C12-bound virus fails to reach the nuclear envelope. In support of this conjecture, analysis of virus infection in the presence of a CRM1 inhibitor clearly showed that 9C12 blocked virus translocation to the MTOC, which is upstream of the nuclear envelope (33). Moreover, our studies of the effect of 9C12 on hexon binding to the nuclear envelope suggest that if the 9C12bound capsid could reach the nuclear envelope, binding to the nuclear pore complex per se would not be affected. Therefore, we favor the model that 9C12 blocks HAdV-5 infection by altering microtubule-dependent translocation to the MTOC. The role of dynein-dependent translocation of the incoming HAdV capsid along microtubules in HAdV entry is well-established. The partially uncoated HAdV capsid that is released into the cytoplasm is approximately 90 nm in diameter, a size that precludes diffusion as an efficient mechanism for longrange transport (20). HAdV has been shown through both EM and biochemical studies to associate with microtubules, MAPs enhance this association, and microtubule disruption inhibits HAdV infection (1, 15, 18, 21, 22, 35, 36, 40). Kinetic studies of HAdV trafficking have demonstrated long-range, perinucleardirected mobility along curvilinear tracks consistent with minus-end-directed transport along microtubules toward the perinuclear MTOC (18, 36). Injection of function-blocking antidynein antibodies or overexpression of a dynactin complex subunit inhibits this movement and the nuclear localization of HAdV (18, 36). Taken together, these studies are consistent
VOL. 82, 2008
POSTENTRY NEUTRALIZATION OF ADENOVIRUS INFECTION
with microtubule-dependent transport mediated by the minusend-directed dynein/dynactin motor complex. The previous study on 9C12 showed that the virus capsid is enclosed in a meshwork of antibody density, which was particularly apparent on the facet hexons (39). As this cryo-EM density represents the spatial average of many alternate bivalent antibody interactions, the functional consequence of these antibody interactions is not clear. The findings described in the present study suggest that 9C12 binding alters the association of the capsid with microtubules in both MAP-dependent and -independent manners. Currently, the capsid component that interacts with microtubules or dynein and the role of the dynactin complex in this interaction are not well-defined. 9C12 could block the interaction with dynein and instead enhance a direct interaction with microtubules. Alternatively, 9C12 binding could redirect the capsid to an alternative and nonproductive MAP. Finally, 9C12-bound HAdV could bind to dynein in a manner that prevents processive translocation by the dynein motor. Future studies to address these hypotheses will be feasible once the nature of the HAdV capsid interaction with microtubules is defined in greater molecular detail. We have identified a previously unappreciated mode by which a neutralizing antibody can inhibit infection by a nonenveloped virus. That microtubules are a common route of entry for many viral pathogens raises the possibility that this mode of neutralization may exist for other viruses. Also unknown is whether MNAbs with 9C12-like functions are elicited upon natural HAdV infection. Analysis of polyclonal neutralizing serum would likely preclude the identification of such specificities due to the predominance of antibodies that neutralize through other mechanisms; however, the availability of recombinant antibody libraries from infected individuals that can be screened, amplified, cloned, and studied in isolation for 9C12-like activity provides a feasible system to address this possibility for HAdV as well as other viral pathogens (3, 23, 48). ACKNOWLEDGMENTS This work was supported by NIH grants 1 F32 AI072936-01 to J.G.S., 5 T32 NS41219, EY01131, and HL054352 to G.R.N., and AI055729 to L.G. This is manuscript number 19451 at the Scripps Research Institute. We thank William Kiosses for help with microscopy. We declare no competing financial interests. REFERENCES 1. Bailey, C. J., R. G. Crystal, and P. L. Leopold. 2003. Association of adenovirus with the microtubule organizing center. J. Virol. 77:13275–13287. 2. Boudin, M. L., and P. Boulanger. 1981. Antibody-triggered dissociation of adenovirus penton capsomer. Virology 113:781–786. 3. Burton, D. R. 1995. Phage display. Immunotechnology 1:87–94. 4. Carey, B., M. K. Staudt, D. Bonaminio, J. C. van der Loo, and B. C. Trapnell. 2007. PU.1 redirects adenovirus to lysosomes in alveolar macrophages, uncoupling internalization from infection. J. Immunol. 178:2440–2447. 5. Chi, N. C., E. J. Adam, and S. A. Adam. 1995. Sequence and characterization of cytoplasmic nuclear protein import factor p97. J. Cell Biol. 130:265–274. 6. Everitt, E., A. de Luca, and Y. Blixt. 1992. Antibody-mediated uncoating of adenovirus in vitro. FEMS Microbiol. Lett. 77:21–27. 7. Fender, P., A. H. Kidd, R. Brebant, M. Oberg, E. Drouet, and J. Chroboczek. 1995. Antigenic sites on the receptor-binding domain of human adenovirus type 2 fiber. Virology 214:110–117. 8. Gahery-Segard, H., F. Farace, D. Godfrin, J. Gaston, R. Lengagne, T. Tursz, P. Boulanger, and J. G. Guillet. 1998. Immune response to recombinant capsid proteins of adenovirus in humans: antifiber and anti-penton base antibodies have a synergistic effect on neutralizing activity. J. Virol. 72:2388– 2397.
6499
9. Gray, G. C., T. McCarthy, M. G. Lebeck, D. P. Schnurr, K. L. Russell, A. E. Kajon, M. L. Landry, D. S. Leland, G. A. Storch, C. C. Ginocchio, C. C. Robinson, G. J. Demmler, M. A. Saubolle, S. C. Kehl, R. Selvarangan, M. B. Miller, J. D. Chappell, D. M. Zerr, D. L. Kiska, D. C. Halstead, A. W. Capuano, S. F. Setterquist, M. L. Chorazy, J. D. Dawson, and D. D. Erdman. 2007. Genotype prevalence and risk factors for severe clinical adenovirus infection, United States 2004–2006. Clin. Infect. Dis. 45:1120–1131. 10. Greber, U. F., and M. Fornerod. 2005. Nuclear import in viral infections. Curr. Top. Microbiol. Immunol. 285:109–138. 11. Greber, U. F., and M. Way. 2006. A superhighway to virus infection. Cell 124:741–754. 12. Greber, U. F., M. Willetts, P. Webster, and A. Helenius. 1993. Stepwise dismantling of adenovirus 2 during entry into cells. Cell 75:477–486. 13. Hong, S. S., M. Bardy, M. Monteil, B. Gay, C. Denesvre, J. Tournier, G. Martin, M. Eloit, and P. Boulanger. 2000. Immunoreactive domains and integrin-binding motifs in adenovirus penton base capsomer. Viral Immunol. 13:353–371. 14. Hong, S. S., N. A. Habib, L. Franqueville, S. Jensen, and P. A. Boulanger. 2003. Identification of adenovirus (Ad) penton base neutralizing epitopes by use of sera from patients who had received conditionally replicative ad (addl1520) for treatment of liver tumors. J. Virol. 77:10366–10375. 15. Kelkar, S. A., K. K. Pfister, R. G. Crystal, and P. L. Leopold. 2004. Cytoplasmic dynein mediates adenovirus binding to microtubules. J. Virol. 78: 10122–10132. 16. Kojaoghlanian, T., P. Flomenberg, and M. S. Horwitz. 2003. The impact of adenovirus infection on the immunocompromised host. Rev. Med. Virol. 13:155–171. 17. Kolavic-Gray, S. A., L. N. Binn, J. L. Sanchez, S. B. Cersovsky, C. S. Polyak, F. Mitchell-Raymundo, L. V. Asher, D. W. Vaughn, B. H. Feighner, and B. L. Innis. 2002. Large epidemic of adenovirus type 4 infection among military trainees: epidemiological, clinical, and laboratory studies. Clin. Infect. Dis. 35:808–818. 18. Leopold, P. L., G. Kreitzer, N. Miyazawa, S. Rempel, K. K. Pfister, E. Rodriguez-Boulan, and R. G. Crystal. 2000. Dynein- and microtubule-mediated translocation of adenovirus serotype 5 occurs after endosomal lysis. Hum. Gene Ther. 11:151–165. 19. Liebermann, H., R. Mentel, L. Dohner, S. Modrow, and W. Seidel. 1996. Inhibition of cell adhesion to the virus by synthetic peptides of fiber knob of human adenovirus serotypes 2 and 3 and virus neutralisation by anti-peptide antibodies. Virus Res. 45:111–122. 20. Luby-Phelps, K. 2000. Cytoarchitecture and physical properties of cytoplasm: volume, viscosity, diffusion, intracellular surface area. Int. Rev. Cytol. 192:189–221. 21. Luftig, R. B., and R. R. Weihing. 1975. Adenovirus binds to rat brain microtubules in vitro. J. Virol. 16:696–706. 22. Mabit, H., M. Y. Nakano, U. Prank, B. Saam, K. Dohner, B. Sodeik, and U. F. Greber. 2002. Intact microtubules support adenovirus and herpes simplex virus infections. J. Virol. 76:9962–9971. 23. Maruyama, T., L. L. Rodriguez, P. B. Jahrling, A. Sanchez, A. S. Khan, S. T. Nichol, C. J. Peters, P. W. Parren, and D. R. Burton. 1999. Ebola virus can be effectively neutralized by antibody produced in natural human infection. J. Virol. 73:6024–6030. 24. Metzgar, D., M. Osuna, A. E. Kajon, A. W. Hawksworth, M. Irvine, and K. L. Russell. 2007. Abrupt emergence of diverse species B adenoviruses at US military recruit training centers. J. Infect. Dis. 196:1465–1473. 25. Nakano, M. Y., K. Boucke, M. Suomalainen, R. P. Stidwill, and U. F. Greber. 2000. The first step of adenovirus type 2 disassembly occurs at the cell surface, independently of endocytosis and escape to the cytosol. J. Virol. 74:7085–7095. 26. Rancourt, C., H. Keyvani-Amineh, S. Sircar, P. Labrecque, and J. M. Weber. 1995. Proline 137 is critical for adenovirus protease encapsidation and activation but not enzyme activity. Virology 209:167–173. 27. Reading, S. A., and N. J. Dimmock. 2007. Neutralization of animal virus infectivity by antibody. Arch. Virol. 152:1047–1059. 28. Roy, S., D. S. Clawson, R. Calcedo, C. Lebherz, J. Sanmiguel, D. Wu, and J. M. Wilson. 2005. Use of chimeric adenoviral vectors to assess capsid neutralization determinants. Virology 333:207–214. 29. Saban, S. D., M. Silvestry, G. R. Nemerow, and P. L. Stewart. 2006. Visualization of ␣-helices in a 6-angstrom resolution cryoelectron microscopy structure of adenovirus allows refinement of capsid protein assignments. J. Virol. 80:12049–12059. 30. Saphire, A. C., T. Guan, E. C. Schirmer, G. R. Nemerow, and L. Gerace. 2000. Nuclear import of adenovirus DNA in vitro involves the nuclear protein import pathway and hsc70. J. Biol. Chem. 275:4298–4304. 31. Smith, J. G., and G. R. Nemerow. 2008. Mechanism of adenovirus neutralization by human alpha-defensins. Cell Host Microbe 3:11–19. 32. Stewart, P. L., C. Y. Chiu, S. Huang, T. Muir, Y. Zhao, B. Chait, P. Mathias, and G. R. Nemerow. 1997. Cryo-EM visualization of an exposed RGD epitope on adenovirus that escapes antibody neutralization. EMBO J. 16: 1189–1198. 33. Strunze, S., L. C. Trotman, K. Boucke, and U. F. Greber. 2005. Nuclear
6500
34.
35.
36.
37.
38.
39.
40.
SMITH ET AL.
targeting of adenovirus type 2 requires CRM1-mediated nuclear export. Mol. Biol. Cell 16:2999–3009. Sumida, S. M., D. M. Truitt, A. A. Lemckert, R. Vogels, J. H. Custers, M. M. Addo, S. Lockman, T. Peter, F. W. Peyerl, M. G. Kishko, S. S. Jackson, D. A. Gorgone, M. A. Lifton, M. Essex, B. D. Walker, J. Goudsmit, M. J. Havenga, and D. H. Barouch. 2005. Neutralizing antibodies to adenovirus serotype 5 vaccine vectors are directed primarily against the adenovirus hexon protein. J. Immunol. 174:7179–7185. Suomalainen, M., M. Y. Nakano, K. Boucke, S. Keller, and U. F. Greber. 2001. Adenovirus-activated PKA and p38/MAPK pathways boost microtubule-mediated nuclear targeting of virus. EMBO J. 20:1310–1319. Suomalainen, M., M. Y. Nakano, S. Keller, K. Boucke, R. P. Stidwill, and U. F. Greber. 1999. Microtubule-dependent plus- and minus end-directed motilities are competing processes for nuclear targeting of adenovirus. J. Cell Biol. 144:657–672. Toogood, C. I., J. Crompton, and R. T. Hay. 1992. Antipeptide antisera define neutralizing epitopes on the adenovirus hexon. J. Gen. Virol. 73:1429– 1435. Trotman, L. C., N. Mosberger, M. Fornerod, R. P. Stidwill, and U. F. Greber. 2001. Import of adenovirus DNA involves the nuclear pore complex receptor CAN/Nup214 and histone H1. Nat. Cell Biol. 3:1092–1100. Varghese, R., Y. Mikyas, P. L. Stewart, and R. Ralston. 2004. Postentry neutralization of adenovirus type 5 by an antihexon antibody. J. Virol. 78: 12320–12332. Weatherbee, J. A., R. B. Luftig, and R. R. Weihing. 1977. Binding of adenovirus to microtubules. II. Depletion of high-molecular-weight microtubule-associated protein content reduces specificity of in vitro binding. J. Virol. 21:732–742.
J. VIROL. 41. Wickham, T. J., P. Mathias, D. A. Cheresh, and G. R. Nemerow. 1993. Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 73:309–319. 42. Wiethoff, C. M., H. Wodrich, L. Gerace, and G. R. Nemerow. 2005. Adenovirus protein VI mediates membrane disruption following capsid disassembly. J. Virol. 79:1992–2000. 43. Wodrich, H., A. Cassany, M. A. D’Angelo, T. Guan, G. Nemerow, and L. Gerace. 2006. Adenovirus core protein pVII is translocated into the nucleus by multiple import receptor pathways. J. Virol. 80:9608–9618. 44. Wodrich, H., T. Guan, G. Cingolani, D. Von Seggern, G. Nemerow, and L. Gerace. 2003. Switch from capsid protein import to adenovirus assembly by cleavage of nuclear transport signals. EMBO J. 22:6245–6255. 45. Wohlfart, C. 1988. Neutralization of adenoviruses: kinetics, stoichiometry, and mechanisms. J. Virol. 62:2321–2328. 46. Wohlfart, C. E., U. K. Svensson, and E. Everitt. 1985. Interaction between HeLa cells and adenovirus type 2 virions neutralized by different antisera. J. Virol. 56:896–903. 47. Wu, E., S. A. Trauger, L. Pache, T. M. Mullen, D. J. von Seggern, G. Siuzdak, and G. R. Nemerow. 2004. Membrane cofactor protein is a receptor for adenoviruses associated with epidemic keratoconjunctivitis. J. Virol. 78:3897–3905. 48. Zebedee, S. L., C. F. Barbas III, Y. L. Hom, R. H. Caothien, R. Graff, J. DeGraw, J. Pyati, R. LaPolla, D. R. Burton, R. A. Lerner, et al. 1992. Human combinatorial antibody libraries to hepatitis B surface antigen. Proc. Natl. Acad. Sci. USA 89:3175–3179. 49. Zhang, Y., and J. M. Bergelson. 2005. Adenovirus receptors. J. Virol. 79: 12125–12131.
