INTRODUCTION
Enterovirus type 71 (EV71) is a member of the
Enterovirus genus of the family
Picornaviridae and is one of the causative viral agents of hand, foot, and mouth disease (HFMD) (
6,
7,
14,
16,
41). HFMD is largely a common self-limited childhood illness but may have complications of severe to fatal neurological symptoms in some children (
1,
5,
6,
16,
21). In the past decade, the frequency of EV71 outbreaks associated with severe neurological illness appeared to have increased in the Pacific region, most notably in China, where large outbreaks have been occurring annually since 2007 (
24,
53). While the epidemiological or virological mechanism underlying this regional focus of severe EV71 infection remains largely unknown, the impact of EV71 infection is a global concern, as evidenced by the increase in virological surveillance and studies of EV71 infection in many regions of the world (
29,
46,
53).
EV71 is composed of a single-stranded, positive-polarity RNA molecule surrounded by a nonenveloped, pentameric icosahedral capsid (
3,
35), which consists of 60 copies of the four structural proteins VP1 to VP4. While there is no animal model for EV71 infection in humans, intraperitoneal injection of EV71 is lethal to suckling mice. Suckling mice born to mothers previously immunized with VP1 subunit vaccines acquire resistance to lethal EV71 challenge (
8,
50), and administration of a VP1-based antigen, either protein or DNA, to mice could elicit a neutralizing antibody against EV71 infection
in vitro (
8,
42,
44). The serum collected from EV71-infected individuals during the convalescent phase could neutralize EV71 infection
in vitro, and the neutralization epitope was mapped to a fragment of VP1 (
8,
43). Lactoferrin, a protein component of mammalian breast milk, has been shown to bind to VP1 and to block EV71 infection (
48). These observations suggest that VP1 plays a pivotal role in the life cycle of EV71.
Identification of host factors that interact with EV71 might provide insights into the virus-host interactions that contribute to pathogenesis. Studies focusing on the cellular factors that assist the entry of EV71 into the cell have identified two factors: human scavenger receptor class B (SCARB2) and P-selectin glycoprotein ligand-1 (PSGL-1, also known as CD162) (
27,
51). PSGL-1 is not utilized by all EV71 strains tested; it is expressed mainly on neutrophils, monocytes, and most lymphocytes, which are not preferred sites for the replication of some EV71 strains (
27). SCARB2 is mainly a lysosomal receptor and is expressed ubiquitously on most types of cells (
51); however, blocking of SCARB2 by its antibody can only partially reduce the entry of EV71 into some cell lines. The roles of these two proteins in the neuropathogenesis of EV71 infection in humans are generally considered to be unclear. Thus, pathways of cellular entry via receptors other than PSGL-1 or SCARB2 need to be elucidated (
27,
51).
In this study, we have identified annexin II (Anx2) as a cellular adherent factor that interacts with EV71 via VP1 binding. Anx2 is a member of the annexin gene family. In addition to the interaction with phospholipid membranes via its calcium ion-binding alpha-helical core domain (
39), it can interact with multiple cellular factors and is involved in regulating cellular functions, including endo-and exocytotic pathways (
10), calcium-dependent F-actin filament bundling (
13), and the profibrinolytic generation of plasmin (
17). The present study demonstrates that Anx2 on the cell surface can interact with EV71 virions and that this binding can enhance the infectivity of EV71.
MATERIALS AND METHODS
Cells and viruses.
RD cells (human embryonal rhabdomyosarcoma) were cultured in Dulbecco's modified Eagle medium (DMEM) supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% fetal bovine serum (FBS). Human hepatoblastoma cell lines—HepG2 and its cognate lineage C3A—were cultured in minimal essential medium (MEM) supplemented with 1 mM sodium pyruvate, 0.1 mM nonessential amino acids (NEAA), 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% FBS. A stable line of Anx2-expressing HepG2 cells, referred to below as HepG2-Anx2 cells, was isolated by transfecting a pcDNA3.1 plasmid containing the full-length Anx2 gene by use of SuperFect transfection reagent (Qiagen, Valencia, CA) and was maintained in a selection medium containing 400 μg/ml G418 (Gibco BRL, Grand Island, NY). A clone that expresses Anx2 was selected for use.
The
neu strain (genotype C2) of EV71, a human isolate from a spinal cord sample taken at necropsy (
50), was amplified in RD cells, purified, quantified by determination of the 50% tissue culture infective dose (TCID
50) per 1 ml in RD cells as described previously (
20), and used as the prototype EV71 strain for all experiments unless stated otherwise. [
35S]-labeled EV71 was obtained by growing the virus in RD cells incubated in a medium containing 10 μM unlabeled methionine and 100 μCi/ml
l-[
35S]methionine (specific activity, 400 Ci/mmol; Amersham Pharmacia Biotech) for 24 h at 37°C. Other strains of human EV71 were clinical isolates recovered in 2004, 2005, and 2008 in Taiwan and had not been adapted to any cell line.
Antibodies.
A mouse anti-EV71 monoclonal antibody (MAb) (Chemicon, Temecula, CA) was used to detect the virus in all experiments. Mouse anti-Anx2 MAbs raised against Anx2 (amino acids [aa] 123 to 339) (BD Transduction Labs, Lexington, KY) and against a peptide near the N terminus of Anx2 (Santa Cruz Biotechnology, Santa Cruz, CA) were used for Western blotting, flow cytometry, and inhibition of virus infection. Fluorescein isothiocyanate (FITC)-conjugated rabbit anti-mouse IgG (Zymed Laboratories, San Francisco, CA) was used in flow cytometry. A mouse IgG1 isotype (Miltenyi Biotec, Auburn, CA) was used as an internal control in the infection inhibition assay and flow cytometry. Horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (Santa Cruz Biotechnology, Santa Cruz, CA) was used as a secondary antibody for enhanced chemiluminescence (ECL) detection in Western blotting. Rhodamine-conjugated goat anti-mouse IgG (Vector Labs, Burlingame, CA) was used for detection by confocal microscopy. A monoclonal antipolyhistidine antibody (Sigma-Aldrich, St. Louis, MO) was used to detect truncated fragments of Anx2. A polyclonal antibody against coxsackievirus A16 (CA16) was obtained from the Taiwan Centers for Disease Control. Mouse MAbs against glutathione S-transferase (GST) (Santa Cruz Biotechnology, Santa Cruz, CA) were used to detect the GST-VP1 fusion protein.
Recombinant proteins. (i) Anx2.
Plasmid pET23a (Novagen, Madison, WI), containing the full-length human Anx2 cDNA (ATCC MGC-2257), and plasmid pET21a, containing truncated Anx2 cDNA sequences (encoding aa 1 to 267 or aa 268 to 339), each with a C-terminal hexahistidine tag, were constructed and expressed in Escherichia coli BL21(DE3) (Stratagene, La Jolla, CA). The Anx2 cDNA and the insertion of the Anx2 cDNA product were verified by sequence analysis. Full-length recombinant Anx2 (rAnx2) protein and its truncated fragments were prepared by growing E. coli BL21 harboring pET23a-Anx2 or pET21a-Anx2 at 37°C to an optical density at 600 nm (OD600) of 0.6, induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at room temperature for 5 h, and then spun down at 12,000 × g for 10 min. Native soluble rAnx2 protein was purified from the supernatant of the cell lysate by using BugBuster extraction reagent (Novagen, Madison, WI) and PureProteome nickel magnetic beads (Millipore) according to the manufacturers' instructions. The purified Anx2 (∼36 kDa) was identified by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (10% SDS-PAGE) after Coomassie blue staining and Western blotting with an anti-Anx2 MAb.
(ii) Viral capsid protein VP1.
The full-length VP1 protein fused with a calmodulin-binding peptide (CBP) tag at the C terminus, termed VP1-CBP, was prepared as described previously (
50). The truncated VP1 cDNA fragment encoding amino acids 40 to 100 was generated by PCR and was cloned into the pGEX-4T-1 GST expression vector (GE Healthcare, London, United Kingdom) for expression of a GST-VP1 (aa 40 to 100) fusion protein. The cell lysate of transformed
E. coli BL21(DE3) contained a major protein of ∼34 kDa upon IPTG induction, which was detectable with an anti-GST antibody.
(iii) Solubilizing and refolding of the truncated recombinant proteins in the inclusion body.
The denatured recombinant proteins in the inclusion body were solubilized in 100 mM Tris containing 8 M urea and 20 mM dithiothreitol (DTT) (pH 8.5) to a concentration of 10 mg/ml and were dialyzed in 1 M urea–20 mM Tris–0.05% Tween 20 (pH 8.0) at 4°C with slow stirring overnight; the dialysis buffer was changed every 6 h with serially diluted urea concentrations to a final concentration of 0.01 M urea.
Preparation of soluble cellular proteins and the membrane fraction of RD cells.
To obtain total cellular proteins, RD cells were scraped into cold lysis buffer containing 10 mM Tris HCl (pH 7.5), 1 mM EDTA, 150 mM NaCl, 10 mM dithiothreitol, 0.5% NP-40, 5% glycerol, and Complete protease inhibitor cocktail (Roche, Mannheim, Germany), set on ice for 30 min, and centrifuged at 10,000 × g for 10 min at 4°C to remove cell debris. Membrane proteins of RD cells were extracted from the total cellular proteins with a BioVision (Mountain View, CA) plasma membrane protein extraction kit according to the manufacturer's instructions. The protein concentration of each sample was determined by using the Bradford protein assay reagent according to the manufacturer's instructions (Bio-Rad). The membrane protein was used for further analyses, including Western blotting, matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry, and a virus overlay protein-binding assay (VOPBA).
VOPBA and Western blot analysis.
Proteins (0.5 μg/lane) were separated on 10% SDS-PAGE gels, transferred to a nylon membrane, blocked in a blocking solution (Tris-buffered saline–Tween 20 [TBST] and 5% skim milk) for 1 h at room temperature (RT), and processed for immunoblot analysis. For the VOPBA, the nylon membrane was incubated with EV71 (10 μg total protein/ml of blocking solution) at 4°C for 18 h, washed three times with TBST, incubated with the respective anti-EV71 MAb or anti-Anx2 MAb in blocking solution for 3 h at RT, incubated with HRP-conjugated goat anti-mouse IgG as a secondary antibody for 1 h at RT, and then visualized using an ECL system.
Preparation of protein samples for analysis by MALDI-TOF mass spectrometry.
RD soluble proteins (200 μg) or the membrane fraction (20 μg) was resolved by SDS-PAGE and stained with Coomassie blue. The protein spot of interest was excised, washed with 25 mM NH4HCO3–50% acetonitrile, and digested with sequencing-grade trypsin (Promega, Madison, WI) in 5 mM NH4HCO3 at 37°C for 18 h. The digested peptides were extracted with 50% acetonitrile–5% trifluoroacetic acid (TFA). For MALDI-TOF mass spectrometry analysis, a mixture of 1 μl of the extracted peptides in a 1:1 ratio with a 1:10 dilution of a saturated α-cyano-4-hydroxycinnamic acid (ACCA) matrix in 0.25% TFA, 50% acetonitrile, and 50% water was spotted onto MALDI target plates, and spectra were acquired using a Voyager-DE STR Biospectrometry workstation (Applied Biosystems, Foster City, CA). Peptide masses were compared with those available in the NCBI database (National Institutes of Health, Bethesda, MD) using the MASCOT search engine (Matrix Science, London, United Kingdom).
Pulldown assy.
The calmodulin-coupled affinity resin (10 μl) (Stratagene, La Jolla, CA) was incubated in 0.5 ml calmodulin binding buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 10 mM 2-mercaptoethanol, 1 mM MgCl2, 2 mM CaCl2, 1 mM imidazole, and Complete protease inhibitor cocktail) containing 180 μg of the crude bacterial lysate expressing the VP1-CBP fusion protein on a rotating wheel for 1 h at RT. The complex was washed five times with binding buffer and was incubated with 0.5 μg rAnx2 in 0.5 ml binding buffer for 1 h at RT. The beads were washed five times with calmodulin binding buffer before being spun down; the resultant pellet was resuspended in SDS-PAGE sample buffer (Bio-Rad), heated, and loaded onto a 12% SDS-PAGE gel for Western blot analysis. To obtain the CBP-calmodulin resin for control purposes, VP1 was removed by incubating the VP1-CBP-calmodulin resin with 3 U thrombin (Novagen, CA) in a cleavage buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 2.5 mM CaCl2, 1 mM DTT) for 16 h at RT. By considering the total input of 0.5 μg rAnx2 to be 100%, the percentage of rAnx2 captured was calculated based on the integrated intensity of the Anx2 detected on the Western blot as measured by a ScanMaker 8700 instrument (Microtek) using MetaMorph software.
Glutathione-Sepharose 4B beads (5 μl) (GE Healthcare, London, United Kingdom), either untreated or preincubated with the cell lysate of plasmid pGEX-VP1 (40-100)-transformed E. coli BL21, were mixed with 6 μg rAnx2 in 400 μl of binding buffer (140 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4 [pH 7.3]) and were incubated with rotation at 4°C for 1 h. The beads were washed five times with binding buffer before being spun down to obtain the captured proteins in the pellet, which was resuspended in SDS-PAGE sample buffer (Bio-Rad), heated, and loaded onto a 15% SDS-PAGE gel for Western blot analysis using an antipolyhistidine antibody to detect rAnx2 or an anti-GST antibody to detect the GST-VP1 fusion protein.
Inhibition of infection. (i) Inhibition by an anti-Anx2 antibody.
Near-confluent RD cells (1 × 105 cells per well of a 12-well tissue culture plate) were treated with an anti-Anx2 antibody or a control antibody for 1 h at 37°C prior to adsorption of the virus (multiplicity of infection [MOI], 0.2) at 4°C for a further 1 h, washed twice with DMEM, and incubated in 2% FBS-DMEM at 37°C for 42 h. The culture medium was collected for determination of the virus titer by a TCID50 assay.
(ii) Inhibition by soluble rAnx2.
EV71 was incubated with soluble rAnx2 at 37°C for 1 h, inoculated onto RD cells (1 × 106 cells per well of a 6-well culture plate), incubated for 1 h at 37°C to allow adsorption to RD cells, washed twice with DMEM to remove unbound virus, and incubated in 2% FBS-DMEM at 37°C for 3 days. Subsequently, the culture medium was collected for determination of the virus titer by a TCID50 assay.
Assessment of soluble rAnx2 as an inhibitor of cell attachment.
[35S]-labeled EV71 was pretreated with soluble rAnx2 (at various concentrations at 4°C for 1 h) or a heat-denatured bacterial lysate and was incubated with 2 × 105 RD cells for 1 h on ice. The cells were washed twice with cold DMEM, detached by scraping into cold TBS buffer, and assayed for bound radioactivity in a liquid scintillation counter.
Confocal microscopy.
RD cells (1 × 105) were plated on a coverslip, allowed to adhere, incubated with FITC-labeled EV71 for 1 h at 4°C, washed twice with DMEM, fixed with 4% paraformaldehyde for 20 min, blocked with 1% bovine serum albumin (BSA) for 1 h, and then stained with an anti-Anx2 antibody for 1 h at 37°C. After washing, cells were stained with rhodamine-labeled goat anti-mouse IgG, washed, mounted, and examined under a confocal microscope or Radiance 2100 [Bio-Rad]) equipped with an argon/krypton laser.
Flow cytometry.
Cells were trypsinized with 0.25% trypsin-EDTA for 1 min, and at least 5 × 105 cells were collected for each experiment, washed with fluorescence-activated cell sorter (FACS) buffer (10 mM HEPES, 1% FBS, 0.1% NaN3, 1× Hanks balanced salt solution [HBSS]), blocked with 3% BSA (in FACS buffer) for 30 min, washed with FACS buffer, incubated with the anti-Anx2 antibody (25 μg/ml of FACS buffer) or mouse IgG1 isotype antibody at 4°C for 2 h, washed with FACS buffer, centrifuged (at 300 × g for 5 min), resuspended in 0.1 ml of FITC-conjugated rabbit anti-mouse IgG for an additional 1 h at 4°C, washed twice, resuspended in phosphate-buffered saline (PBS), and analyzed by flow cytometry (FACSCalibur; Beckman, San Jose, CA).
Yeast two-hybrid analysis.
To map the VP1-Anx2 interacting domains, truncated VP1 and Anx2 DNA fragments were amplified by PCR. The VP1 DNA fragments were cloned into the SmaI site of the pBTM116 vector (a generous gift from Hsiu-Ming Shih), resulting in an N-terminal in-frame fusion with a LexA DNA-binding domain (DBD) coding sequence. The Anx2 DNA fragments were cloned into the SmaI site of the pACT2 vector (Clontech), resulting in an N-terminal in-frame fusion with the Gal4 activating domain (AD). Yeast two-hybrid analysis was performed according to the user's manual (Clontech). Interactions between the VP1 and Anx2 proteins were initially indicated by colony formation on the selective medium and were confirmed by β-galactosidase activity in a filter colony lift assay with 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) as a substrate.
Molecular docking.
We used the PatchDock server (
http://bioinfo3d.cs.tau.ac.il/PatchDock) to predict the putative binding sites for VP1 and Anx2. The structure of VP1 (aa 27 to 297) was created from homology modeling based on poliovirus (Protein Data Bank [PDB] code 1PVC) and was used to localize the ligand to Anx2, assuming the X-ray crystal structure already determined (PDB code 2HYW). The clustering RMSD (root mean square deviation) was set for 4.0 Å, and the complex type was set as the default type. The top 10 results were transferred to FireDock for refinement. The top-ranked FireDock solution according to the contribution of the atomic contact energy was chosen for the study of binding sites, and the global binding energy of the solution was −1.63.
Statistical analysis.
Data on virus titers were presented as means ± standard errors (SE), and statistical significance was tested by a paired t test using SAS/STAT. P values of <0.05 were considered significant.
DISCUSSION
Based on the consensus that VP1 is an important site for interaction with host cells (
8,
42,
44,
48,
50), our study focused on the VP1-binding protein Anx2, which was identified using the laboratory-adapted strain
neu5. The virus-Anx2 interaction was shown to be common to EV71 clinical isolates of other selected genotypes—C2, C5, and B5—that had not yet been adapted to
in vitro infection in the laboratory, suggesting that interaction with Anx2 could be a common feature of circulating EV71 strains. To further elaborate the biological function of the VP1-Anx2 interaction, we investigated the location of the Anx2-interacting domain (aa 40 to 100) of VP1 within the 3-dimensional viral structure and its accessibility to host cell binding sites. While the structure of the EV71 virion has not been delineated, picornaviruses have a shared structure for the icosahedral capsid, consisting of 60 copies of proteins VP1 to VP4 (
35). Each of proteins VP1 to VP3 folds into an eight-stranded antiparallel β-sheet, with the strands designated B through I. Based on this structural paradigm (
35) and an alignment of the amino acid sequences encoded by the EV71 genes with those of poliovirus types 1 to 3 (
Fig. 10A), we propose that the Anx2-binding region consists of β-sheet B and the partial BC loop (i.e., a coiled region between β-sheets B and C; aa 97 to 106). The BC loop of VP1, which is part of the exposed outer capsid surrounding the 5-fold axis, can react with type-specific antibodies to polioviruses (
9,
19,
26), possibly participates in conformational changes of the viral particle during virus uncoating (
9,
19), and contains neutralizing epitopes for poliovirus types 2 and 3 (
25). Our data show that either the binding of soluble Anx2 to the virion or the binding of an anti-Anx2 antibody to RD cells can result in reduced attachment of the virus to RD cells; this is consistent with an externally located domain of the EV71 virion that interacts with Anx2. However, on the virus surface topography, this VP1 domain is not within the putative receptor binding site of the virion (for a comprehensive review, see reference
45). We further constructed a protein-protein interaction model (
Fig. 10B), which also suggests a potential interaction between domain IV and the BC loop of VP1, where a hydrogen bond between VP1 L95 and Anx2 R309 was found. Interestingly, in this model, domain III may also interact with the C terminus of VP1, which might explain how soluble Anx2 (aa 1 to 267) can partially interfere with EV71 infection of RD cells.
Functionally, we have provided several lines of evidence to suggest that the Anx2-virion interaction impacts virus infectivity; the most definitive evidence is the increase in virus yield following the introduction of an Anx2 expression plasmid into HepG2 cells and the reversal of this increase by the blockade of Anx2 with antibody. Nevertheless, the evidence also argued against Anx2 being the essential cellular receptor for EV71. HepG2 cells, which express no Anx2, remain permissive to EV71 infection. Adult mice are not permissive to EV71 infection despite 100% identity of the amino acid sequences of the VP1-binding domains between the mouse and human Anx2 proteins. Moreover, mouse L929 cells express abundant Anx2 but are barely permissive to EV71 infection (only at a very high MOI); EV71 virions can attach to the surfaces of mouse L929 cells, and the attachment can be blocked by an antibody raised against the C terminus of human Anx2 (unpublished data). Thus, Anx2 alone is very inefficient for virus entry. Anx2 is not a transmembrane protein, and its interaction with VP1 is not within the canyon—-a structure typically involved in receptor binding by picornaviruses (
34). These characteristics of Anx2 are reminiscent of decay-accelerating factor (DAF), which serves as a primary receptor (
32) for the initial attachment of some enteroviruses and other viruses in preparation for further contact with a secondary receptor, which mediates conformational changes conducive to uncoating and cell entry. In this regard, it would be pertinent to investigate the role of Anx2 in conjunction with the recently identified EV71 receptor SCARB2 during the entry of EV71 into cells, although investigation into the mechanisms of virus entry at the molecular level is beyond the scope of this report. Furthermore, the question of which of the many cellular functions of Anx2 (
10,
13) was exploited by EV71, and the mechanisms involved, remains to be answered.
Our study is the first to demonstrate that Anx2 can enhance EV71 infectivity, especially at lower virus titers. Recent reports have shown that Anx2 is a receptor or cofactor for several types of viral infections and plays a role in the entry of human cytomegalovirus (CMV) into endothelial cells (
37,
49), of respiratory syncytial virus (RSV) into a hepatoma cell line (
23), and of HIV into macrophages (
22,
40). In addition, Anx2 could also contribute to the formation of infectious hepatitis C virus (HCV) particles (
2), and it activates plasminogen so as to promote influenza A virus replication in the absence of neuraminidase (
18). In all these cases, however, Anx2 enhances virus infectivity in a nonessential manner (
31,
36). Thus, the functional significance of Anx2 in natural infection by these important human viruses remains to be elucidated. To this end, comprehensive information on the distribution of Anx2 in tissues would be pertinent (
11); Anx2 expression has been reported in the suprabasal and basal cell layers of the normal human nasopharyngeal mucosa (
30), in proliferating cells but not in quiescent cells (
12), and in the human brain during fetal development (but Anx2 expression is silenced in normal adult brains) (
38). In rat cerebellum, Anx2 expression declines during development (
4). Further elucidation of the biological significance of Anx2 in EV71 infection would need to take into account the specific tissue distribution of Anx2 expression in relation to clinical observations such as the neurotropism of EV71 in children (
5,
52) and the fact that throat swabs are the clinical specimens most favorable (>90% of diagnosed cases) for EV71 isolation (
12,
47).
Because EV71 is mainly a human pathogen and has no animal model resembling natural EV71 infection in humans, our work on Anx2 so far has been based solely on the available
in vitro infection and neutralization assays; these assays are also commonly applied in the research and development of EV71 vaccines (
15,
20,
28,
50). While Anx2 is widely expressed in cell lines commonly used for EV71 amplification in the laboratory, immunohistochemical study of fatal EV71 infection has demonstrated that EV71 antigen was selectively detected only in human neurons at the time of death (
5,
52). Given the disparity between the cell type selectivity of EV71 in human infection and the highly promiscuous nature of EV71 in infecting numerous human cell lines in experimental settings, we caution that data collected with the
in vitro neutralization assay during vaccine testing might be potentially confounded by the role that Anx2 plays in the cell culture system.
In summary, our data show that Anx2 expressed on the cell surface can interact with EV71 and enhance its infectivity. While not an essential host factor for EV71 infection in general, Anx2 significantly enhances viral infectivity under unfavorable conditions, such as a low infective dose of virus. Because the true biological significance of Anx2-EV71 interaction during natural infection in humans remains to be elucidated, our work on the role of Anx2 in relation to EV71 infection in cell culture has served to highlight some unsolved issues that might be crucial and pertinent to the ongoing efforts at EV71 vaccine research and development.