Single Neutralizing Monoclonal Antibodies Targeting the VP1 GH Loop of Enterovirus 71 Inhibit both Virus Attachment and Internalization during Viral Entry

ABSTRACT

Antibodies play a critical role in immunity against enterovirus 71 (EV71). However, how EV71-specific antibodies neutralize infections remains poorly understood. Here we report the working mechanism for a group of three monoclonal antibodies (MAbs) that potently neutralize EV71. We found that these three MAbs (termed D5, H7, and C4, respectively) recognize the same conserved neutralizing epitope within the VP1 GH loop of EV71. Single MAbs in this group, exemplified by D5, could inhibit EV71 infection in cell cultures at both the pre- and postattachment stages in a cell type-independent manner. Specifically, MAb treatment resulted in the blockade of multiple steps of EV71 entry, including virus attachment, internalization, and subsequent uncoating and RNA release. Furthermore, we show that the D5 and C4 antibodies can interfere with EV71 binding to its key receptors, including heparan sulfate, SCARB2, and PSGL-1, thus providing a possible explanation for the observed multi-inhibitory function of the MAbs. Collectively, our study unravels the mechanism of neutralization by a unique group of anti-EV71 MAbs targeting the conserved VP1 GH loop. The findings should enhance our understanding of MAb-mediated immunity against enterovirus infections and accelerate the development of MAb-based anti-EV71 therapeutic drugs.

IMPORTANCE Enterovirus 71 (EV71) is a major causative agent of hand, foot, and mouth disease (HFMD), which has caused significant morbidities and mortalities in young children. Neither a vaccine nor an antiviral drug is available. Neutralizing antibodies are major protective components in EV71 immunity. Here, we unraveled an unusual mechanism of EV71 neutralization by a group of three neutralizing monoclonal antibodies (MAbs). All of these MAbs bound the same conserved epitope located at the VP1 GH loop of EV71. Interestingly, mechanistic studies showed that single antibodies in this MAb group could block EV71 attachment and internalization during the viral entry process and interfere with EV71 binding to heparan sulfate, SCARB2, and PSGL-1 molecules, which are key receptors involved in different steps of EV71 entry. Our findings greatly enhance the understanding of the interplays among EV71, neutralizing antibodies, and host receptors, which in turn should facilitate the development of an MAb-based anti-EV71 therapy.

INTRODUCTION

Enterovirus 71 (EV71) is one of the major causative agents of hand, foot, and mouth disease (HFMD), which has been prevalent in Southeast Asia (1). Patients with severe HFMD cases manifest neurological complications, such as brainstem encephalitis and pulmonary edema, resulting in death, and such cases are often associated with EV71 infection (2). No prophylactic vaccine against EV71 infection or therapeutic drug for EV71 infection is available. Even though inactivated whole-virus-based EV71 vaccines have progressed into clinical trials (3, 4), their licensure faces serious challenges (5). Thus, the development of efficacious antiviral drugs for the treatment of EV71-infected patients with critical clinical conditions is urgently needed (6).

EV71 belongs to the Enterovirus genus of the Picornaviridae family. It possesses a single-stranded positive-sense RNA genome, which is encapsidated within an icosahedral protein shell made up of 60 copies of each of the VP1, VP2, VP3, and VP4 capsid subunit proteins (7, 8). Similar to other picornaviruses, EV71 entry into susceptible cells may involve multiple consecutive steps, including attachment, internalization, uncoating, and RNA release (9), which require coordinated interactions of the virus with compatible receptors and are often accompanied by serial conformational changes of the viral capsid (7, 10). A number of molecules have been identified to be receptors for EV71, including heparan sulfate, SCARB2, and PSGL-1 (11). Specifically, heparan sulfate, which is universally distributed on the surface of all animal cells (12), facilitates initial attachment during EV71 entry (13), whereas SCARB2, which locates primarily in the membranes of lysosomes and endosomes, mediates virus uncoating (14–17). Upon SCARB2 binding under acidic conditions, viral capsids undergo conformational changes, shifting from 160S particles to 135S particles and further to 80S particles, with these changes ultimately leading to capsid dissociation and the release of viral RNA (10, 18). PSGL-1 facilitates virus infection of leukocytes in a strain-specific manner (19, 20); however, its exact role in viral entry in nonleukocytes remains unclear. Each of the EV71-receptor interactions is critical for the successful entry of EV71 into host cells and the establishment of infection.

Passive transfer of neutralizing antiserum protects mice from lethal EV71 infection in vivo (21–23), indicating that neutralizing antibodies are major protective components in anti-EV71 immunity. The use of monoclonal antibodies (MAbs) with neutralization capabilities represents an excellent strategy in the development of antiviral drugs due to their high specificity and potency (24), as exemplified by the successful commercialization of palivizumab, a humanized MAb against respiratory syncytial virus (25). Recently, a number of anti-EV71 neutralizing MAbs have been generated (26–28). Plevka et al. showed that an anti-EV71 MAb could mediate neutralization by induction of a conformational change, resulting in genome release (29). However, how neutralizing antibodies inhibit EV71 entry, especially the interplay among the neutralizing antibody, the virus, and the virus receptors, remains largely unknown. Therefore, in order to accelerate the development of an effective MAb-based drug for the prevention and treatment of EV71 infections, it is important to fully characterize anti-EV71 neutralizing MAbs and determine their modes of action.

We previously generated three MAbs (termed D5, H7, and C4) that potently neutralize EV71 (30). In the present study, we comprehensively characterized these MAbs and investigated their inhibitory mechanisms. Our results show that all three MAbs bind the same epitope located within the VP1 GH loop of EV71. Moreover, single MAbs were found to be capable of inhibiting both the virus attachment and internalization steps of EV71 entry and interfering with multiple interactions between the virus and its receptors.

MATERIALS AND METHODS

Cells and viruses.

RD and Vero cells were maintained as previously described (21). Jurkat T cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 μg/ml streptomycin. EV71 clinical isolate G082 and a mouse-adapted EV71 strain termed EV71 MAV-W were described previously (31). Virus titers were determined and expressed as the 50% tissue culture infective dose (TCID₅₀), as previously described (21).

Antigens and peptides.

Mature EV71 particles (also termed “F particles” [EV71-F] and immature capsids (also termed “E particles” [EV71-E]) were prepared as previously described (21). Virus-like particles (VLP) of EV71 were prepared as previously described (32).

A series of 58 peptides spanning the entire VP1 region of EV71 was synthesized in a previous study (32). Each peptide consists of 15 amino acid residues and has 10 residues that overlap with the adjacent peptides (32). In addition, for fine mapping of the epitope, a set of 10 peptides, each of which covers amino acids 211 to 220 of the VP1 protein and which also carries a single alanine substitution at 1 of the 10 positions, was synthesized.

Antibodies.

Three murine anti-EV71 MAbs, namely, D5, H7, and C4, were developed in a previous study (30). The anti-HIV MAb 3F5 was used as a control and was generated in-house from a mouse immunized with N46FdFc. Rabbit polyclonal antisera against the EV71 VP1 protein and mouse polyclonal antisera against EV71 VLP were generated in a previous study (32). Rabbit polyclonal antiserum against EV71 VLP was generated in-house. The animal immunization experiments were approved by the Institutional Animal Care and Use Committee at the Institut Pasteur of Shanghai.

ELISAs.

The binding specificity of the MAbs was determined by an enzyme-linked immunosorbent assay (ELISA) with a variety of capture antigens. Briefly, ELISA plates were coated overnight at 4°C with one of the following antigens: 20 ng/well of the F particle, E particle, or VLP of EV71 or 400 ng/well of each of the peptides described above. Then, the wells were blocked with 5% milk in phosphate-buffered saline (PBS)–Tween 20 (PBST) for 2 h and incubated with 50 μl (1 μg/ml) of each MAb for 2 h, followed by incubation with a horseradish peroxidase (HRP)-conjugated secondary antibody for 1 h. After each incubation step, the plates were rinsed with PBS three times. After color development, the absorbance (the optical density at 450 nm [OD₄₅₀]) was determined using a 96-well plate reader.

Biolayer interferometry assay.

Antibody binding to EV71 was analyzed by biolayer interferometry on an Octet Red 96 system (ForteBio) in kinetics buffer (PBS buffer supplemented with 0.1% bovine serum albumin [BSA] and 0.02% Tween 20) at room temperature. The inactivated EV71 (F particle) was labeled with biotin using an EZ-Link sulfo-NHS-LC-LC-biotin kit (Thermo Scientific). After a brief rinse in kinetics buffer, the streptavidin (SA) biosensor tips were dipped into 0.066 μg/ml of EV71-biotin solution for 10 min. Following a rinse in kinetics buffer, the EV71-immobilized biosensor tips were allowed to associate with antibody at different concentrations (40, 8, 1.6, 0.32, 0.064, and 0.0128 μg/ml) for 25 min and then dissociate in kinetics buffer for 10 min. The EV71-bound biosensor was also allowed to associate with kinetics buffer alone (without antibody) to serve as a loading control. In addition, an empty sensor tip (without EV71 loading) was allowed to associate with 40 μg/ml of antibody to assess nonspecific antibody binding. Data were processed using Octet data analysis (v6.4) software (ForteBio), and affinity values were calculated by steady-state analysis.

Generation and sequencing of MAb-resistant mutant viruses.

To generate MAb-resistant mutants, EV71 strain G082 (1 × 10⁸ TCID₅₀s) was mixed with 400 μg/ml of an MAb (D5 or C4), and the mixtures were incubated at room temperature for 1 h and then at 37°C for 2 h before addition to 2 × 10⁶ RD cells. After incubation for 2 days at 37°C, the infected cell cultures were harvested, analyzed for virus titer, and then used for second-round infections in the presence of 1 mg/ml of the MAbs. After two rounds of selection, the MAb-resistant viruses were plaque purified. Individual mutant clones were subjected to reverse transcription-PCR (RT-PCR) amplification of the P1 region using forward primer 5′-GGCCATCCGGTGTGCAACAG-3′ and reverse primer 5′-AGCAAGTCGCGAGAGCTGTC-3′ (32) and subsequent sequencing analysis.

Neutralization assays.

All neutralization assays were performed using 96-well plates. In the standard neutralization assays, 50 μl of diluted MAbs was mixed with 50 μl of EV71 G082 (100 TCID₅₀s) in Dulbecco modified Eagle medium (DMEM), and the mixtures were added to RD cells. After incubation at 37°C for 72 h, the cells were analyzed for viability by a 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT)-based method as described previously (32). For a given sample, the neutralization efficiency was calculated by normalization of its OD₄₉₀ value to that of the sample treated with virus only as follows: percent neutralization = [(OD₄₉₀ of the given sample − OD₄₉₀ of the sample treated with virus only)/(OD₄₉₀ of the mock-treated sample − OD₄₉₀ of the sample treated with virus only)] × 100. The 50% inhibitory concentration (IC₅₀) was defined as the MAb concentration that reduced cell death by 50%.

In the preattachment neutralization assays, MAb and virus were mixed as described above for the standard neutralization assay; however, the mixtures were then incubated with cells at 4°C instead of 37°C for 1 h to allow the attachment of virus or the virus-MAb complex to the cell surface. Next, the cells were washed with DMEM three times and then incubated at 37°C for 72 h. Cell viability and the neutralization efficiency for each treatment were determined as described above.

The postattachment neutralization assays were performed with both RD and Jurkat T cells. For RD cells, 50 μl of EV71 G082 (100 TCID₅₀s) in DMEM was first incubated with RD cells at 4°C for 1 h to allow attachment. The cells were washed three times to remove unbound virus and then immediately treated with 50 μl of diluted MAb at 37°C for 1 h. After another three washes, the cells were incubated at 37°C for 72 h and then analyzed for viability and inhibitory efficiency as described above. For Jurkat T cells, the postattachment neutralization assays were performed with the following modifications: 2 × 10⁴ Jurkat T cells per well were seeded in 96-well plates; 1,000 TCID₅₀s of EV71 G082 was used for each treatment. After incubation at 37°C for 72 h, the samples were collected and analyzed by quantitative RT-PCR (qRT-PCR) with a pair of EV71-specific primers described previously (21) and a pair of β-actin-specific primers (forward primer, GGACTTCGAGCAAGAGATGG; reverse primer, AGCACTGTGTTGGCGTACAG). Data analysis was performed using the 2^−ΔΔCT method as described previously (21). The threshold cycle (C_T) values of all samples were first normalized to the C_T value of β-actin and then compared to the C_T value of the control (samples treated with virus only). The postattachment neutralization efficiency for a given treatment group was defined as the EV71 genome copy number for the group as a percentage of that for the control (which was treated with the virus only).

To determine the time window at the postattachment stage within which the MAbs remained functional, the postattachment neutralization assay was performed as described above, except that the virus-bound cells were incubated at 37°C for 0, 30, 60, 90, or 120 min prior to addition of the MAbs.

Neutralization inhibition assay.

Inhibition of MAb-mediated neutralization of EV71 infection by synthetic peptides was determined by a neutralization inhibition assay as previously described (32).

Inhibition of virus attachment by MAbs.

Different amounts of MAbs were mixed with 9 × 10⁶ TCID₅₀s of EV71 G082 in DMEM containing 2% FBS, followed by incubation at 37°C for 1 h. The mixtures were added to RD, Vero, or Jurkat T cells (3.5 × 10⁵ cells per treatment). Then, the cells were incubated at 4°C for 1 h to allow virus attachment. After three washes with cold PBS to remove unbound virus, the cells were harvested and analyzed by Western blotting or qRT-PCR assays as described previously (21, 32). The qRT-PCR assays were performed with a slight modification: the amount of β-actin mRNA, used as an internal control, was measured.

Immunofluorescence staining and confocal microscopy.

Vero cells were seeded (5 × 10⁴ cells per well) onto 24-well plates 1 day prior to the experiment. On the second day, 7.8 × 10⁶ TCID₅₀s of EV71 F particles diluted in DMEM was added to each well, followed by incubation at 4°C for 1 h to allow virus attachment. The cells were washed three times with PBS to remove unbound virus. Then, 10 μg of MAb (D5 or 3F5) diluted in DMEM was added to each well. The plates were incubated at 4°C for 1 h to allow MAb binding to virus. After the cells were washed three times with PBS to remove free antibodies, 300 μl DMEM containing 2% FBS was added to each well, and the cells were incubated at 37°C for 0, 30, 120, or 300 min. Then, the cells were subjected to immunostaining. Briefly, cells were fixed using 4% paraformaldehyde at room temperature for 20 min, followed by treatment with permeabilization buffer (0.02% Triton X-100 and 0.01% NP-40 in PBS) for 5 min. The fixed cells were blocked with PBS containing 10% FBS plus 10% BSA for 1 h and incubated with 1/100-diluted anti-EV71 VLP rabbit polyclonal antibody for 1 h. Next, the cells were incubated with Alexa Fluor 488-conjugated anti-mouse IgG and Alexa Fluor 555-conjugated anti-rabbit IgG (Life Technologies, Carlsbad, CA) for 1 h, followed by 4′,6-diamidino-2-phenylindole (DAPI) staining for 5 min. All incubation steps were carried out at 37°C, and the cells were rinsed three times with PBS between steps. The stained samples were examined using a Leica DM6000B laser scanning confocal microscope (Leica, Germany), and images were collected and analyzed using the Leica Advanced Fluorescence Lite application suite.

Analysis of virus uncoating and RNA release.

Vero cells were seeded (2.5 × 10⁶ cells per well) onto 6-well plates 1 day prior to the assays. EV71 F particles (3.9 × 10⁷ TCID₅₀s) in 1 ml DMEM containing 2% FBS were added to each well, followed by incubation at 4°C for 1 h to allow virus attachment. The cells were washed three times with PBS to remove unbound virus. Then, the cells were either immediately lysed with lysis buffer (0.2% Triton X-100 and 0.1% NP-40 in PBS) or incubated with 50 μg/well of MAb at 4°C for 1 h, followed by three washes with DMEM and subsequent incubation at 37°C for 2 h to allow uncoating/RNA release to occur. Finally, the cells were harvested and lysed. The cell lysates were subjected to three freeze-thaw cycles and subsequent ultrasonication for 3 min. The lysates were clarified by centrifugation at 12,000 rpm for 5 min, and the resultant supernatant was subjected to a 15% to 55% discontinuous sucrose gradient ultracentrifugation at 39,000 rpm for 3 h. After centrifugation, 24 sucrose gradient fractions were collected and analyzed by ELISA, in which a rabbit anti-EV71 VLP polyclonal antibody was used as the detection antibody. The lysate from Vero cells without any treatment was also fractionated and used in the ELISA to eliminate the background signal of each fraction. The fractions were also subjected to Western blotting with the rabbit anti-EV71 VP1 polyclonal antibody.

To examine viral RNA release, sucrose gradient fractions corresponding to the uncoated virus were pooled and subjected to RNA extraction using a PureLink RNA minikit (Life Technologies). Subsequently, 40 ng of total RNA was used for reverse transcription with SuperScript II reverse transcriptase (Life Technologies), and the resultant cDNA was used as a template for qRT-PCR assays. A plasmid, pIEXBac-P1, containing the EV71 P1 gene (32) served as a standard template for calculation of absolute virus copy numbers. Data are given as the EV71 genome copy number for the treatment groups as a percentage of that for the control (which was treated with the virus only).

Pulldown assays.

For heparin pulldown assays, different amounts of MAb were mixed with 3.4 × 10⁶ TCID₅₀s of EV71 G082 in 400 μl DMEM containing 2% FBS, and the mixtures were incubated at 37°C for 1 h, followed by addition of 20 μl of heparin-agarose (catalog number H6508-5ML; Sigma) and subsequent incubation at 4°C for 2 h. The antibody-virus-bead mixtures were subjected to centrifugation at 3,000 rpm for 3 min. The supernatants were discarded, and the beads were washed three times. Virus-bound beads were analyzed by SDS-PAGE and Western blotting with a rabbit anti-VP1 polyclonal antibody.

For the SCARB2 and PSGL-1 pulldown assays, 1 μg of recombinant SCARB2-Fc (catalog number 1966-LM-050; R&D Systems), PSGL-1–Fc (catalog number 3345-PS-050; R&D Systems), or control human Fc (catalog number 10702-HNAH; Sino-Biotech) was mixed with 15 μl of anti-human Fc beads (catalog number A3316-5ML; Sigma) in PBS. The mixtures were incubated at 4°C overnight with gentle rotation. Unbound protein was removed by washing with PBS. EV71 G082 (1 × 10⁷ TCID₅₀s) was mixed with different amounts (1, 10, or 100 μg) of antibody in 1 ml DMEM containing 2% FBS, and the mixture was then incubated at 37°C for 1 h. The virus-antibody mixtures were added to the beads prepared as described above, followed by incubation at 4°C for 2 h with gentle rotation. Unbound virus and antibody were removed by three washes with PBS. The remaining beads were resuspended in 100 μl SDS-PAGE sample buffer and then boiled for 10 min. The resulting protein samples were subjected to SDS-PAGE and Western blotting, using either an anti-VP1 rabbit polyclonal antibody or an HRP-conjugated goat polyclonal antibody to human IgG (catalog number ab6858; Abcam).

Generation and characterization of SCARB2-KD cells.

SCARB2-knockdown (KD) cells were generated according to a previously described protocol (16) with slight modifications. Briefly, a short hairpin RNA (shRNA) targeting human and simian SCARB2 mRNA was cloned into the pLKO.1-puro vector (Sigma), yielding the plasmid pLKO.1-shSCARB2. Lentivirus was packaged by cotransfection of 293T cells with packaging plasmids pVSV-G, pCDH-Δ8.9, and pLKO.1-shSCARB2. The cell culture supernatant was harvested at 48 h posttransfection, and cell debris was removed by centrifugation at 3,800 rpm for 10 min, yielding the crude lentivirus. The crude lentivirus preparation was then concentrated by ultracentrifugation at 25,000 rpm for 3 h, resulting in the lentivirus stock. Wild-type RD cells or Vero cells were infected with the shSCARB2-carrying lentivirus. After incubation for 4 days, the infected cells were screened using 2.0 μg/ml of puromycin (Sigma). Individual puromycin-resistant clones were isolated and then tested for SCARB2 expression by a relative qRT-PCR assay with primers hSCARB2-F (CCAATACGTCAGACAATGCC) and hSCARB2-R (ACCATTCTTGCAGATGCTGA). The GAPDH (glyceraldehyde-3-phosphate dehydrogenase) gene was also assayed with primers GAPDH-F (GAAGGTGAAGGTCGGAGTC) and GAPDH-R (GAAGATGGTGATGGGATTTC) and served as the internal control. Virus attachment to the SCARB2-knockdown cells was analyzed by qRT-PCR or MTT assays as described above.

Statistics.

All statistical analyses were performed using GraphPad Prism (v5) software. Kaplan-Meier survival curves were compared by the log-rank test. A two-way analysis of variance was used to compare the neutralization curves at the pre- and postattachment stages, the neutralization inhibition efficiencies of the peptides, the viral growth curves, and cell viabilities. Other results were analyzed by the Student two-tailed t test unless otherwise stated. Statistical significance was shown as a P value of <0.05.

RESULTS

Characterization of the three anti-EV71-neutralizing MAbs.

Three murine anti-EV71 MAbs (D5, H7, and C4) developed previously (30) were analyzed for their binding specificity by ELISA. As shown in Fig. 1A, all of the MAbs bound three different forms of EV71 particles, including the authentic F particle (EV71-F), E particle (EV71-E), and recombinant VLP; in contrast, none of them reacted with the other two HFMD-causing enteroviruses, namely, coxsackievirus A16 (CA16) and coxsackievirus A10 (CA10), demonstrating that the D5, H7, and C4 antibodies are EV71 specific.

FIG 1.

Characterization of D5, H7, and C4. (A) Binding specificities of the MAbs. Three forms of EV71 particles (EV71-F, EV71-E, and EV71 VLP) and two inactivated viruses (CA16 and CA10) were examined for their reactivity to the MAbs by ELISA. (B) Affinity of binding of the MAbs to the EV71 F particle determined by biolayer interferometry analysis. K_D, equilibrium dissociation constant. (C) Abilities of D5, H7, and C4 to neutralize the EV71 G082 strain. The neutralization IC₅₀ for each MAb is shown.

Biolayer interferometry analysis showed that the affinities of the MAbs to the EV71 F particle were the same, with the equilibrium dissociation constant values being 7.2, 7.3, and 8 nM for antibodies D5, H7, and C4, respectively (Fig. 1B). However, the standard neutralization assay revealed that the three MAbs had distinct neutralization capabilities, with IC₅₀s being 0.203, 0.287, and 0.952 μg/ml for D5, H7, and C4, respectively (Fig. 1C). These results indicate that the EV71 neutralization potency of C4 is less than that of D5 and H7, even though their affinities for the virus are very similar.

D5, H7, and C4 recognize the same epitope, located at the VP1 GH loop of EV71.

D5, H7, and C4 MAbs reacted with the VP1 but not the VP0 capsid protein in both ELISA and Western blot assays (30), suggesting that their binding epitopes reside within VP1. To define the corresponding epitopes, we screened a panel of 58 overlapping peptides that span the entire VP1 region for reactivity with the MAbs. These peptides are 15 amino acids in length and have 10 residues overlapping with the adjacent peptides (32). As shown in Fig. 2A, D5 and H7 reacted strongly with peptides 42 and 43 but less so with four other peptide clusters represented by peptides 1, 5, 20, and 29, respectively, whereas C4 reacted only with peptides 42 and 43. To determine whether these MAb-binding peptides are indeed involved in MAb recognition and function, we performed neutralization inhibition assays using the five representative peptides 1, 5, 20, 29, and 42. As shown in Fig. 2B, a low concentration (0.031 μg/ml) of peptide 42 significantly reduced the neutralization activities of D5 MAb, whereas the other peptides had no significant inhibitory effect even when the highest concentration (20 μg/ml) was used. In addition, the inhibitory effect exerted by peptide 43 was similar to that exerted by peptide 42 (Fig. 2B). Only peptides 42 and 43 and not the other peptides were consistently found to inhibit the neutralization mediated by the H7 or C4 antibody (data not shown).

FIG 2.

Epitope mapping. (A) VP1 peptide ELISA. A panel of 58 peptides spanning the entire VP1 region was screened for reactivity to D5, H7, and C4 by ELISA. (B) Inhibition of D5-mediated neutralization by the selected peptides. VP1-derived peptides 1, 5, 20, 29, 42, and 43, as well as an irrelevant control peptide Fd, were used for incubation with D5 or for 1 h at 37°C. The peptide-MAb mixtures were subjected to neutralization testing. Neutralization inhibition by the peptides was determined using an MTT method. Data are means ± SDs of the OD₄₉₀ readings for triplicate wells. n.s., not significant (P ≥ 0.05); ***, P < 0.001. (C) Amino acid sequence alignment of peptides 41 to 44 and the GH loop of VP1. The core sequence for MAb binding is highlighted in red. (D) Binding of MAbs to peptide mutants. A panel of 10 peptides with a single Ala amino acid replacement at each site from amino acids 211 to 220 of the GH loop was screened for reactivity to D5, H7, and C4 by ELISA. The ELISA data are means ± SDs of the OD₄₅₀ readings from triplicate wells. (E) Locations of the MAb epitope and SCARB2 binding regions. Light blue, pale green, and salmon, EV71 capsid proteins VP1, VP2, and VP3, respectively. Five SCARB2 binding clusters, including three in VP1 (marked in red) and one each in VP2 and VP3 (both marked in yellow), are shown in one pentamer (left) and a zoomed-out icosahedral asymmetric unit (right). Residue Lys149 within VP2 (VP2-K149) is marked in blue.

Alignment of the amino acid sequences of peptides 42 and 43 and the adjacent peptides 41 and 44 mapped the MAb epitopes to residues 211 to 220 (FGEHKQEKDL) of VP1 (Fig. 2C). To identify the residues critical for MAb binding, alanine scanning analysis was performed. As shown in Fig. 2D, alanine (A) substitution of residue 211, 212, 213, 214, or 220 did not affect the binding of the corresponding peptides to the three MAbs, whereas point mutations at positions 215 to 219 resulted in the complete loss of binding, indicating that residues 215 to 219 are essential for MAb binding.

In addition, we generated EV71 mutants which were resistant to D5- or C4-mediated neutralization. These mutant viruses lost the ability to react with D5 or C4 in the ELISA (Table 1). Sequencing of the P1 structural region of these escape mutants revealed that each of them carried a single amino acid change at position 218 of VP1 (K218E, K218T, or K218N), and no further mutation was found at other sites (Table 1).

TABLE 1.

Characterizations of the MAb-resistant mutants

Virus	Total no. of mutants	Mutation site (frequencya)	Binding with MAbb	Concn for neutralization by MAb (μg/ml)
Wild-type virus	NAc	NA	++	≤1.25
D5-resistant mutant	9	K218E (3/9), K218T (6/9)	−	>80
C4-resistant mutant	9	K218E (4/9), K218T (2/9), K218N (3/9)	−	>80

^aFrequencies are number of mutants with the mutation at the indicated mutation site/number of mutants tested.

^bThe abilities of the MAbs to bind to the mutant viruses were analyzed by indirect ELISA, and the results were expressed as follows: −, OD₄₅₀ < 0.2; ++, OD₄₅₀ ≥ 0.6.

^cNA, not applicable.

Collectively, the results presented above demonstrate that the D5, H7, and C4 antibodies share the same epitope mapped to amino acids 211 to 220 of VP1 (Fig. 2E). This epitope is completely conserved among all EV71 subgenotypes and is located at the surface-exposed GH loop of VP1 and close to the canyon region, according to the crystal structure of EV71 (PDB accession number 3VBH) (7) (Fig. 2E).

Because D5 and H7 bound the same epitope and exhibited comparable neutralization potencies, it is possible that they might have been derived from the same B cell clone. We thus determined the sequences of the variable heavy chain (V_H) and variable light chain (V_L) regions of the three MAbs and found that none of them was identical (data not shown). Therefore, D5, H7, and C4 are distinct antibodies, yet they constitute a group of anti-EV71 MAbs targeting the same epitope on the VP1 GH loop. In the following studies, we explored the working mechanism of this MAb group, using the highly neutralizing MAb D5 and the modestly neutralizing MAb C4 as models.

D5 and C4 block EV71 attachment to susceptible cells.

To determine the mode of action of the MAbs, we first examined whether D5 or C4 could inhibit virus attachment, the first step of viral entry, to three susceptible cell lines, including RD, Vero, and Jurkat T cells. Preincubation of EV71 with D5 reduced the amount of attached virus in a dose-dependent manner in all three cell lines, as demonstrated by a Western blot analysis which detects the VP1 protein of EV71 (Fig. 3A). A similar trend of dose-dependent inhibition of virus attachment was also observed for C4 (Fig. 3B) but not for a control MAb, 3F5 (Fig. 3A and B). Real-time PCR analysis, which quantifies viral RNA, also consistently showed antibody dose-dependent inhibition of EV71 attachment by D5 and C4 in the three cell lines (Fig. 3C to E). These results demonstrate that the attachment inhibition by D5 and C4 was cell type independent. It is worth noting that, in both assays, the amount of cell-bound virus in the D5-treated samples was significantly lower than that in the C4-treated ones when the same antibody doses were used (Fig. 3A to E), indicating that D5 was more potent than C4 in preventing virus attachment to susceptible cells, consistent with their neutralization potency (Fig. 1C) determined using the standard neutralization protocol. We then performed preattachment neutralization assays to determine the IC₅₀s of the MAbs applied prior to virus attachment and found that D5 (IC₅₀ = 0.324 μg/ml) was also more potent than C4 (IC₅₀ = 4.32 μg/ml) in preattachment neutralization (Fig. 4A and B). Thus, the efficiency with which D5 and C4 inhibited virus attachment correlated well with the preattachment neutralization potency of these two MAbs.

FIG 3.

D5 and C4 block virus attachment to susceptible cells. Different amounts of D5 or C4 were mixed with 9 × 10⁶ TCID₅₀s of EV71 G082, and then the mixture was incubated at 37°C for 1 h before it was added to RD, Vero, or Jurkat T cells. The cells were incubated at 4°C for 1 h to allow virus attachment and then washed 3 times with cold PBS to remove unbound virus. The cell-attached virus was analyzed by Western blotting (immunoblotting [IB]) (A and B) or qRT-PCR (C to E). β-Actin was used as the internal control. Means ± SDs for triplicate wells are shown. Dashed lines, the viral genome level of the control group (treated with virus only). n.s., not significant (P ≥ 0.05); **, P < 0.01; ***, P < 0.001.

FIG 4.

D5 and C4 inhibit EV71 infection at the postattachment stage. (A) Neutralization efficacies of D5 at the pre- and postattachment stages. (B) Neutralization efficacies of C4 at the pre- and postattachment stages. (C and D) Neutralization efficacies of D5 (C) and C4 (D) at different time points postattachment when added to RD cells to which virus was attached. The cells to which virus was attached were incubated without MAb at 37°C for the indicated periods of time and then treated with MAbs. Cell viability was measured, and the percent neutralization for each treatment was calculated. Means ± SDs for triplicate wells are shown. Dashed lines, 50% neutralization. (E) Postattachment neutralization efficiencies of D5 and C4 in Jurkat T cells. Different concentrations of D5 or C4, as indicated, were used to treat Jurkat T cells, to which EV71 was preattached. (F) Effect of times of antibody addition on the postattachment neutralization efficacy of D5 or C4. Jurkat T cells to which virus was attached were incubated without MAb at 37°C for the periods of time indicated and then treated with MAbs. For each treatment, the viral RNA levels relative to those for the group treated with virus only are presented. Means ± SDs for triplicate wells are shown. Dashed line, 50% level of relative amount of viral RNA. n.s., not significant (P ≥ 0.05); **, P < 0.01; ***, P < 0.001.

D5 inhibits EV71 infection at the postattachment stage as efficiently as at the preattachment stage.

Antibody-mediated inhibition of viral entry at the postattachment stage is an important neutralization mechanism, which has been commonly reported for enveloped viruses but rarely for nonenveloped viruses (33–37). We therefore investigated whether D5 or C4 was capable of neutralizing EV71 infection at the postattachment stage. D5 was found to efficiently inhibit EV71 infection in RD cells at the postattachment stage, with the neutralization curve at the postattachment stage being almost identical to that at the preattachment stage; the IC₅₀s at both stages were comparable (0.324 μg/ml versus 0.539 μg/ml) (Fig. 4A). However, the neutralization curves for C4 at the two stages were significantly different, with C4 having a much higher neutralization potency at the preattachment stage than at the postattachment stage (4.32 μg/ml versus 12.43 μg/ml) (Fig. 4B).

To define the time frame within which the MAbs were able to exert postattachment inhibition, virus-bound RD cells were brought to 37°C for different periods of time before D5 or C4 antibody was applied. As shown in Fig. 4C and D, for both D5 and C4, a longer exposure at 37°C resulted in less neutralization, regardless of the antibody concentrations. For example, both D5 and C4 were about 50% effective at inhibition when administered at 30 min postexposure to 37°C but much less effective when administered at later time points. Together, these results indicate that the early administration of the MAb, e.g., within 30 min of exposure to 37°C, was critical for its inhibitory effect on the postattachment entry process.

In Jurkat T cells, D5 and C4 also exhibited neutralization at the postattachment stage, and similar trends in the time-dependent decrease in postattachment inhibition were observed (Fig. 4E and F). Because Jurkat T cells use PSGL-1 but RD cells use SCARB2 as a receptor to support EV71 infection, the results presented above indicate that our MAbs could effectively mediate postattachment neutralization in both PSGL-1- and SCARB2-expressing cells.

D5 inhibits EV71 internalization, uncoating, and RNA release.

We then used D5 as a representative antibody to investigate how these MAbs exert their inhibitory effect at the postattachment stage. EV71-preadsorbed cells were treated with D5 or the control MAb (3F5) at 37°C for different periods of time and then subjected to immunostaining and confocal microscopy to examine the localization of the virus and MAb. For the control MAb 3F5-treated cells not incubated at 37°C, the virus was detected at the cell periphery, whereas for cells incubated at 37°C for 30 min, positive signals were found in the cytoplasm, suggesting that the virus was rapidly internalized. Similarly, intracellular localization of the virus was evident in the cells with longer incubation times (120 and 300 min) (Fig. 5A, right). For the D5-treated cells, D5 and virus were colocalized at the cell periphery regardless of the time of incubation at 37°C (Fig. 5A, left), indicating that D5 and EV71 formed antibody-virus complexes which were prohibited from being internalized.

FIG 5.

D5 inhibits virus internalization, uncoating, and RNA release at the postattachment stage. (A) Localization of EV71 and MAbs revealed by confocal microscopy. EV71 F particles were preattached to Vero cells, and then the cells were treated with D5 or 3F5. The cells were then incubated at 37°C for 0, 30, 120, or 300 min, as indicated. Then, the cells were subjected to immunostaining and confocal microscopy as described in Materials and Methods. Green signal, D5 or the control 3F5 mouse MAb; red signal, the presence of EV71, which was recognized by a rabbit anti-VLP polyclonal antibody; blue signal, DAPI staining; Merge 1, merge of the green, red, and blue channels; Merge 2, merge of the green, red, blue, and white-field channels. (B to D) Inhibition of virus uncoating and RNA release by D5 treatment at the postattachment stage. Vero cells to which EV71 F particles were preattached were either immediately lysed to generate the EV71, 4°C only samples or incubated at 37°C for 2 h to allow uncoating/RNA release with the presence of MAbs. The samples were harvested and lysed to generate the EV71 + D5 and EV71 + 3F5 samples. Finally, all the samples described above were subjected to sucrose gradient ultracentrifugation as described in Materials and Methods. The resulting sucrose gradient fractions were analyzed by ELISA (B), Western blotting (C), and qRT-PCR (D), as described in Materials and Methods. The ELISA data are reported as means ± SDs of the OD₄₅₀ values for triplicate wells. The qRT-PCR data are reported as means ± SDs of the relative virus genome numbers. Dashed line, the background level of the assay. Representative results from two independent experiments are shown. n.s., not significant (P ≥ 0.05); **, P < 0.01.

For picornaviruses, mature virions undergo an uncoating process which involves capsid conformational changes at the postattachment stage to facilitate the release of viral RNA (38, 39). Specifically, it has been reported that the uncoating of EV71 virions results in two types of altered particles, A (or 135S) particles (10, 16) and empty capsids (or 80S particles) (14). To determine whether EV71 uncoating is affected by D5 binding, we performed sucrose gradient sedimentation to analyze the particle characteristics of the antibody-treated or nontreated viruses. Vero cells were incubated with mature virions (also termed F particles [14, 40]) at 4°C for 1 h to allow virus attachment. Vero cells to which F particles had been preadsorbed were then treated with D5 or 3F5 at 4°C for 1 h, followed by incubation at 37°C for 2 h. The lysates from antibody-treated and nontreated cells were subjected to sucrose gradient analyses. For the sample without antibody treatment and incubation at 37°C (the EV71, 4°C only sample; Fig. 5B and C), the virus was mainly detected in gradient factions 17 to 21, representing the infectious F particle (14, 40). The MAb 3F5-treated sample (the EV71 + 3F5 sample; Fig. 5B and C) did not show significant reactivity by ELISA, but the viral protein was detected in slowly sedimenting fractions 3 to 6 by Western blotting, suggesting that the viral particles had dissociated. In contrast, for the D5-treated sample (the EV71 + D5 sample; Fig. 5B and C), the viral proteins were mainly detected in fractions 17 to 19 by ELISA and by Western blotting, suggesting that D5-bound virus retained the F-particle conformation.

We further investigated whether the release of RNA from the virions had occurred. Gradient fractions 17 to 21, which corresponded to the F particle, were pooled for each treatment and analyzed for viral genomic RNA by real-time PCR. We found that the amount of viral RNA of the EV71 + D5 samples was significantly higher than that of the control EV71 + 3F5 samples and was comparable to that of the EV71, 4°C only samples (Fig. 5D), suggesting an abortive RNA release in the EV71 + D5 samples. Collectively, the results presented above demonstrate that the internalization and subsequent uncoating and RNA release during EV71 entry were blocked by D5 at the postattachment stage.

MAbs interfere with EV71 binding to heparin.

To unravel the molecular basis of the multiple inhibitory functions of the MAbs, we investigated whether D5 or C4 could interfere with EV71 binding to its receptors.

Heparan sulfate glycosaminoglycan is a recently identified attachment receptor for EV71 (13). It has been shown that EV71 interacts with heparin in vitro and can therefore be pulled down using heparin-conjugated agarose (13). We thus determined whether MAb treatment could block this interaction. D5 treatment greatly reduced the amount of virus pulled down in a dose-dependent manner, with 95% and 99% reductions occurring when 1 μg and 10 μg of D5 were applied, respectively (Fig. 6, top). Similarly, C4 also showed the same trend of a dose-dependent inhibition, although only 68% and 86% reductions were observed when 1 μg and 10 μg of antibody were applied, respectively (Fig. 6, bottom). For control antibody 3F5, no significant inhibition was detected, regardless of the antibody doses. These results indicate that both D5 and C4 can block EV71 binding to heparin and the blocking efficiency of D5 is higher than that of C4, which may explain the observation that D5 was more potent than C4 in inhibiting EV71 attachment to susceptible cells (Fig. 3).

FIG 6.

D5 or C4 treatment interferes with EV71 binding to heparin. Different amounts of D5 or C4 were mixed with EV71 G082, and then the mixture was incubated at 37°C for 1 h. Subsequently, the antibody-virus mixtures were incubated with the immobile heparin on agarose beads. After washing of the beads to remove unbound virus, the agarose beads were precipitated by centrifugation, and the binding of the virus was detected by Western blotting with a virus-specific antibody.

SCARB2 does not play a significant role in EV71 attachment to susceptible cells.

SCARB2 is an identified uncoating receptor for EV71 (15). Previous studies have demonstrated that SCARB2 is capable of binding EV71 in vitro (14). However, SCARB2 is primarily located at the membrane of the endosome and lysosome but not the plasma membrane (17); it is thus questionable whether SCARB2 plays a role in the initial attachment of EV71 to cells. To determine the necessity of SCARB2 for EV71 attachment, we generated SCARB2-knockdown (KD) RD and Vero cell lines, designated RD-SCARB2-KD and Vero-SCARB2-KD, respectively (Fig. 7A and B). Both RD-SCARB2-KD and Vero-SCARB2-KD cells were found to be much less permissive to EV71 infection than the wild-type counterparts in viral growth kinetics (Fig. 7C and D) and cell viability (Fig. 7E and F) assays. However, no significant difference in EV71 attachment between the wild-type and knockdown cells was detected when either viral RNA (Fig. 7G and H) or VP1 protein (Fig. 7I and J) was examined. These results strongly suggest that SCARB2 does not play a significant role in the initial attachment of EV71; rather, it mainly mediates EV71 entry at the postattachment steps, such as uncoating (14).

FIG 7.

Knockdown of SCARB2 decreases EV71 infectivity but does not affect attachment. (A) SCARB2 expression in wild-type (WT) and SCARB2-knockdown RD cells (RD-SCARB2-KD) determined by qRT-PCR. (B) SCARB2 expression in wild-type and SCARB2-knockdown Vero cells (Vero-SCARB2-KD) determined by Western blotting assay. (C and D) EV71 growth curves for wild-type or SCARB2-KD cells. EV71 was added to wild-type or SCARB2-KD RD (C) or Vero (D) cells for attachment at 4°C for 1 h. The unbound viruses were washed away. Samples were collected at different times postinfection, as indicated, and the relative levels of virus RNA to β-actin RNA were analyzed by qRT-PCR assay. (E and F) Susceptibility of wild-type or SCARB2-KD cells to EV71 infection. Wild-type or knockdown RD (E) or Vero (F) cells were infected by EV71 at different multiplicities of infection (MOIs), as indicated. After 48 h, the viabilities of infected cells relative to those of mock-infected cells were analyzed as described in Materials and Methods. (G and J) EV71 attachment to wild-type or SCARB2-KD cells. EV71 G082 was allowed to attach to wild-type or knockdown RD (G, I) or Vero (H, J) cells at 4°C for 1 h. The attached viruses were analyzed by qRT-PCR (G, H) or Western blot (I, J) assays as described in Materials and Methods. Statistical significance was determined as described in Materials and Methods. n.s., not significant (P ≥ 0.05); **, P < 0.01; ***, P < 0.001.

MAbs interfere with EV71 binding to SCARB2.

We then determined whether D5 and C4 could disrupt the EV71-SCARB2 interaction by an in vitro pulldown assay with recombinant SCARB2-Fc or control Fc-immobilized agarose beads. As shown in Fig. 8A and B, SCARB2-Fc but not the control Fc efficiently pulled down EV71 in the absence of MAbs; the amount of virus pulled down from the 3F5-treated sample was comparable to that pulled down from the nontreated sample (SCARB2-Fc); in contrast, treatment with D5 or C4 reduced the amount of SCARB2-bound virus in an antibody dose-dependent manner. This result demonstrates that MAb binding to EV71 affects the subsequent EV71 interaction with SCARB2.

FIG 8.

D5 or C4 treatment interferes with EV71 binding to SCARB2 and PSGL-1. Different amounts of D5 or C4 were mixed with EV71 G082 and then incubated at 37°C for 1 h. Subsequently, the antibody-virus mixtures were incubated with immobile SCARB2 (A and B) or PSGL-1 (C and D) on agarose beads. After washing to remove unbound virus, the agarose beads were precipitated by centrifugation, and the binding of virus was detected by Western blotting with a virus-specific antibody or an anti-Fc antibody.

MAbs interfere with EV71 binding to PSGL-1.

Since PSGL-1 is another identified EV71 receptor (20), we also performed pulldown assays to evaluate the effect of MAb treatment on the EV71 interaction with PSGL-1 in vitro. A dose-dependent inhibitory effect was observed with D5 and C4 but not with 3F5 (Fig. 8C and D), indicating that D5 and C4 are able to interfere with the interaction between EV71 and PSGL-1.

DISCUSSION

In the present study, we demonstrate that a group of protective MAbs targeting the VP1 GH loop of EV71 are able to block multiple steps of EV71 entry through interference with manifold interactions between EV71 and its receptors. Our results elucidate an unusual mechanism of MAb-mediated neutralization of EV71 infection.

Virus entry is a multistep process that begins with virus attachment to the cell surface and ends with the delivery of the viral genome to the cytoplasm (41). For enveloped viruses, such as dengue virus, West Nile virus, and influenza virus, neutralizing MAbs can inhibit initial viral attachment or inhibit the postattachment fusion steps, or both, during viral entry (33–37), which may be dependent on the epitope and occupancy of antibodies (42, 43). However, for nonenveloped viruses, neutralizing MAbs usually target one particular step of entry (41, 44). For example, MAbs against enteroviruses mostly inhibited the attachment step (41, 45, 46), whereas a few antibodies which target the internal epitopes within VP1 or VP4 were suggested to exhibit neutralization only at the postattachment stage by inhibiting conformation transitions (47–49). Interestingly, in the present study we found that a group of anti-EV71 MAbs (represented by D5) exhibited neutralization at both the pre- and postattachment stages in a cell type-independent manner. First, D5 inhibited EV71 attachment to susceptible cells, including RD, Vero, and Jurkat T cells (Fig. 3). Second, treatment of the preattached EV71 with D5 resulted in the arrest of the virus at the cell periphery even after 300 min of exposure at 37°C, whereas the control MAb-treated EV71 was readily detected in the cytoplasm at 30 min postinfection (Fig. 5A). To further understand the process of EV71 internalization, we performed double staining of EV71 (in the absence of neutralizing MAbs) with the early and late endosome markers EEA1 and LAMP1, respectively. Colocalization of EV71 with EEA1 was observed as early as 5 min postinfection, increased afterwards, and was sustained until 60 min postinfection (the last time point tested), whereas colocalization of EV71 with LAMP1 was clearly seen from 30 to 60 min postinfection (data not shown), indicating that internalization and subsequent transport of EV71 virions to the early and late endosomes occur mainly within 15 to 30 min after infection, in agreement with the similar findings reported previously (50). By comparing the localization pattern and timing of EV71 in the presence of D5 to those in the absence of the antibody, it was concluded that the internalization of EV71 is prohibited by the D5 antibody treatment. As a consequence, the uncoating of EV71 virions and the subsequent release of viral RNA were also inhibited (Fig. 5B to D).

The IC₅₀s of D5 at the pre- and postattachment stages were comparable (0.324 μg/ml versus 0.539 μg/ml), suggesting that stoichiometry unlikely dictates at what stage and how D5 exerts neutralization. Rather, the observed dual inhibitory function of D5 may involve other mechanisms. We therefore attempted to explore the molecular basis of the inhibition of EV71 entry by this group of MAbs. Surprisingly, we found that these MAbs could interfere with multiple interactions between EV71 and its receptors. First, D5 and C4 blocked EV71 binding to heparan sulfate, an identified attachment receptor for EV71 (13), providing a logical explanation for the observed attachment inhibition by D5 or C4. Second, the interaction between EV71 and SCARB2 could be blocked by D5 or C4 in a dose-dependent manner (Fig. 8). Because SCARB2 mediates the internalization and uncoating of EV71 (14), the observed postattachment neutralization by D5 or C4 in the SCARB2-expressing cells is most likely attributed to the blockade of the EV71-SCARB2 interaction. Third, D5 or C4 treatment also impaired EV71 binding to another identified receptor, PSGL-1 (Fig. 8), thus explaining the postattachment neutralization by the MAbs in the PSGL-1-bearing Jurkat T cells. Clearly, D5 and C4, as a single MAb, can inhibit multiple virus-receptor interactions and therefore block each of the receptor-mediated entry steps.

The unusual ability of our MAbs to interfere with multiple EV71-receptor interactions is likely attributed to their epitope location. Epitope mapping revealed that the surface-exposed GH loop (core residues 211 to 220) of VP1 is the binding epitope for D5, H7, and C4 (Fig. 2). It was recently suggested that the canyon region of EV71 mediates its interaction with SCARB2 (18). Specifically, five synthetic peptides whose sequences were located around the canyon, according to the three-dimensional structure of EV71, including VP1-2 (residues 92 to 106), VP1-6 (residues 212 to 225), VP1-8 (residues 280 to 293), VP2-2 (residues 136 to 150), and VP3-4 (residues 175 to 188), were found to bind SCARB2 in pulldown assays (18) (Fig. 2E). Among them, VP1-6 encompasses the binding epitope (residues 211 to 220) of our MAbs; therefore, binding of our MAbs to EV71 may simply mask the VP1-6 site essential for SCARB2 binding. In addition, the MAb epitope resides sterically between the VP2-2 and VP3-4 sites (marked in yellow in Fig. 2E), suggesting that the MAbs may block SCARB2 binding to VP2-2 and VP3-4 through steric hindrance. For picornaviruses, it is generally recognized that heparan sulfate mediates virus attachment via electrostatic interactions between its negatively charged moieties and the clustered, positively charged residues on the virus surface (51–53). The exact binding site of heparan sulfate on EV71 is unclear, although it was proposed that clustering of positively charged residues (A166, K242, and K244 of VP1) at the 5-fold axis of EV71 capsids is responsible for heparan sulfate binding (13). Interestingly, the VP1 GH loop is also rich in positively charged residues (H214, K215, and K218) and, thus, could be part, if not all, of the binding site for heparan sulfate. We therefore speculate that our MAbs may interfere with the heparan sulfate-EV71 interaction through the preoccupancy of the heparan sulfate binding site or the creation of steric hindrance. In this study, we also found that D5 or C4 pretreatment blocked the interaction between EV71 and PSGL-1 in a dose-dependent manner. The epitope of our MAbs is sterically close to residue K149 of VP2 (shown as a blue dot in Fig. 2E), which has been implicated to be involved in the EV71 interaction with PSGL-1 (54). It is thus possible that steric hindrance is the major mode by which our MAbs block the EV71–PSGL-1 interaction. Nonetheless, further structural investigations of the receptor binding sites are required to unravel the precise mechanisms by which our MAbs interfere with the EV71-receptor interactions.

In conclusion, our work presented here reports that a group of anti-EV71 MAbs targeting the VP1 GH loop can inhibit the attachment and internalization steps of EV71 entry in vitro through interference with multiple interactions between the virus and its receptors. Our findings shed new light on the mechanism of MAb-mediated neutralization of EV71 infection and should facilitate the development of MAb-based anti-EV71 drugs.