Identification of a Critical and Conformational Neutralizing Epitope in Human Adenovirus Type 4 Hexon

ABSTRACT Human adenovirus type 4 (HAdV-4) is an epidemic virus that contributes to serious acute respiratory disease (ARD) in both pediatric and adult patients. However, no licensed drug or vaccine is currently available to the civilian population. The identification of neutralizing epitopes of HAdV-4 should allow the development of a novel antiviral vaccine and a novel gene transfer vector, and an effective neutralizing monoclonal antibody (MAb) will be useful in developing appropriate antiviral drugs. In this study, we report that MAb MN4b shows strong neutralizing activity against HAdV-4. MN4b recognizes a conformational epitope (418AGSEK422) within hypervariable region 7 (HVR7). Mutations within this site permitted HAdV-4 mutants to escape neutralization by MN4b and to resist neutralization by animal and human anti-HAdV-4 sera. A recombinant virus, rAd3-A4R7-1, containing the identified neutralizing epitope in the HVR7 region of HAdV-3 hexon, successfully induced antiserum that inhibited HAdV-4 infection. These results indicate that a small surface loop of HAdV-4 hexon is a critical neutralization epitope for this virus. The generation of MN4b and the identification of this neutralizing epitope may be useful in developing therapeutic treatment, a subunit vaccine, and a novel vector that can escape preexisting neutralization for HAdV-4. IMPORTANCE Neutralizing antibodies are considered good tools for the prevention of human adenovirus type 4 (HAdV-4) infections. The identification of the epitopes recognized by such neutralizing antibodies is important for the generation of recombinant antiviral vaccines. However, until now, no neutralizing epitope has been reported for HAdV-4. Here, we developed a serotype-specific neutralizing MAb directed against HAdV-4, MN4b. We provide evidence that MN4b recognizes a conformational epitope within HVR7 of HAdV-4 hexon. Antisera generated to this conformational epitope displayed on HAdV-3 hexon inhibited infection of AD293 cells by HAdV-4. Our findings are very important for the development of therapeutic treatment, a subunit vaccine, and a novel vector for HAdV-4.

Approved live oral vaccines based on HAdV-4 and HAdV-7 derived from human diploid cells have been used in U.S. military recruits to reduce febrile respiratory illness (FRI) and have significantly reduced the risk of FRI in the U.S. military (14,16). However, no vaccine is currently approved for general use in children and adults, and no efficient antiviral therapy for adenoviruses is available.
Neutralizing monoclonal antibodies (MAbs) are promising prophylactic or therapeutic drugs against viral disease. The generation of neutralizing MAbs is also useful in identifying neutralizing epitopes, a highly important step in the molecular design of vaccines. HAdV-4 has also been proposed as an alternative vector for human immunodeficiency virus (HIV)/simian immunodeficiency virus (SIV) and candidate influenza vaccines (17). To escape the neutralization reaction in vivo, the neutralizing epitope is modified in a viral gene delivery vector. Well-characterized neutralization epitopes are useful for the construction of such vectors (18)(19)(20). The adenoviral capsid is composed of three major structural proteins: hexon, penton base, and fiber. Adenovirusneutralizing antibodies can be raised against any of these major capsid proteins (21,22), although the hexon protein is the predominant target of serotype-specific neutralizing antibodies (NAbs) (22)(23)(24)(25)(26)(27). The serotype-specific neutralization epitopes on hexon are located mainly on the tower region, which consists of seven hypervariable regions (HVRs), among which HVR7 can be further subdivided into three additional highly variable regions (28)(29)(30)(31)(32). The specific locations of the neutralizing epitopes of HAdV-4 have not been identified.
In this study, we report a neutralizing MAb with strong neutralizing activity against HAdV-4, MN4b, which recognizes a conformational epitope within HVR7. Neutralization assays confirmed that mutants within this site escaped neutralization by MN4b, by antisera from animals immunized with HAdV-4, and by sera from humans infected with HAdV-4. These results indicate that this epitope is a critical neutralization site in HAdV-4 hexon.

RESULTS
Identification of neutralizing MAb MN4b directed against HAdV-4 hexon. Of the six anti-HAdV-4 MAbs generated, only the monoclonal IgG1 isotype antibody MN4b had high neutralizing activity against HAdV-4 in AD293 cells. The ascites titer of MN4b was determined by an enzyme-linked immunosorbent assay (ELISA) against HAdV-4 virions and was about 800,000. MN4b had a high neutralization titer of up to 3,200 (about 0.6 g/ml) against HAdV-4. MN4b did not neutralize HAdV-3, HAdV-7, or HAdV-5, indicating that it is a serotype-specific NAb against HAdV-4 (data not shown). An indirect ELISA indicated that MN4b reacted with its parental antigen, whole HAdV-4 GZ01 virus particles, suggesting that the epitope recognized by MN4b is presented on the surface of the virions (Fig. 1A). MN4b did not react with purified HAdV-3 virions. MN4b also did not react with the recombinant HAdV-4 hexon (Ad4hexon) peptide (amino acids 112 to 491 of hexon) or the fiber knob peptide (Ad4FK) expressed in Escherichia coli (Fig. 1A).
Native and denatured Western blot analyses were used to determine whether MN4b recognizes a conformation-dependent antigen. Western blotting was performed with purified HAdV-4 particles that had been separated electrophoretically after exposure to 1% SDS at room temperature (native) (Fig. 1B, lanes 3 and 7) or at 98°C (denatured) (lanes 4 and 8). At room temperature, the hexon protein maintains its trimeric form in SDS. MN4b recognized only the native trimeric HAdV-4 antigen (Fig. 1B, lane 3) and not the denatured monomeric HAdV-4 antigen (lane 4) that had been heated to 98°C in the presence of SDS. Anti-HAdV-4 serum reacted strongly with the native HAdV-4 antigen The reaction of MN4b with wild-type HAdV-4 (wAd4), rAd3, a recombinant hexon fragment (Ad4hexon), or fiber knob (Ad4FK) of HAdV-4 was detected by ELISAs. Antiserum from mice immunized with HAdV-4 (anti-Ad4) was used as the positive control, and antiserum from mice immunized with PBS was used as the negative control. Each experiment was repeated independently at least three times, and the means Ϯ standard deviations are shown. OD450nm, optical density at 450 nm. (B) Immunoblot analysis indicates that MN4b recognizes HAdV-4 hexon in its trimeric form but not the hexon monomer. Purified HAdV-4 virions were stored at room temperature (native [N]) (lanes 3 and 7) or heated at 98°C (denatured [D]) (lanes 4 and 8) for 5 min in the presence of loading buffer. Purified HAdV-3 virions at room temperature (native) (lanes 1, 5, and 9) or boiled at 98°C (denatured) (lanes 2, 6, and 10) were used as the controls. The samples were then separated by SDS-PAGE and transferred onto polyvinylidene difluoride membranes. The membranes were incubated with MN4b (lanes 1 to 4) or anti-Ad4 antiserum (lanes 5 to 8). Purified HAdV-3 virions that disintegrated at room temperature (lane 9) or at 98°C (lane 10) were incubated with mouse anti-HAdV-3 serum as controls.   (Fig. 1B, lanes 1, 5, and 9) or denatured (lanes 2, 6, and 10) HAdV-3 virions were used as the controls. Interestingly, anti-HAdV-3 serum reacted with the native HAdV-3 antigen (Fig. 1B, lane 9) and also with the denatured HAdV-4 antigen (lane 10). The molecular weight of the antigen recognized by MN4b (Fig. 1B, lane 3) was similar to that of the major capsid protein, the hexon homotrimer, which was recognized by anti-HAdV-4 serum and anti-HAdV-3 serum (Fig. 1B, lanes 5 and 9, respectively). Further Western blot analyses with antiserum from a mouse immunized with the recombinant HAdV-4 hexon expressed in E. coli confirmed that MN4b detected the hexon homotrimer (Fig. 1C). These results suggest that MN4b recognizes a conformationdependent epitope on the hexon homotrimer.
Four epitope-incorporated recombinant adenoviruses were constructed by incorporating the predicted HAdV-4 epitope peptides into the corresponding surface loops of HAdV-3 hexon (Fig. 3A). These epitopes were displayed on hexon trimers assembled in virus particles. An ELISA showed that MN4b recognized rAd3-A4R7-1 but did not react with the other recombinant viruses (Fig. 3B). Further tests with serially diluted MN4b confirmed these results (Fig. 3C), which suggest that MN4b detects a conformation-dependent epitope in HVR7.
Confirmation of the epitope recognized by MN4b using HAdV-4 HVR7 mutations. HAdV-4 HVR7 can be divided into two independent regions, R7-1 and R7-2. To map the neutralizing epitope recognized by MN4b with an independent approach, three recombinant type 4 adenoviruses were generated, in which mutations were introduced into the long loop of HVR7 (Fig. 5A). These mutant viruses replicated to similar titers and showed similar particle/PFU ratios as those of the wild-type virus (data not shown), suggesting that mutations in R7 did not significantly alter the virion structure. These mutant viruses were then screened with MN4b by using an ELISA, Western blotting, and a neutralization test. A mutation in the R7-2 region (mutant rAd4-3R2) did not affect the recognition and neutralization of the recombinant virus by MN4b (Fig. 5). However, mutants rAd4-3R7-1 and rAd4-3R7, with altered R7-1 residues (AGSEK to DANG), escaped neutralization by MN4b (Fig. 5D). Data from ELISAs and Western blotting also indicated that rAd4-3R7-1 and rAd4-3R7 completely escaped recognition by MN4b ( Fig. 5B and C). MAb 3D7, a neutralizing antibody against HAdV-3 prepared previously (33), was used as the control. These results confirm that R7-1 residues (AGSEK) form the critical epitope recognized by MN4b.
R7 mutations reduce virus susceptibility to neutralization by HAdV-4 antisera. The susceptibility of R7 mutants to neutralization by anti-HAdV-4 serum was measured with neutralization tests. Pooled anti-HAdV-4 serum from mice showed 8-fold less neutralization activity against the rAd4-3R7-1 mutant and 16-fold less activity against the rAd4-3R7 mutant than against the wild-type virus. However, pooled anti-HAdV-4 serum from mice showed activity against rAd4-3R7-2 that was similar to that against wild-type HAdV-4. Similar results were observed with rabbit anti-HAdV-4 serum, which was about 16-to 32-fold less efficient in neutralizing the rAd4-3R7 and rAd4-3R7-1 mutants and about 2-fold less efficient in neutralizing rAd4-3R7-2 than in neutralizing the wild-type virus (Fig. 6). These results indicate that R7-1 contains a critical neutralization epitope, which accounts for most of the neutralizing antibody response to HAdV-4 in both mouse and rabbit.
The susceptibility of the R7 mutants to neutralization by sera from humans infected with HAdV-4 was also tested. All six anti-HAdV-4-positive serum samples showed 2-to 4-fold less neutralization activity against rAd4-3R7-1 and rAd4-3R7 than against wildtype HAdV-4, whereas rAd4-3R7-2 showed similar susceptibility to all four anti-HAdV-4-positive serum samples as that of wild-type HAdV-4 (Fig. 7). These results demonstrate that R7-1 is a critical target of the neutralizing antibodies directed against HAdV-4 in human sera.

DISCUSSION
In this study, we report a neutralizing MAb with high neutralizing activity against HAdV-4, which recognizes a conformational epitope (amino acid positions 418 to 422 in HAdV-4 hexon) within HVR7. This epitope is a critical target for NAbs directed against HAdV-4.
Viruses with mutations within the R7-1 epitope were less susceptible than wild-type HAdV-4 to neutralization by anti-HAdV-4 serum from mice, rabbits, and humans ( Fig. 6   FIG 4 Neutralization assay of HAdV-4 with polyclonal antisera from mice immunized with epitopechimeric adenoviruses. The antibody-virus mixtures were transferred to 96-well plates containing 85% to 95% confluent monolayers of AD293 cells. After culture for 96 h, the HAdV-4 genome copy numbers were determined by Q-PCR. Mouse anti-rAd3 and anti-Ad4 sera were used as the controls.  4). (D) Neutralizing titers of MN4b that inhibited infection of AD293 cells by HAdV-4, rAd4-3R7-1, rAd4-3R-2, or rAd3E were determined by neutralization assays. MAb 3D7, a neutralizing antibody directed against HAdV-3, was used as a control. Antisera that showed no reactivity at a 1:16 dilution (the lowest dilution tested) were assigned a titer of 1:4. and 7). These results suggest that R7-1 is the critical site in HAdV-4 for the induction of NAbs. We also found that mouse antiserum had a lower neutralizing titer against rAd4-3R7 than against rAd4-3R7-1 (Fig. 6), which indicates that other parts of R7 also affect neutralization. The great resistance of mutants with a few altered amino acids to neutralization highlights the need for continuous surveillance of HAdV-4 field strains. Our data suggest that hexon is the predominant HAdV-4 antigen that induces NAbs, which is consistent with data from previous reports (25)(26)(27). However, HAdV-4 antisera still neutralized the mutated viruses. Therefore, further studies should investigate whether other epitopes within hexon or other capsid proteins, such as fiber or penton base, are also targets of NAbs in human sera.
MN4b recognizes HAdV-4 and inhibits infection by HAdV-4 but not HAdV-3 (Fig. 1). Therefore, MN4b may be used to distinguish HAdV-4 from other HAdV serotypes. An alignment of the hexon protein sequences available in GenBank revealed that the epitope recognized by MN4b is unique to HAdV-4. However, the distribution of this epitope among HAdV-4 strains, which might affect the detection and neutralization abilities of MN4b, must be determined. A phylogenetic analysis of the genomes and hexon sequences of global HAdV-4 strains available in GenBank showed that HAdV-4 can be classified into two subtypes (not shown here) and that some highly variable sites in the hexon protein vary between subtypes (Fig. 8). These sites were conserved within each of the subtypes. The epitope in R7-1 of different HAdV-4 strains (418AGSEK422) is highly conserved, and the sequences corresponding to this epitope are identical in both subtypes.
MN4b, which had a high neutralization titer of 3,200 (about 0.6 g/ml), could potentially be humanized as a therapeutic or prophylactic agent against HAdV-4 infection (38). The identification of this important epitope may be useful for developing capsid-modified HAdV-4 vectors that would be permitted to escape preexisting antivector immunity (17,20,39).
The mechanism by which MN4b neutralizes HAdV-4 is unclear. It is also puzzling that HVR7-1 of HAdV-4 is a predominant epitope. Neutralization is defined as the reduction in viral infectivity caused by the binding of antibodies to the surfaces of viral particles (VPs), thereby blocking a step in the viral replication cycle that precedes virally encoded transcription or protein synthesis (40). HAdV infection is initiated by fiber binding, and the virus is then internalized via clathrin-coated vesicles after interactions occur between cellular integrins and penton base. An antibody may act at many steps of the adenovirus infection process (41). The interaction of MN4b with HAdV-4 hexon may block a step after virus entry, such as capsid disassembly, endosome penetration, translocation, or genome entry into the nucleus. Until now, the role of NAbs in inhibiting late steps in the entry process has been poorly characterized, although several studies have shown that an antihexon antibody inhibits adenoviral infection by blocking the microtubule-dependent translocation of the virus or by a TRIM21-dependent mechanism (42)(43)(44). The mechanism by which antihexon antibodies neutralize adenoviruses requires further investigation.
Defining the potential neutralizing epitope region of HAdV-4 hexon. The primary amino acid sequence of HAdV-4 hexon was submitted to the SWISS-MODEL workspace for homology modeling. The SWISS-MODEL template library (SMTL version 2017-07-26, PDB release 2017-07-21) was searched with BLAST (46) and HHBlits (47) for evolutionarily related structures that matched the target sequence (48)(49)(50). The highest-quality templates were selected for model building. The models were built based on the target-template alignment using ProMod3. Coordinates that were conserved between the target and the template were copied from the template to the model. Insertions and deletions were remodeled by using a fragment library. Side chains were then rebuilt. Finally, the geometry of the resulting model was regularized with a force field. When loop modeling with ProMod3 failed, an alternative model was built with PROMOD-II (51). PyMOL v0.99 was used to generate the space-filling representation of the HAdV-4 hexon structure. The antigenic epitopes with lengths of 6 to 15 amino acids that were predicted to be exposed on the capsid surface and to be located in HVRs were selected as potential sites for recognition by NAbs.

FIG 8
Multiple-sequence alignment of hexon proteins from HAdV-4 strains. The AdC68 sequence was included as a reference. Six hypervariable regions (HVR1 and HVR3 to -7) were identified. "*" indicates conserved amino acids, "." indicates either conserved size or conserved hydropathy, and ":" indicates that both size and hydropathy are conserved. Gaps used to optimize the alignments are indicated by dashes. Variable sites between HAdV-4 subtypes are indicated by a gray frame.
to transfect AD293 cells with Lipofectamine LTX with Plus reagent (Invitrogen, USA), according to the manufacturer's instructions. The transfected cells were cultured at 37°C under 5% CO 2 for 6 to 10 days and were examined daily for evidence of a cytopathic effect (CPE). The cells were frozen and thawed for three cycles. The cell suspensions were collected and then used to infect the cells. At 96 h postinfection, the viruses were harvested and designated rAd3-A4R1, rAd3-A4R2, rAd3-A4R5, and rAd3-A4R7-1. Finally, the mutant viruses were cultured with AD293 cells in 20 dishes (100 mm), harvested, and purified with standard CsCl gradient centrifugation, as described above. The modified hexon genes of the viruses were confirmed by DNA sequencing.
Recombinant peptides and polyclonal antisera. A recombinant HAdV-4 hexon peptide (amino acids 112 to 491 of hexon) expressed in E. coli and mouse antiserum directed against the HAdV-4 hexon peptide were prepared as described previously (52). Mouse antiserum against HAdV-3 was prepared as described previously (33). A recombinant HAdV-4 fiber knob peptide (amino acids 234 to 425 of fiber) was expressed in E. coli and purified with affinity chromatography using Ni-nitrilotriacetic acid (NTA) resin (Novagen, USA) under native conditions. Rabbit antiserum was obtained from animals that had been injected intramuscularly with 5 ϫ 10 11 VPs of HAdV-4 and then boosted subcutaneously with the same dose after 3 weeks. Groups of five female BALB/c mice aged 4 to 6 weeks were immunized intramuscularly with 1 ϫ 10 10 VPs of HAdV-4, rAd3-A4R1, rAd3-A4R2, rAd3-A4R5, or rAd3-A4R7-1. Two booster doses were given at 2-week intervals with the same dose of antigen. Blood was collected from the anesthetized mice via the retro-orbital lobe 10 days after the final immunization, and the sera were isolated, heat inactivated, and stored frozen for serology tests.
The animal procedures used in this work were evaluated and approved by the Ethics Committee of the First Affiliated Hospital of Guangzhou Medical University (Guangzhou, China). They complied with all relevant guidelines and the National Law for Laboratory Animal Experimentation of China. The animal experiments were conducted in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health of United States (53). All animals were housed individually and received humane care. During injection and sample collection, the mice were anesthetized with 1.5% isoflurane or 1 ml/kg body weight of 3% pentobarbital sodium to minimize their suffering.
Serum samples from healthy human donors from Guangzhou (China) were collected after approval by the Ethics Committee of the First Affiliated Hospital of Guangzhou Medical University, and informed consent was obtained from each volunteer.
Generation of NAbs. Purified HAdV-4 virions were used to immunize BALB/c mice and to screen the resulting MAbs. The production and screening of mouse MAbs directed against HAdV-4 were performed as described previously (34). Briefly, BALB/c mice (6 to 8 weeks old) were injected intramuscularly with 1 ϫ 10 10 VPs of CsCl-purified HAdV-4 and then boosted twice intramuscularly with 5 ϫ 10 9 VPs of CsCl-purified HAdV-4 at 2-week intervals. Three days after intravenous boosting with 1 ϫ 10 10 VPs of HAdV-4 per mouse in phosphate-buffered saline (PBS), the mice were killed, and hybridoma fusion was performed by using a standard protocol (54). Hybridomas secreting HAdV-4-specific antibodies were screened with an ELISA with HAdV-4 virions. Positive hybridomas were then screened with a virus neutralization assay and subcloned with limiting dilution. Ascites were generated by injecting hybridoma cells into mice primed with Freund's incomplete adjuvant. The ascites titers against HAdV-4 were then determined with an ELISA, and the ascites were purified by octanoic acid-ammonium sulfate precipitation. The IgG concentrations were determined spectrophotometrically with a factor of 1.4 mg/ml per absorbance unit at 260 nm. The antibody isotypes were determined with a mouse hybridoma subtyping kit (Roche, Germany) according to the manufacturer's instructions.
Indirect ELISA. For ELISAs, 96-well Nunc MaxiSorp flat-bottom plates (Thermo, China) were coated with recombinant peptides (about 2 g/ml) or virus particles (about 10 10 VPs/ml) in PBS (pH 7.4) overnight at 4°C. They were then washed once with 0.05% Tween 20 in phosphate-buffered saline (PBST) and blocked for 2 h with 2% bovine serum albumin (BSA) in PBST. MAb ascites (100 l/well) or antiserum in a series of dilutions from 1:100 to 1:1,000,000 was then added to each well and incubated for 1 h at 37°C. The plates were washed three times with PBST and incubated for 1 h with a 1:10,000 dilution of horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG(HϩL) affinity-purified secondary antibody or HRP-conjugated goat anti-rabbit IgG(HϩL) secondary antibody. After the plates were washed four times with PBST, the products were visualized with the tetramethylbenzidine (TMB) substrate. The reaction was stopped with 1 M H 2 SO 4 , and the results were analyzed with an ELISA plate reader (Multiskan MK3; Thermo Scientific) at 450 nm.
Virus neutralization test. For in vitro adenovirus neutralization experiments, MAb ascites or antiserum pretreated at 56°C for 30 min was serially diluted 2-fold with DMEM (Gibco, China), and 50-l aliquots of each dilution were mixed with an equal volume (100 50% tissue culture infective doses [TCID 50 ]) of wild-type or recombinant adenovirus. The antibody-virus mixtures were incubated for 1 h at FIG 9 Introduction of a mutated site into the hexon gene in the HAdV-4 genome. Fragment H4A3R7 (here designated hexon-PCR), with a mutation in HVR7 of the HAdV-4 hexon region, was generated by overlapping PCR extension mutagenesis. The mutated fragment was then cloned into the MluI-digested pBRAd4 vector to generate the mutagenesis vector pBRAd4-A3R7 with homologous recombination in vitro by using Exnase II.
37°C and transferred to 96-well plates containing 85% to 95% confluent monolayers of AD293 cells. After culture for 96 h, the monolayers were observed microscopically, and the neutralization titers were determined as the reciprocal of the highest dilution of mouse ascites or antiserum that protected the monolayer from a visually observable CPE. To confirm the results of the microneutralization assays, neutralization assays were performed under the same conditions, and the copies of the adenovirus genome were quantified with an adenovirus quantitative PCR (Q-PCR) kit (HYSMed, China) to measure the inhibition of viral infection by different sera (30).
Immunoblot analysis. AD293 cells were infected with wild-type adenovirus or the recombinant adenoviruses. At 72 h or 96 h postinfection, cells were harvested and freeze-thawed three times. The viral suspensions or purified virions were then mixed with 5ϫ loading buffer (10% sodium dodecyl sulfate [SDS], 5% 2-mercaptoethanol, 0.5% bromophenol blue, and 50% glycerol in 250 mM Tris-HCl [pH 6.8]), kept at room temperature for 5 min, and then incubated on ice (native) or heated for 5 min at 98°C (denatured). The samples were then separated on 10% SDS-polyacrylamide gels and transferred electrophoretically onto polyvinylidene difluoride (PVDF) membranes. The membranes were blocked with 5% skim milk in PBS and then incubated with MAb ascites or mouse antiserum at final dilutions of 1:500 or 1:1,000. The membranes were washed again and exposed to a 1:10,000 dilution of HRP-conjugated goat anti-mouse IgG secondary antibody or HRP-conjugated goat anti-rabbit IgG secondary antibody. After washing, the blots were developed with the 1-Step Ultra TMB Blotting Solution substrate (Thermo Scientific, USA) for 5 min at room temperature.