Serotype-Specific Neutralizing Antibody Epitopes of Human Adenovirus Type 3 (HAdV-3) and HAdV-7 Reside in Multiple Hexon Hypervariable Regions

Adenovirus Type 3 (HAdV-3) and HAdV-7 Reside in Multiple Hexon

Hypervariable Regions

Hongling Qiu,a,bXiao Li,aXingui Tian,aZhichao Zhou,aKe Xing,aHaitao Li,aNi Tang,aWenkuan Liu,aPeisheng Bai,aand Rong Zhoua

State Key Laboratory of Respiratory Diseases (Guangzhou Medical University), The First Affiliated Hospital of Guangzhou Medical College, Guangzhou, China,a and Institute for Nutritional Science, Shanghai Institute for Biological Science, Chinese Academy of Science, Shanghai, Chinab

Human adenovirus types 3 and 7 (HAdV-3 and HAdV-7) occur epidemically and contribute greatly to respiratory diseases, but

there is no currently available licensed recombinant HAdV-3/HAdV-7 bivalent vaccine. Identification of serotype-specific

neu-tralizing antibody (NAb) epitopes for HAdV-3 and HAdV-7 will be beneficial for development of recombinant HAdV-3/HAdV-7

bivalent vaccines. In this study, four NAb epitopes within hexon hypervariable regions (HVRs) were predicted for HAdV-3 and

HAdV-7, respectively, by using bioinformatics. Eight hexon chimeric adenovirus vectors with the alternation of only one

pre-dicted neutralizing epitope were constructed. Further

in vitro

and

in vivo

neutralization assays indicated that E2 (residing in

HVR2) and E3 (residing in HVR5) are NAb epitopes for HAdV-7, and E3 plays a more important role in generating NAb

re-sponses. Cross-neutralization assays indicated that all four predicted epitopes, R1 to R4, are NAb epitopes for HAdV-3, and R1

(residing in HVR1) plays the most important role in generating NAb responses. Humoral immune responses elicited by the

re-combinant rAdH7R1 (containing the R1 epitope) were significantly and durably suppressed by HAdV-3-specific NAbs.

Surpris-ingly, the rAd

⌬

E3GFP-specific neutralizing epitope responses induced by rAdMHE3 (R3 replaced by E3) and rAdMHE4 (R4

re-placed by E4) were weaker than those of rAdMHE1 (R1 rere-placed by E1) or rAdMHE2 (R2 relaced by E2)

in vitro

and

in vivo

.

Furthermore, rAdMHE4 replicated more slowly in HEp-2 cells, and the final yield was about 10-fold lower than that of

rAd

⌬

E3GFP. The current findings contribute not only to the development of new adenovirus vaccine candidates, but also to the

construction of new gene delivery vectors.

T

o date, 51 human adenovirus (HAdV) serotypes have been

identified and divided into six groups (A to F) based on their

morphological, immunogenic, and genomic information (

7

). A

novel human adenovirus, HAdV-52, was identified recently, but

whether it should be considered a novel “serotype” remains

dis-puted (

6

,

12

). Most HAdV can cause respiratory disease,

gastro-enteritis, pneumonia, and keratoconjunctivitis, among other

dis-eases (

3

,

10

,

13

). Of these, the acute respiratory diseases induced

by the subspecies B1 adenoviruses HAdV-3 and HAdV-7 have

occurred epidemically and have caused outbreaks in Asia, Europe,

and North America (

2

,

11

,

15

,

21

,

26

,

31

). However, there is

cur-rently no efficient antiviral therapy against adenoviruses and,

therefore, it is necessary to develop licensed HAdV-3 and HAdV-7

vaccines to help prevent adenovirus infection outbreaks. A simple

combined formulation of HAdV-3 and HAdV-7 monovalent

vac-cines combined to produce a bivalent vaccine might introduce

considerable constraints, such as serotype competition and/or

dominance. Therefore, a recombinant vaccine containing

HAdV-3 and HAdV-7 neutralizing epitopes that elicits protective

immunity simultaneously to HAdV-3 and HAdV-7 is required.

However, no neutralizing epitopes have been identified for

HAdV-3 and HAdV-7 until now.

All adenovirus particles contain a common structure

consist-ing of the nonenveloped capsid, which is 60 to 90 nm in diameter

with icosahedral symmetry. Twelve copies of a trimeric hexon

form each of the 20 triangular facets of the icosahedron, and

tri-meric fibers are inserted into the 12 pentatri-meric penton bases at

the 12 vertices. All three major structural proteins, the hexon,

fiber, and penton, can generate neutralizing antibodies (NAbs)

against adenovirus, but the major protein component of the

ico-sahedral capsid, the hexon, is the major target of serotype-specific

NAbs (

17

,

19

,

24

). Alignment of available HAdV hexon protein

sequences reveals high homology except within the seven

hyper-variable regions (HVRs). The highly conserved domains of the

hexon protein form the core tertiary structure, a pseudohexagonal

shape that allows formation of tightly packed hexon arrays within

the facets of the icosahedral capsid, and the seven HVRs are

local-ized on the exterior surface of the virion (

20

). It is therefore

pos-sible to change amino acids in the HVRs without destroying the

core tertiary structure of the hexon protein.

A hexon chimeric adenovirus serotype 5 vector was reported to

circumvent preexisting antivector immunity by replacing the

seven short HVRs on the surface of the HAdV-5 hexon protein

with the corresponding HVRs from HAdV-48 without perturbing

its core structure (

19

). These data indicated that the seven HVRs

contained epitopes recognized by serotype-specific NAbs.

There-fore, if serotype-specific NAb epitopes for different adenoviruses

can be identified, it would be possible to generate polyvalent

vac-cines by coexpressing different serotype-specific NAb epitopes in

the HVRs of hexon proteins. In this study, we sought to identify

the serotype-specific NAb epitopes for HAdV-3 and HAdV-7 by

Received15 December 2011Accepted8 May 2012

Published ahead of print23 May 2012

Address correspondence to Rong Zhou, [email protected]. Q.H., L.X., and T.X. contributed equally to this work.

Copyright © 2012, American Society for Microbiology. All Rights Reserved.

doi:10.1128/JVI.07076-11

on November 7, 2019 by guest

http://jvi.asm.org/

construction of a gene replacement strategy. Our study will

con-tribute to the development of new HAdV-3/HAdV-7

recombi-nant vaccines and gene delivery vectors.

MATERIALS AND METHODS

Defining the potential NAb epitopes of HAdV-3 and HAdV-7 hexons. The nonconserved domains of HAdV-3 and HAdV-7 hexon proteins were determined by aligning them with the available human hexon protein sequences using the protein BLAST program (Basic Local Alignment Tool; http://blast.ncbi.nlm.nih.gov/Blast.cgi). The EMBOSS antigenic tool (http://binfo.ym.edu.tw/emboss/Apps/antigenic.html) was used to predict antigenic epitopes of the HAdV-3 and HAdV-7 hexon proteins. The potential template structures for modeling HAdV-3 and HAdV-7 hexon proteins were identified by searching them against information in the Protein Data Bank (PDB) by BLAST-P analysis. The Modeler 9v8 tool was used to model the three-dimensional (3D) structure of HAdV-3 and HAdV-7 hexon proteins (8). The antigenic epitopes that were predicted to be exposed on the capsid surface and located in nonconserved regions were selected as potential sites for recognition by NAbs.

Construction of the mutagenesis vectors MHE1, pBRAd-MHE2, pBRAd-MHE3, pBRAd-MHE4, pBRAd3H7R1, pBRAd3H7R2, pBRAd3H7R3, and pBRAd3H7R4.The plasmid pBRAd⌬E3GFP encod-ing a HAdV-3 GZ-01 vector (GenBank accession no.DQ099432) and enhanced green fluorescent protein (GFP) with an E3 region deletion was constructed as previously described (32). The hexon chimeric adenovirus vector pAd3egf/H7, in which the HAdV-3 hexon gene in the pBRAd⌬E3GFP vector was replaced with the hexon gene from HAdV-7, was constructed as previously described (27). The mutated hexon genes were cloned into small pBRLR shuttle vectors that facilitated transfer of the mutated hexon genes into full-length viral plasmids pBRAd⌬E3GFP or pAd3egf/H7.

(i) Construction of the shuttle vector pBRLR.The L fragment (from nucleotide [nt] 16748 to nt 18435) was amplified from plasmid pBRAd⌬E3GFP using the primers HexonLU and HexonLD (Table 1) and subsequently digested with EcoRI and ClaI and inserted into the EcoRI/ ClaI-digested pBR322 vector to generate the pBRL vector. Fragment R (from nt 21119 to nt 24526) was amplified from plasmid pBRAd⌬E3GFP using the primers HexonRU and HexonRD (Table 1) before being

di-TABLE 1Primers used in this study

Primer Sequencea

HexonLU 5=-AAAGAATTCTATCCAGCCTGAGGTCAAAGTAAGA-3=

HexonLD 5=-GTCATCGATGGGGTGGCCATCTTGAAAG-3=

HexonRU 5=-AAAGGATCCCATGGATGAGCCCACCCTG-3=

HexonRD 5=-AAGGTCGACTGCATGTTCTGGTTTTCTGAAAG-3=

HexU 5=-CCCATCGATGATGCCCCAATGGGCATAC-3=

HexD 5=-CCGGGATCCACCTCAAAAGTCATGT-3=

MHE1D 5=-GTATGTGGTGGCATTGTCTTCTCCCGTTGTAACTATCCA-3=

MHE1U 5=-GGAGAAGACAATGCCACCACATACACATTTGGCATTGCT-3=

MHE2D 5=GTTGTCTGCGGTAATGTCTTTCCCAATTTCCAAACCTTCTTTAG T3=

MHE2U 5=-GAAATTGGGAAAGACATTACCGCAGACAACAAGCCCATTTATGCCGAT-3=

MHE3D 5=-CGAAAAAGCGTCAGCAGCTTC CCTACCATCGAAAAA-3=

MHE3U 5=-GAAGCTGCTGACGCTTTTTCG CCTGAAATTGTGCTT-3=

MHE4U 5=-CCTAGAGACACTGCATGGGAAAAAGATACTAAAGTTGATACAGCTAAT-3=

MHE4D 5=-TTTAGTATCTTTTTCCCATGCAGTGTCTCTAGGTTTAATGCCTTGATA-3=

IMHE1a 5=-GTATGTGGTGGCATTGTCTTCTCCCG-3=

IMHE2a 5=-GGCATAAATGGGCTTGTTGTCTGCGG-3=

IMHE3a 5=-CGAAAAAGCGTCAGCAGCTTCCCT-3=

IMHE4a 5=-TTTAGTATCTTTTTCCCATGCAGTGTCTCTAGGT-3=

A3HVR1u 5=-ACATCTCAGTGGATAGTTACAACGAATCGAGACAATGCAGTAACTACCACCACAAACAC-3=

A3HVR1r 5=-AGCAATGCCAAATGTGTTTGTGGTGGTAGTTACTGCATTGTC TCGATTCGTTGTAACTATCCACTGAGATGTATTGG-3=

A3HVR2u 5=-ATTGGGAAAGACATTACTACTACTGAAGGAGAAGAAAAGCCCATTTATGCCGATAAAAC-3=

A3HVR2r 5=-CTGATATGTTTTATCGGCATAAATGGGCTTTTCTTCTCCTTCAGTAGTAGTAATGTCTTTCCCAATTTCTAAACCTTC-3=

A3HVR5u 5=-TTTGATGGTAGAGATGCTGTTGCAGGAGCTTTAGCGCCTGAAATTGTGCTTTACACGGA-3=

A3HVR5r 5=-TTCCGTGTAAAGCACAATTTCAGGCGCTAAAGCTCCTGCAAC AGCATCTCTACCATCAAAAAATTCCATATCAATATC-3=

A3HVR7u 5=-AAGTTAAAACCGATGACGCTAATGGATGGGAAAAAGATGCTAATGTTGATACAGCTAATGAAATAGCCATAGGCAAC-3=

A3HVR7r 5=-AGCTGTATCAACATTAGCATCTTTTTCCCATCCATTAGCGTCATCGGTTTTAACTTTAATGCCTTGATATGTTTTCGC-3=

HR1r 5=-TACTGCATTGTCTCGATTCGT-3=

HR2r 5=-GCTTTTCTTCTCCTTCAGT-3=

HR5r 5=-CGCTAAAGCTCCTGCAACAGCATC-3=

HR7r 5=-CCATTAGCGTCATCGGTTTTAAC-3=

HU 5=-GGGTATTCCATATGGCCACCCCATCGATGATGCCCC-3=

HD 5=-CCGCTCGAGTTATTGAGGCTCTGGCTGATATG-3=

AD7SPR1 5=-GGCAGTTGTGCTGGGCATG-3=

AD7SPR2 5=-TTCTCTGGAATTAACATTGAAAGGATAG-3=

AD7SPS 5=-CCACTTAATCATAAATCAGGGCAAAA-3=

AD7SPBH 5=-FAM-TGGCTACTGGTGCCCTTACTAATGCTAAAGG-BHQ1-3=b

AD3GZ01F 5=-CGAGACCTCCTACCCATGAACTAA-3=

AD3GZ01R 5=-GACTGTATTGCTGATTTCAAGTAAGTGTCT-3=

AD3GZ01FB 5=-FAM-TCATTGCCCCTACCTTACCCAATCCAA-BHQ1 a_{Underlined nucleotides indicate restriction sites used for cloning.}

b

FAM, 6-carboxyfluorescein; BHQ1, Black Hole Quencher 1.

on November 7, 2019 by guest

http://jvi.asm.org/

gested with BamHI and SalI and insertion into the BamHI/SalI-digested pBRL vector to generate the pBRLR vector.

(ii) Construction of HAdV-3 hexon mutagenesis vectors pBRAd-MHE1, pBRAd-MHE2, pBRAd-MHE3, and pBRAd-MHE4.HAdV-3 hexon mutagenesis vectors MHE1, MHE2, pBRLR-MHE3, and pBRLR-MHE4 were created to facilitate the transfer of mutated hexon fragments from the mutagenesis vectors into the full-length plasmid pBRAd⌬E3GFP. The mutated HAdV-3 hexon gene fragment MHE1 (or MHE2, MHE3, or MHE4) was mutated by replac-ing the region encodreplac-ing the potential neutralizreplac-ing epitope R1 (or R2, R3, or R4) of the HAdV-3 hexon gene with the region encoding the potential neutralizing epitope E1 (or E2, E3, or E4) of the HAdV-7 hexon gene. These mutated fragments were produced by overlapping PCR extension mutagenesis. For example, the shuttle vector pBRLR-MHE1 was constructed as follows. The primer pair pBRLR-MHE1U and MHE1D contained the complementary mutated region that encoded the E1 epitope from the HAdV-7 hexon gene. A 2,279-bp segment was amplified from plasmid pBRAd⌬E3GFP by using the primers MHE1U and HexD (Table 1). In a separate PCR, a 433-bp segment was ampli-fied from plasmid pBRAd⌬E3GFP using the primers MHE1D and HexU (Table 1). Subsequently, the two PCR products were purified and annealed as a template in a second round of PCR with the primer pair HexU and HexD, to generate a 2,697-bp PCR product, MHE1. This fragment was digested with ClaI/BamHI and inserted into ClaI/ BamHI-digested pBRLR to construct the pBRLR-MHE1 vector. The vectors pBRLR-MHE2, pBRLR-MHE3, and pBRLR-MHE4 were con-structed using the same strategy.

The vectors MHE1, MHE2, MER3, pBRLR-MHE4, were digested with EcoRI and SalI, and the digested fragments (LRMHE1, LRMHE2, LRMHE3, and LRMHE4) containing mutated HAdV-3 hexon genes were purified using a gel extract kit (Qiagen). The plasmid pBRAd⌬E3GFP was digested with PacI and AvrII, and the 35-kb fragment without the wild-type hexon gene was purified and served as the backbone, which was used to transformed BJ5183 cells with the linear fragment MHE1, MHE2, MHE3, or MHE4 to reconstruct the mutant viral genome. The presence of mutations in the viral vectors pBRAd-MHE1, pBRAd-MHE2, pBRAd-MHE3, and pBRAd-MHE4 was verified by se-quencing cDNA clones for hexon transcripts.

(iii) Construction of HAdV-7 hexon mutagenesis vectors pBRAd-H7R1, pBRAd-H7R2, pBRAd-H7R3,and pBRAd-H7R4.We used the same strategy to construct the HAdV-7 hexon mutagenesis vectors pBRAd-H7R1, pBRAd-H7R2, pBRAd-H7R3, and pBRAd-H7R4. For ex-ample, the mutated fragment H7R1 was produced by overlapping PCR extension mutagenesis with primer pair A3HVR1u/HexD, A3HVR1r/ HexU, and HexU/HexD, using pAd3egf/H7 as DNA templates. Then, the fragment H7R1 was cloned into pBRLR to generate shuttle vector pBRLR-H7R1. Finally, the LR-H7R1 fragment was cloned into the pAd3egf/H7 vector to generate HAdV-7 hexon mutagenesis vector pBRAd-MH7R1 by using recombinant DNA technology. The presence of mutations in the viral vectors H7R1, H7R2, H7R3, and pBRAd-H7R4 was verified by sequencing cDNA clones for hexon transcripts.

Before transfection, 4⫻105_{HEK-293 cells were added to each well of} a 24-well plate (Corning) and grown overnight at 37°C in a 5% CO2 environment. The chimeric viral genomes were released by using AsiSI (New England BioLabs) from the vector backbones and transfected into HEK-293 using the Sofast transfection kit (Xiamen Sunma Biotechnol-ogy, China). After 5 days, the transfected cells were harvested, freeze-thawed for three cycles, and used to infect HEp-2 cells. The recombinant adenoviruses rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, rAdH7R1, rAdH7R2, rAdH7R3, and rAdH7R4 were harvested when a 95 to 100% cytopathic effect (CPE) was observed after several passages.

Animals, virus strains, and cells.Female BALB/c mice were pur-chased from the Animal Center at Sun Yat-sen University and housed under pathogen-free conditions. HAdV-7 strain GZ08 (GenBank acces-sion no.GQ478341), HAdV-3 strain GZ01, rAd⌬E3GFP, and eight

re-combinant viruses (rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, rAdH7R1, rAdH7R2, rAdH7R3, and rAdH7R4) were cultured in HEp-2 cells or 293 cells. Virus particles were purified by using CsCl gradient centrifugation. After purification, the viruses were quantified by real-time PCR and stored at⫺80°C. HEp-2 cells and 293 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 100 IU/ml penicillin, 100 mg/ml streptomycin, and 10% newborn calf serum. RT-PCR. To confirm the generation of chimeric viruses, we per-formed reverse transcription-PCR (RT-PCR) assays with specific PCR primers for the eight chimeric viruses, 1, 2, 3, and 4 and rAdH7R1, rAdH7R2, rAdH7R3, and rAdH7R4. The common sense primer HU and antisense primer HD were located in the conserved regions of HAdV-3 and HAdV-7 hexon protein sequences, and the antisense primers (IMHE1a, IMHE2a, IMHE3a, IMHE4a, HR1r, HR2r, HR3r, and HR4r) were located in the eight mutant regions (E1, E2, E3, and E4 from HAdV-7 strain GZ08 hexon and R1, R2, R3, and R4 from HAdV-3 strain GZ01) of chimeric hexons. Total RNA was extracted from HEp-2 cells infected with rAd⌬E3GFP, HAdV-7, rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, rAdH7R1, rAdH7R2, rAdH7R3, or rAdH7R4 and reverse transcribed to cDNA templates. PCR controls included rAd⌬E3GFP cDNA, HAdV-7 cDNA, and total RNA. A PCR using the primer pair HU and HD was performed to test cDNA quality. A PCR using the primer pair HU and IMHE1a (or IMHE2a, IMHE3a, IMHE4a, HR1r, HR2r, HR3r, or HR4r) was performed to examine the stability of chimeric viruses.

Western blotting.HEp-2 cells were infected with the recombinant viruses rAd⌬E3GFP, rAdMHE1, rAdMHE2, rAdMHE3, or rAdMHE4. At 96 h postinfection, cells were harvested and freeze-thawed three times; then the virus suspensions were mixed with 2⫻loading buffer (1% so-dium dodecyl sulfate [SDS], 10% 2-mercaptoethanol, 0.0025% phenol red, and 10% glycerol in 50 mM Tris-HCl [pH 6.8]) and boiled for 5 min at 100°C. Samples were then separated on 10% SDS-polyacrylamide gels and electrophoretically transferred onto nitrocellulose membranes. The membranes were blocked in blocking buffer (1% [wt/vol] bovine serum albumin [BSA], 5% [wt/vol] skim milk in 10 mM phosphate-buffered saline [PBS]) for 2 h and probed with the monoclonal antibody 5D4, which specifically recognizes hexons from HAdV-3 and HAdV-7, as pre-viously described (18). The membranes were further incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies. The color reactions were performed with 0.1 M Tris-HCl (pH 6.8) containing 0.5 mg/ml diaminobenzidine (Sigma) and 0.01% H2O2.

Quantitative PCR.To evaluate the replication characteristics of the recombinant viruses rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, and rAd⌬E3GFP, real-time quantitative PCR (Q-PCR) was performed using an adenovirus Q-PCR kit (Huayin Corporation, Guangzhou, China). HEp-2 cells were infected with 1⫻108_{to 1⫻10}9_{genome copies of} rAd⌬E3GFP, rAdMHE1, rAdMHE2, rAdMHE3, or rAdMHE4. The in-fected cells were harvested after 12, 24, 36, 48, 60, and 72 h postinfection, and viral genome copies were determined using adenovirus Q-PCR kits. Forin vivoneutralization assays, the rAd⌬E3GFP, rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, rAdH7R1, rAdH7R2, rAdH7R3, rAdH7R4, and pAd3egf/H7 genome copy numbers were determined using primers AD3GZ01F, AD3GZ01R, and AD3GZ01FB, and the HAdV-7 genome copy numbers were determined using primers AD7SPR1, AD7SPR2, AD7SPS, and AD7SPBH.

As a standard for the determination of adenovirus genome copies, the rAd⌬E3GFP genome was purified and quantified. Amplification was per-formed using an Applied Biosystems 7500 real-time PCR system. To as-sess reproducibility, each assay was performed independently three times in duplicate.

Fluorescence-based microneutralization assay.Except for HAdV-7, all recombinant viruses (rAd⌬E3GFP, pAd3egf/H7, rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, rAdH7R1, rAdH7R2, rAdH7R3, and rAdH7R4) used in this study encoded enhanced GFP. Fluorescence read-ings were made using a microplate spectrophotometer (Thermo

on November 7, 2019 by guest

http://jvi.asm.org/

skanFlash) to measure the ability of different sera to neutralize these vi-ruses (28).

Six- to 8-week-old BALB/c mice received 1010_{genome copies of CsCl} gradient-purified rAd⌬E3GFP, pAd3egf/H7, HAdV-7, rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, rAdH7R1, rAdH7R2, rAdH7R3,or rAdH7R4 intraperitoneally five times at 2-week intervals (seven animals per virus group). Mice were sacrificed, and antisera were used forin vitro neutralization assays. The sera from naïve BALB/c mice were used as neg-ative controls. Nonspecific protection was observed when the dilution of the serum from naïve mice was lower than 1:64. Therefore, all sera used in the neutralization assays were used at an initial dilution of 1:64.

Each polyclonal serum was inactivated at 56°C for 30 min and then serially diluted in PBS, and 150␮l of each dilution was mixed with 150␮l rAd⌬E3GFP, Ad3/H7, HAdV-7, rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, rAdH7R1, rAdH7R2,rAdH7R3, or rAdH7R4 recombinant vi-ruses (⬃1⫻107virus particles). The antibody-virus mixtures were incu-bated at 37°C and 5% CO2and then transferred into the HEp-2 cells in triplicate (1⫻104_{cells per well) in 96-well plates and incubated for 72 h.} Enhanced GFP expression in HEp-2 cells was measured using a micro-plate spectrophotometer (Thermo VarioskanFlash [488-nm excitation, 507-nm emission]) (28). Background fluorescence was equalized to wells containing cells only, and maximum fluorescence was determined based on wells with cells infected only with virus. Due to edge effects on fluores-cence readings, the outer 36 wells of 96-well plates were not used in any assay. Each assay included control wells of HEp-2 cells only, HEp-2 cells infected with virus, and HEp-2 cells infected with virus mixed with sera from naïve mice. The neutralizing ability of the serum was calculated as follows: (fluorescence in cells incubated with the antibody-virus mixture

⫺background fluorescence)/(maximum fluorescence). The NAbs titers were defined as the serum dilutions that resulted in a 70% reduction of maximum fluorescence (cells infected only with viruses).

To confirm the results of fluorescence-based microneutralization as-says, neutralization assays were performed under the same conditions, and genome copies of adenovirus were quantified to measure the inhibi-tion of virus infecinhibi-tion by different sera using adenovirus Q-PCR kits. In addition, assays were performed independently, three times in duplicate, to assess reproducibility. Furthermore, for further standardization, neu-tralization assays were performed with the same virus batch and serum batch by the same operator. The NAb titers were defined as the serum dilution that resulted in a 90% reduction of maximum genome copy number (cells infected only with viruses).

CMN assay.The HAdV-7 virus vector does not encode GFP; there-fore, colorimetric microneutralization (CMN) assays were used to exam-ine the neutralization abilities of different sera against HAdV-7, as previ-ously described (5). In brief, sera were inactivated at 56°C for 30 min, then serially diluted in PBS, and 150␮l of each dilution was mixed with 150␮l of HAdV-7 virus (100 TCID50[50% tissue culture infectious doses]) at 37°C for 1 h in 5% CO2. Then, 100␮l was transferred into 96-well plates containing HEp-2 cells in triplicate. The assays included control wells of HEp-2 cells only, HEp-2 cells infected with virus, and HEp-2 cells infected with virus that was mixed with sera from naive mice. After incubation for 72 h, the culture medium was removed from the wells. One hundred microliters of a 1:5,000 dilution of Finter’s neutral red in PBS was added, and the plates were incubated at 37°C in 5% CO₂for 1 h. All medium was removed, and the wells were gently washed twice with 150␮l PBS. A final solution of acid alcohol (50% ethanol, 1% acetic acid in distilled water) was added (100␮l per well). TheA550of the solubilized dye was read on a microplate spectrophotometer (Thermo Varyskan Flash). The assays were performed three times in duplicate to assess reproducibility. The NAb titers were defined as the serum dilution that protected 90% of cells from HAdV-7 infection. The results of the CMN assay were confirmed using Q-PCR assays as described above.

In vivoneutralization assay.Female BALB/c mice with no previous exposure to rAd⌬E3GFP and HAdV-7 were tested by Q-PCR and en-zyme-linked immunosorbent assay (ELISA) with blood from tail veins. A

total of 168 mice were divided into two groups: 96 mice in group A and 72 mice in group B (seeFig. 5A, below). In group A, 72 mice were intraperito-neally vaccinated with 1010_{genome copies of different viruses (HAdV-7,} rAd⌬E3GFP, rAdMHE1, rAdMHE2, rAdMHE3, or rAdMHE4) four times at 2-week intervals, with 12 mice per virus group. Forty-eight mice in group B received 1010_{genome copies of rAd⌬E3GFP, rAdMHE3,} rAd-MHE4, or HAdV-7, with the same vaccination schedule and with 12 mice per virus group. Twenty-four nonvaccinated mice in each group were used as negative controls. Two weeks after the last vaccination, vaccinated mice in group A were anesthetized with diethyl ether and challenged in-tranasally with 50␮l (about 2⫻109_{genome copies) of live HAdV-7 virus,} and vaccinated mice in group B were anesthetized and challenged intra-nasally with 50␮l (about 2⫻109_{genome copies) of live rAd⌬E3GFP} virus. Twelve nonvaccinated mice in each group were challenged with live HAdV-7 or rAd⌬E3GFP virus and used as negative controls. Twelve naïve mice without virus challenge in each group were used as additional neg-ative controls.

Prior to receiving the virus challenge (day 0), three mice in each sub-group were sacrificed, and the lungs were used to examine the genome copies of challenge virus, HAdV-7 or rAd⌬E3GFP. The remaining mice were sacrificed on days 3, 5, and 7 (three mice per time point) after virus challenge. The lungs from the sacrificed mice were harvested and homog-enized in a Tissuemizer (Tekmar) in 600␮l of homogenization buffer (10 mM Tris-HCl, 2% SDS, and 1 mM EDTA; pH 8.0). The numbers of HAdV-7 or rAd⌬E3GFP copies in the lung tissues were determined by Q-PCR, as described above.

Adoptive transfer assays.To study the inhibitory effects of HAdV-3-specific NAbs to four rAd3/H7-based hexon chimeric viruses, adoptive transfer studies were performed as previously described (25). BALB/c mice were preimmunized with 1010_{virus particles of wild-type HAdV-3} GZ01 three times, and then IgG was purified from 2 ml of sera pooled from five immunized mice. The purified antibody was then injected into naive mice via the tail vein. For passive antibody transfer, mice were in-jected via the tail vein with 40␮l HAdV-3-specific antibody (with a neu-tralizing titer of 1:16,000). On the day following adoptive transfer, the neutralizing antibody titers of the sera from recipient mice were measured in a fluorescence-based microneutralization assay as described above, and then the recipient mice were immunized intramuscularly with 5⫻109 virus particles of rAd⌬E3GFP, rAd3/H7, rAdH7R1, rAdH7R2, rAdH7R3, and rAdH7R4 vectors, which encoded EGFP. Control mice received adoptive transfer of IgG purified from PBS-injected mice.

At days 12 and 18 postimmunization, the EGFP-specific antibody ti-ters from immunized mice were measured by ELISA. Ninety-six-well plates were coated overnight with 100␮l/well of 1␮g/ml recombinant GFP (Millipore) and then blocked for 2 h with blocking buffer (PBS, 2% BSA,0.05% Tween 20). Sera dilutions (100-fold dilution of sera obtained at day 12 postimmunization, 40-fold dilution of sera obtained at day 18 postimmunization) were then added to the wells and incubated for 1 h. The plates were washed three times with PBST (PBS, 0.05% Tween 20) and incubated for 1 h with a 1/10,000 dilution of rabbit anti-mouse IgG plus IgM (H⫹L)–peroxidase– conjugated affinity-purified secondary an-tibodies (Jackson ImmunoResearch). The plates were washed four times, and reactions were developed with tetramethylbenzidine, stopped with 2 M H2SO4, and analyzed at 450 nm with an ELISA plate reader (Thermo Scientific Multiskan MK3).

RESULTS

Analysis of potential serotype-specific NAb epitopes for

HAdV-3 and HAdV-7.

Previous studies indicated that epitopes

recognized by serotype-specific NAbs may be exposed on the

vi-rion surface and reside within the seven HVRs of hexon proteins

(

1

,

17

). Seven HVRs that reside within HAdV-3 and HAdV-7

hexon proteins were identified by alignment of the available

hu-man adenovirus hexon protein sequences (

4

). Using sequence

alignment data and the known 3D structure of hexon proteins

on November 7, 2019 by guest

http://jvi.asm.org/

from HAdV-2, HAdV-5, and AdC68 as models, four regions

within the HAdV-3 hexon (R1, residues 136 to 151; R2, residues

164 to 187; R3, residues 265 to 284; R4, residues 422 to 437) and

four regions within the HAdV-7 hexon (E1, residues 136 to 148;

E2, residues 161 to 181; E3, residues 258 to 276; E4, residues 415 to

426) were predicted as potential NAb targets (

Fig. 1A

). These

po-tential neutralization sites for HAdV-3 and HAdV-7 were located

within the HVR1 (R1 and E1), HVR2 (R2 and E2), HVR5 (R3 and

E3), and HVR7 of hexon proteins (R4 and E4) (

Fig. 1A

).

Con-served amino acids were found between the predicted epitopes of

HAdV-3 and the corresponding epitopes of HAdV-7 by sequence

alignment (

Fig. 1B

).

Four hexon chimeric adenovirus vectors were constructed

by replacing the predicted epitopes within the HAdV-3 hexon

protein with the corresponding epitopes from the HAdV-7

hexon.

To identify the neutralizing epitopes for HAdV-7, four

hexon chimeric adenoviruses containing predicted NAb epitopes

from HAdV-7 were constructed. A recombinant HAdV-3-based

vector, rAd

⌬

E3GFP expressing GFP, was constructed in a

previ-ous study (

32

). Four recombinant virus vectors, rAdMHE1 (with

the predicted R1 epitope of the HAdV-3 hexon protein replaced

with the E1 epitope from the HAdV-7 hexon protein), rAdMHE2

(R2 replaced with E2), rAdMHE3 (R3 replaced with E3), and

rAdMHE4 (R4 replaced with E4), were constructed by switching

one of the predicted epitopes within the hexon protein of

rAd

⌬

E3GFP with the corresponding epitope from the

corre-sponding HAdV-7 hexon (see

Fig. 4A

, below). Transmission

elec-tron microscope observation indicated that all four recombinant

viruses were rescued and packaged successfully in HEp-2 cells

(data not shown).

The stabilities of four chimeric hexon sequences were

exam-ined in RT-PCR assays. Total RNA was isolated from HEp-2 cells

infected with rAdMHE1, rAdMHE2, rAdMHE3, or rAdMHE4 for

10 serial passages, and no virus genomic DNA contamination was

detected in PCR assays (

Fig. 2A

, lanes 20 to 25). Then, cDNA

templates were prepared, and all of them were suitable for the

RT-PCR assays (

Fig. 2A

, lanes 14 to 19). The cDNA template

gen-erated from HAdV-7-infected HEp-2 cells was used as a positive

control, and the cDNA template generated from rAd

⌬

E3GFP-infected HEp-2 cells was used as a negative control. RT-PCR

as-says indicated that all four hexon chimeric vectors remained stable

for at least 10 serial passages

in vitro

without detectable loss of

chimeric hexon sequences (

Fig. 2A

, lanes 9 to 12). No changes to

the chimeric hexon sequence were detected by sequencing the

cloned PCR fragments. The expression of chimeric hexons was

confirmed by Western blotting with the monoclonal antibody

5D4, which recognizes a common epitope of the HAdV-3 and

HAdV-7 hexon proteins (

Fig. 2B

) (

18

).

Replication efficiencies of four chimeric viruses and

rAd

⌬

E3GFP.

During the rescue and purification processes for

recombinant viruses, we observed that under the same conditions,

a shorter time was needed for rAdMHE4-infected HEp-2 cells to

reach a complete CPE than cells infected with other chimeric

FIG 1Analysis of the potential serotype-specific neutralizing epitopes for HAdV-3 and HAdV-7. (A) Modeling of 3D structures of the HAdV-3 hexon and HAdV-7 hexon and map of the potential epitopes on the 3D structures. Potential epitopes of HAdV-3 and HAdV-7 hexons are indicated at the tops of the molecules. (B) Sequence alignment of potential epitopes from HAdV-3 and HAdV-7 hexon proteins.

FIG 2Confirmation of expression of the four chimeric hexons based on RT-PCR and Western blot assays. (A) RT-RT-PCR assays. RT-RT-PCRs were performed to amplify the hexon chimeric regions by using chimeric viral cDNA, and products were analyzed by 1% agarose gel electrophoresis. Lanes 9 and 16: rAdMHE1 cDNA; lanes 10 and 17, rAdMHE2 cDNA; lanes 11 and 18, rAdMHE3 cDNA; lanes 12 and 19, rAdMHE4 cDNA. RT-PCR controls in-cluded rAd⌬E3GFP viral cDNA (lanes 1 to 4 and 14) and total RNA extracted from cells infected with rAd⌬E3GFP (lane 20), HAdV-7 viral cDNA and total RNA extracted from cells infected with HAdV-7 (lane 21), four chimeric virus cDNAs (lanes 16 to 19), and total RNA extracted from cells infected with rAdMHE1, rAdMHE2, rAdMHE3, or rAdMHE4 (lanes 22 to 25). (B) Western blot assay results. Total protein extracts of HEp-2 cells infected with rAdMHE1 (lane 1), rAdMHE2 (lane 2), rAdMHE3 (lane 3), rAdMHE4 (lane 4), or rAd⌬E3GFP (lane 5) were separated by SDS-PAGE and then analyzed by Western blotting using monoclonal antibody 5D4. MK, protein molecular size

marker.

on November 7, 2019 by guest

http://jvi.asm.org/

viruses or rAd

⌬

E3GFP virus, but it required 2-fold more

rAdMHE4-infected HEp-2 cells to acquire 1

⫻

10

rAdMHE4

virus particles compared to cells infected with other viruses. We

concluded that the R4 epitope replaced by E4 from the HAdV-7

hexon protein had an impact on the efficiency of rAdMHE4

rep-lication. Q-PCR assays were performed to gain a quantitative

un-derstanding of the replication efficiencies of rAd

⌬

E3GFP and the

four recombinant viruses (

Fig. 3

). Approximately 10

to 10

virus

particles were used to infect HEp-2 cells. Two days postinfection,

the CPE ratios for rAdMHE4-infected cells and

rAdMHE2-in-fected cells was 100% and 80%, respectively, while the CPE ratios

for rAdMHE1-, rAdMHE5-, or rAd

⌬

E3GFP-infected cells were

very low (less than 10%). All infected cells were found to reach

CPE 3 days postinfection. In addition, the final yield of rAdMHE4

was 10 times lower than that of rAd

⌬

E3GFP, while rAdMHE1 and

rAdMHE3 replicated as well as the parental virus, rAd

⌬

E3GFP, in

HEp-2 cells. The replication efficiency of rAdMHE2 was the

high-est before 36 h postinfection, but the final yield was similar to that

for rAd

⌬

E3GFP, rAdMHE1, and rAdMHE3. These results

dem-onstrated that all of the chimeric viruses could encapsidate the

recombinant genomes, while the replication efficiency of

rAdMHE4 was lower than that of the other three chimeric viruses.

Of interest, a lower virus titer was required for rAdMHE4 than

other chimeric viruses or rAd

⌬

E3GFP to cause CPE. This may

have been be due to the replaced R4 epitope being very important

for HAdV-3 virus replication and infection.

Identification of serotype-specific NAb epitopes for HAdV-7

in vitro

.

In vitro

neutralization assays were performed to detect

whether anti-HAdV-7 sera could cross-neutralize four chimeric

viruses, and the cross-neutralization responses of

anti-rAd

⌬

E3GFP sera to four chimeric viruses were analyzed too (

Fig.

4B

to

E

). The NAb titer was defined as the serum dilution that

resulted in a 70% reduction of the maximum fluorescence signal

(fluorescence signal occurs only in cells infected with viruses that

encode GFP), as assessed in a fluorescence-based

microneutraliza-tion assay as previously described (

23

,

28

). Sera from naive mice

were used as negative controls, and sera from mice preimmunized

with the four chimeric viruses were used as positive controls. With

sera from HAdV-7-preimmunized mice in neutralization assays,

high rAdMHE2-specific NAb titers (log

titer, 6) and

rAdMHE3-specific NAb titers (log

titer, 8) were detected, but

rAdMHE1-and rAdMHE4-specific NAb titers were not observed (

Fig. 4B

).

These data indicate that E2 and E3, but not E1 and E4, are HAdV-7

NAbs targets. With sera from rAd

⌬

E3GFP-preimmunized mice

in neutralization assays, high NAb titers were observed against all

four chimeric viruses, and the rAdMHE4 NAb titer was 2 log

to 4

log

lower than the NAb titers of the other three chimeric viruses

(

P

⬍

0.05, Duncan’s multiple range test) (

Fig. 4B

).

To confirm the results of the inhibition of virus infection by

measuring GFP expression, the microneutralization assays were

repeated and the chimeric virus genome copy number was

deter-mined by Q-PCR (

Fig. 4C

). The NAb titer was defined as the

serum dilution that resulted in a 90% reduction of maximum of

genome copy number (genome copies in cell infected only with

viruses). rAdMHE2-specific (log

titer, 6) and rAdMHE3-specific

(log

titer, 8) NAb titers, but not rAdMHE1- and

rAdMHE4-spe-cific NAb titers, induced by HAdV-7 were observed. As expected,

high NAb titers against all four chimeric viruses were observed.

The rAdMHE4 NAb titer was the lowest (

P

⬍

0.005, Duncan’s

multiple range test). These results supported those of the

fluores-cence-based microneutralization assays (

Fig. 4B

).

Further microneutralization assays were performed to determine

whether rAd

⌬

E3GFP and HAdV-7 could escape the host

neutraliza-tion response generated by four chimeric viruses (

Fig. 4D

and

E

). We

found that sera from all four groups of mice preimmunized with

chimeric virus inhibited rAd

⌬

E3GFP infection; however, there were

meaningful differences in the NAb titers. rAd

⌬

E3GFP-specific NAb

titers induced by rAdMHE4 were 8 log

lower than those induced by

rAd

⌬

E3GFP (

P

⬍

0.005, Duncan’s multiple range test).

rAd

⌬

E3GFP-specific NAb titers induced by rAdMHE3 were 4 log

lower than those induced by rAd

⌬

E3GFP (

P

⬍

0.005, Duncan’s

mul-tiple range test), and rAd

⌬

E3GFP-specific NAb titers induced by

rAdMHE2 were 2 log

lower than those induced by rAd

⌬

E3GFP (

P

⬍

0.05, Duncan’s multiple range test). No statistical differences in

rAd

⌬

E3GFP NAbs titers were found between anti-rAdMHE1 sera

and anti-rAd

⌬

E3GFP sera in fluorescence-based

microneutraliza-tion assays (

P

⬍

0.1, Duncan’s multiple range test). In the

Q-PCR-based microneutralization assay, no statistical differences in

rAd

⌬

E3GFP Nabs titers were found between rAdMHE1,

anti-rAdMHE2, and anti-rAd

⌬

E3GFP sera (

Fig. 4C

). These data indicate

that rAd

⌬

E3GFP NAbs may be directed partially, but not exclusively,

against only one of the four predicted epitopes.

HAdV-7 virus does not encode GFP; therefore, CMN assays

were performed to determine whether HAdV-7 virus could

escape the host neutralizing response generated by the four

chimeric viruses (

Fig. 4D

and

E

). Anti-rAdMHE2 and

anti-rAdMHE3 sera, but not anti-rAdMHE1 or rAdMHE4 sera,

protected HEp-2 cells from HAdV-7 infection.

HAdV-7-spe-cific NAb titers generated by anti-rAdMHE2 sera were 2 log

lower than those of anti-HAdV-7. HAdV-7-specific NAb titers

generated by anti-rAdMHE3 sera and anti-HAdV-7 sera

showed no significant differences (

P

⬍

0.05, Duncan’s multiple

range test). The Q-PCR results were in accordance with the

results of the CMN assays. These results further confirmed that

FIG 3DNA replication curves of rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, and rAd⌬E3GFP. Cells infected with rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, or rAd⌬E3GFP were harvested at 12, 24, 36, 48, 60, and 72 h, and the genome copy numbers were determined by Q-PCR with Q-PCR kits. Each experiment was repeated independently three times, and the mean values and standard deviations are shown.

on November 7, 2019 by guest

http://jvi.asm.org/

the predicted epitopes E2 and E3 from the HAdV-7 hexon

pro-tein neutralized epitopes for HAdV-7, and E3 played a more

important role in generating NAb responses.

Confirmation of neutralizing epitopes for HAdV-7

in vivo

.

In vivo

neutralization assays were performed to investigate

whether the chimeric viruses could generate a host neutralizing

response against HAdV-7 and HAdV-3 infection (

Fig. 5

). The

study design is shown in

Fig. 5A

. The target organ of HAdV-3 and

HAdV-7 is the respiratory system; therefore, lung tissue was

cho-sen for the detection of adenovirus infection. Groups of mice were

immunized four times with rAdMHE1, rAdMHE2, rAdMHE3,

rAdMHE4, rAd

⌬

E3GFP, or HAdV-7. Two weeks after the last

vaccination, the mice in group A were challenged intranasally with

live HAdV-7 virus, and the mice in group B were challenged

in-tranasally with live rAd

⌬

E3GFP virus. Mice were then sacrificed

after virus infection on days 0 (before HAdV-7 or rAd

⌬

E3GFP

FIG 4Cross-neutralization responses between hexon chimeric vectors and rAd⌬E3GFP or HAdV-7. (A) Schematic depiction of four rAd⌬E3GFP-based hexon chimeric vectors. Epitopes derived from HAdV-7 (E1, E2, E3, and E4) are shown in the empty boxes. (B and C) Assessment of rAdMHE1-, rAdMHE2-, rAdMHE3-, and rAdMHE4-specific NAb titers induced by anti-rAd⌬E3GFP or anti-HAdV-7 sera. Sera from mice preimmunized by the hexon chimeric virus itself were used as positive controls, and sera from naïve mice were used as negative controls. (D and E) Assessment of rAd⌬E3GFP and HAdV-7-specific NAb titers induced by anti-rAdMHE1, anti-rAdMHE2, anti-rAdMHE3, or anti-rAdMHE4 sera. Sera from rAd⌬E3GFP- or HAdV-7-preimmunized mice were used as positive/negative controls. The neutralization assays depicted in panels B to E were performed under the same conditions. (B and D) Hexon chimeric virus NAb titers and rAd⌬E3GFP NAb titers were measured in fluorescence-based neutralization assays. HAdV-7 Nab titers in panel D were measured by CMN assay. (C and E) NAb titers were measured in Q-PCR-based neutralization assays. *,P⬍0.05; **,P⬍0.005.

on November 7, 2019 by guest

http://jvi.asm.org/

challenge), 3, 5, and 7, and the lung tissues were examined for

HAdV-7 or rAd

⌬

E3GFP infection.

In group A, HAdV-7 genome copies could only be detected in

the lung tissues from naïve mice or mice preimmunized with the

four chimeric viruses or with rAd

⌬

E3GFP on days 3 and 5, with

higher HAdV-7 genome copies present on day 3 than on day 5. No

HAdV-7 genome copies could be detected before HAdV-7

chal-lenge (day 0) or after HAdV-7 chalchal-lenge on day 7 (

Fig. 5B

). No

HAdV-7 genome copies were detected in the lungs of mice

preim-munized with HAdV-7 on day 3. Very low HAdV-7 genome copy

numbers (near the detection limit of the Q-PCR) were detected in

mice preimmunized with rAdMHE3 on day 3, and no genome

copies were detected on day 5; thus, preimmunization with

rAdMHE3 contributed to rapid clearance of HAdV-7

in vivo

.

HAdV-7 genome copies in mice preimmunized with rAdMHE2

were lower than those in mice preimmunized with rAdH1,

rAdMHE4, or rAd

⌬

E3GFP or in naïve mice on day 3 (

P

⬍

0.05,

Duncan’s multiple range test). Thus, preimmunization with

rAdMHE2 also contributed to the clearance of HAdV-7 in BALB/c

mice

in vivo

. In conclusion, rAdMHE2 and rAdMHE3 generate a

host neutralizing response against HAdV-7. The E2 and E3

epitopes were NAb epitopes for HAdV-7, and E3 played a more

important role in generating NAb responses; these results were in

accordance with the results of the neutralization assays

in vitro

.

Further neutralization assays

in vivo

were also performed to

investigate whether the chimeric viruses could generate a host

neutralizing response against rAd

⌬

E3GFP. We found that

preim-munization of mice with rAdMHE1 or rAdMHE2 contributed

equally to the clearance of rAd

⌬

E3GFP in BALB/c mice, while

preimmunization of rAdMHE3 or rAdMHE4 contributed little to

the clearance of rAd

⌬

E3GFP in BALB/c mice (data not shown).

To confirm these results, we repeated the neutralization assays

in

vivo

again and focused on the neutralizing responses of rAdMHE3

and rAdMHE4. As shown in

Fig. 5C

, no significant differences in

the numbers of rAd

⌬

E3GFP genome copies were detected

be-tween the mice preimmunized with rAdMHE4 versus HAdV-7 on

day 3. The number of rAd

⌬

E3GFP genome copies in mice

preim-munized with rAdMHE3 was lower than that of mice

preimmu-nized with rAdMHE4 but higher than that of mice preimmunizied

with rAd

⌬

E3GFP on day 3 (

P

⬍

0.05, Duncan’s multiple range

test). The rAd

⌬

E3GFP genome copy numbers in the lungs of mice

preimmunized with rAdMHE3 or HAdV-7 decreased to almost

the same level as that of mice vaccinated with rAd

⌬

E3GFP on days

5 and 7, while the rAd

⌬

E3GFP genome copy numbers remained

relatively high in lung tissues of mice vaccinated with rAdMHE4

and in naive mice. These data indicated that replacement of either

the R3 or R4 epitope in the rAd

⌬

E3GFP vector has a great effect on

generating the host neutralization response against rAd

⌬

E3GFP

in vivo

, and these results were in accordance with the results of

neutralization assays

in vitro

.

Construction of hexon chimeric adenovirus vectors by

re-placing the predicted epitopes within the HAdV-7 hexon

pro-tein with the corresponding epitopes from the HAdV-3 hexon.

The rAd

⌬

E3GFP-specific NAb responses induced by rAdMHE3 and

rAdMHE4 were weaker than those against rAdMHE1 and

rAd-MHE2. We doubted that these would play a dominant role in

gener-ating the host neutralization response against rAd

⌬

E3GFP.

There-fore, four recombinant adenovirus vectors containing HAdV-7

hexon chimeras were constructed (

Fig. 6A

). A recombinant

FIG 5Analysis of NAb responses of BALB/c mice to rAd⌬E3GFP or HAdV-7 infection generated by preimmunization with rAdMHE1, rAdMHE2, rAdMHE3, or rAdMHE4. (A) Study design. (B) Mice were preimmunized with rAdMHE1, rAdMHE2, rAdMHE3, rAdMHE4, rAd⌬E3GFP, or HAdV-7 and then challenged with HAdV-7. Q-PCR assays were performed to detect genome copy numbers of HAdV-7 in lung tissues. (C) Mice were preimmunized with rAdMHE3, rAdMHE4, rAd⌬E3GFP, or HAdV-7 and then challenged with rAd_⌬E3GFP. Q-PCR assays were per-formed to detect genome copy numbers of rAd⌬E3GFP in lung tissues. Controls included mice preimmunized with rAd_⌬E3GFP or HAdV-7 or naïve mice *,P⬍0.05.

on November 7, 2019 by guest

http://jvi.asm.org/

rAd

⌬

E3GFP-based hexon chimeric vector, rAd3/H7, in which the

HAdV-3 hexon was replaced with that of HAdV-7, was constructed

in a previous study (

27

). Previous work showed that the majority of

HAdV-3- and HAdV-7-specific NAbs target their hexon proteins

(

27

). Four rAd3/H7-based, hexon chimeric virus vectors, rAdH7R1

(with the E1 epitope from the HAdV-7 hexon protein replaced by the

R1 epitope from the HAdV-3 hexon protein), rAdH7R2 (E2 replaced

by R2), rAdH7R3 (E3 replaced by R3), and rAdH7R4 (E4 replaced by

R4), were constructed as described above. All four recombinant

vi-ruses were rescued and packaged successfully in HEp-2 cells. cDNA

FIG 6Cross-neutralization NAb responses between rAd3/H7-based hexon chimeric vectors and rAd_⌬E3GFP. (A) Schematic depiction of four rAd3/H7-based hexon chimeric vectors. Epitopes derived from the HAdV-3 hexon (R1, R2, R3, and R4) are shown as empty boxes. (B and C) Sera from rAd⌬ E3GFP-preimmunized mice were assessed for NAb titers specific to rAdH7R1, rAdH7R2, rAdH7R3, rAdH7R4, or rAd3/H7. rAd_⌬E3GFP was used as a positive control. (D and E) Sera from mice preimmunized with rAdH7R1, rAdH7R2, rAdH7R3, rAdH7R4, or rAd3/H7 were assessed for rAd⌬E3GFP-specific NAb titers. Control sera included anti-rAd_⌬E3GFP and naïve serum. The neutralization assays depicted in panels B to E were performed under the identifical conditions. (B and D) NAb titers as measured in fluorescence-based neutralization assays. (C and E) NAb titers measured in Q-PCR based neutralization assays.*,P⬍0.05; **,P⬍ 0.005.

on November 7, 2019 by guest

http://jvi.asm.org/

sequences in the hexon chimeric domains of the recombinant viruses

were confirmed by sequencing.

Further identification of serotype-specific NAb epitopes

for HAdV-3.

Sera from BALB/c mice preimmunized with

rAd

⌬

E3GFP, rAd3/H7, rAdH7R1, rAdH7R2, rAdH7R3, or

rAdH7R4 were prepared, and NAb titers of these serum samples

ranged from 14 log

to 16 log

titer in fluorescence-based

neutral-ization assays.

With sera from mice preimmunized with rAd

⌬

E3GFP,

in vitro

neutralization assays were performed to evaluate

HAdV-3-spe-cific NAb responses against four rAd3/H7-based, hexon chimeric

viruses. Sera from naïve mice were used as negative controls. As

shown in

Fig. 6B

, high rAd

⌬

E3GFP NAb titers (16 log

) and low

rAd/H7 NAb titers (9 log

) were detected, as described previously

(

27

). rAdH7R1 NAb titers were 4 log

higher than rAd/H7 NAb

titers (

P

⬍

0.005, Duncan’s multiple range test). The NAb titers

for rAdH7R2, rAdH7R3, and rAdH7R4 showed no significant

dif-ference (

P

⬍

0.05, Duncan’s multiple range test), and they were

about 2 log

higher than the rAd/H7 NAb titers (

P

⬍

0.05,

Dun-can’s multiple range test). Further neutralization assays were

car-ried out to evaluate the cross-neutralization responses of sera

from mice preimmunized with the four chimeric viruses to

rAd

⌬

E3GFP. As shown in

Fig. 6D

, the titer of anti-rAdMH1 sera

was 6 log

higher than that of anti-rAd3/H7 sera against

rAd

⌬

E3GFP (

P

⬍

0.005, Duncan’s multiple range test). The

rAd

⌬

E3GFP NAb titers generated by rAdMH2,

anti-rAdMH3, and anti-rAdMH4 sera showed no significant

differ-ence, and they were at least 2 log

higher than that of anti-rAd3/H7

sera (

P

⬍

0.05, Duncan’s multiple range test). As a positive

con-trol, rAd

⌬

E3GFP NAb titers generated by anti-rAd

⌬

E3GFP sera

were the highest. The Q-PCR results correlated with the results of

fluorescence-based neutralization assays (

Fig. 6C

and

E

). These

data indicate that all four epitopes, R1 to R4, are NAb epitopes of

HAdV-3, and R1 played the most important role in generating

NAb responses.

Confirmation of serotype-specific NAb epitopes for HAdV-3

in vivo

.

The functional significance of four NAb epitopes, R1 to

R4, was assessed by determining the suppression capacity of

HAdV-3-specific NAbs to the immunogenicity of the four

chime-ric viruses. Purified IgG from mice with wild-type HAdV-3 GZ-01

immunity were confirmed for HAdV-3-specific NAb activity

be-fore adoptive transfer to recipient mice. Groups of naive recipient

mice (

n

⫽

3/group) received 40

␮

l of purified IgG by tail vein

injection before immunization with 5

⫻

10

virus particles of

rAd

⌬

E3GFP, rAd3/H7, rAdH7R1, rAdH7R2, rAdH7R3, or

rAdMH7R4 vectors (all these viruses encode GFP). To verify

cir-culating NAb titers in these animals, serum was obtained after

adoptive transfer and immediately before vaccination. As

ex-pected, serum NAbs in the recipient mice following adoptive

transfer exhibited similar HAdV-3-specific NAb titers that ranged

from 64 to 128. Vaccine-elicited immune responses were assessed

by enhanced GFP-specific ELISAs on day 12 and day 18. As shown

in

Fig. 7B

and

C

, rAdH7R1- and rAd

⌬

E3GFP-elicited humoral

immune responses were significantly and durably suppressed by

HAdV-3-specific NAbs (

P

⬍

0.05,

t

test), while no significant

sup-pression of humoral immune responses elicited by rAd3/H7,

rAdH7R2, rAdH7R3, or rAdH7R4 were detected by

HAdV-3-spe-cific NAbs (

P

⬍

0.05,

t

test). Together, the results of the

neutral-ization assays and adoptive transfer assays show that R1 appears to

be a major NAb epitope, while R2 to R4 are relative minor NAb

epitopes for rAd

⌬

E3GFP.

DISCUSSION

The subspecies B1 adenoviruses HAdV-3 and HAdV-7 have

occurred epidemically and contribute greatly to pediatric acute

respiratory diseases (

9

,

11

,

14

,

26

). Identification of the NAb

epitopes of HAdV-3 and HAdV-7 will be beneficial for

devel-oping recombinant bivalent HAdV-3/HAdV-7 vaccines and

preventing outbreak of HAdV-3 and HAdV-7 infection. In this

study, four NAb epitopes within hexon HVRs were predicted

for HAdV-3 and HAdV-7 by bioinformatics. Eight hexon

chi-meric adenovirus vectors with the alteration of only one

pre-dicted neutralizing epitope were constructed. All of them were

rescued and packaged successfully in HEp-2 cells. The

replica-tion efficiency of four rAd

⌬

E3GFP-based hexon chimeras

demonstrated that a lower virus titer was required for

rAdMHE4 than the other three chimeric viruses (rAdMHE1,

rAdMHE2, and rAdMHE3) or the parental virus rAd

⌬

E3GFP

FIG 7Adoptive transfer of purified IgG from mice with high levels of anti-HAdV-3 immunity. Purified IgG antibody levels are shown for mice preimmunized with wild-type HAdV-3 adoptively transferred into naive recipient BALB/c mice (n⫽3/group) by tail vein injection before immunization with 5⫻109_virus

particles of rAdE3GFP, rAd3/H7, rAdH7R1, rAdH7R2, rAdH7R3, rAdH7R4, or PBS (none). Vaccine-elicited immune responses following immunization were determined by EGFP-specific ELISAs at 12 days (A) or 18 days (B). Controls included mice that received adoptive transfer of naive IgG purified from saline-injected mice (naive). *,P⬍0.05.

on November 7, 2019 by guest

http://jvi.asm.org/

to cause CPE, but the final virion yield was the lowest. We

specu-late that exchanging the R4 epitope on the surface of the HAdV-3

hexon disturbs its interaction with other cell components and

boosts cell death, and cell death in turn decreases the replication

efficiency of rAdMHE4 in later stages. A recent study indicated

that adenovirus transport occurs through a direct cytoplasmic

dy-nein-hexon interaction (

2

). Exchanging the R4 epitope on the

surface of the HAdV-3 hexon may have an adverse effect on the

dynein-hexon interaction, resulting in the retention of viruses in

the cytoplasm and cell death. However, the role of the hexon in the

CPE needs to be further investigated.

We further identified four serotype-specific NAb epitopes (R1

to R4) for HAdV-3 and two serotype-specific NAb epitopes (E2

and E3) for HAdV-7. Yuan et al. previously reported five NAb

epitopes which reside in HVR1, HVR2, HVR4, HVR5, and HVR7

of the HAdV-3 hexon based on molecular modeling (

30

). Four

NAb epitopes (R1 to R4) identified in this study overlap with four

of the five NAb epitopes the other authors reported. Considering

that HAdV-3 NAbs target partially, but not exclusively, to all four

identified NAb epitopes, further studies will be necessary to

exam-ine whether the remaining three HVRs contribute to HAdV-3 or

HAdV-7 NAb responses.

Cross-neutralization NAb assays between four rAd3/H7-based

hexon chimeric vectors (rAdH7R1, rAdH7R2, rAdH7R3, and

rAdH7R4) and rAd

⌬

E3GFP indicated that all four predicted

epitopes are NAb epitopes of HAdV-3, while R1 play the most

important role in generating NAb responses (

Fig. 6

). Further

adoptive transfer studies confirmed that R1 was a dominant NAb

epitope (

Fig. 7

). We were surprised to find that the rAd

⌬

E3GFP-specific neutralizing epitope responses induced by rAdMHE3 (R3

replaced by E3) and rAdMHE4 (R4 replaced by E4) were weaker

than those of rAdMHE1 (R1 replaced by E1) and rAdMHE2 (R2

relaced by E2)

in vitro

and

in vivo

(

Fig. 4

and

5

). In accordance, E2

and E3 were identified for HAdV-7 serotype-specific NAb

epitopes, but HAdV-7-specific NAb neutralizing epitope

re-sponses induced by rAdH7R2 (E2 replaced by R2) and rAdH7R3

were weaker than those for rAdH7R1 (E1 replaced by R1) and

rAdH7R4 (E4 replaced by R4) (data not shown). Thus, a

domi-nant NAb epitope candidate may be a good candidate for

gener-ating NAb responses, but it may not be a good switched candidate

to evade preexisting host neutralization responses generated by

parental vectors.

In vitro

neutralization assays indicated that E2 and E3 are NAb

epitopes for HAdV-7, and E3 plays a more important role in

gen-erating NAb responses. Results of neutralization assays

in vivo

were in accordance with these results. HAdV-7 virus loads

de-tected in different groups of mice preimmunized with different

viruses were significantly different on day 3. HAdV-7 virus loads

in mice preimmunized with rAdMHE2 and rAdMHE3 were

ob-viously lower than those of mice preimmunized with rAdMHE1,

rAdMHE2, or rAd

⌬

E3GFP (

P

⬍

0.05, Duncan’s multiple range

test). Thus, both E2 and E3 are good candidates for generation of

recombinant vaccines against HAdV-7 infection.

In summary, our studies have identified two serotype-specific

NAb epitopes for HAdV-7 and four serotype-specific NAb

epitopes for HAdV-3. All of them are located in hexon HVRs.

These findings will contribute not only to the development of new

vaccines, but also to the development of new gene delivery vectors.

ACKNOWLEDGMENTS

This study was supported by a foundation of bioMérieux, the National Nature Science Foundation of China (31070150), the Nature Science Foundation of Guangdong (10151009504000006), and School of Guang-zhou research projects (10A006D).

We are grateful for the kind assistance from all of our colleagues.

REFERENCES

1.Bradley RR, et al.2012. Adenovirus serotype 5-specific neutralizing an-tibodies target multiple hexon hypervariable regions. J. Virol.86:1267– 1272.

2.Choi EH, et al.2005. Adenovirus type 7 peptide diversity during out-break, Korea, 1995–2000. Emerg. Infect. Dis.11:649 – 654.

3.Chuang YY, et al.2003. Severe adenovirus infection in children. J. Mi-crobiol. Immunol. Infect.36:37– 40.

4.Crawford-Miksza L, Schnurr DP.1996. Analysis of 15 adenovirus hexon proteins reveals the location and structure of seven hypervariable regions containing serotype-specific residues. J. Virol.70:1836 –1844.

5.Crawford-Miksza LK, Schnurr DP.1994. Quantitative colorimetric mi-croneutralization assay for characterization of adenoviruses. J. Clin. Mi-crobiol.32:2331–2334.

6.de Jong JC, Osterhaus AD, Jones MS, Harrach B.2008. Human adeno-virus type 52: a type 41 in disguise? J. Virol.82:3809. (Letter.)

7.Ebner K, Pinsker W, Lion T.2005. Comparative sequence analysis of the hexon gene in the entire spectrum of human adenovirus serotypes: phy-logenetic, taxonomic, and clinical implications. J. Virol.79:12635–12642. 8.Eswar N, Eramian D, Webb B, Shen MY, Sali A.2008. Protein structure

modeling with MODELLER. Methods Mol. Biol.426:145–159. 9.Guo L, et al.2012. Detection of three human adenovirus species in adults

with acute respiratory infection in China. Eur. J. Clin. Microbiol. Infect. Dis.31:1051–1058.

10. Hayashi S, Hogg JC.2007. Adenovirus infections and lung disease. Curr. Opin. Pharmacol.7:237–243.

11. James L, et al.2007. Outbreak of human adenovirus type 3 infection in a pediatric long-term care facility—Illinois, 2005. Clin. Infect. Dis.45:416 – 420.

12. Jones MS, II, et al.2007. New adenovirus species found in a patient presenting with gastroenteritis. J. Virol.81:5978 –5984.

13. Knopf HL, Hierholzer JC.1975. Clinical and immunologic responses in patients with viral keratoconjunctivitis. Am. J. Ophthalmol.80:661– 672.

14. Lee J, Choi EH, Lee HJ.2010. Clinical severity of respiratory adenoviral infection by serotypes in Korean children over 17 consecutive years (1991–2007). J. Clin. Virol.49:115–120.

15. Mitchell LS, Taylor B, Reimels W, Barrett FF, Devincenzo JP.2000. Adenovirus 7a: a community-acquired outbreak in a children’s hospital. Pediatr. Infect. Dis. J.19:996 –1000.

16. Reference deleted.

17. Pichla-Gollon SL, et al.2007. Structure-based identification of a major neutralizing site in an adenovirus hexon. J. Virol.81:1680 –1689. 18. Qiu H, et al.2009. Epitope mapping and cross-reactivity analysis of the

monoclonal antibodies against hexon protein of human adenovirus type 3. Virus Res.146:58 – 65.

19. Roberts DM, et al.2006. Hexon-chimaeric adenovirus serotype 5 vectors circumvent pre-existing anti-vector immunity. Nature441:239 –243. 20. Rux JJ, Burnett RM.2000. Type-specific epitope locations revealed by

X-ray crystallographic study of adenovirus type 5 hexon. Mol. Ther.

1:18 –30.

21. Selvaraju SB, Kovac M, Dickson LM, Kajon AE, Selvarangan R.2011. Molecular epidemiology and clinical presentation of human adenovirus infections in Kansas City children. J. Clin. Virol.51:126 –131.

22. Reference deleted.

23. Sprangers MC, et al.2003. Quantifying adenovirus-neutralizing antibod-ies by luciferase transgene detection: addressing preexisting immunity to vaccine and gene therapy vectors. J. Clin. Microbiol.41:5046 –5052. 24. Sumida SM, et al.2005. Neutralizing antibodies to adenovirus serotype 5

vaccine vectors are directed primarily against the adenovirus hexon pro-tein. J. Immunol.174:7179 –7185.

25. Sumida SM, et al.2004. Neutralizing antibodies and CD8⫹T lympho-cytes both contribute to immunity to adenovirus serotype 5 vaccine vec-tors. J. Virol.78:2666 –2673.

26. Tang L, Wang L, Tan X, Xu W.2011. Adenovirus serotype 7 associated

on November 7, 2019 by guest

http://jvi.asm.org/

with a severe lower respiratory tract disease outbreak in infants in Shaanxi Province, China. Virol. J.8:23.

27. Tian X, et al. 2011. Construction and characterization of human adenovirus serotype 3 packaged by serotype 7 hexon. Virus Res.160: 214 –220.

28. Wang Z, Mo C, Kemble G, Duke G.2004. Development of an efficient fluorescence-based microneutralization assay using recombinant human cytomegalovirus strains expressing green fluorescent protein. J. Virol. Methods120:207–215.

29. Reference deleted.

30. Yuan X, et al.2009. Molecular modeling and epitopes mapping of human adenovirus type 3 hexon protein. Vaccine27:5103–5110.

31. Zhang Q, et al.2006. Comparative genomic analysis of two strains of human adenovirus type 3 isolated from children with acute respiratory infection in southern China. J. Gen. Virol.87:1531–1541.

32. Zhang Q, et al.2009. Construction and characterization of a replication-competent human adenovirus type 3-based vector as a live-vaccine can-didate and a viral delivery vector. Vaccine27:1145–1153.

on November 7, 2019 by guest

http://jvi.asm.org/