Journal of Medical Virology 78:379–383 (2006)

Genetic Heterogeneity of the Hexon Gene ofAdenovirus Type 3 Over a 9-year Period in Korea

Eun Hwa Choi,1,2 Hee Sup Kim,3 Ki Ho Park,4 and Hoan Jong Lee1,4*1Department of Pediatrics, Seoul National University College of Medicine, Seoul, Korea2Medical Science Research Institute, Seoul National University Bundang Hospital, Seongnam, Korea3Department of Pediatrics, Dongkuk University School of Medicine, Gyeongju, Korea4Virus Research Center, Clinical Research Institute, Seoul National University Hospital, Seoul, Korea

Hexon sequences were analyzed in 29 epidemio-logically unrelated adenovirus type 3 (Ad3) iso-lates from the 7 genome types to understand themolecular basis of the genome-type diversity ofAd3 associated with childhood pneumonia inKorea during the period 1991–1999. Nine nucleo-tide substitutions were observed among the29 Ad3 strains. Five of the 9 involved amino acidchanges in loops 1 (Gly to Val at codon 205 andThr to Ile at 211) and loop 2 (His to Asn at 417, ThrtoAla at 429, andAla toAsp at 439). The predictedhydropathic characteristics of this region havebeen affected by these amino acid changes. Theregion surrounding codons from 417 to 439 ofAd3a16 and Ad3a18 manifested greater hydro-phobicity than the region of other genome types(Ad3a, Ad3a13, Ad3a14, Ad3a15, and Ad3a17). Inparticular, three amino acid changes in loop 2were associated with two new genome types,namely, Ad3a16 and Ad3a18, which were recog-nized during later epidemics in 1998–1999.Phylogenetic relatedness revealed that thesetwogenome types clustered intodistinct lineagesin the phylogenetic tree. This result suggests thatthe genetic heterogeneity of Ad3 hexon couldplay a potential role in the appearance of newgenome types and that it could affect theantigenic characteristics of Ad3. J. Med. Virol.78:379–383, 2006. � 2006 Wiley-Liss, Inc.

KEY WORDS: human adenovirus serotype 3;hexon; genetic diversity; mole-cular epidemiology

INTRODUCTION

Human adenovirus is one of the major etiologic agentsof lower respiratory tract infections in childhood [Yunet al., 1995; Horwitz, 2001]. And, of its 51 knownserotypes, Ad3 and Ad7 are often associated with severepneumonia [Kajon et al., 1990; Li et al., 1996; Hong et al.,2001; Ryan et al., 2002]. Nationwide outbreaks of severe

pneumonia caused by adenovirus type 3 (Ad3) and Ad7were reported in Korea during the period 1995–1999[Hong et al., 2001; Kim et al., 2003].

According to the database at a children’s hospital inKorea, a total of 270 adenoviruses was identified fromchildhood pneumonia over 10 years, 1990–2000. Ofthese 270 strains, Ad3 was one of the commonly isolatedserotypes. However, relative frequency of Ad3 isolatedfrom the children with lower respiratory tract infectionsfluctuates by year with a range from the nadir toreaching more than half of the total isolates annually.Currently, Ad3 is the predominant serotype associatedwith adenoviral lower respiratory tract infections inKorea. This observation suggests that change of Ad3epidemiology may be caused either by the level ofpopulation immunity or by viral diversity of Ad3.

A previous study on the genome type distributions ofAd3 and Ad7 revealed that Ad3 strains show morediverse genome types than Ad7 [Kim et al., 2003].Multiple genome types of Ad3 strains co-circulated andsome of these were isolated persistently throughout theperiod, 1991–1999, while others only lasted for a briefperiod or appeared for the first time later on. However, itis unknown whether some unique feature in the genometype distribution could play a crucial role in theevolution and molecular epidemiology of Ad3 strains.So far, studies on the genetic heterogeneity of Ad3 havebeen limited to an analysis of genome types by restric-tion pattern analysis [Li and Wadell, 1988; Fu et al.,1989; Golovina et al., 1991].

Given this background, it was hypothesized that thegenetic heterogeneity of the hexon gene might contribute

Grant sponsor: Korean Ministry of Health and Welfare; Grantnumber: HMP-99-M-04-002.

*Correspondence to: Hoan Jong Lee, MD, Professor of Pedia-trics, Seoul National University College of Medicine, 28 Yeongeon-dong Jongno-gu, Seoul 110-744, Korea.E-mail: hoanlee@snu.ac.kr

Accepted 28 November 2005

DOI 10.1002/jmv.20550

Published online in Wiley InterScience(www.interscience.wiley.com)

� 2006 WILEY-LISS, INC.

to the molecular diversity of adenoviruses. The hexongene sequence contains seven serotype-specific hyper-variable regions [Crawford-Miksza and Schnurr, 1996],and it was suggested that mutations of hexon might playan importantrole innew outbreaks ofadenoviral infection[Wadell et al., 1985; Larranaga et al., 2000]. This studywas undertaken to identify variations in the hexon geneamong the different genome types of Ad3 strains isolatedin Korea over a 9-year period, with the aim of under-standing the influence of hexon on the prevalence of Ad3genome types during this period.

MATERIALS AND METHODS

Virus Isolates

Fifty-six strains of Ad3 have been isolated fromKorean children with lower respiratory tract infectionsduring 1991–1999. The distributions of Ad3 genometypes among the 56 strains were 4 strains of Ad3a, 13 ofAd3a13, 6 of Ad3a14, 5 of Ad3a15, 17 of Ad3a16, 8 ofAd3a17, and 3 of Ad3a18. Twenty-nine of the 56 strainsof Ad3 were used for this genetic study of the hexon gene.Representative viral strains of diverse genome typeswere selected from various places and at different timesto represent epidemiologically unrelated strains. Full-length adenoviral DNA was purified from infectedHEp-2 cell lysates by using a modified Hirt procedure,as described previously [Shinagawa et al., 1983].Genome types were assigned by restriction fragmentanalysis using 12 enzymes as described in a previousstudy [Kim et al., 2003].

Sequencing of the Hexon Gene

Viral genomic DNA (1� 2 mg) was used as a directtemplate for sequencing. The sequencing primers weredesigned using reference sequences (GenBank acces-sion No. X76549, Ad3p, GB strain). Sequencing analysiswas carried out using Big Dye version 3 (AppliedBiosystems, Inc., Foster City, CA) and the followingconditions: 968C for 2 min followed by 25 cycles of 968Cfor 10 sec, 508C for 10 sec, and 608C for 4 min. Sequenc-ing reactions were analyzed on a ABI 3700 platformusing Sequencher software version 4.0.5 (Gene CodesCorporation, Ann Arbor, MI). Nucleotide sequences

were confirmed by duplicate reactions using the primersshown in Table I. Both nucleotide and predicted aminoacid sequences were aligned using the Clustal W 1.4method (http://www.ebi.ac.uk/clustalw/).

GenBank accession numbers for representativestrains of each genome type are as follows: Ad3a_92393,AY854179; Ad3a_98500, AY854180; Ad3a13_91353,AY854173; Ad3a14_92159, AY854174; Ad3a15_92165,AY854175; Ad3a16_98494, AY854176; Ad3a17_98640,AY854177; and Ad3a18_99453, AY854178.

Phylogenetic Analysis

MEGA version 2.1 [Kumar et al., 2001] was used toconstruct phylogenetic trees based on evolutionarydistances between sequences across the entire hexongene (2,835 bp). The trees were rooted using the X76549sequence. Kimura two-parameters were used for thedistance method using the neighbor-joining algorithm.A thousand additional bootstrap analyses were carriedout on each phylogenetic tree.

RESULTS

Annual Distributions of Ad3 Genome Types

Seven genome types (Ad3a, Ad3a13 through Ad3a18)had been identified previously from restriction fragmentpatterns using 12 enzymes. Annual distributions ofgenome types during the period 1990–1999 were foundto vary markedly by year. During the study period, twoepidemics of Ad3 pneumonia were identified. Ad3 wasprevalent at low activity and constituted 23% of alladenoviral isolates from 1991 to 1994. No Ad3 strainswere isolated during the 3 years 1995–1997. Beginningfrom September 1998, large outbreak of Ad3 associatedchildhood pneumonia occurred and lasted for 1 year,accounting for 54% of all adenoviral isolates.

Several changes in the temporal distribution patternsof the Ad3 strains were noted when strains wereanalyzed in the context of Ad3 genome type. Genometypes Ad3a, Ad3a13, and Ad3a14 were isolated consis-tently during the 9-year observation period. In contrast,new genome types Ad3a16, Ad3a17, and Ad3a18 thatwere not identified previously during 1991–1994, pre-dominated during the Ad3 outbreak in 1998–1999.

J. Med. Virol. DOI 10.1002/jmv

TABLE I. Sequences of Primers used for Adenovirus Type 3 Hexon Sequencing Analysis

Primers Sense Sequences (50 !30) Position

hx1 Sense ATG GCC ACC CCA TCG ATG AT 1–20hx2 Antisense GAA TGT GCT GGC CAT GTC AA 281–300hx3 Sense GCG ACA ACA GAG TGC TTG AC 266–285hx4 Sense CAG CCA GAG CCT CAA GTT GG 568–587hx5 Sense GTT ATG ATC CCG ATG TTC GC 1157–1176hx6 Antisense CAA GGG AAC TTT GCA GGA CC 1749–1768hx7 Sense GTG AAC ATG GTC CTG CAA AG 1741–1760hx8 Sense ACC AAG GAG ACT CCA TCT CT 2029–2048hx9 Antisense ACA TGC GAT CCT TGT ATC CC 2331–2350hx10 Sense ACA TCC CTG AGG GAT ACA AG 2321–2340hx11 Sense CCT GTG TGA CAG GAC CAT GT 2568–2587

The numbering of nucleotides is based on that of Ad3p (strain GB, X76549).The underlined sequence of primer hx5 mismatches that of the reference sequence.

380 Choi et al.

Ad3a15, which was one of the most common genometypes in 1992–1993, was not observed since 1994.

Identification of Hexon Sequence Variations

The complete sequence of the hexon gene (2,835 bp)was determined, and comparisons of the nucleotide andamino acid sequences of hexon are shown in Table II.Thirty-seven common nucleotide substitutions werefound among the 29 Korean Ad3 strains versus thepreviously reported sequence (GenBank X76549). Asshown in Table II, nine substitutions were observedwithin Korean Ad3 strains.

Identical hexon gene sequences were observed fordifferent genome types, for example, Ad3a, Ad3a14,and Ad3a17, and no sequence differences were foundbetween Ad3a13 and Ad3a15, or between Ad3a16 andAd3a18. However, one Ad3a strain showed a singleamino acid change of Thr to Ile at position 254.

Within the hexon gene, no restriction site was alteredby four enzymes (SmaI, BglI, BglII, and XbaI), whichcould discriminate genome types of Ad3.

Five out of the nine nucleotide substitutions in Ad3strains resulted in amino acid changes, that is, two inthe loop 1 and three in loop 2 (Table II). Three amino acidchanges in loop 2 were observed only in strains of theAd3a16 and Ad3a18 genome types, which appeared inlate 1998.

The hydrophilicity plots (Kyte–Doolittle method) ofhexon loops 1 and 2 were compared between the aminoacids of the seven-genome types of Ad3. The hydro-philicity plots have predicted the change of hydropho-bicity of the strains with substitutions of Val at codon205 and of Ile at codon 254 in loop 1 compared to otherstrains without two substitutions. Three amino acidsubstitutions in loop 2 (codons 417, 429, and 439)affected the predicted hydrophilicity in this region.The region surrounding codons from 417 to 439 ofAd3a16 and Ad3a18 manifested greater hydrophobicitythan the region of other genome types (Ad3a, Ad3a13,Ad3a14, Ad3a15, and Ad3a17).

Phylogenetic Relationship

Phylogenetic trees were inferred based on wholehexon sequences (Fig. 1), in order to understand theevolutionary relationships between the Ad3 strains. Thephylogenetic trees revealed that Ad3 isolates wereclustered into two different lineages. Ad3a16 andAd3a18 formed one cluster. Other genome types, Ad3a,Ad3a13, Ad3a14, Ad3a15, and Ad3a17, were included inanother lineage which was divided subsequently intotwo subclusters (Ad3a, Ad3a14, and Ad3a17; Ad3a13and Ad3a15).

DISCUSSION

In this study, five amino acid changes were identifiedwhich were confined to loops 1 and 2 of the hexon gene ofAd3 isolated from Korean children with lower respira-tory tract infections over a 9-year period from 1991 to

J. Med. Virol. DOI 10.1002/jmv

TA

BL

EII

.C

omp

ari

son

sof

the

En

tire

Hex

onG

ene

Seq

uen

ces

ofA

den

ovir

us

Typ

e3

Str

ain

sIs

olate

dF

rom

Ch

ild

ren

Wit

hA

cute

Low

erR

esp

irato

ryT

ract

Infe

ctio

ns

Gen

ome

typ

es(n

o.of

stra

ins

sequ

ence

d)

Pos

itio

ns

ofn

ucl

eoti

de

(am

ino

aci

d)

614

(205)

633

(211)

761

(254)

1249

(417)

1285

(429)

1316

(439)

1731

(577)

2061

(687)

2067

(689)

Ad

3p

aG

(Gly

)T

(Gly

)C

(Th

r)C

(His

)A

(Th

r)C

(Ala

)C

(Ph

e)C

(Ph

e)C

(Pro

)A

d3a

(3)

T(V

al)

T(G

ly)

C/T

b(T

hr/

Ile)

A(A

sn)

A(T

hr)

C(A

la)

T(P

he)

T(P

he)

T(P

ro)

Ad

3a13

(7)

G(G

ly)

C(G

ly)

C(T

hr)

A(A

sn)

A(T

hr)

C(A

la)

C(P

he)

C(P

he)

C(P

ro)

Ad

3a14

(5)

T(V

al)

T(G

ly)

C(T

hr)

A(A

sn)

A(T

hr)

C(A

la)

T(P

he)

T(P

he)

T(P

ro)

Ad

3a15

(4)

G(G

ly)

C(G

ly)

C(T

hr)

A(A

sn)

A(T

hr)

C(A

la)

C(P

he)

C(P

he)

C(P

ro)

Ad

3a16

(5)

G(G

ly)

T(G

ly)

C(T

hr)

C(H

is)

G(A

la)

A(A

sp)

C(P

he)

C(P

he)

C(P

ro)

Ad

3a17

(3)

T(V

al)

T(G

ly)

C(T

hr)

A(A

sn)

A(T

hr)

C(A

la)

T(P

he)

T(P

he)

T(P

ro)

Ad

3a18

(2)

G(G

ly)

T(G

ly)

C(T

hr)

C(H

is)

G(A

la)

A(A

sp)

C(P

he)

C(P

he)

C(P

ro)

Th

ep

osit

ion

sof

the

am

ino

aci

dch

an

ges

are

un

der

lin

ed.

aT

he

pro

toty

pe

Ad

3(s

train

GB

)w

as

sequ

ence

d.S

equ

ence

sare

iden

tica

las

Gen

Ban

kX

76549.

bO

ne

(str

ain

98-5

00)

ofth

eth

ree

stra

ins

ofA

d3a

gen

ome

typ

esh

owed

ach

an

ge

ofT

hr

toIl

eat

pos

itio

n254.

Hexon Diversity of Adenovirus Type 3 381

1999. Interestingly, three amino acid changes in loop 2 ofhexon appear to be significantly related to the epide-miology of Ad3 in Korea.

Multiple Ad3 genome types were found to have co-circulated over this 9-year period, and several wereidentified throughout the study period, whereas otherswere restricted to sporadic isolates. In particular, theAd3a16 and Ad3a18 genome types, which involve threeamino acid changes in loop 2, were not isolated before1998 but became the major genome types during the Ad3outbreak in 1998–1999. Phylogenetic analysis revealedthat two genome types associated with three amino acidchanges in loop 2 clustered into a distinct lineage in thephylogenetic tree.

Comparisons of the sequences of Ad3 and Ad7 strains,which caused successive epidemics from 1995 to 1999 inKorea, demonstrated remarkably different character-istics in hexon variations. The higher genetic variabilityof Ad3, as compared with Ad7, is attributable to hexonsequence variations. In a recent study, we found thatAd7 had only two genome types, Ad7d and Ad7l, and novariations were found in the hexon sequences [Choiet al., 2005]. However, nine nucleotide variations werefound in the hexon sequences among Korean Ad3 strainsin the present study.

It is unclear whether nucleotide changes influence theantigenic characteristics or pathogenicity of Ad3. Fiveamino acid substitutions of nine nucleotide changes alloccurred in the loop 1 and loop 2 regions which subject tochange the predicted hydrophobicity of this region.Hexon loop 1 and loop 2 are known to form the mainneutralization determinant ‘‘epsilon’’ [Shenk, 2001].This finding suggests a possible association betweenamino acid changes and antigenic characteristics, andthat hexon heterogeneity has a key role in theemergence of novel Ad3 genome types.

The present study is the first to demonstratethe diversity of the hexon gene in Ad3 by sequence vari-ation. Although several studies have described thegenetic diversity of Ad3 associated with conjunctivitisor respiratory tract infections by restriction fragmentanalysis, most studies have focused on the distribution

patterns of the different genome types of Ad3 [Adrianet al., 1986, 1989; O’Donnell et al., 1986; Fu et al., 1989;Itakura et al., 1990; Kajon et al., 1990; Golovina et al.,1991; Mizuta et al., 1994; Li et al., 1996; Itoh et al., 1999].Little is known about the sequence variations of hexonwith respect to genome type within an epidemic orbetween epidemics.

Several genome types of Ad3 co-circulate simulta-neously [Golovina et al., 1991], whereas a single genometype can cause an epidemic in Glasgow, Scotland[O’Donnell et al., 1986]. Mizuta et al. [1994] isolatedAd3 from children with acute respiratory tract infec-tions from 1986 to 1991 in Yamagata, Japan, and foundthat outbreaks of Ad3 were due to endemic genome typesrather than new types. In China, Ad3a2 was found to bethe most prevalent genome type in children withpneumonia over the 27-year period 1962–1988, exceptin 1984 when Ad3a4 predominated [Li et al., 1996]. Incontrast, Itakura et al. [1990] observed an apparent shiftfrom a predominant Ad3g to a new Ad3f in patients withconjunctivitis during 1983–1986 in Japan. More inter-estingly, the predominant serotype Ad3 in Japaneseconjunctivitis patients in 1990 was replaced by Ad4 in1992, and during this transition period, the emergenceof new Ad3 genome types were identified [Itoh et al.,1999].

In this study, two new genome types, Ad3a16 andAd3a18, were found to have replaced the previouslypredominating types during the Ad3 outbreak in Koreain 1998–1999. This finding suggests that viral, environ-mental, or host factors might promote this change.Alternatively, these new Ad3 genome types may havebeen introduced from outside Korea. New genome typesseem to be closely related to prior genome types bypairwise comigrating restriction fragment analysis[Kim et al., 2003], which suggests they were derivedfrom previous genome types, rather than having beenintroduced from other countries. In addition, these twogenome types have not been described previously.

The prevalence of adenoviral serotypes was found tohave changed markedly throughout the study period. Aprevious major serotype, Ad3, was completely replacedby Ad7 during the outbreak in 1995–1998, and 3 years ofAd7 epidemic were finally terminated by its exclusivereplacement with serotype 3. Ad3 then increaseddramatically and caused an outbreak that involvedthree new Ad3 genome types in 1998–1999. Althoughthe molecular basis of the observed genetic hetero-geneity in Ad3 is not understood, it is our belief thathexon diversity in loop 2 could have contributed to theemergence of the new genome types, Ad3a16 andAd3a18, and an Ad3-associated epidemic of childhoodpneumonia. However, it is possible that genetic changesin other regions of the genome are more important in theepidemiological aspect. This issue has to be addressed infuture studies.

The influence of Ad7 on successive Ad3 outbreaks bydifferent genome types remains unknown. Genome typechanges of Ad3 may be influenced by Ad7 throughinteraction between the two serotypes, or alternatively,

J. Med. Virol. DOI 10.1002/jmv

Fig. 1. Phylogenetic analyses of the entire hexon (2,835 bp)sequences of each genome type in adenovirus type 3 (Ad3). Phylogenetictrees were constructed using the neighbor-joining algorithm. Branchlengths are proportional to the number of nucleotide substitutes andbootstrap probabilities �60 are shown at each adjacent node. Thesequence of Ad3 (X76549) was defined as an outgroup.

382 Choi et al.

Ad3 may have acquired adaptive strategies during anepidemiologically quiescent period. The possibility thata specific serotype contributes to the appearance of newgenome types of other serotypes should be elucidated byfuture studies.

REFERENCES

Adrian T, Wadell G, Hierholzer JC, Wigand R. 1986. DNA restrictionanalysis of adenovirus prototypes 1 to 41. Arch Virol 91:277–290.

Adrian T, Best B, Hierholzer JC, Wigand R. 1989. Molecularepidemiology and restriction site mapping of adenovirus type3 genome types. J Clin Microbiol 27:1329–1334.

Choi EH, Kim HS, Eun BW, Kim BI, Choi JY, Lee HJ, Inada T. 2005.Adenovirus type 7 peptide diversity during outbreak, Korea, 1995–2000. Emerg Infect Dis 11:649–654.

Crawford-Miksza L, Schnurr DP. 1996. Analysis of 15 adenovirushexon proteins reveals the location and structure of sevenhypervariable regions containing serotype-specific residues. J Virol70:1836–1844.

Fu WY, Liang D, Zheng YC, Liu WM, Xu Z, Guo HJ, Wang ZL. 1989. Astudy of molecular epidemiology of adenovirus of types 3 and 7 oninfant pneumonia in northern China. Chin Med J (Engl) 102:857–861.

Golovina GI, Zolotaryov FN, Yurlova TI. 1991. Sensitive analysis ofgenetic heterogeneity of adenovirus types 3 and 7 in the SovietUnion. J Clin Microbiol 29:2313–2321.

Hong JY, Lee HJ, Piedra PA, Choi EH, Park KH, Koh YY, Kim WS.2001. Lower respiratory tract infections due to adenovirus inhospitalized Korean children: Epidemiology, clinical features, andprognosis. Clin Infect Dis 32:1423–1429.

Horwitz MS. 2001. Adenoviruses. In: Knipe DM, Howley PM, editors.Fields virology, 4th edition. Vol. 2, Philadelphia: Lippincott.Williams & Wilkins. pp 2301–2326.

Itakura S, Aoki K, Sawada H, Shinagawa M. 1990. Analysis withrestriction endonucleases recognizing 4- or 5-base-pair sequences ofhuman adenovirus type 3 isolated from ocular diseases in Sapporo,Japan. J Clin Microbiol 28:2365–2369.

Itoh N, Tanaka K, Aoki K, Hinokuma R, Nakagawa H, Takeuchi S,Uchio E, Shiao S, Ohno S. 1999. Four new genotypes of adenovirustype 3 isolated from patients with conjunctivitis in Japan. J MedVirol 59:73–77.

Kajon AE, Murtagh P, Garcia Franco S, Freire MC, Weissenbacher MC,Zorzopulos J. 1990. A new genome type of adenovirus 3 associatedwith severe lower acute respiratory infection in children. J MedVirol 30:73–76.

Kim YJ, Hong JY, Lee HJ, Shin SH, Kim YK, Inada T, Hashido M,Piedra PA. 2003. Genome type analysis of adenovirus types 3 and 7isolated during successive outbreaks of lower respiratory tractinfections in children. J Clin Microbiol 41:4594–4599.

Kumar S, Tamura K, Jakobsen IB, Nei M. 2001. MEGA2: Molecularevolutionary genetics analysis software. Bioinformatics 17:1244–1245.

Larranaga C, Kajon A, Villagra E, Avendano LF. 2000. Adenovirussurveillance on children hospitalized for acute lower respiratoryinfections in Chile (1988–1996). J Med Virol 60:342–346.

Li QG, Wadell G. 1988. Comparison of 17 genome types of adenovirustype 3 identified among strains recovered from six continents. J ClinMicrobiol 26:1009–1015.

Li QG, Zheng QJ, Liu YH, Wadell G. 1996. Molecular epidemiology ofadenovirus types 3 and 7 isolated from children with pneumonia inBeijing. J Med Virol 49:170–177.

Mizuta K, Suzuki H, Ina Y, Yazaki N, Sakamoto M, Katsushima N,Numazaki Y. 1994. Six-year longitudinal analysis of adenovirustype 3 genome types isolated in Yamagata, Japan. J Med Virol42:198–202.

O’Donnell B, Bell E, Payne SB, Mautner V, Desselberger U. 1986.Genome analysis of species 3 adenoviruses isolated during summeroutbreaks of conjunctivitis and pharyngoconjunctival fever in theGlasgow and London areas in 1981. J Med Virol 18:213–227.

Ryan MA, Gray GC, Smith B, McKeehan JA, Hawksworth AW, MalasigMD. 2002. Large epidemic of respiratory illness due to adenovirustypes 7 and 3 in healthy young adults. Clin Infect Dis 34:577–582.

Shenk T. 2001. Adenoviridae: The viruses and their replication. In:Knipe DM, Howley PM, editors. Fields virology. 4th edition. Vol. 2.Philadelphia: Lippincott. Williams & Wilkins. pp 2265–2300.

Shinagawa M, Matsuda A, Ishiyama T, Goto H, Sato G. 1983. A rapidand simple method for preparation of adenovirus DNA frominfected cells. Microbiol Immunol 27:817–822.

Wadell G, Cooney MK, da Costa Linhares A, de Silva L, Kennett ML,Kono R, Gui-Fang R, Lindman K, Nascimento JP, Schoub BD. 1985.Molecular epidemiology of adenoviruses: Global distribution ofadenovirus 7 genome types. J Clin Microbiol 21:403–408.

Yun BY, Kim MR, Park JY, Choi EH, Lee HJ, Yun CK. 1995. Viraletiology and epidemiology of acute lower respiratory tract infectionsin Korean children. Pediatr Infect Dis J 14:1054–1059.

J. Med. Virol. DOI 10.1002/jmv

Hexon Diversity of Adenovirus Type 3 383