Molecular identification of adenovirus sequences: a rapid scheme for early typing of human adenoviruses in diagnostic samples of immunocompetent and immunodeficient patients.

Journal of Medical Virology 78:1210–1217 (2006)
Molecular Identification of Adenovirus Sequences:
A Rapid Scheme for Early Typing of Human
Adenoviruses in Diagnostic Samples of
Immunocompetent and Immunodeficient Patients
Ijad Madisch,1 Roman Wo¨lfel,2 Gabi Harste,1 Heidi Pommer,1 and Albert Heim1*
1Institut fu¨r Virologie, Medizinische Hochschule Hannover, Hannover Germany
2Bundeswehr Institute of Microbiology, Munich, Germany
Precise typing of human adenoviruses (HAdV)
is fundamental for epidemiology and the detec-
tion of infection chains. As only few of the
51 adenovirus types are associated with life-
threatening disseminated diseases in immuno-
deficient patients, detection of one of these
types may have prognostic value and lead to
immediate therapeutic intervention. A recently
published molecular typing scheme consisting
of two steps (sequencing of a generic PCR
product closely adjacent to loop 1 of the main
neutralization determinant e, and for species
HAdV-B, -C, and -D the sequencing of loop 2
[Madisch et al., 2005]) was applied to 119 clinical
samples. HAdV DNA was typed unequivocally
even in cases of culture negative samples, for
example in immunodeficient patients before
HAdV causes high virus loads and disseminated
disease. Direct typing results demonstrated
the predominance of HAdV-1, -2, -5, and -31 in
immunodeficient patients suggesting the sig-
nificance of the persistence of these viruses for
the pathogenesis of disseminated disease. In
contrast, HAdV-3 predominated in immuno-
competent patients and cocirculation of four
subtypes was demonstrated. Typing of samples
from a conjunctivitis outbreak in multiple mili-
tary barracks demonstrated various HAdV types
(2, 4, 8, 19) and not the suspected unique
adenovirus etiology. This suggests that our
molecular typing scheme will be also useful for
epidemiological investigations. In conclusion,
our two-step molecular typing system will
permit the precise and rapid typing of clinical
HAdV isolates and even of HAdV DNA in clinical
samples without the need of time-consuming
virus isolation prior to typing. J. Med. Virol.
78:1210–1217, 2006.
� 2006 Wiley-Liss, Inc.
KEY WORDS: human adenovirus; molecular
typing; immunodeficiency;
persistence; keratoconjuncti-
vits outbreak
INTRODUCTION
Adenoviridae are non-enveloped, double-stranded DNA
viruses with icosahedrals capsids [Swenson et al., 2003].
Human adenoviruses (HAdV) are classified into six
species (HAdV-A to HAdV-F) with 51 types, and have
long been recognized as pathogens causing a broad
spectrum of diseases. Organotropism and virulence are
clearly type related. The types HAdV-A31, -F40, and
-F41 cause severe gastroenteritis and diarrhea in
infants and young children. However, HAdV-A31 also
causes a frequently fatal disseminated disease in
immunosuppressed children [Venard et al., 2000;
Seidemann et al., 2004; Kampmann et al., 2005]. Lower
respiratory tract illness that can present as acute
respiratory distress syndrome is primarily associated
with HAdV-B3, -B7, -B21, and -E4 infection [Bhat et al.,
1984; Crawford-Miksza et al., 1999; Erdman et al., 2002;
Chuang et al., 2003]. Less severe upper respiratory
illness is frequently caused by HAdV-B3, -B7, -C1, -C2,
and -C5 [Dakhama et al., 1999; Erdman et al., 2002].
Thus, typing of HAdV isolated from a respiratory sample
or feces sample may predict the clinical course. Infec-
tions of the eye with HAdV-D8, -D19, and -D37 may
cause severe epidemic keratoconjunctivitis, whereas
*Correspondence to: Dr. Albert Heim, Institut fr Virologie,
Medizinische Hochschule Hannover, Carl-Neuberg-Str. 1,
D-30625 Hannover, Germany.
E-mail: Heim.Albert@mh-hannover.de
Accepted 16 May 2006
DOI 10.1002/jmv.20683
Published online in Wiley InterScience
(www.interscience.wiley.com)
� 2006 WILEY-LISS, INC.

mild follicular conjunctivitis and pharyngoconjuntival
fever are most frequently caused by HAdV- B3, -B7, and
-E4 [Mellman-Rubin et al., 1995; Aoki and Tagawa,
2002]. Disseminated disease in highly immunosup-
pressed patients, for example allogenic stem cell
transplant recipients, is associated with HAdV-A31,
-C1, -C2, and -C5, most frequently with the latter three
types [Venard et al., 2000; Lion et al., 2003; Seidemann
et al., 2004; Kampmann et al., 2005]. Acute hemorrhagic
cystitis and acute renal failure in infants, young
children, and immunosuppressed patients are primarily
associated with the serotypes HAdV-B11, -B21, -B34,
and -B35 [Stalder et al., 1977; Carrigan, 1997;
Echavarria et al., 1999; Akiyama et al., 2001; Friedrichs
et al., 2003; Mori et al., 2003]. Due to the wide spectrum
of diseases that are caused by HAdV of different
organotropism and virulence, rapid and precise typing
can be important for the prediction of the clinical course
and in determining the importance of therapeutic
intervention. For example, the early identification of a
HAdV type related to disseminated disease (e.g., HAdV-
A31) in an immunosuppressed patient may require an
immediate reduction of immunosuppressive therapy
and the instigation of experimental antiviral therapy.
Although treatment modalities remain limited, there
have been encouraging reports on cidofovir and new
antiviral agents [Ljungman, 2004; Kampmann et al.,
2005; Naesens et al., 2005; Stock et al., 2006]. HAdV
typing is also essential for epidemiological studies and
quality control of diagnostics.
Conventional diagnostic methods, such as isolation of
the virus in cell culture, direct detection by immunos-
pecific methods, or visualization by electron microscopy,
are lacking in rapidity or sensitivity compared to HAdV
DNA detection by PCR. Therefore, diagnosis of HAdV
infection is nowadays usually made by using generic
PCR protocols [Allard et al., 2001; Heim et al., 2003].
Moreover, a rapid and simple diagnosis of disseminated
adenovirus disease in immunosuppressed patients
can be achieved by monitoring adenovirus load in
peripheral blood by quantitative HAdV PCR [Heim
et al., 2003; Lion et al., 2003; Seidemann et al., 2004].
Virus isolation, which is needed for classical typing
methods, usually requires weeks to obtain a result
and may fail in cases with low virus loads, from
which virus can still be detected by sensitive PCR. As
fatal disseminated disease usually starts with low virus
loads, molecular typing of HAdV DNA positive clinical
samples with failing or not yet completed virus isolation
would be a significant advance.
Therefore, a rapid two-step molecular PCR-based
typing system was developed and extensively validated
with prototype viruses and intermediate strains
[Madisch et al., 2005]. The first step consists of generic
hexon PCR protocols frequently used for diagnostic
purposes, either conventional [Allard et al., 2001] or real
time [Heim et al., 2003]. Sequencing of these amplicons
permits reliable species identification and type identi-
fication of six HAdV types [Madisch et al., 2005]. Typing
of the remaining 45 HAdV types is achieved by PCR
amplification and sequencing the highly variable hexon
loop-2 (L2) of the neutralization determinant e. The
advantage of L2 compared to loop-1 (L1) sequence-based
typing approaches is that L2 has a smaller amplicons
size, resulting in higher PCR sensitivity and easier
sequencing. Typing results of all 51 HAdV prototypes
achieved by L2 sequencing were identical to neutraliza-
tion tests as this sequence also encodes a part of the
neutralization epitope [Madisch et al., 2005].
In the present study, HAdV was typed unequivocally
in 119 clinical samples, in most of these (98) HAdV DNA
was typed directly without virus isolation in cell
cultures. This is the first report of molecular typing
of a broad spectrum of different clinical specimens
from both immunosuppressed and immunocompetent
patients by hexon sequencing.
MATERIALS AND METHODS
Clinical Samples
The molecular HAdV typing system was applied to
119 HAdV positive diagnostic samples originating from
multiple hospitals in Germany (Table I). These samples
originated from 99 patients who had presented
with symptoms suggestive of adenovirus infection
(e.g., keratoconjunctivitis, pharyngoconjunctival fever,
pneumonia, gastroenteritis, cystitis, or septic symptoms
suggestive of disseminated disease in immunosup-
pressed patients). Only for immunodeficient patients
with a suspicion of disseminated disease were multiple
samples analysed, and these samples originated from
different body sites. Eye swabs from an outbreak of
conjunctivitis in different locations of the German
armed forces were also included. DNA was extracted
from all clinical samples with the Qiagen Blood Mini kit
(Hilden, Germany) and concentration of HAdV DNA
was quantified using a real time PCR protocol [Heim
et al., 2003]. Samples with a HAdV DNA concentration
>2�103 copies/ml were included for direct typing
without virus isolation on cell cultures. In 21 of the
119 diagnostic samples virus isolation on A549 cells was
successful before molecular typing was started. There-
fore, molecular typing was performed with the cell
culture isolate of these 21 diagnostic samples. In case of
blood samples virus isolation was not carried out
routinely. Medical School of Hannover guidelines on
ethics in health research were followed. Individual
consent was not required.
PCR Amplification of Region Adjacent
to Hexon L1
HAdV DNA was amplified for typing by a conventional
generic PCR (only first round of a nested PCR protocol)
[Allard et al., 2001] or a real time PCR protocol [Heim
et al., 2003].
PCR Amplification of Highly Variable Hexon L2
In case of species B, C, and D, PCR amplification of L2
was performed as recently described [Madisch et al.,
2005]. The amplicon size of species HAdV-B amplicons is
J. Med. Virol. DOI 10.1002/jmv
Molecular Typing Scheme of Human Adenoviruses 1211

590 bp and of species HAdV-C and -D amplicons 322 bp.
The sensitivity of the L2 PCR protocols was determined
by testing serial dilutions of cell culture derived HAdV-
B3, -C2, and -D8 previously that had been quantified by
real time PCR [Heim et al., 2003].
PCR Amplification of the
Fiber (g Determinant)
For g determinant PCR amplification and subsequent
sequencing primer sets and PCR conditions were used as
previously described [Madisch et al., 2005].
Sequencing
Both strands were cycle sequenced with rhodamine-
labeled dideoxynucleotide chain terminators (Applied
Biosystems (ABI), Warrington, England) and the same
primers as used for PCR. Analysis was performed on an
ABI Prism 310 automatic sequencer.
Direct Molecular Typing Scheme
The molecular typing scheme is comprised of two steps
(Fig. 1). The first step consists of a generic PCR (either
conventional or real time) amplifying a conserved region
J. Med. Virol. DOI 10.1002/jmv
Fig. 1. Flow chart of the two step molecular typing scheme. Percent values indicate nucleotide divergence compared to prototype databank
sequences if a conventional generic PCR is used as the first step [Allard et al., 2001]. In this case species HAdV-C viruses can be typed by the first
step. However, if real time PCR is used as a first step [Heim et al., 2003], step 2 (L2 sequencing) for species C has to be performed. The criteria for the
real time PCR as a first step are as follows: species identification is achieved with a nucleic acid divergence of<4.8% and type identification in case
of species HAdV-A, -E, and -F with a nucleic acid divergence of �1.2%.
TABLE I. Molecular HAdV Typing of 119 Clinical Samples With Our Two Step Molecular Typing Concept
Specimen Numbers of specimens
HadV- Blood Urine Eye-swabs Feces N-PSW TS & BAL Total (n¼ 119)
From immunosuppressed
patients
A12a) 1 1 1
A18a) 1 1 0
A31a) 2 1 7 1 11 10
B3c) 1 5 2 14 22 3
B11c) 2 2 4 4
B21c) 1 1 0
C1a) 10 9 3 2 24 21
C2a) 10 3 7 3 2 25 19
C5a) 1 2 3 3
C6a) 3 3 3
D8c) 3 3 0
D19c) 3 3 0
D30c) 1 1 1
E4a) 1 1 1 3 2
F40a) 1 1 0
F41a) 13 13 1
The types marked with an a) were identified by applying a generic diagnostic PCR (step 1 of the molecular typing scheme) Allard et al. [2001], or by
L2 species-specific primers BL/BR b),or CDL/CDR c) Madisch et al. [2005].
In 98 of these samples (including all blood and feces samples) molecular typing was performed directly without time-consuming virus isolation.
N-PSW, Nasopharyngeal swabs and washes; TS & BAL, tracheal swabs and broncho-alveolar lavage.
1212 Madisch et al.

at the 50 end of the hexon gene closely adjacent to the L1
coding region [Allard et al., 2001; Heim et al., 2003]. By
sequencing these amplicons, final species identification
can be achieved and type identification of all members of
species A, E, and F. For the conventional PCR protocol
(amplicon size: 301bp) [Allard et al., 2001] a nucleic acid
sequence divergence �9% from the closest databank
prototype sequence was deduced as a criterion for
species definition, whereas in case of the diagnostic real
time PCR protocol [Heim et al., 2003] with its smaller
amplicon size (140 bp), a <4.8% sequence divergence is
required [Madisch et al., 2005]. A �1.5% nucleic acid
sequence divergence to the next homologous prototype
sequence is suggested when making a final typing
decision for members of the species A, E, and F (1.2%
for real time PCR amplicons) [Madisch et al., 2005]. For
viruses of species HAdV-C, typing may be feasible with
the same criteria if the conventional PCR is used as step
one, otherwise the second step (L2 amplification and
sequencing) has to be performed (L2 amplicon size
species HAdV-B 590 bp and for species HAdV-D 322 bp).
Due to the low sequence divergence between viruses of
the species B and D (41 of 51 HAdV serotypes), L2
sequencing as a second step is always required for final
typing. A sequence divergence of<2.5% in the L2 region
compared to the next homologous prototype sequence in
the databank was proposed as a criterion for molecular
identification of a serotype [Madisch et al., 2005].
Sequence Homology Searches
A local database for blast searches containing
sequences of all 51 HAdV prototypes was generated
from GenBank sequence data files and cross checked
with our newly produced prototype sequences [Madisch
et al., 2005]. All sequences were checked for mislabeling
or incorrect annotation [Davison et al., 2003; Madisch
et al., 2005], which both complicate HAdV sequence
homology searches in the GenBank database. This local
database was built with the BioEdit Software Package
(version 6.0.5., Tom Hall, Ibis Therapeutics, Carlsbad,
CA) and can be downloaded from the national adeno-
virus advisory laboratory webpage (http://www99.mh-
hannover.de/institute/virologie/section/adeno/adeno_
en.html). For typing purposes based on our two-step
molecular system, this local database facilitates, expe-
dites and improves the matching process to HAdV
prototypes. In addition, this database was implemented
in an online blast search (http://www.vmri.hu/blast_
d.htm) including only sequences relevant for the two
step molecular typing scheme. Alternatively, the
nucleotide blast (blastn) service provided by NCBI can
be used (http://www.ncbi.nlm.nih.gov/BLAST/).
Phylogenetic Analysis
Phylogenetic analysis was performed by using the
MEGA Software package (version 3.1). The phyloge-
netic trees were constructed with the neighbor-joining
method (Kimura-two parameter matrix) with a transi-
tion/transversion ratio of 2.0. The ratio of non-synon-
ymous to synonymous mutations were calculated with
the help of the Nej–Gojobori method implemented in the
MEGA software.
Hydrophobicity Analysis
Hydrophobicity along the L2 deduced amino acid
sequence was determined by using the BioEdit Package
(version 6.0.5) including the Eisenberg hydrophobicity
moment plot calculation with a window size of five amino
acids [Eisenberg et al., 1984].
Statistical Analysis
Statistical analysis was performed with the SPSS
Software Package (version 13.0).
RESULTS
Sensitivity of PCR Protocols of the
Proposed Molecular Typing System
Sensitivity of PCRs is crucial for typing of clinical
samples containing low concentrations of HAdV DNA.
The sensitivity of the first step (generic PCR closely
adjacent to L1) has been published previously [Allard
et al., 2001; Heim et al., 2003], whereas the sensitivity
of the second step was determined in this article.
Amplicons of HAdV-B were visible on an ethidium
bromide-stained agarose gel with virus concentrations
�103, amplicons of HAdV-C with virus concentrations
�104, and of HAdV-D�103 genome equivalents (copies)/
ml (Fig. 2). Thus, the new designed two step system
can achieve typing results if the virus concentration is
�103 genome equivalents (copies)/ml.
J. Med. Virol. DOI 10.1002/jmv
Fig. 2. Gel electrophoresis analysis of PCR products obtained after L2 amplification of HAdV DNA (ethidium bromide stain). A: HAdV-B3
amplified with primer set BL/BR (amplicon size 590 bp), (B) HAdV-C2 amplified with primer set CDL/CDR, (C) HAdV-D8 amplified with primer set
CDL/CDR (amplicon size 322 bp) [Madisch et al., 2005]. Numbers indicate template concentration (genome equivalents/ml), ‘‘-’’ negative control,
and ‘‘MW’’: molecular weight marker (50-bp ladder).
Molecular Typing Scheme of Human Adenoviruses 1213

Application of the Two Step Molecular
Typing Scheme
Species identification was achieved unequivocally in
all clinical samples by step one. Moreover, typing of all
85 samples containing HAdV sequences of species A, C,
E, F was accomplished by step one using a generic,
conventional PCR protocol (Table I) [Allard et al., 2001].
This PCR amplifies a conserved region at the 50 end of
the hexon gene adjacent to the L1 region coding for the
major part of the neutralization determinant e. There-
fore, recombination events between the amplified region
and the neutralization epitope e, which may cause
misleading typing results, are quite unlikely. For
10 clinical samples containing species HAdV-C viruses
both step one and step two (L2 sequencing) were
performed. L2 sequencing always confirmed step one
typing results. This indicates that typing results based
on step one sequence data are reliable, although these
PCRs do not amplify the neutralization determinant e.
All 34 clinical samples containing HAdV sequences of
species B and D were typed unequivocally by L2
sequencing (Table I). Altogether, definite molecular
typing was achieved successfully for 119 clinical
samples.
Direct Typing of HAdV in Clinical Samples
Without Virus Isolation
Diagnostic samples (119) positive for HAdV were used
for the evaluation of the two-step molecular typing
scheme. Molecular typing was achieved directly in 98
samples either before virus isolation in cell cultures
was positive or failed, probably because of low virus
concentrations, neutralizing antibodies, or insufficient
quality of the diagnostic material. In the remaining 21
samples HAdV was isolated and molecular typing was
performed subsequently. HAdV DNA concentrations in
clinical materials were in the range of 2�103 to
6.5� 1011 copies/ml as determined by quantitative
PCR [Heim et al., 2003]. Typing of HAdV DNA was also
obtained directly from blood samples, with viruses of the
species HAdV-C most prevalent (Table I). All blood
samples originated from immunosuppressed patients.
The species HAdV-C predominated significantly in
immunosuppressed patients (68% positive for species
C) compared to immunocompetent patients (17%,
P<0.001, w2 test), as did HAdV-A31 (P¼ 0.04). In
contrast, infections with HAdV-B3 and the species
HAdV-F were significantly associated with immuno-
competent patients (P< 0.001, w2 test).
Investigation of a Conjunctivitis
Outbreak in a Military Setting
From February to April 2004 an outbreak of con-
junctivitis occurred in different locations of the German
armed forces causing the closure of several military
facilities. A total of 940 eye swab samples were examined
using a real time PCR protocol [Heim et al., 2003] but
only 12 HAdV DNA positive samples could be identified.
Although this result was already contradictory to the
suspected outbreak of HAdV keratoconjunctivits, mole-
cular typing was carried out in order to exclude
inadequate sampling. In two cases HAdV-D8 was
identified and HAdV-D19 DNA was found in another
sample. All three cases were not associated in time or
place of illness although these HAdV types are usually
associated with epidemic keratoconjunctivitis. Unre-
lated to these cases, HAdV-C2 DNA was found in two
samples and HAdV-E4 in a single sample, all from
different locations in Germany. These serotypes are
associated occasionally with keratoconjunctivitis. Typ-
ing failed in six eye swabs because of low HAdV DNA
concentrations (<1,000 copies/ml). In conclusion, mole-
cular typing suggested that eye swabs from single HAdV
conjunctivitis patients or small outbreaks of HAdV
conjunctivitis in single barracks had been sampled from
a majority of patients with another, as yet unidentified
etiology of conjunctivitis, for example unencapsulated
streptococci [Crum et al., 2004].
Genetic Variability of Circulating
HAdV Types
Several HAdV DNA sequences obtained from patient
samples had significant nucleotide variability when
compared with prototype sequences but fulfilled the
criteria for precise typing. For example, L2 sequences of
HAdV-B and -D were 97.6% to 100% identical to the
sequences of their respective prototype strains (Table II)
and 75%–92% identical to the sequences of the highest-
scoring heterologous prototype (data not shown). In the
case of HAdV-B3, four phylogenetic clusters were found,
three of these representing multiple isolates (Fig. 3).
J. Med. Virol. DOI 10.1002/jmv
TABLE II. Median and Maximum Divergence (%) of HAdV
DNA of Clinical Samples Compared to the Homologous
Prototype Databank Sequences
HAdV
type n
Divergence to
prototype sequence (%)
Closest phylogenetic
prototype sequence
Median Maximum GeneBank number
A12 1 0.4a) 0.4 AT12CGA
A18 1 0.6a) 0.6 AF161575
A31 11 1.2a) 1.2 AF161576
B3 22 0.0b) 1.6 see Figure 3
B11 4 0.4b) 0.4 AF542104
B21 1 0.0b) 0.0 AJ005537
C1 24 0.4a) 1 AF534906
C2 25 0.0a) 0.8 J01917
C5 3 0.8a) 0.8 M73260
C6 3 0.4a) 0.5 AY375455
D8 4 0.0c) 0.0 X74663
D19 3 1.7c) 2.4 X98359
D30 1 0.0c) 0.0 AJ745888
E4 3 0.0a) 0.0 AF065064
F40 1 0.4a) 0.4 X51782
F41 12 0.0a) 1.1 AY375456
The accession numbers represent the closest prototype sequence in the
Genbank.
Numbers marked with an a) were typed by step 1, b) with step 2 primer
set for species B (BL/BR), and c) with step 2 primer set for species C and
D (CDL/CDR) Madisch et al. [2005].
1214 Madisch et al.

Most (5 of 9) of the L2 mutations observed in clusters
two, three and four were non-silent in comparison to the
HAdV-B3 prototype sequence, indicating positive selec-
tion of these mutations in the neutralization determi-
nant by the immune system (ratio of non-synonymous to
synonymous mutations 1.49). These non-synonymous
mutations resulted in changes of the Eisenberg hydro-
phobicity profile (Fig. 4), potentially interfering with the
recognition of the neutralization epitope by antibodies.
Amino acid substitutions at positions 46 and 76 affected
the hydrophobicity in all three HAdV-B3 subclusters
divergent from the prototype strain and may represent a
first step of immune escape. Subsequently, additional
mutations probably affecting recognition by the immune
system were selected in cluster 3 and cluster 4 (Fig. 4).
These findings suggested positive selection of these
mutations by an immune escape mechanism. Although
one of the four clusters of HAdV-B3 (cluster 1, Fig. 3) was
identical to the prototype sequence, all the other clusters
probably represented unsequenced subtypes (Fig. 3).
Therefore, additional fiber sequencing was performed.
In case of cluster four (containing two clinical isolates
and four other HAdV DNA positive samples), the fiber g
determinant was identical to the intermediate strain
HAdV-B7H3 (GenBank #Z48954) [Kajon and Wadell,
1996] with 0.0% sequence diversity whereas the L2
sequence (#DQ280366) was 1.5% divergent in compar-
ison to the HAdV-B3 prototype (#X01998) sequence and
fulfilled typing criteria. Therefore, these two clinical
isolates represent a not previously isolated HAdV-B3
recombination ‘‘donor’’ strain of the g determinant (fiber
knob region) for the HAdV-B7H3 intermediate strain
[Kajon and Wadell, 1996].
In contrast, only one or two sequence variants were
found for most other HAdV types (Table II). Several
HAdV types, which were only isolated occasionally
(HAdV-B21, -D8, -D30, -E4), had sequences completely
J. Med. Virol. DOI 10.1002/jmv
Fig. 3. Phylogenetic analysis (neighbor joining tree) of HAdV-B3
nucleic acid sequences. The clinical samples positive for HAdV-B3
clustered in four different phylogenetic groups, cluster 1 contained also
the prototype sequence. HAdV-B3 subtypes isolated in Korea (strain
KNIH Ad99/6 and Ad99/10) were added to the phylogenetic analysis,
which grouped into their own cluster.
Fig. 4. Eisenberg hydrophobicity plots of the different HAdV-B3
clusters with a window size of five amino acid residues [Eisenberg et al.,
1984]. Hydrophobicity plot marked with an A represents members of
cluster 1 (see Fig. 3), B types of cluster 2, C types of cluster 3, and D
types of cluster 4. The arrows mark hydrophobicity changes in the
different HAdV-B3 clusters.
Molecular Typing Scheme of Human Adenoviruses 1215

identical to their GenBank prototype sequences.
Furthermore, all clinical samples positive for HAdV-
A31 had identical sequences which were 1.2% divergent
from the prototype sequence. HAdV-A31 positive sam-
ples originated from six patients, five of these were
immunosuppressed, and one of these died of dissemi-
nated adenovirus disease. Patients were not temporally
or spatially related. This indicates the widespread
circulation of a single HAdV-A31 subtype which is
associated with potentially fatal disseminated infec-
tions in immunodeficient patients.
DISCUSSION
Results from 119 clinical samples clearly supported
molecular typing with the two step typing scheme
suggested recently [Madisch et al., 2005] as a convenient
and sufficient approach (Table I). In this study,
diagnostic samples from different body sites were tested
and direct typing was even achieved in 98 HAdV DNA
positive samples in which virus was not isolated or
before virus isolation was successful. Standard diag-
nostic procedures started with a generic quantitative
PCR [Heim et al., 2003] and all 119 samples containing
>103 copies/ml were successfully typed. A few diagnostic
samples with very low virus loads (e.g., six eye swabs
from the conjunctivitis outbreak in military barracks)
cannot be typed by the two-step molecular typing
scheme. If typing in these cases is required, at least
species identification could be achieved by nested PCR
[Allard et al., 2001]. The results give an impression of
the circulation of HAdV types associated with sympto-
matic, clinically significant infections in Germany
during the sampling period between June 2002 and
November 2005 (Table I). All blood samples originated
from highly immunosuppressed patients, for example,
allogenic stem cell recipients, and the results confirmed
the predominance of species HAdV-C viruses (in this
study mainly HAdV-C2) and HAdV-A31 in disseminated
infections (Table I) [Venard et al., 2000; Seidemann
et al., 2004; Kampmann et al., 2005]. In feces samples
from patients suffering from acute diarrhea, the
fastidious HAdV-F40 and -F41 were typed directly
without virus isolation. HAdV-C1, -C2, and -C5 were
found as frequently as enteric HAdV in feces samples
(Table I). However, virus isolation failed in all of these
samples (except two samples originating from termin-
ally ill immunosuppressed patients), HAdV-C DNA
concentrations were low (<105 copies/ml), or other
pathogens associated with diarrhea were found (e.g.,
rotavirus). Therefore, it is possible that species HAdV-C
viruses were not the cause of the diarrhea, and detection
of species HAdV-C DNA in feces may be virus shedding
in cases of HAdV persistence [Fox et al., 1977;
Garnettet al., 2002; McNees et al., 2004]. The species
HAdV-C and HAdV-A31 were associated significantly
with immunsuppression. However, these types are not
typical opportunistic pathogens and also cause signifi-
cant disease in immunocompetent patients. Therefore,
our findings indicate that the predominance of species
HAdV-C (and HAdV-A31) in immunosuppressed
patients may be related to previous infection of patients
and reactivation of persistent viral DNA [Garnett et al.,
2002; McNees et al., 2004]. In contrast, persistence of
HAdV DNA has not been reported for HAdV types
not associated with immunosuppression but circulating
frequently in general population, as for example HAdV-
B3 and species HAdV-F (Table I).
Molecular typing was achieved unequivocally for all
119 clinical samples and isolates (Table I) by our new
defined molecular typing criteria [Madisch et al., 2005].
Previously, another L2 sequence-based typing scheme
was suggested but precise typing criteria had not yet
been defined, and the scheme was only applied to eight
clinical isolates [Sarantis et al., 2004]. Nucleic acid
divergence of all 119 clinical samples to the nearest
homologous prototype sequence was as low as 0.0% with
a maximum divergence of HAdV-A31 (1.2%), comparing
favorably to the �1.5% criteria of step one. Median
genetic divergence in step 2 (immunogenic L2) was only
0.9% (highest for HAdV-D19, 2.4%), also comparing
favorably to the <2.5% L2 typing criteria. Therefore,
additional molecular typing techniques, for example
sequencing of L1 or phylogenetic analysis of deduced
amino acid sequences were not required, indicating that
the proposed criteria permit rapid decisions and simple
handling of sequence data. If the molecular typing
criteria for step 1 (species HAdV-A, -C, -E, and -F) and
for step 2 (species HAdV-B and -D) can not be fulfilled, a
deduced amino acid sequence of the clinical isolate has to
be aligned with prototype databank sequences, for
example by using the blastx internet server. If the L2
amino acid divergence compared to the next databank
sequence is <1.2%, typing is accomplished [Madisch
et al., 2005]. However, if the L2 amino acid divergence
compared to the next databank sequence is �1.2%,
typing is not feasible and the isolation of a new HAdV
type may be suspected. Additional genetic analysis
(L1 sequencing and fiber knob sequencing) should then
be performed to confirm the identification of a new
HAdV type [Pring-Akerblom et al., 1997; Madisch et al.,
2005; Lu and Erdman, 2006].
In conclusion, hexon L2 sequencing is an easy but
sufficient approach to type HAdV DNA without the
need of time-consuming virus isolation. Statistical
analysis of typing results revealed the predominance
of species C and HAdV-A31 in immunodeficient patients
in contrast to HAdV-B3 and HAdV-F in immunocompe-
tent patients. This supports the hypothesis that
serious disseminated HAdV disease may origin rather
from HAdV DNA persistence than from exogenous
infection.
ACKNOWLEDGMENTS
We thank Prof. Balzas Harrach, Veterinary Medical
Research Institute, Hungarian Academy of Sciences for
providing an online blast search for molecular typing of
human adenoviruses and Dr. Penelope Kay-Jackson for
critical reading of the manuscript.
J. Med. Virol. DOI 10.1002/jmv
1216 Madisch et al.

REFERENCES
Akiyama H, Kurosu T, Sakashita C, Inoue T, Mori S, Ohashi K,
Tanikawa S, Sakamaki H, Onozawa Y, Chen Q, Zheng H, Kitamura
T. 2001. Adenovirus is a key pathogen in hemorrhagic cystitis
associated with bone marrow transplantation. Clin Infect Dis
32:1325–1330.
Allard A, Albinsson B, Wadell G. 2001. Rapid typing of human
adenoviruses by a general PCR combined with restriction endonu-
clease analysis. J Clin Microbiol 39:498–505.
Aoki K, Tagawa Y. 2002. A twenty-one year surveillance of adenoviral
conjunctivitis in Sapporo, Japan. Int Ophthalmol Clin 42:49–54.
Bhat AM, Meny RG, Aranas EA, Yehia F. 1984. Fatal adenoviral (type
7) respiratory disease in neonates. Clin Pediatr (Phila) 23:409–411.
Carrigan D. 1997. Adenovirus infections in immunocompromised
patients. Am J Med 102:71–74.
Chuang YY, Chiu CH, Wong KS, Huang JG, Huang YC, Chang LY, Lin
TY. 2003. Severe adenovirus infection in children. J Microbiol
Immunol Infect 36:37–40.
Crawford-Miksza LK, Nang RN, Schnurr DP. 1999. Strain variation in
adenovirus serotypes 4 and 7a causing acute respiratory disease. J
Clin Microbiol 37:1107–1112.
Crum NF, Barrozo CP, Chapman FA, Ryan MA, Russell KL. 2004. An
outbreak of conjunctivitis due to a novel unencapsulated Strepto-
coccus pneumoniae among military trainees. Clin Infect Dis
39:1148–1154.
Dakhama A, Hegele RG, Laflamme G, Israel-Assayag E, Cormier Y.
1999. Common respiratory viruses in lower airways of patients with
acute hypersensitivity pneumonitis. Am J Respir Crit Care Med
159:1316–1322.
Davison AJ, Benko M, Harrach B. 2003. Genetic content and evolution
of adenoviruses. J Gen Virol 84:2895–2908.
Echavarria MS, Ray SC, Ambinder R, Dumler JS, Charache P. 1999.
PCR detection of adenovirus in a bone marrow transplant recipient:
Hemorrhagic cystitis as a presenting manifestation of disseminated
disease. J Clin Microbiol 37:686–689.
Eisenberg D, Schwarz E, Komaromy M, Wall R. 1984. Analysis of
membrane and surface protein sequences with the hydrophobic
moment plot. J Mol Biol 179:125–142.
Erdman DD, Xu W, Gerber SI, Gray GC, Schnurr D, Kajon AE,
Anderson LJ. 2002. Molecular epidemiology of adenovirus type 7 in
the United States, 1966-2000. Emerg Infect Dis 8:269–277.
Fox JP, Hall CE, Cooney MK. 1977. The Seattle Virus Watch. VII.
Observations of adenovirus infections. Am J Epidemiol 105:362–
386.
Friedrichs N, Eis-Hubinger AM, Heim A, Platen E, Zhou H, Buettner R.
2003. Acute adenoviral infection of a graft by serotype 35 following
renal transplantation. Pathol Res Pract 199:565–570.
Garnett CT, Erdman D, Xu W, Gooding LR. 2002. Prevalence and
quantitation of species C adenovirus DNA in human mucosal
lymphocytes. J Virol 76:10608–10616.
Heim A, Ebnet C, Harste G, Pring-Akerblom P. 2003. Rapid and
quantitative detection of human adenovirus DNA by real-time
PCR. J Med Virol 70:228–239.
Kajon AE, Wadell G. 1996. Sequence analysis of the E3 region and fiber
gene of human adenovirus genome type 7h. Virology 215:190–196.
Kampmann B, Cubitt D, Walls T, Naik P, Depala M, Samarasinghe S,
Robson D, Hassan A, Rao K, Gaspar H, Davies G, Jones A, Cale C,
Gilmour K, Real M, Foo M, Bennett-Rees N, Hewitt A, Amrolia P,
Veys P. 2005. Improved outcome for children with disseminated
adenoviral infection following allogeneic stem cell transplantation.
Br J Haematol 130:595–603.
Lion T, Baumgartinger R, Watzinger F, Matthes-Martin S, Suda M,
Preuner S, Futterknecht B, Lawitschka A, Peters C, Potschger U,
Gadner H. 2003. Molecular monitoring of adenovirus in
peripheral blood after allogeneic bone marrow transplantation
permits early diagnosis of disseminated disease. Blood 102:1114–
1120.
Ljungman P. 2004. Treatment of adenovirus infections in the
immunocompromised host. Eur J Clin Microbiol Infect Dis 23:
583–588.
Lu X, Erdman DD. 2006. Molecular typing of human adenoviruses by
PCR and sequencing of a partial region of the hexon gene. Arch Virol
(DOI 10.1007/s00705-005-0722-7).
Madisch I, Harste G, Pommer H, Heim A. 2005. Phylogenetic analysis
of the main neutralization and hemagglutination determinants of
all human adenovirus prototypes as a basis for molecular
classification and taxonomy. J Virol 79:15265–15276.
McNees AL, Mahr JA, Ornelles D, Gooding LR. 2004. Postinternaliza-
tion inhibition of adenovirus gene expression and infectious virus
production in human T-cell lines. J Virol 78:6955–6966.
Mellman-Rubin TL, Kowalski RP, Uhrin M, Gordon YJ. 1995.
Incidence of adenoviral and chlamydial coinfection in acute
follicular conjunctivitis. Am J Ophthalmol 119:652–654.
Mori K, Yoshihara T, Nishimura Y, Uchida M, Katsura K, Kawase Y,
Hatano I, Ishida H, Chiyonobu T, Kasubuchi Y, Morimoto A,
Teramura T, Imashuku S. 2003. Acute renal failure due to
adenovirus-associated obstructive uropathy and necrotizing tubu-
lointerstitial nephritis in a bone marrow transplant recipient. Bone
Marrow Transplant 31:1173–1176.
Naesens L, Lenaerts L, Andrei G, Snoeck R, Van Beers D, Holy A,
Balzarini J, De Clercq E. 2005. Antiadenovirus activities of several
classes of nucleoside and nucleotide analogues. Antimicrob Agents
Chemother 49:1010–1016.
Pring-Akerblom P, Adrian T, Kostler T. 1997. PCR-based detection and
typing of human adenoviruses in clinical samples. Res Virol
148:225–231.
Sarantis H, Johnson G, Brown M, Petric M, Tellier R. 2004.
Comprehensive detection and serotyping of human adenoviruses
by PCR and sequencing. J Clin Microbiol 42:3963–3969.
Seidemann K, Heim A, Pfister ED, Ko¨ditz H, Beilken A, Sander A,
Melter M, Sykora KW, Sasse M, Wessel A. 2004. Monitoring of
adenovirus infection in pediatric transplant recipients by quanti-
tative PCR: Report of six cases and review of the literature. Am J
Transplant 4:2102–2108.
Stalder H, Hierholzer JC, Oxman MN. 1977. New human adenovirus
(candidate adenovirus type 35) causing fatal disseminated infection
in a renal transplant recipient. J Clin Microbiol 6:257–265.
Stock R, Harste G, Madisch I, Heim A. 2006. A rapid quantitative PCR-
based assay for testing antiviral agents against human adeno-
viruses demonstrates type specific differences in ribavirin activity.
Antiviral Res (DOI 10.1016/j.antiviral.2006.03.009).
Swenson PD, Wadell G, Allard A, Hierholzer JC. 2003. Adenoviruses.
In: Murray PR, Baron EJ, Pfaller MA, Jorgensen JH, Yolken RA,
editors. Manual of clinical microbiology. Washington, DC: ASM
Press. p 1404–1417.
Venard V, Carret A, Corsaro D, Bordigoni P, Le Faou A. 2000.
Genotyping of adenoviruses isolated in an outbreak in a bone
marrow transplant unit shows that diverse strains are involved. J
Hosp Infect 44:71–74.
J. Med. Virol. DOI 10.1002/jmv
Molecular Typing Scheme of Human Adenoviruses 1217