Genetic diversity of novel circular ssDNA viruses in bats in China

Abstract

Novel circular ssDNA genomes have recently been detected in animals and in the environment using metagenomic and high-throughput sequencing approaches. In this study, five full-length circular ssDNA genomes were recovered from bat faecal samples using inverse PCR with sequences designed based on circovirus-related sequences obtained from Solexa sequencing data derived from a random amplification method. These five sequences shared a similar genomic organization to circovirus or the recently proposed cyclovirus of the family Circoviridae. The newly obtained circovirus/cyclovirus-like genomes ranged from 1741 to 2177 bp, and each consisted of two major ORFs, ORF1 and ORF2, encoding putative replicase (Rep) and capsid (Cap) proteins, respectively. The potential stem-loop region was predicted in all five genomes, and three of them had the typical conserved nonanucleotide motif of cycloviruses. A set of primers targeting the conserved Rep region was designed and used to detect the prevalence of circovirus/cyclovirus sequences in individual bats. Among 199 samples tested, 47 were positive (23.6%) for the circovirus genome and two (1.0%) were positive for the cyclovirus genome. In total, 48 partial Rep sequences plus the five full-length genomes were obtained in this study. Detailed analysis indicated that these sequences are distantly related to known circovirus/cyclovirus genomes and may represent 22 novel species that belong to the family Circoviridae.

Full text

Genetic diversity of novel circular ssDNA viruses in
bats in China
Xingyi Ge,
1
Jialu Li,
1
Cheng Peng,
1
Lijun Wu,
1
Xinglou Yang,
1
Yongquan Wu,
1
Yunzhi Zhang
2
and Zhengli Shi
1
Correspondence
Zhengli Shi
zlshi@wh.iov.cn
Received 16 May 2011
Accepted 25 July 2011
1
State Key Laboratory of Virology, Wuhan Institute of Virology, Chinese Academy of Sciences,
Wuhan, PR China
2
Yunnan Institute of Endemic Diseases Control and Prevention, Dali, PR China
Novel circular ssDNA genomes have recently been detected in animals and in the environment
using metagenomic and high-throughput sequencing approaches. In this study, five full-length
circular ssDNA genomes were recovered from bat faecal samples using inverse PCR with
sequences designed based on circovirus-related sequences obtained from Solexa sequencing
data derived from a random amplification method. These five sequences shared a similar genomic
organization to circovirus or the recently proposed cyclovirus of the family Circoviridae. The newly
obtained circovirus/cyclovirus-like genomes ranged from 1741 to 2177 bp, and each consisted
of two major ORFs, ORF1 and ORF2, encoding putative replicase (Rep) and capsid (Cap)
proteins, respectively. The potential stem–loop region was predicted in all five genomes, and three
of them had the typical conserved nonanucleotide motif of cycloviruses. A set of primers targeting
the conserved Rep region was designed and used to detect the prevalence of circovirus/
cyclovirus sequences in individual bats. Among 199 samples tested, 47 were positive (23.6 %) for
the circovirus genome and two (1.0 %) were positive for the cyclovirus genome. In total, 48 partial
Rep sequences plus the five full-length genomes were obtained in this study. Detailed analysis
indicated that these sequences are distantly related to known circovirus/cyclovirus genomes and
may represent 22 novel species that belong to the family Circoviridae.
INTRODUCTION
Members of the virus family Circoviridae, known to infect
vertebrates, are non-enveloped, with a diameter of 17–
24 nm. The capsid consists of a circular ssDNA genome
of 1.7–2.0 kb (Fauquet & Fargette, 2005). Two genera,
Gryovirus and Circovirus, are included in the family by the
International Committee on Taxonomy of Viruses (ICTV),
although the taxonomic position of the former is still being
debated (Hino & Prasetyo, 2009). The genus Gryovirus has
only one member, Chicken anemia virus (CAV), with a
negative-sense genome. CAV is a ubiquitous pathogen of
chickens worldwide and has been linked to specific clinical
diseases and subclinical immunosuppression, which cause
economic losses in commercial chicken production (Schat,
2009; Smyth et al., 2006). In contrast to the gryovirus
genome, members of the genus Circovirus have an
ambisense genome, in which two major ORFs that code
for the replicase (Rep) and capsid (Cap) proteins are
arranged inversely. Circoviruses have been found to be
pathogens of pigs, as well as a broad range of avian species
including canaries, geese, pigeons and parrots (Bassami
et al., 1998; Chae, 2005; Hattermann et al., 2003; Todd
et al., 2001a, b). Avian circoviruses have been linked to
symptoms of avian lymphoid depletion, immunosuppres-
sion and developmental abnormalities such as deformities
of the beak and claws, feathering disorders and growth
retardation (Bassami et al., 1998; Stewart et al., 2006; Todd,
2000). Porcine circovirus (PCV)-1 and PCV-2 infect pigs
and cause porcine circovirus-associated disease (PCVAD)
worldwide, which has a huge impact on swine production.
The most relevant clinical diseases include post-weaning
multisystemic wasting syndrome, porcine dermatitis and
nephropathy syndrome, and reproductive disorders
(Darwich et al., 2004; Ellis et al., 1998, 2004; Firth et al.,
2009).
Among the known circoviruses, CAV, PCV-1 and PCV-2
are the only members that can readily be grown in cell
culture. Metagenomic approaches and high-throughput
sequencing techniques have enabled the recent discovery of
novel circovirus-like genomes from faecal samples of wild
mammals and insects and in the environment (Blinkova
The GenBank/EMBL/DDBJ accession numbers for the sequences
determined in this study are JN377555–JN377581 (contigs of Solexa
data related to circoviruses), JF938078–JF938082 (genomes of YN-
BtCV-1 to -5) and JF938083–JF938130 (partial Rep gene sequences).
Supplementary material is available with the online version of this
paper.
Journal of General Virology (2011), 92, 2646–2653 DOI 10.1099/vir.0.034108-0
2646 034108 G2011 SGM Printed in Great Britain

et al., 2010; Donaldson et al., 2010; L. Li et al., 2010a, b;
Rosario et al., 2009a, b, 2011). Additionally, diverse
circoviruses have been detected in humans and farmed
animals using PCR, suggesting possible cross-species
transmission in farm animals (Li et al., 2011). Some newly
detected circular genomes were genomically similar to
those of circoviruses but phylogenetically different from
known avian and porcine circoviruses. Based on phylo-
genetic distances, these newly detected viruses were
grouped into the genus Cyclovirus, a newly proposed genus
of the family Circoviridae (L. Li et al., 2010a). Compared
with circoviruses, cycloviruses have a smaller genome,
encoding a smaller Rep and Cap protein and with shorter
or no 39intergenic regions between the stop codons of the
two major ORFs and a longer 59intergenic region between
the start codons of the two major ORFs (L. Li et al., 2010a).
Here, we report on the detection of genetically diverse
circovirus-like genomes from bat faecal samples. These
sequences were distantly related to known circovirus/
cyclovirus genomes and may represent 22 novel species
belonging to the family Circoviridae.
RESULTS
Overview of the Solexa sequencing data
In total, 34 666 contigs (.100 bp, mean length 195 bp)
were assembled after Solexa sequencing. All contigs were
analysed using BLAST with the National Center for
Biotechnology Information (NCBI) nr and nt databases.
There were 28 691 (82.8 %) contigs that did not exhibit any
similarities with known sequences, whilst 4661 (13.4 %)
were related to prokaryotic and eukaryotic host sequences,
664 (1.9 %) were related to phages and 650 (1.9 %) were
related to eukaryotic viral sequences. Among the 650
contigs related to eukaryotic viral sequences, 343 (52.8 %)
were related to densoviruses, 58 (8.9 %) were related to
adenoviruses, 44 (6.8 %) were related to dicisitroviruses, 37
(5.7 %) were related to circoviruses or circovirus-like
viruses, and 34 (5.2 %) were related to coronaviruses (see
Supplementary Table S1, available in JGV Online).
Genomic features of the novel circovirus-like
sequences in bats
Five full-length circular ssDNA genomes of the novel bat
circoviruses (BtCVs), ranging from 1741 to 2177 bp,
named YN-BtCV-1 to -5, were identified. The predicted
two ORFs encoding Rep and Cap, were inversely arranged
(Fig. 1). The Rep protein sequences consisted of 267–
286 aa and had an amino acid identity of 51–72 % with
known circoviruses and 25–69 % identity among them-
selves (Table 1). The Cap protein sequences ranged from
219 to 233 aa and had an amino acid identity of 7–56 %
with the known circoviruses and 5–36 % identity among
themselves (Table 1). Three conserved regions in Rep
(WWDGY, DRYP and G-GKS) were found in all five Rep
sequences (see Supplementary Fig. S1, available in JGV
Online); among them, the G-GKS region is related to a
dNTP-binding sequence (Todd et al., 2007). The highly
basic and arginine-rich region at the N terminus, which is
typical of circovirus Cap proteins (Stewart et al., 2006), was
found in YN-BtCV-2, -3, -4 and-5 but not in YN-BtCV-1
(see Supplementary Fig. S2, available in JGV Online).
The genomic structures of YN-BtCV-2, -3, -4 and -5 were
similar to the newly proposed cyclovirus (L. Li et al.,
2010a). YN-BtCV-2 and -5 had a 59intergenic region of
224 and 292 bp, respectively, and their two ORFs had a
4 nt overlap at the 39termini, whilst YN-BtCV-3 and -4
had a 59intergenic region of 236 and 229 bp, respectively,
and a 39intergenic region of 2 and 1 bp, respectively (Table
2). YN-BtCV-1 had an extremely long 59intergenic region
of 560 bp and a normal 39intergenic region of 112 bp,
compared with other members of the genus Circovirus
(L. Li et al., 2010a; Todd et al., 2007).
Potential stem–loop structures were found in all five bat
circular genomes and are thought to be involved in the
Fig. 1. Organization of the five novel circular ssDNA genomes
recovered from bat faeces in Yunnan Province, China. The two
major ORFs encoding the putative Rep and Cap proteins are
shaded. The locations of the potential stem–loop structures are
marked by a filled box.
Novel circular ssDNA viruses in bats
http://vir.sgmjournals.org 2647

initiation of rolling-circle replication, similar to the
replication model established using plasmids (Cheung,
2006). The genomes of YN-BtCV-2, -3 and -4 contained a
cyclovirus conserved nonanucleotide motif (59-TAATA-
CTAT-39) in the loop region of the 59intergenic region. YN-
BtCV-1 and -5 contained a potential loop region of 23 nt
(59-CAGTTAAAACAAGCAGGATTGAA-39) and 12 nt (59-
GGTATATATCGA-39), respectively, in the 59intergenic
region, which were different from the conserved cyclovirus
and circovirus nonanucleotide sequences (Table 2). The
length of the tandem repeat of the five viruses was between 7
and 13 nt (Table 2).
The complete predicted protein sequences encoded by
ORF1 (Rep-like) of YN-BtCV-1 to -5 were used for phy-
logenetic analysis with those of representative and some
newly discovered circoviruses/cycloviruses; CAV was used
as an outgroup, as it is distantly related to other members
of the family Circoviridae (Fig. 2). As shown in the
phylogenetic tree, YN-BtCV-2, -3, -4 and -5 fell into the
cyclovirus clade, members of which were identified in
mammals (human, goat, cow, bat and chimpanzee), a bird
(chicken) and an insect (dragonfly) (L. Li et al., 2010a, b;
Rosario et al., 2011). YN-BtCV-4 was related to GF-4c, a
bat-sourced cyclovirus detected from bat guano in
California using a metagenomic method (Li et al.,
2010b). YN-BtCV-3 was related to DfCyV-A1 and
DfCyV-A11, cycloviruses detected in dragonflies though
viral metagenomics (Rosario et al., 2011). YN-BtCV-2 was
related to cyclovirus PK5034, which was detected from
human stool samples from Pakistan. YN-BtCV-5 was
related to cyclovirus NG14, which was identified from
human stool samples from Nigeria (L. Li et al., 2010a). YN-
BtCV-1 formed a monophyletic branch and was not
grouped with the cycloviruses or circoviruses, suggesting
that this virus is a novel single-stranded circular genome
with a long evolutionary distance from other known
circoviruses and cycloviruses.
Prevalence and genetic diversity of circovirus-like
Rep sequences in bats
A set of generated primers that targeted the Rep sequences
was designed based on an alignment of circular genomes of
Table 1. Pairwise comparison of BtCVs based on amino acid identities (%) shared by the Rep and Cap proteins
Cap protein Rep protein
YN-BtCV-1 YN-BtCV-2 YN-BtCV-3 YN-BtCV-4 YN-BtCV-5
YN-BtCV-1 – 25 25 26 27
YN-BtCV-2 5 – 62 69 63
YN-BtCV-3 7 35 – 65 62
YN-BtCV-4 6 35 36 – 65
YN-BtCV-5 6 28 32 30 –
Table 2. Genomic features of BtCVs, representative members of the family Circoviridae were included for comparison
BFDV, Beak and feather disease virus; Chimp17, chimpanzee stool avian-like circovirus Chimp17; GF-4c, bat cyclovirus GF-4c; CyCV-TB,
cyclovirus bat/USA/2009; Chimp12, cyclovirus Chimp12; NG14, cyclovirus NG14. Conserved loop motifs are underlined.
Virus Genome
size (nt)
Putative
Rep (aa)
Putative
Cap (aa)
5§intergenic
region (nt)
3§intergenic
region (nt)
Loop motif
(3§A5§)
Tandem
repeat
YN-BtCV-1 2177 267 233 560 112 TTCAATCCTGCT
TGTTTTAACTG
ACAGCTGTT
YN-BtCV-5 1818 286 222 292 TCGATATATACC GGCGCCC
YN-BtCV-2 1771 276 219 224 TAATACTAT AAGTGACGGG
YN-BtCV-3 1743 277 223 236 2 GTAATACTATA AGTCGCGG
YN-BtCV-4 1741 279 223 229 1 GTAATACTATA ACGAAGTGGACGG
GF-4c 1866 282 231 326 GTAATACTATA CGAAGTGCCGG
CyCV-TB 1703 279 226 192 GTAATACTATA ACGAAGTGGCGG
NG14 1795 286 230 230 245 GTAATACTATA CGAAGTGACGG
Chimp12 1747 280 220 220 255 GTAATACTATA ACGAAGTGGCTGG
Chimp17 1935 292 233 198 162 GTACAGTATTACT GGTTCCAAGTTGCC
PCV-1 1759 312 233 82 36 CTGTAGTATTAC GAAGTGCGCTG
PCV-2 1768 314 233 83 38 TAAGTATTAC GAAGTGCGCTG
BFD-V 1993 289 247 126 232 TTAGTATTAC GGCGGCGG
X. Ge and others
2648 Journal of General Virology 92

bat sources and those of representative circoviruses. This
was used for the detection of similar viruses in individual
bat samples. In total, 199 bat samples were screened and
47 (23.6 %) were positive (Table 3). The prevalence of
circovirus-like genomes in different bat species ranged
from 2.6 to 66.7 % (Table 3). Positive samples were from
Rousettus leschenaultia (Rol), Rhinolophus pusillus (Rp),
Rhinolophus luctus (Rhl), Hipposideros armiger (Ha),
Myotis spp. (Mys) and Miniopterus schreibersii (Mis). The
samples from Rhinolophus affinis,Rhinolophus sinicus,
Hipposideros larvatus and Myotis chinensis were free of
detected viruses, probably due to limited sampling
numbers.
Products with a size of approximately 430–480 bp were
purified and sequenced. Samples were named by sampling
number, province (YN, Yunnan Province; HN, Henan
Province; HB, Hubei Province) and bat species abbre-
viation. One sample from Hipposideros armiger (2152-YN-
Ha) showed sequence heterogeneity. The PCR product of
this sample was cloned into a pGEM-T Easy vector
(Promega) and eight clones were sequenced. These clones
contained two consensus sequences, named 2152-YN-
Ha-A and 2152-YN-Ha-B. In total, 48 consensus Rep
sequences ranging from 400 to 467 bp were obtained.
These partial Rep sequences had an amino acid identity of
17–53 % with those of known circoviruses and 19–100 %
Fig. 2. Phylogenetic analysis of the five novel circular ssDNA genomes using Rep protein sequences. Representative members
of the genus Circovirus were included in the analysis, and GenBank accession numbers are indicated. SwCV, Swan circovirus;
GoCV, goose circovirus; DuCV, duck circovirus; GuCV, gull circovirus; CoCV, columbid circovirus; CaCV, canary circovirus;
FiCV, finch circovirus; StCV, starling circovirus; RaCV, raven circovirus; TM-6c, bat circovirus-like virus; Chimp11, cyclovirus
Chimp11; NG12, cyclovirus NG12; PK5006, cyclovirus PK5006; PK5034, cyclovirus PK5034; PK5222, cyclovirus PK5222;
PK5510, cyclovirus PK5510; TN25, cyclovirus TN2; DfCyV-A1, dragonfly cyclovirus DfCyV-A1; DfCyV-A11, dragonfly
cyclovirus DfCyV-A11; NG13, human stool-associated circular virus NG13; NG8, cyclovirus NGchicken8/NGA/2009; NG15,
cyclovirus NGchicken15/NGA/2009; PKbeef23, cyclovirus PKbeef23/PAK/2009; PKgoat11, cyclovirus PKgoat11/PAK/2009;
PKgoat21, cyclovirus PKgoat21/PAK/2009. See Table 2 for other abbreviations. Bar, 0.1 substitutions per site.
Novel circular ssDNA viruses in bats
http://vir.sgmjournals.org 2649

among themselves. Three samples, 2101-YN-Rhl, 2114-
YN-Mys and 2152-YN-Ha-A, were identical in amino acid
sequences but their nucleotide sequence identity was 98 %.
A phylogenetic tree was constructed based on alignment of
the 48 Rep sequences (128 aa) with those of representative
circoviruses and some of the newly detected cycloviruses;
CAV was included as an outgroup (Fig. 3). As shown in the
tree and with the exception of two sequences, 2125-YN-
Mys and 2121-YN-Mys, which fell into the clade of
cycloviruses, all other partial Rep sequences formed five
groups distinct from those formed by the known
circoviruses, suggesting that these viruses had the same
host origin, probably from bats. Using the criterion that
was recently used in a circovirus diversity analysis (L. Li
et al., 2010a), where an amino acid identity of .85 % in
the highly conserved Rep protein region indicates an
individual viral species, we obtained 22 species from bat
samples, including five cycloviruses and 17 circoviruses
(numbered 1–22 in Fig. 3). We arbitrarily divided the
potential bat-sourced circoviruses into five groups (I–V)
for further analysis. Group I, comprising 31 Rep sequences
with an amino acid identity of 48–100 %, was closest to the
known cycloviruses and circoviruses and could be divided
into ten species (numbered 6–15). Group II, comprising six
Rep sequences with an amino acid identity of 86–98 %,
belonged to species 16 and clustered with TM-6c, a bat-
sourced circovirus-like virus. Group III, comprising four
Rep sequences with an amino acid identity of 43–96 %,
could be divided into three species (17–19). Group IV,
comprising five Rep sequences with an amino acid identity
of 23–97 %, could be divided into two species (20 and 21).
Group V, containing one sequence, comprised species 22.
DISCUSSION
We have reported the discovery of five novel full-length
circular ssDNA genomic sequences from grouped bat
faeces in Yunnan Province, China. All of the sequences
(YN-BtCV-2, -3, -4 and -5) except for one (YN-BtCV-1)
had a typical cyclovirus genome organization with two
major opposite ORFs, a long 59intergenic region and a
short or absent 39intergenic region. In addition, the 59
intergenic region of YN-BtCV-2, -3 and -4, but not YN-
BtCV-5, shared a cyclovirus-conserved nonanucleotide
motif (59-TAATACTAT-39) in the loop region (Table 2).
Phylogenetic analysis based on the full-length Rep sequence
strongly supports the suggestion that YN-BtCV-2, -3, -4
and -5 should be clustered into the newly proposed genus
Cyclovirus (Fig. 2). YN-BtCV-1, containing an extremely
long 59intergenic region of 560 bp and the absence of the
arginine-rich region at the N terminus of Cap, had a low
amino acid identity with Rep (,30 %) and Cap (,10 %)
sequences of four other bat-sourced sequences and known
circoviruses (Table 1). Thus, it formed a monophyletic
clade (Fig. 2). These results imply that numerous unknown
circoviruses exist in bat faecal samples.
A high prevalence of the circovirus-like genome was detected
using degenerated primers in individual bats in three
sampling locations. With the exception of a few bat species
that had limited sampling numbers, most of the bat species
were positive for the detected virus, with a detection rate of
6.5–62.1 % in individual bats. A high detection rate (22.9–
61.5 %) was found in five out of seven bat species in Yunnan
Province: Rousettus leschenaultii,Rhinolophus pusillus,
Rhinolophus luctus,H. armiger and Myotis spp. The high
detection rate in bat samples in Yunnan Province was
probably due to the primers being based on viral sequences
obtained from samples from the same location. Similar to the
full-length Rep sequences, the partial Rep sequences obtained
from individual bats showed great genetic diversity and were
distantly related to those of known circoviruses. Further
investigation is needed into the prevalence of circovirus in
some bat species including Rhinolophus affinis,Rhinolophus
sinicus,H. larvatus and Myotis chinensis for which there were
fewer than five samples.
Interestingly, in our study, most cycloviruses were found
via metagenomics and most members of the new circovirus
Table 3. Prevalence of circoviruses among ten bat species in three provinces of China
Results are shown as the number of circoviruses out of the total number of samples screened (%) for each bat species.
Bat species Geographical origin
Hubei Province Henan Province Yunnan Province
Rousettus leschenaultii – – 11/48 (22.9)
Rhinolophus pusillus 3/15 (20) – 1/3 (33.3)
Rhinolophus affinis 0/1 0/3 –
Rhinolophus luctus – – 7/13 (53.8)
Rhinolophus sinicus 0/1 – –
Hipposideros armiger 2/10 (20) 1/39 (2.6) 2/4 (50)
Hipposideros larvatus – – 0/1
Myotis chinensis 0/1 – –
Myotis spp. 2/3 (66.7) – 16/26 (61.5)
Miniopterus schreibersii – 2/30 (6.7) 0/1
X. Ge and others
2650 Journal of General Virology 92

clade were found by consensus PCR. Furthermore, we did
not find identical consensus-PCR-derived bat circovirus
Rep sequences in the initial Solexa data. These disparate
results indicate that there may still be a wide diversity of
circoviruses to be found, most probably outside the family
Circoviridae, such as YN-BtCV-1 found in this study.
The host of the newly detected bat ssDNA genome is still
uncertain. Based on the epidemiology data and phylogen-
etic analysis, we suspect that most sequences represent a
bat origin. However, we cannot exclude the possibility of
contamination from virus sequences of other animals. For
example, YN-BtCV-3 shared 60 % amino acid identity with
DfCyV-A1, which was recently identified from a dragonfly
and probably originated from an insect fed on by bats. YN-
BtCV-1 formed a monophyletic branch from isolates of
cyclovirus and circovirus. More information is needed
to determine its host. Future research such as virus isola-
tion, epidemiology studies and bioinformatics analysis on
circular ssDNA viruses in vertebrates, invertebrates and
plants will be helpful in making clearer evolutionary links
among novel identified circular ssDNA viruses from
environmental and animal samples.
Among the known members of the family Circoviridae,
PCV-2 is the only pathogen related to mammalian disease,
causing PCVAD (Firth et al., 2009). Cycloviruses have been
found in the muscle tissue of farm animals, including
goats, cows, sheep, camels and chickens, but there was no
obvious disease syndrome related to the viruses (L. Li et al.,
2010a, 2011). Closely related cyclovirus Rep sequences have
been detected in cows and goats from farms in Pakistan
(L. Li et al., 2010a), and PCV-2 nucleic acids have been
detected in cattle (Kappe et al., 2009) and rodents (Lo˝rincz
et al., 2010) in PCV-2-infected farms. These reports
indicate that circoviruses have potential cross-species
transmission. Here, we detected 17 novel circoviruses and
five novel cycloviruses in bat faecal samples. Currently, the
pathogenesis of these circoviruses not yet clear. However,
considering the high diversity of bat circoviruses and the
high rate of mutation and recombination of ssDNA
circular genomes, the relationships between these small
viruses and bats, as well as potential effects on other
mammals, need to be investigated further.
METHODS
Bat samples. Bat samples were collected during September 2009 and
September 2010 (n5199) in the provinces of Henan (one roosting
site), Hubei (two roosting sites) and Yunnan (two roosting sites),
China, and sampling of individual bats was performed as described
previously (Li et al., 2005). More information on the bat species is
shown in Table 3. Bat faecal samples (a mixture of Myotis spp.) used
for Solexa sequencing was collected in Yunnan Province in 2009 (one
roosting site). To collect faecal samples from grouped bats, clean
plastic sheets measuring 2.062.0 m were placed under known bat
roosting sites at about 18 :00. Fresh faecal samples were collected at
approximately 5 : 30–6 : 00 the next morning and stored in liquid
nitrogen. Samples were transported to the laboratory and stored at
280 uC until they were analysed.
Fig. 3. Phylogenetic analysis of partial Rep protein sequences
obtained from individual bat samples. The bat species Circovirus
is defined as having an amino acid identity of .85 % in the
conserved Rep region and members are labelled by vertical bars
and numbers from 1 to 22. The Rep sequences of bat-sourced
circoviruses formed five groups, labelled I–V. Individual samples
(indicated in bold) are named by the sampling number, location
and bat species. For abbreviations of virus strains, see Fig. 2
legend. Bar, 0.1 substitutions per site.
Novel circular ssDNA viruses in bats
http://vir.sgmjournals.org 2651

Inverse PCR. In a previous study, we performed metagenomic analysis
of viromes in grouped bat faecal samples collected in Yunnan Province
using a high-throughput sequencing technique (Solexa; unpublished
data). The viral particles were purified from grouped bat guano.
Unprotected nucleic acids were removed by digestion with a mixture
of DNase and RNase. Viral nucleic acid was then extracted using a
QIAamp Viral RNA Mini kit (Qiagen). A viral nucleic acid library was
constructed by random PCR amplification as described by Y. Li et al.
(2010). Sequences similar to the Rep or Cap gene sequences of known
circoviruses or circovirus-like genomes were chosen for designing
outward-pointing primers to perform inverse PCR. The inverse PCR
was conducted in a 50 ml reaction mixture volume containing 10 ml
PCR buffer, 1 mM MgSO
4
, 0.2 mM dNTPs, 50 pmol universal primer,
1 U KOD-Plus DNA polymerase (Toyobo) and 1 ml viral nucleic acid.
The reaction was conducted for 35 cycles of 94 uC for 30 s, 50 uC for
40 s and 68 uC for 2 min, followed by incubation for 10 min at 68 uC.
The products were loaded on a 1% agarose gel and bands of 1–2 kb
were purified with an EZNA Gel Extraction kit (Omega Bio-Tek) and
cloned into a pGEM-T Easy vector (Promega) after adding adenine at
the 39termini of the PCR products using Taq DNA polymerase. Insert-
containing plasmids were sequenced using M13 forward and reverse
primers with an ABI Prism 3730 DNA analyser (Applied Biosystems).
The consensus sequences of the PCR products were assembled with the
contigs.
Detection of circovirus-like genomic fragments. Viral DNA was
extracted from individual bat faecal swabs using a QIAamp DNA Mini
kit (Qiagen). Degenerate semi-nested PCR primers were designed
based on Rep sequences conserved for circoviruses from birds and pigs
and from those obtained in this study. The motif HLQGF region was
chosen for the forward primer, and WWDGY and VIIDDFYG regions
for the reverse primers. The primer sequences were: BtCV-F1, 59-
CIGIWYICCICAYYKICARG-39; BtCV-R1, 59-AWCCAICCRYRRA-
ARTCRTC-39; and BtCV-R2, 59-TGYTIYTIRTAICCRTCCCA-39.
PCR amplification was carried out in a 25 ml reaction mix
containing 1 ml extracted DNA, 2.5 ml buffer, 10 pmol each primer,
0.2 mM dNTPs and 0.5 U Taq DNA polymerase (Promega). After
an initial incubation at 94 uC for 5 min, 35 cycles of amplification
were carried out consisting of 94 uC for 30 s, 50 uC for 30 s (for the
first- and second-round PCR) and 72 uC for 30 s, with a final
extension at 72 uC for 10 min. Standard precautions were taken to
avoid PCR contamination, and a negative control was included in
each PCR assay.
Sequence analysis. Putative ORFs with a minimum size of 100 aa
and a coding capacity were predicted by the NCBI ORF Finder
(http://www.ncbi.nlm.nih.gov/gorf/gorf.html). Hairpin and stem–loop
structures were identified using the Mfold web server version 3.2
(Zuker, 2003) (http://mfold.rna.albany.edu/?q=mfold). The maps of
bat circular ssDNA genomes were constructed with BVTech plasmid-
drawing software version 5.0 (http://www.biovisualtech.com). Refe-
rence sequences of different viral families were obtained from the
NCBI. Amino acids alignments were generated with CLUSTAL W
implemented in MEGA4.1 (Kumar et al., 2008) with the default settings
and edited in GENEDOC and BioEdit software. Aligned sequences were
trimmed to match the genomic region of the sequences obtained in this
study and used to generate phylogenetic trees in MEGA4.1 using the
neighbour-joining method with amino acid Poisson correction and
1000 bootstrap replicates.
ACKNOWLEDGEMENTS
This work was jointly funded by the National Natural Science
Foundation of China (30970137) and the State Key Program for Basic
Research Grant (2011CB504701).
REFERENCES
Bassami, M. R., Berryman, D., Wilcox, G. E. & Raidal, S. R. (1998).
Psittacine beak and feather disease virus nucleotide sequence analysis
and its relationship to porcine circovirus, plant circoviruses, and
chicken anaemia virus. Virology 249, 453–459.
Blinkova, O., Victoria, J., Li, Y., Keele, B. F., Sanz, C., Ndjango, J.-B. N.,
Peeters, M., Travis, D., Lonsdorf, E. V. & other authors (2010). Novel
circular DNA viruses in stool samplesof wild-living chimpanzees. J Gen
Virol 91, 74–86.
Chae, C. (2005). A review of porcine circovirus 2-associated
syndromes and diseases. Vet J 169, 326–336.
Cheung, A. K. (2006). Rolling-circle replication of an animal
circovirus genome in a theta-replicating bacterial plasmid in
Escherichia coli.J Virol 80, 8686–8694.
Darwich, L., Segale
´s, J. & Mateu, E. (2004). Pathogenesis of
postweaning multisystemic wasting syndrome caused by Porcine
circovirus 2: an immune riddle. Arch Virol 149, 857–874.
Donaldson, E. F., Haskew, A. N., Gates, J. E., Huynh, J., Moore, C. J. &
Frieman, M. B. (2010). Metagenomic analysis of the viromes of three
North American bat species: viral diversity among different bat
species that share a common habitat. J Virol 84, 13004–13018.
Ellis, J., Hassard, L., Clark, E., Harding, J., Allan, G., Willson, P.,
Strokappe, J., Martin, K., McNeilly, F. & other authors (1998).
Isolation of circovirus from lesions of pigs with postweaning
multisystemic wasting syndrome. Can Vet J 39, 44–51.
Ellis, J., Clark, E., Haines, D., West, K., Krakowka, S., Kennedy, S. &
Allan, G. M. (2004). Porcine circovirus-2 and concurrent infections in
the field. Vet Microbiol 98, 159–163.
Fauquet, C. M. & Fargette, D. (2005). International Committee on
Taxonomy of Viruses and the 3,142 unassigned species. Virol J 2, 64.
Firth, C., Charleston, M. A., Duffy, S., Shapiro, B. & Holmes, E. C.
(2009). Insights into the evolutionary history of an emerging livestock
pathogen: porcine circovirus 2. J Virol 83, 12813–12821.
Hattermann, K., Schmitt, C., Soike, D. & Mankertz, A. (2003).
Cloning and sequencing of Duck circovirus (DuCV). Arch Virol 148,
2471–2480.
Hino, S. & Prasetyo, A. A. (2009). Relationship of Torque teno
virus to chicken anemia virus. Curr Top Microbiol Immunol 331, 117–
130.
Kappe, E. C., Halami, M., Schade, B., Bauer, J., Dekant, W.,
Buitkamp, J., Boettcher, J. & Mueller, H. (2009). Fatal aplastic
anaemia with haemorrhagic disease in calves in Germany. J Comp
Pathol 141, 293.
Kumar, S., Nei, M., Dudley, J. & Tamura, K. (2008). MEGA: a biologist-
centric software for evolutionary analysis of DNA and protein
sequences. Brief Bioinform 9, 299–306.
Li, W., Shi, Z., Yu, M., Ren, W., Smith, C., Epstein, J. H., Wang, H.,
Crameri, G., Hu, Z. & other authors (2005). Bats are natural reservoirs
of SARS-like coronaviruses. Science 310, 676–679.
Li, L., Kapoor, A., Slikas, B., Bamidele, O. S., Wang, C., Shaukat, S.,
Masroor,M. A., Wilson, M. L., Ndjango, J.-B. N. & other authors (2010a).
Multiple diverse circoviruses infect farm animals and are commonly
found in human and chimpanzee feces. JVirol84, 1674–1682.
Li, L., Victoria, J. G., Wang, C., Jones, M., Fellers, G. M., Kunz, T. H. &
Delwart, E. (2010b). Bat guano virome: predominance of dietary
viruses from insects and plants plus novel mammalian viruses. J Virol
84, 6955–6965.
Li, L., Shan, T., Soji, O. B., Alam, M. M., Kunz, T. H., Zaidi, S. Z. &
Delwart, E. (2011). Possible cross-species transmission of circoviruses
and cycloviruses among farm animals. J Gen Virol 92, 768–772.
X. Ge and others
2652 Journal of General Virology 92

Li, Y., Ge, X., Zhang, H., Zhou, P., Zhu, Y., Zhang, Y., Yuan, J., Wang,
L.-F. & Shi, Z. (2010). Host range, prevalence, and genetic diversity of
adenoviruses in bats. J Virol 84, 3889–3897.
Lo˝ rincz, M., Csa
´gola, A., Biksi, I., Szeredi, L., Da
´n, A. & Tuboly, T.
(2010). Detection of porcine circovirus in rodents – short
communication. Acta Vet Hung 58, 265–268.
Rosario, K., Duffy, S. & Breitbart, M. (2009a). Diverse circovirus-like
genome architectures revealed by environmental metagenomics. J Gen
Virol 90, 2418–2424.
Rosario, K., Nilsson, C., Lim, Y. W., Ruan, Y. J. & Breitbart, M. (2009b).
Metagenomic analysis of viruses in reclaimed water. Environ Microbiol
11, 2806–2820.
Rosario, K., Marinov, M., Stainton, D., Kraberger, S., Wiltshire, E. J.,
Collings, D. A., Walters, M., Martin, D. P., Breitbart, M. & Varsani, A.
(2011). Dragonfly cyclovirus, a novel single-stranded DNA virus
discovered in dragonflies (Odonata: Anisoptera). J Gen Virol 92,
1302–1308.
Schat, K. A. (2009). Chicken anemia virus. In TT Viruses – the Still
Elusive Human Pathogens, pp. 151–183. Edited by E. M. de Villiers &
H. zur Hausen. Berlin: Springer-Verlag Berlin.
Smyth, J. A., Moffett, D. A., Connor, T. J. & McNulty, M. S. (2006).
Chicken anaemia virus inoculated by the oral route causes
lymphocyte depletion in the thymus in 3-week-old and 6-week-old
chickens. Avian Pathol 35, 254–259.
Stewart, M. E., Perry, R. & Raidal, S. R. (2006). Identification of a
novel circovirus in Australian ravens (Corvus coronoides) with feather
disease. Avian Pathol 35, 86–92.
Todd, D. (2000). Circoviruses: immunosuppressive threats to avian
species: a review. Avian Pathol 29, 373–394.
Todd, D., Weston, J., Ball, N. W., Borghmans, B. J., Smyth, J. A.,
Gelmini, L. & Lavazza, A. (2001a). Nucleotide sequence-based identi-
fication of a novel circovirus of canaries. Avian Pathol 30, 321–325.
Todd, D., Weston, J. H., Soike, D. & Smyth, J. A. (2001b). Genome
sequence determinations and analyses of novel circoviruses from
goose and pigeon. Virology 286, 354–362.
Todd, D., Scott, A. N. J., Fringuelli, E., Shivraprasad, H. L., Gavier-
Widen, D. & Smyth, J. A. (2007). Molecular characterization of novel
circoviruses from finch and gull. Avian Pathol 36, 75–81.
Zuker, M. (2003). Mfold web server for nucleic acid folding and
hybridization prediction. Nucleic Acids Res 31, 3406–3415.
Novel circular ssDNA viruses in bats
http://vir.sgmjournals.org 2653