The "pressure pan" evolution of human erythrovirus B19 in the Amazon, Brazil.

Abstract

To understand the evolutionary dynamics of human parvovirus B19, we analyzed VP1 and VP2 gene sequences of B19 sampled from Belém (Amazon), the city of São Paulo, Brazil and globally. Our analysis revealed a strikingly different pattern of evolutionary change for those viral lineages introduced into Belém, which exhibited a higher rate of nonsynonymous substitutions compared to those viruses sampled from other locations. We propose that difference this is due to the high prevalence of B19 in Belém (up to 85%) compared to other locations (prevalences of approximately 50%), which imposes a more intense selection pressure. Hence, those B19 lineages introduced into Belém experienced an elevated rate of amino acid change, driven by positive selection, in order to generate serial re-infections in a small web of transmission, which can be thought of as an evolutionary "pressure pan".

Full text

Page 1

The “pressure pan” evolution of human erythrovirus
B19 in the Amazon, Brazil
Ronaldo Barros de Freitasa, Edison Luiz Durigonb, Darleise de Souza Oliveiraa,
Camila Malta Romanoc, Maria Rute Castro de Freitasa, Alexandre da Costa Linharesa,
Fernando Lucas Meloc, Lílian Walshkellerb, Maria Luisa Barbosad, Egma M. Mayta Huatucoe,
Edward C. Holmesf, Paolo Marinho de A. Zanottoc,∗
aSeção de Virologia, Instituto Evandro Chagas, Secretaria de Vigilância em Saúde, Ministério da Saúde, Belém, Brazil
bLaboratorio de Virologia Clinica, Departamento de Micobiologia, Instituto de Ciências Biomédicas – ICB II, Universidade de São Paulo USP,
São Paulo – SP, Brazil
cLaboratorio de Evolução Molecular e Bioinformática, Departamento de Microbiologia, Instituto de Ciências Biomédicas – ICB II,
Universidade de São Paulo USP, São Paulo – SP, Brazil
dInstituto Adolfo Lutz – São Paulo, Brazil
eUniversity of San Marcos, Lima, Peru
fCenter for Infectious Disease Dynamics, Department of Biology, The Pennsylvania State University, Mueller Laboratory, University Park, PA 16802, USA
Received 23 March 2007; returned to author for revision 2 May 2007; accepted 3 July 2007
Available online 15 August 2007
Abstract
To understand the evolutionary dynamics of human parvovirus B19, we analyzed VP1 and VP2 gene sequences of B19 sampled from Belém
(Amazon), the city of São Paulo, Brazil and globally. Our analysis revealed a strikingly different pattern of evolutionary change for those viral
lineages introduced into Belém, which exhibited a higher rate of nonsynonymous substitutions compared to those viruses sampled from other
locations. We propose that difference this is due to the high prevalence of B19 in Belém (up to 85%) compared to other locations (prevalences of
approximately50%),whichimposesamoreintenseselectionpressure.Hence,thoseB19lineagesintroducedintoBelémexperiencedanelevatedrate
of aminoacid change, driven by positive selection, inorder to generate serial re-infections ina small web oftransmission, which can be thoughtof as
an evolutionary “pressure pan”.
© 2007 Elsevier Inc. All rights reserved.
Keywords: Erythrovirus evolution; Natural selection; B19 phylogeny
Introduction
A complex interplay of factors influences the emergence and
re-emergence of viral diseases, including virus genetic variation
(itself generated by mutation, recombination and reassortment)
and environmental factors (ecological, social and behavioral
influences). Most emerging diseases are caused by RNAviruses
(Cleaveland et al., 2001), which can often quickly adapt to
varying ecological conditions, including new host species, due
to the high error rate of the virus RNA polymerase and their
immense population sizes. In contrast, DNA viruses, which
replicate by means of DNA polymerase, are less prone to
mutational error, generally experience lower long-term rates of
nucleotide substitution and, as a consequence, are less often
associated with cross-species transmission.
However, not all DNA viruses evolve slowly. Most notably,
canine parvovirus (CPV), a single-strand DNA virus, was
recently found to possess an evolutionary rate of approxi-
mately 10−4substitutions/site/year, broadly similar to that
observed in RNA viruses, and to experience strong positive
selection following its emergence from feline panleukopenia
parvovirus (FPLV) or a closely related virus (Shackelton et al.,
Available online at www.sciencedirect.com
Virology 369 (2007) 281–287
www.elsevier.com/locate/yviro
∗Corresponding author.
E-mail address: pzanotto@usp.br (P.M.A. Zanotto).
0042-6822/$ - see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.virol.2007.07.007

Page 2

2005). The erythrovirus B19, a human-associated member of
the family Parvoviridae, evolves at a similar rate, suggesting
that parvoviruses, and perhaps all single-stranded DNAviruses,
experience an elevated mutation burden (Shackelton and
Holmes, 2006). Parvovirus B19 is the only known human
pathogenic erythrovirus and has been detected globally in
human populations. The virus causes a wide spectrum of
clinical conditions, mainly erythema infectious in children and
arthropathy/arthralgia in adults. In immunocompetent indivi-
duals a self-limited infection is most common, while in
immunocompromised patients a persistent infection may
occur (Pattison, 1990; Brown et al., 1994). However, despite
the clinical importance of B19 and the observation that it
evolves at an elevated rate, little is known about the selection
pressures acting on this virus and how they may differ among
localities with differing epidemiological profiles. To this end,
we conducted an intensive analysis of the evolutionary
dynamics of genotype 1 of B19 in two populations in Brazil,
i.e., Belém in the Amazon region and the metropolis of São
Paulo.
Results and discussion
Phylogenetic relationships of B19 virus
In total, we obtained 46 individual sequences from 46
patientsinBelém,sampledfrom1995to2005,and48individual
sequences from 48 patients from São Paulo, sampled from 1990
to 2003. These sequences were analyzed together with the refe-
rence sequences in a codon-based alignment of 139 taxa and
covering 476 bp. The maximum a posteriori (MAP) tree for B19
with branch lengths corrected by time of isolation is shown in
Fig. 1 (with the outgroup sequence removed to increase re-
solution).AsimilartopologywasobtainedusingtheMLmethod
(data not shown). Although sequences from Belém (shown in
magentainFig.1)fallthroughoutthetree,mostclusterintothree
Fig. 1. Maximum a posteriori (MAP) tree for 133 sequences of the common region of VP1 and VP2 gene (477 bp) for human parvovirus B19. The tree has branch
lengths set by date of sampling and a time-scale to a hypothetical Most Recent Common Ancestor (MRCA) before 1973. Three clusters of lineages including samples
collected in Belém (Amazon) are shownby asterisks. Shared sites with elevated dN/dS(yellow cross dashes), unique elevated dN/dSratios (red cross dashes) and amino
acid changes taking place under purifying selection (green cross dashes), are shown along branches of the tree (i.e., lineages).
282R.B. de Freitas et al. / Virology 369 (2007) 281–287

Page 3

main lineages (shown by numbers followed by the ∗ symbol).
Cluster 1∗ was largely comprised of sequences from isolates
from Brazil (including Belém) and suggested a significant ex-
change of viruses within Brazil, although it also contained 4
sequences sampled from elsewhere: a single US isolate from
1994 (U38515/US), a 2001 isolate from Sweden (AY028237/
sw), a 1994 isolate from Germany (Z70528/ge) and a 1996
isolate from Finland (AF161224/fi). In contrast, Clusters 2∗ and
3∗ are dominated by lineages from Belém and therefore appear
to indicate independent introductions of lineages from the
Northern hemisphere into Belém (although because of the
relative small sampled size, we cannot exclude that the initial
introduction was from southern Brazil). Nevertheless, both in-
stances suggest the independent introduction of lineages of B19
genotype 1 into the Amazon region. To test for independent
entries more formally, we compared the best tree for these data
(log likelihood=−2721.907), to that obtained under a topolo-
gical constraint that forces samples from Belém to become
monophyletic (log likelihood=−2758.270), which would imply
a single introduction event. The difference in likelihood of
36.363 was significant with a one-tailed Shimoidara–Hasegawa
test (p=0.045), supporting the hypothesis of multiple and in-
dependentintroductionsofgenotype1lineagesintotheAmazon
region.
Patterns of evolutionary change
Since the phylogenetic analysis suggested independent
introductions of B19 virus into the Amazon, we divided the
sequences into 2 groups – Belém and ‘Cosmopolitan’ (i.e., all
sequences from elsewhere, including São Paulo) – to determine
whether they differ in evolutionary dynamics. First, to obtain a
graphical overview of the data, we explored patterns of genetic
divergence at different codon positions. This revealed far higher
genetic distances at the 1st and 2nd codon positions of B19
isolates in Belém compared to the Cosmopolitan population, as
indicated by the slopes of regressions in Fig. 2. In marked
contrast, the 3rd codon positions of B19 isolates in Belém had
lower levels of genetic divergence than that of the Cosmopolitan
sequences (Fig. 2). This highly distinctive pattern indicates that
B19 in Belém is characterized by a higher rate of nonsynon-
ymous substitution than those viral isolates sampled from other
locations,butthatlevelsofgeneticdiversityatsynonymoussites
are consistently reduced.
Fig. 2. Plot comparing genetic distances at 1st and 2nd with 3rd codon positions versus total distance among common regions of VP1 and VP2 gene of human
parvovirus B19. Regression slopes indicated that sequences from Belém had a higher rate of change at 1st and 2nd codon positions compared to the cosmopolitan
sequences (i.e., all other sequences not from Belém) but a lower rate of change at 3rd codon position.
283R.B. de Freitas et al. / Virology 369 (2007) 281–287

Page 4

Selection pressures on B19
To further explore the evolutionary processes acting on B19
in Belém, we estimated values of Tajima's D statistics that tests
the null hypothesis of a mutation–drift equilibrium (i.e., of
neutral evolution). Negative values of Tajima's D suggest that
positive selection may be acting on a population. Again, a
strikinglydifferent pattern was observed between Belém and the
Cosmopolitan populations: (i) Cosmopolitan isolates, including
sequences from Erdman et al. (1996) and those from the
University Hospital of São Paulo City, D=−0.8776, S.E.=
−0.0455 (87 sequences, 10 segregating amino acid sites); and
(ii) Belém isolates, D=−4.1684, S.E.=−1.6532 (47 sequences,
24 segregating amino acid sites). The strongly negative D value
obtained for the Belém sequences further supports the idea that
theincreaseingeneticdiversity at1stand2nd codonpositions is
due to a departure from neutral evolution. Similar results were
obtained through an analysis of the relative rates of synonymous
(dS) and nonsynonymous (dN) substitution per site. In this case,
dN was 3.9 times greater in Belém (dN=0.0043±0.0011)
compared with sequences from elsewhere (dN=0.0011±
0.003). However, there was no significant difference in dS
among populations; Cosmopolitan, dS=0.0548±0.0105 and
Belém, dS=0.0490±0.0139.
We next investigated whether selection pressures acting on
different codon sites could be responsible for the difference
among the Belém and Cosmopolitan populations. Using the
CODEML method, we found no evidence for positive selection
atanyindividualcodonineithertheBelémorCosmopolitandata
sets (in that neutral models of codon evolution could not be
rejected by those more complex models that support positive
selection). However, this analysis also indicated that 15 codons
in the B19 isolates from Belém had an elevated dN/dSratio, with
a mean ratio of 1.50. Further, six of these sites were detected as
being subject to weak positive selection in HyPhy (Bayes factor
values b20) with a mean dN/dS ratio of 1.70. In contrast,
CODEML detected 4 sites in the Cosmopolitan group with an
elevated mean dN/dSratio of 2.25, three of which were also
observed in HyPhy, but with a lower mean dN/dSratio of 1.10.
Considered together, the data revealed an elevation in
nonsynonymous rates among the B19 lineages sampled from
Belém, several of which could possibly be under positive
selection, although this latter conclusion will need to be
confirmed with a larger sample of sequences. Moreover, it
remains to be determined if similar results of higher nonsynon-
ymous rates among samples from Belém are also to be found in
the entire VP1 and VP2 genes, where immune-dominant
epitopes are known to be present.
Fig. 1 depictsthemost parsimoniousreconstructions (MPRs)
on the MAP tree, showing the excess of nonsynonymous
changes along the independent radiation of lineages in Belém.
Sites 4, 27, 72 and 105 changed independently within the Belém
clusters, while site 155 was shared by a group of 3 sequences,
suggestive of possible propagation among the human popula-
tion. A single site coding a lysine at codon 19 (shown in yellow
in Fig. 1) was shared by both Belém (once) and Cosmopolitan
(twice) sequences. Other than codon 19, no other sites in the
Cosmopolitan sequences appeared more than once.
Population dynamics of B19
To investigate other aspects of the population dynamics of
B19 in the Belém and Cosmopolitan populations, we estimated
several population parameters using the Bayesian coalescent
method available in the BEAST package. This analysis revealed
that B19 appears to experience equivalent evolutionary rates in
all populations, growing according to a logistic model, with
similar growth rates and doubling times (λ) (Table 1). For all
groups, rates of nucleotide substitution exhibited overlapping
HPD values, with an average rate of 2.572×10−4nucleotide
substitutions per site, per year (subs/site/year) for the Belém
lineages, of 1.9×10−4subs/site/year for the Cosmopolitan
Table 1
Bayesian estimates of population dynamic and evolutionary parameters for B19 in samples from around the world (Cosmopolitan), Belém (Amazon) and the city of
São Paulo
Cosmopolitan BelémSão Paulo
Sample size
Date range
Best-fit demographic model
Mean ka
95% HPDbk
39
1973–2003
Logistic
2.884e−4
Lower=2.224e−5
Upper=5.641e−4
66.5
Lower=30
Upper=166.093
0.353
Lower=4.749e−3
Upper=0.887
1.9636
Lower=145.95
Upper=0.7814
46
1995–2005
Logistic
2.572e−4
Lower=7.467e−5
Upper=4.507e−4
45.104
Lower=17.756
Upper=89.528
0.186
Lower=3.021e−4
Upper=0.405
3.7266
Lower=2294.43
Upper=1.7115
48
1990–2003
Logistic
8.809e−5
Lower=2.683e−5
Upper=1.681e−5
73.223
Lower=23.918
Upper=143.451
7.168e−2
Lower=1.053e−5
Upper=0.193
9.6700
Lower=65,825
Upper=3.59
95% mean age (years)
95% HPD age (years)
Mean growth ratec
95% HPD growth rate
Mean epidemic doubling time (l) in years
95% HPD doubling time (l) in years
aRate of nucleotide substitution per site, per year.
bHPD=high probability density.
cNumber of new infections per individual, per year.
284R.B. de Freitas et al. / Virology 369 (2007) 281–287

Page 5

isolates, and of 8.809×10−5subs/site/year for the São Paulo
viruses. These results are similar to those estimated previously
for B19 (Shackelton and Holmes, 2006).
The most striking result of this analysis is that those B19
lineages independently introduced into Belém have all
experienced an elevated rate of nonsynonymous evolution.
Since we did not find any features of viral epidemiological
dynamics (i.e., population growth rates) consistent with this
pattern, we argue that unique differences in the host population,
particularly in terms of selection pressure, represent the most
likely explanation.
TheprevalenceofB19inSouthAmericamaybearound50%
(Abarca et al., 2002; Anderson et al., 1986; Cohen and Buckley,
1988; Gay et al., 1994; Nunoue et al., 1985), which are close to
those obtained globally. B19 is endemic in the Amazon region
and may undergo fluctuations in incidence in intervals ranging
from 3 to 5 years (Freitas et al., 1993). More notably, there was
an increase in prevalence from 43% in 1990 to 85% in 2002 in
Belém (Freitas et al., 2002). Consequently, the current
prevalence in Belém is roughly 70% higher than in other
regions. The clinical resolution of acute B19 infection is
associated with the emergence of persistent antiviral IgG,
mainly directed to both B19 structural proteins, VP1 and VP2
(Azzi et al., 2004; Rosenfeld et al., 1994). In addition, B19
infectioninducesapersistentactivatedCD8+Tcellresponse(Isa
etal.,2005).BoththepresenceofpreviousIgGandastrongCTL
may play a significant role upon reinfection in immune-
competent individuals (Nikkari et al., 1996; Zakrzewska et al.,
2001), and B19 has been shown to reinfect and persist in
immune-compromisedpatients(Flunkeretal.,1998).Therefore,
a plausible evolutionary scenario for B19 in the Amazon would
entail the reintroduction of viral lineages, which must then
negotiate, by way of adaptive evolution, higher levels of herd
immunity; in other words there is an increased necessity to
infect, and reinfect, a small susceptible population. Under this
type of selective pressure, B19 lineages present in Belém would
show an increase in the number of nonsynonymous changes
through adaptive evolution to facilitate serial re-infections.
Further, since the host population in the Amazon is smaller than
that of other regions providing it with new B19 lineages, a
relatively small web of transmission would prevent genetic
diversity from accumulating at synonymous positions follow-
ing repeated selective sweeps (and reflected in measures of
genetic diversity at 1st and 2nd codon positions). Hence, we
propose that Belém acts as an evolutionary “pressure pan” for
B19. It is clearly important to determine whether B19 is under
similar pressures in other regions of the world possessing
epidemiological conditions similar to those of Belém, or if other
endemic pathogens in the Amazon exhibit similar evolutionary
patterns.
Materials and methods
Clinical samples
The clinical samples used in this study were collected by the
Department of Virology, Evandro Chagas Institute, Belém
(n=46) between 1995 and 2005, and at the University of São
Paulo Hospital and several other hospitals in the city (n=48)
between 1990 and 2003. Blood samples were collected during
acutephaseofdiseasebyantecubitalvenepuncture.Weincluded
patients diagnosed with: (i) exanthematous diseases, (ii) acute
arthropathy, (iii) hematological disorders, (iv) encephalitis, (v)
myocarditis and, (vi) systemic lupus erythematosus. All sera
were frozen at −20 °C until processed anonymously upon the
approval by an ethics committee at both, Evandro Chagas
Institute and University of São Paulo (Resolução No. 196,
1996).
DNA extraction, PCR and semi-nested PCR
Viral DNA was extracted from sera samples by the phenol–
chloroform method (Umene and Nunoue, 1993). The PCR and
semi-nested PCR was performed according to Durigon et al.
(1993). The expected molecular size was 563 and 476 bp for the
first and semi-nested PCR, respectively.
Sequencing
The nucleotide sequences of the amplified fragment were
determined by the dideoxy chain termination method using the
sequencing ABI PRISM Dye Terminator kit (Applied Biosys-
tems, Foster City, CA) and resolved in an ABI 3100 DNA
sequencer.Bothcomplementarystrandsweredirectlysequenced
using the same primers used in semi-nested PCR.
Sequences used in this study
As well as the sequences newly generated here (with
GenBank accession numbers in parenthesis) (EF089178,
EF089179, EF089180, EF089181, EF089182, EF089183,
EF089184, EF089185, EF089186, EF089187, EF089188,
EF089189, EF089190, EF089191, EF089192, EF089193,
EF089194, EF089195, EF089196, EF089198, EF089199,
EF089200, EF089201, EF089202, EF089203, EF089204,
EF089205, EF089206, EF089207, EF089208, EF089209,
EF089210, EF089211, EF089212, EF089213, EF089214,
EF089215, EF089216, EF089217, EF089218, EF089219,
EF089220, EF089221, EF089222, EF089223, EF089224,
EF154284, EF154285, EF154286, EF154287, EF154288,
EF154289, EF154290, EF154291, EF154292, EF154293,
EF154294, EF154295, EF154296, EF154297, EF154298,
EF154299, EF154300, EF154301, EF154302, EF154303,
EF154304, EF154306, EF154308, EF154309, EF154310,
EF154311, EF154312, EF154313, EF154314, EF154315,
EF154316, EF154317, EF154318, EF154319, EF154320,
EF154321, EF154322, EF154324, EF154325, EF154326,
EF154327, EF154328, EF154329, EF154330, EF154331,
EF154332, EF154333, EF491000), we collected a number of
reference sequences from GenBank to expand the scope of our
study. These latter sequences had following accession numbers:
U31358, U38506, U38507, U38509, U38510, U38511,
U38512, U38513, U38514, U38515, U38516, U38517,
U38518, U38546, U53600, AB030673, AB030693,
285R.B. de Freitas et al. / Virology 369 (2007) 281–287

Page 6

AB030694, AB126262, AB126263, AB126264, AB126265,
AB126267, AB126268, AB126271, AF162273, AF161223,
AF161224, AF161225, AF161226, AY386330, AY028237,
M24682, M13178, Z68146, Z70560, Z70528, Z70599,
NC000883. A number of outgroup sequences were also utilized;
genotype 2 (AY044266, AY064475), genotype 3 (AY083234,
NC004295) and the Simian parvovirus (U26342). This repre-
senteda total of44 sequencesadditional sequences. Therefore, a
total of 138 sequences were compiled for analysis.
Evolutionary analysis
To compare B19 from different regions we inferred maxi-
mum likelihood (ML) phylogenetic trees using PAUP∗ (Swof-
ford, 2003), using the best-fit evolutionary model as determined
by Modeltest 3.7 (Posada and Crandall, 1998) which was found
tobeHKY+Γ(withatransition/transversionratioof3.213anda
shape parameter, α, of 0.493). Using this model, pairwise
geneticdistances wereestimated for1stand2ndcodonpositions
taken together(mostly nonsynonymous changes)and forthe3rd
codon position separately (mostly synonymous changes), and
plotted against total nucleotide distance.
We used a number of methods to determine the selective
pressures acting on B19 virus. First, we obtained pairwise esti-
mates of the value of the Tajima's D statistic, and rates of
nonsynonymous(dN)andsynonymous(dS)substitutionspersite
(ratio dN/dS) using the MEGA v3.1 package (Kumar et al.,
2004). For the analysis of dN/dS we used the codon-based
method of Li–Wu–Luo, with standard errors estimated after
1000 bootstrap replicates. We also inferred site-specific dNand
dSusing both the CODEML (Yang, 1997) and HyPhy (Pond et
al., 2005) programs. In CODEML we employed models 0
(invariant), 1 (neutral), 2 (negative selection), 3 (positive
selection), 7 (ten categories of rates for codon sites) and 8
(eleven categories of rates for codon sites) and with HyPhy we
used the “MG94xHKY85x3_4x2_Rates” allowing for rate he-
terogeneity in the model parameters and 4 categories per rate
parameter.
To estimate rates of evolutionary change of B19 in both
Brazilian populations we employed the Bayesian–Markov
Chain–Monte Carlo (MCMC) method available in the BEAST
v1.3package(DrummondandRambaut,2003).Sequenceswere
dated according to the year of sampling and run with a chain
length of 40 million under the HKY+Γ substitution model,
under the assumption of a relaxed molecular clock and a variety
of different models of demographic history; constant population
size, exponential population growth, logistic growth, expansion
growth, and a Bayesian skyline plot which gives a piece-wise
graphical picture of demographic history. Only the parameters
for the best model (i.e., the one showing the maximum a
posteriori probability) are shown here. Confidence intervals
were given by the 95% highest probability density (HPD). The
BEASTanalysiswasalsousedtoinferthemaximumaposteriori
(MAP) tree, which branch-lengths scaled in time. Finally, the
most parsimonious reconstructions (MPR) of amino acid
changes along lineages of the B19 phylongey were estimated
using the MacClade v. 4.07 program (Maddison and Maddison,
2003) using both the ML and MAP trees. MPRs of sites with
elevated dN/dSwere then mapped to specific lineages.
Acknowledgments
This study was supported by CNPq, MCT and FAPESP
projects 00/04205-06 and 00/11511-06-VGDN program. ELD
and PMAZ hold a PQ-CNPq scholarship. CMR and FLM hold a
CAPES scholarship.
References
Abarca, K., Cohen, B.J., Vial, P.A., 2002. Seroprevalence of parvovirus B19 in
urban Chilean children and young adults, 1990 and 1996. Epidemiol. Infect.
128, 59–62.
Anderson, L.J., Tsou, C., Parker, R.A., Chorba, T.L., Wulf, H., Tattersall, P.,
Mortimer, P.P., 1986. Detection of antibodies and antigens of human
parvovirus B19 by enzyme-linked immunosorbent assay. J. Clin. Microbiol.
24, 522–526.
Azzi, A., Manaresi, E., Zakrzewska, K., DeSantis, R., Musiani, M., Zerbini, M.,
2004. Antibody response to B19 parvovirus VP1 and VP2 linear epitopes in
patients with haemophilic arthritis. J. Med. Virol. 72, 679–682.
Brown, K.E., Young, N.S., Liu, J.M., 1994. Molecular, cellular and clinical
aspects of parvovirus B19 infection. Crit. Rev. Oncol. Hematol. 16 (1),
1–31.
Cleaveland, S., Laurenson, M.K., Taylor, L.H., 2001. Diseases of humans and
their domestic mammals: pathogen characteristics, host range and the risk of
emergence. Philos. Trans. R. Soc. Lond., B 356, 991–999.
Cohen, B.J., Buckley, M.M., 1988. The prevalence of antibody to human
parvovirus B19 in England and Wales. J. Med. Microbiol. 25, 151–153.
Drummond, A.J., Rambaut, A., 2003. BEAST v1.0, Available from http://
evolve.zoo.ox.ac.uk/beast/.
Durigon, E.L., Erdman, D.D., Gary, G.W., Pallansch, M.A., Torok, T.J.,
Anderson, L.J., 1993. Multiple primer pairs for polymerase chain reaction
(PCR) amplification of human parvovirus B19 DNA. J. Virol. Methods 44,
155–165.
Erdman, D.D., Durigon, E.L., Wang, Q.Y., Anderson, L.J., 1996. Genetic
diversity of human parvovirus B19: sequence analysis of the VP1/VP2 gene
from multiple isolates. J. Gen. Virol. 77, 2767–2774.
Flunker, G., Peters, A., Wiersbitzky, S., Modrow, S., Seidel, W., 1998. Persistent
parvovirus B19 infections in immunocompromised children. Med. Micro-
biol. Immunol. 186, 189–194.
Freitas, R.B., Miranda, M.F., Shirley, J., Tudor, R., Desselberger, U., Linhares,
A.C., 1993. Parvovirus B19 antibodies in sera of patients with unexplained
exanthemata from Belem, Para, Brazil. Mem. Inst. Oswaldo Cruz 88,
497–499.
Freitas, R.B., Monteiro, T.A., Silva Filho, M.G., Linhares, A.C., 2002.
Association between human parvovirus B19 and arthropathy in Belem,
Para, North Brazil. Rev. Inst. Med. Trop. Sao Paulo 44, 17–22.
Gay, N.J., Hesketh, L.M., Cohen, B.J., Rush, M., Bates, C., Morgan-Capner, P.,
Miller, E., 1994. Age specific antibody prevalence to parvovirus B19: how
many women are infected in pregnancy? Commun. Dis. Rep. CDR Rev. 4
(9), R104–R107.
Isa, A., Kasprowicz, V., Norbeck, O., Loughry, A., Jeffery, K., Broliden, K.,
Klenerman, P., Tolfvenstam, T., Bowness, P., 2005. Prolonged activation of
virus-specific CD8+T cells after acute B19 infection. PLoS Med. 2, e343.
Kumar, S., Tamura, A., Nei, M., 2004. MEGA 3: integrated software for
molecular evolutionary genetic analysis and sequence alignment. Brief
Bioinform. 5, 150–160.
Maddison, D.R., Maddison, W.P., 2003. MacClade 4: Analysis of Parsimony
and Character Evolution, 4.06th edition. Sunderland, Sinauer Associates.
Nikkari, S., Lappalainen, H., Saario, R., Lammintausta, K., Kotilainen, P.,
1996. Detection of parvovirus B19 in skin biopsy, serum, and bone marrow
of a patient with fever, rash, and polyarthritis followed by pneumonia,
pericardial effusion, and hepatitis. Eur. J. Clin. Microbiol. Infect. Dis. 15,
954–957.
286R.B. de Freitas et al. / Virology 369 (2007) 281–287

Page 7

Nunoue, T., Okochi, K., Mortimer, P.P., Cohen, B.J., 1985. Human parvovirus
(B19) and erythema infectiosum. J. Pediatr. 107, 38–40.
Pattison, J.R., 1990. The pathogenesis of diseases associated with B19 virus.
Behring-Inst.-Mitt. 85, 55–59.
Pond, S.L., Frost, S.D., Muse, S.V., 2005. HyPhy: hypothesis testing using
phylogenies. Bioinformatics 21 (5), 676–679.
Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA
substitution. Bioinformatics 14, 817–818.
Rosenfeld, S.J., Young, N.S., Alling, D., Ayub, J., Saxinger, C., 1994. Subunit
interaction in B19 parvovirus empty capsids. Arch. Virol. 136, 9–18.
Shackelton, L.A., Holmes, E.C., 2006. Phylogenetic evidence for the rapid
evolution of human B19 erythrovirus. J. Virol. 80, 3666–3669.
Shackelton, L.A., Parrish, C.R., Truyen, U., Holmes, E.C., 2005. High rate of
viral evolution associated with the emergence of carnivore parvovirus. Proc.
Natl. Acad. Sci. U.S.A. 102, 379–384.
Swofford, D.L., 2003. PAUP∗. Phylogenetic Analysis Using Parsimony (∗and
Other Methods) Version 4. Sinauer Associates, Sunderland, Massachusetts.
Umene, K., Nunoue, T., 1993. Partial nucleotide sequencing and
characterization of human parvovirus B19 genome DNAs from damaged
human fetuses and from patients with leukemia. J. Med. Virol. 39,
333–339.
Yang, Z., 1997. PAML: a program package for phylogenetic analysis by
maximum likelihood. Comput. Appl. Biosci. 13, 555–556.
Zakrzewska, K., Azzi, A., De Biasi, E., Radossi, P., De Santis, R., Davoli, P.G.,
Tagariello, G., 2001. Persistence of parvovirus B19 DNA in synovium of
patients with haemophilic arthritis. J. Med. Virol. 65, 402–407.
287 R.B. de Freitas et al. / Virology 369 (2007) 281–287