Internally deleted WNV genomes isolated from exotic birds in New Mexico: function in cells, mosquitoes, and mice.

Abstract

Most RNA viruses exist in their hosts as a heterogeneous population of related variants. Due to error prone replication, mutants are constantly generated which may differ in individual fitness from the population as a whole. Here we characterize three WNV isolates that contain, along with full-length genomes, mutants with large internal deletions to structural and nonstructural protein-coding regions. The isolates were all obtained from lorikeets that died from WNV at the Rio Grande Zoo in Albuquerque, NM between 2005 and 2007. The deletions are approximately 2kb, in frame, and result in the elimination of the complete envelope, and portions of the prM and NS-1 proteins. In Vero cell culture, these internally deleted WNV genomes function as defective interfering particles, reducing the production of full-length virus when introduced at high multiplicities of infection. In mosquitoes, the shortened WNV genomes reduced infection and dissemination rates, and virus titers overall, and were not detected in legs or salivary secretions at 14 or 21 days post-infection. In mice, inoculation with internally deleted genomes did not attenuate pathogenesis relative to full-length or infectious clone derived virus, and shortened genomes were not detected in mice at the time of death. These observations provide evidence that large deletions may occur within flavivirus populations more frequently than has generally been appreciated and suggest that they impact population phenotype minimally. Additionally, our findings suggest that highly similar mutants may frequently occur in particular vertebrate hosts.

Full text

Page 1

Internally deleted WNV genomes isolated from exotic birds in New Mexico:
Function in cells, mosquitoes, and mice
Kendra N. Peskoa,⁎, Kelly A. Fitzpatricka, Elizabeth M. Ryanb, Pei-Yong Shic, Bo Zhangd, Niall J. Lennonb,
Ruchi M. Newmanb, Matthew R. Hennb, Gregory D. Ebela,⁎⁎
aDepartment of Pathology, University of New Mexico School of Medicine, Albuquerque, NM, USA
bBroad Institute of MIT and Harvard, Cambridge, MA, USA
cWadsworth Center, New York State Department of Health, Albany, NY, USA
dState Key laboratory of Virology, Wuhan Institute of Virology, Chinese Academy of Sciences, China
a b s t r a c t a r t i c l ei n f o
Article history:
Received 8 November 2011
Returned to author for revision
13 December 2011
Accepted 26 January 2012
Available online 23 February 2012
Keywords:
West Nile virus
Genetic variation
Defective interfering particles
Most RNA viruses exist in their hosts as a heterogeneous population of related variants. Due to error prone
replication, mutants are constantly generated which may differ in individual fitness from the population as
a whole. Here we characterize three WNV isolates that contain, along with full-length genomes, mutants
with large internal deletions to structural and nonstructural protein-coding regions. The isolates were all
obtained from lorikeets that died from WNV at the Rio Grande Zoo in Albuquerque, NM between 2005 and
2007. The deletions are approximately 2 kb, in frame, and result in the elimination of the complete envelope,
and portions of the prM and NS-1 proteins. In Vero cell culture, these internally deleted WNV genomes
function as defective interfering particles, reducing the production of full-length virus when introduced at
high multiplicities of infection. In mosquitoes, the shortened WNV genomes reduced infection and dissemina-
tion rates, and virus titers overall, and were not detected in legs or salivary secretions at 14 or 21 days post-in-
fection.Inmice,inoculationwithinternallydeletedgenomesdidnotattenuatepathogenesisrelativetofull-length
or infectious clone derived virus, and shortened genomes were not detected in mice at the time of death. These
observations provide evidence that large deletions may occur within flavivirus populations more frequently
than hasgenerally been appreciatedandsuggestthat theyimpactpopulation phenotypeminimally.Additionally,
our findings suggest that highly similar mutants may frequently occur in particular vertebrate hosts.
© 2012 Elsevier Inc. All rights reserved.
Introduction
West Nile virus, (WNV), (Flaviviridae: Flavivirus) is a mosquito
borne virus now endemic to the United States. It circulates between
avian hosts and various mosquito vectors, mostly in the genus Culex.
Since its first detection in the United States in 1999, WNV has dis-
persed throughout the western hemisphere (Artsob et al., 2009;
Komar and Clark, 2006; Murray et al., 2010). WNV is an RNA virus,
and as such accumulates mutations readily during replication because
of the lack of a proof reading mechanism in the virus-encoded RNA
dependent RNA polymerase, that can lead to the production of
defective RNA genomes (Holland et al., 1982; Holmes, 2009). WNV
and other arthropod-borne viruses (arboviruses) seem to have slower
rates of evolution than single-host RNA viruses, which may be a result
of evolutionary constraint imposed by the requirement for replication
in both vertebrate and invertebrate cells (Jenkins et al., 2002). Sup-
porting this, arbovirus populations in nature are subject to strong pu-
rifying selection (Weaver, 2006). Nonetheless, WNV and other
arboviruses exist in mosquito and vertebrate hosts as a genetically
heterogeneous mixture, with many individual mutants making up a
genetically diverse population (Aaskov et al., 2006; Jerzak et al.,
2005). Error-prone replication and consequent intrahost viral genetic
diversity are, therefore, a central feature of the population biology of
RNA viruses, including some arboviruses.
Defective interfering particles (DIPs) are subgenomic viral parti-
cles that arise during the course of viral infection, replicate through
complementation with full-length homologous viruses, and have an
inhibitory effect on virus growth (Thompson et al., 2009). DIPs have
been detected after serial passage of the mosquito borne flaviviruses,
WNV and Japanese encephalitis virus (JEV) in cell culture (Brinton,
1983, 2001; Debnath et al., 1991; Tsai et al., 2007; Yoon et al.,
2006). Until recently, whether they exist in naturally acquired infec-
tions, and their role in shaping the outcome of virus–host interactions
has been unclear. Importantly, isolates from patients with acute
Virology 427 (2012) 10–17
⁎ Correspondenceto:K.N.Pesko,DepartmentofEcologyandEvolution,YaleUniversity,
New Haven, CT 06520, USA.
⁎⁎ Correspondence to: G.D. Ebel, Department of Microbiology, Immunology and Pa-
thology, College of Veterinary Medicine and Biomedical Sciences, Colorado State Uni-
versity, Fort Collins, CO, USA.
E-mail addresses: Kendra.Pesko@yale.edu (K.N. Pesko), Gregory.Ebel@colostate.edu
(G.D. Ebel).
0042-6822/$ – see front matter © 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.virol.2012.01.028
Contents lists available at SciVerse ScienceDirect
Virology
journal homepage: www.elsevier.com/locate/yviro

Page 2

Dengue infections were recently shown to contain deletion mutants
made up of the 5′ and 3′ untranslated regions of the genome (with
the entire protein-coding sequence deleted), which behave in vitro
as DIPs (Li et al., 2011). Isolates of another member of Flaviviridae,
hepatitis C virus, containing large deletions to structural coding re-
gions have been made from chronically infected individuals. There-
fore, DIPs may occur in flavivirus infections more frequently than
previously appreciated, and their presence within a virus population
may significantly impact the outcome of infection in terms of trans-
mission and pathogenesis.
Recently, we identified three WNV populations from lorikeets that
died at the Rio Grande Zoo in two separate years that contain mutants
with large internal deletions. We therefore sought to characterize
these mutants and determine whether they function as WNV DIPs
in vitro. We then determined the extent to which they influence
transmission by mosquitoes and pathogenesis in mice to evaluate
the possibility that they may contribute to the WNV perpetuation
and/or disease outcomes in vertebrates. Our results suggest that nat-
urally occurring WNV deletion mutants function as DIPs, and that
these DIPs interfere with virus transmission by mosquitoes. However,
the WNV DIPs we examined have surprisingly little impact on virus
pathogenesis in mice.
Results
Identification of deletion mutants
RT-PCR targeting structural protein-coding regions of the WNV
genome (Table 1: DH1F/DH1R) resulted in amplicons approximately
2 kb smaller than expected for samples from three birds that died of
WNV infection (Fig. 1A). All three samples were taken from Rainbow
Lorikeets (Trichoglossus haematodus) as confirmed by sequencing of
the cytochrome oxidase (COI) gene. Isolates were made by inoculat-
ing Vero cells with kidney tissue homogenate of birds found dead at
the Albuquerque Zoo in August of 2005 and 2007. Smaller genomes
were confirmed directly using northern analysis of RNAs from one
isolate (2774) infecting Vero cells, at 48 h post-infection and com-
pared to infectious clone derived virus infected cells harvested 48 h
post-infection. A probe to the NS-5 coding region annealed to a single
band for RNA extracted from infectious clone derived virus infected
cells but identified two bands in RNA extracted from cells infected
with isolate 2774 (Fig. 1B), in contrast, a probe to the envelope coding
region resulted in a single band produced for each RNA sample
(Fig. 1C). Thus, shortened genomes could be detected directly with
probes to NS-5 and appeared to make up around half of the viral
RNA present within infected cells, but only full-length genomes
were detected with envelope probe, as expected.
Genomic location of internal deletions
PCR amplicons from bird kidneys were cloned and sequenced to
reveal large deletions within the structural and NS-1 coding regions.
Deletions were in-frame and resulted in the loss of 3′ portions of
prM, complete envelope, and 5′ portions of NS-1 encoding RNA
(Fig. 1D). All eight clones sequenced from one kidney sample
(2774) contained the same deletion, whereas the other two samples
(3336, 3337) had two types of deletions present, labeled A and B in
Fig. 1D. For 2774, the same deletion mutant was detected by cloning
and sequencing of RT-PCR amplicons from the original kidney sample
as well as Vero passages one and two. A probe was designed across
the junction produced from the prM-NS-1 fusion that was detected
in 2774, and was used in a qRT-PCR assay to directly detect the pres-
ence of this particular deletion mutant (Table 1). This assay estimated
4.6×103mutant genome copies/0.1 ml in the original tissue sample
of 2774, 2.7×105mutant genome copies/0.1 ml in Vero passage 1
2774 and 2.6×106mutant genome copies/0.1 ml in Vero passage 2
2774. Plaque assays estimated 2.4×104pfu/0.1 ml in 2774 original
kidney tissue, 4.7×104pfu/0.1 ml in Vero passage one, and 3.7×106
pfu/0.1 ml in Vero passage 2.
Comparison of full genome sequences
Full genome sequencing of isolate 2774 revealed a predicted
amino acid substitution, 2K-V9M, that had previously been detected
in virus that was capable of bypassing superinfection exclusion (Zou
et al., 2009b). Sequencing of the original sample to confirm this sub-
stitution showed a polymorphism at this site, both A and G appeared
to be present at nucleotide (nt) position 6871 (determining 2K-V9 or
2K-M9, respectively). Sequencing of the other two isolates, 3336 and
3337, showed the same polymorphism present. To further character-
ize the presence of this polymorphism, RNA was extracted from indi-
vidual plaques of 2774vp1, 2774vp3, 3336 (unpassed), and 3337
(unpassed) and subject to sequencing across the location of this mu-
tation (nt.6871). The 6871G variant was more commonly detected
and made up 7/10 plaques from 2774vp1, 2774vp3, and 3336, and
10/10 of the plaques from 3337. The 6871A variant was present as
3/10 plaques from 2774vp1, 2774vp3, and 3336, but not detected in
any of the ten plaques from 3337, indicating it could be present
only in the deletion mutant portion of this isolate. Nested PCR using
primers designed to amplify solely internally deleted WNV genomes
showed that 6871A was present for 2774. Bayesian analysis of partial
genome sequences of NM isolates 3337 (NM05) and 2774 (NM07)
did not group these two isolates together relative to other genomes
sampled from the southwestern United States, indicating the muta-
tion leading to this polymorphism may have arisen independently
for each isolate (Fig. 2). Three of the consensus sequences shown in
this phylogenetic tree (CA03, CA04, and NM07) had an M encoded
at the 2K-9 site, whereas the remainder consensus sequences encode
a V. However, it is currently not clear whether this variant may be
present in other strains as a minority subpopulation.
Isolates inhibit production of full-length virus in Vero cells
To assess whether deletion mutants behave as defective interfer-
ing particles (DIPs), we evaluated full-length virus production from
Vero cells infected over a range of multiplicity of infection (MOI)
with isolate 2774 and with infectious clone derived WNV (WT).
Full-length virus production was measured by plaque assay of cell
culture supernatants after 3 days of growth in Vero cells. The full-
length virus yield from each supernatant varied with virus isolate
and viral dose, and these two factors had a significant interaction
influencing virus production (Fig. 3A, F=28.51, p=0.0014). Super-
natants from 2774 infected cells were tested by RNA extraction and
quantitative RT-PCR to quantify the presence of deletion mutant
Table 1
Primer and probe sequences, DH = demi-hemi; numbers refer to location in full
genome; qPCR MJ = quantitative reverse transcription polymerase chain reaction
across mutant junction.
NameSequence
DH1F
DH1R
10296rT7
ACTACTTCGGCTGTGTGAGCT
ATGGGCCCTGGTTTTGTGTCT
CTA ATA CGA CTC ACT ATA GGG AGA TCC GAT GAT TGC
TCT GAC TT
CTA ATA CGA CTC ACT ATA GGG AGA CGT CCT TCA CTG
CTT CCC AGA
AGATCCTGTGTTCTCGCACCA
ATC CGA GTG CTG GTG AGA CCA AAT
TTC CAA GGG AAG GTG ATG ATG ACG
/56-FAM/GGA AAG AAC /ZEN/CTA AGC TTA GAA GTG
GAG GA/3IABkFQ/
2666rT7
3′ end
qRT-PCR MJ F
qRT-PCR MJ R
qRT-PCR MJ probe
11
K.N. Pesko et al. / Virology 427 (2012) 10–17

Page 3

genomes at 3 dpi (Fig. 3B). No internally deleted genomes were
detected in supernatants of 2774 at 3 dpi at the lowest MOI (0.001),
and only one of three replicates at MOI 0.01 contained detectable de-
letion mutant, but the remaining treatments produced comparable
amounts of internally deleted genomes (Fig. 3B).
Deletion mutants decrease titers and spread of virus in mosquitoes
Mosquitoes were given blood meals containing either 2774 after
passage two times on Vero cells and thus still containing deletion mu-
tant and full-length virus (DM+FL), or a plaque purified isolate from
2774 containing full-length virus only (FL), as confirmed by RT-PCR
and Northern analysis. From Northern blot and qRT-PCR analysis, it
appeared that the DM+FL treatment contained an equal proportion
of DM:FL. Both blood meals contained approximately 2×107PFU/ml,
as determined by plaque assay, thus were matched by FL virus capa-
ble of productive infection, but not by genomes or particles present.
Mosquitoes fed on full-length virus alone had significantly higher
infection and dissemination rates (Fig. 4A, infection rate: X2=10.44,
p=0.0012; dissemination rate: X2=5.26, p=0.022), and log trans-
formed body titers (Fig. 4B, t=3.189, p=0.0018; difference=0.69
log10±0.21). Deletion mutants were detected by qRT-PCR in bodies
of 7/44 and 5/26 mosquitoes tested on days 14 and 21 post-infection
(Table 2), but not in legs or salivary secretions at either time point.
Deletion mutants do not attenuate morbidity and mortality in mice
To assess whether deletion mutants could influence morbidity and
mortality in vertebrates, we tested two infecting viral doses in C3H
mice, which are susceptible to mortality from infection with WNV,
and one dose in C57Bl/6 mice, which are more resistant to mortality
(Brown et al., 2007). Inoculation with internally deleted genomes
did not attenuate morbidity or mortality in either mouse strain
(Fig. 5). Internally deleted genomes were not detected in mice either
at time of death, or at 28 days post-infection when surviving mice
were sacrificed. Mortality rates were not significantly different by
inoculating strain, however, more C3H mice inoculated with deletion
mutant containing virus died than those inoculated with infectious
clone derived or plaque purified full-length virus (Fig. 5). C3H mice
inoculated with plaque purified full-length or deletion mutant con-
taining 2774 stocks lost significantly more weight than infectious
clone exposed and mock mice in both trials (High dose, F=57.2,
pb0.00001, Bonferroni multiple comparison t-tests were significant
for all comparisons, pb0.05; Low dose, F=55.9, pb0.00001, Bonferroni
multiple comparison t-tests were significant for all comparisons at pb
0.05, except for deletion mutant vs. full length: t=0.56, ns). Infection
was confirmed in surviving mice by plaque reduction neutralization
tests.
Discussion
As part of our ongoing studies of WNV population genomics in
North America, we identified a group of virus strains collected in
two separate years from the same location and host type that pro-
duced atypical RT-PCR products. Cloning and sequencing these prod-
ucts revealed a group of highly similar large in-frame deletions that
removed a portion of the prM, all of the E, and a portion of the NS1
coding sequences. RT-PCR analysis of RNA extracted from primary
kidney tissue indicated that full-length WNV was present in the
infecting population along with the shortened genomes. Northern
analysis of RNA produced by one of these isolates confirmed that
both full-length and shortened RNAs are produced during infection
of Vero cells. Further, after a single passage in Vero cells, the short-
ened genomes are retained as a portion of the WNV population.
Fig. 1. Identification of deletion mutants. A) PCR amplicons produced by primers to the structural coding regions of WNV genome to each of three samples, WT target size is 2.5 kb,
actual size observed from isolates is 0.5–0.7 kb, B) Northern blot of RNAs from cells infected with infectious clone derived virus (WT) or isolate 2774 from infected birds, and probe
to envelope. C) Northern analysis of RNAs as in B but with probe to NS-5. D) Genomic locations of deletions found in isolates 2774, 3336, and 3337. At top, diagram of West Nile
virus genome, showing structural protein-coding regions for a nucleocapsid (C), membrane precursor (prM), and envelope (E) towards the 5′ end, and nonstructural proteins
(NS1–5) towards the 3′ end. Locations of deletions found in isolates 2774, 3337, and 3336 are shown by grey boxes, with numbers at start and finish indicating distance in
nucleotide bases from start of genome where deletions begin and end. Three letter codes are given for the resultant amino acid sequence at the deletion sites. For isolates 3337
and 3336 the deletion mutant population was mixed evenly between two variants (A and B).
12
K.N. Pesko et al. / Virology 427 (2012) 10–17

Page 4

Collectively, these data suggest that genome deletions occur as a por-
tion of the population in WNV in nature, and that these genomes can
be complemented by full-length genomes and packaged by structural
proteins provided in trans.
The collection of several similar deletions in WNV from the same
place over two years raised the possibility that the deletions were
identical by descent, i.e. had a common origin and perpetuated
through time by complementation. Phylogenetic and sequence analy-
sis of the deletion events demonstrated that all were slightly different
from one another, with deletions beginning at nucleotide positions
529–717 and ending at positions 2627–3093. We reasoned that size
differences might be apparent over time if the deletions shared a
common origin, with specimens collected later in time possessing
larger deletions than those collected earlier. However, no logical rela-
tionship was detected between the size of the deletion and when the
specimen was collected. In fact, some of the smallest deletions were
collected in 2007, two years after the first specimen was collected in
2005. Phylogenetic analysis of the intact portions of the genomes in-
dicated that they were distantly related compared with other isolates
made nearby in time and space. Taken together, these results strongly
suggest that the deletions arose independently. These results differ
from previous descriptions of defective dengue genomes, which con-
tain a premature stop codon and were present in isolates from both
human and mosquito samples over the course of two years. These
Fig. 2. 50% majority-ruleconsensus phylogramisshown basedonpartial genomesequences (nt 789–10395)with posteriorprobabilities given asvaluesatnodes.Straindesignationsare
given in Table 3.
Fig. 3. A) Titers produced at 72 h post-infection of Vero cells with infectious clone derived virus (WT, black circles) or deletion mutant containing virus (2774, grey squares) at
indicated multiplicities of infection. Individual replicates are represented, with the mean and 95% confidence interval given as accompanying bars. B) Deletion mutant genomes
detected by qRT-PCR in supernatants at 72 h post-infection of Vero cells with deletion mutant containing virus.
13
K.N. Pesko et al. / Virology 427 (2012) 10–17

Page 5

dengue genomes also had an increased accumulation of nonsynon-
ymous mutations downstream of the stop codon which led re-
searchers toconcludethatthese
complemented by full-length virus and transmitted from vector to
host over the course of several years (Aaskov et al., 2006). Since the
WNV mutants we describe apparently arose independently and
share a common host origin, it may be that some vertebrates (in
this case lorikeets) contribute to WNV population biology by select-
ing similar genomic variants or allowing particular commonly arising
variants to persist, although the mechanism through which this
would occur is presently unclear.
Comparison of these deletions to previously reported Flavivirus
deletion mutants revealed a high degree of similarity to genome var-
iants known to be associated with DIPs found in vitro for related vi-
ruses JEV and Murray Valley encephalitis virus (MVEV) (Lancaster
et al., 1998; Yoon et al., 2006). We therefore sought to determine
whether they might interfere with replication of full-length WNV in
Vero cells. We found that at higher MOIs, production of full-length
WNV is suppressed in comparison to an infectious clone derived
WNV, indicating interference can occur when deletion mutants are
added at a ratio where coinfection of the same cell may occur. At
the lowest MOI, production of full-length WNV did not differ between
these two treatments. These observations establish that the deletion
mutants detected in this study function as DIPs in vitro, similar to de-
letion mutants isolated from acutely infected dengue patient sera that
contained only the untranslated regions of the genome (Li et al.,
2011).
To assess the ability of WNV DIPs to interfere with transmission by
mosquitoes, we fed mosquitoes with an isolate of WNV that con-
tained internally deleted genomes. Oral exposure to isolates contain-
ing defective genomes reduced infection and dissemination rates and
body titers in Culex quinquefasciatus mosquitoes, relative to rates esti-
mated for mosquitoes fed on a full-length only control. For mosqui-
toes fed on deletion mutants, defective RNAs were detected in
mosquito bodies but not legs or salivary secretions, at 14 and
21 days post-infection. This could be a result of resource competition
truncatedgenomes are
occurring in those tissues where coinfection and complementation has
occurred, which has previously been shown to decrease replication of
superinfecting strains of WNV (Zou et al., 2009b). Dissemination rate
has been correlated with body titers in WNV infected mosquitoes, so a
reduction of body titer through resource competition could have led
to the reduced dissemination rates observed in mosquitoes exposed
to virus containing internally deleted genomes, despite the fact that
no internally deleted genomes were detected in disseminated tissues
(Anderson et al., 2010). The restriction of this deletion mutant to mos-
quito midgut may be due to physiological barriers, such as the midgut
escape barrier, but other studies have found deleterious mutants in
multiple mosquito tissues or multiple vectors and hosts over time
(Aaskovetal.,2006;Brackneyetal.2011).Overalltheseresultsindicate
the internally deleted genomes are able to interfere with the first stage
ofmosquitoinfection,whichcouldhaveadownstreamimpactontrans-
mission.However,theobservedinefficiencyofcomplementationacross
variousmosquito tissues furthersupportsourhypothesisthat eachmu-
tant arose independently and may not be transmitted efficiently from
vector to host.
To assess the influence of deletion mutants on morbidity and mor-
tality in vertebrates, we inoculated two different strains of mice with
a virus population containing internally deleted mutants. For both
mouse strains, mortality was not attenuated in the mice inoculated
with internally deleted mutants, relative to a full-length only control
from that isolate, and an infectious clone produced control. For the
more susceptible mouse strain, C3H, inoculation with deletion mu-
tant containing and full-length virus alone produced significantly
more morbidity, as assessed by weight loss, than infectious clone de-
rived WNV. The results of these infections in mice are in contrast to
other studies which have used Semliki forest virus DI particles to
reduce mortality during concurrent or prior immunization (Barrett
and Dimmock 1984). In the infections presented here, internally de-
leted genomes were not detected in any of the tissues sampled at
time of death, suggesting that the lack of interference could be due
to a lack of complementation. It may be that inoculating at a higher
dose or using a different vertebrate host, such as a bird, would influ-
ence the outcome.
Each of the isolates that contained internally deleted genomes as
part of the infecting virus population was also polymorphic at a pre-
dicted amino acid substitution in the 2K peptide, 2K-V9M. This substi-
tution was present in plaque picks from 2774vp1, 2774vp3, and 3336.
It may be that this substitution predisposes the emergence of deletion
mutants. Interestingly, the same amino acid substitution evolved in
populations of WNV that were able to overcome superinfection ex-
clusion and replicate in WNV replicon bearing cells (Zou et al.,
2009b). This mutation has also been associated with West Nile virus
resistance to 2′5′-oligoadenylate synthetase 1b, and to the flavivirus
Fig. 4. A) Infection, dissemination, and transmission rates for Culex pipiens quinquefasciatus mosquitoes fed on full-length only (FL, grey) or deletion mutant containing virus (DM+
FL, white). Data from days 14 and 21 are combined, as day of sampling had no significant influence on rates, by chi-squared comparison. **pb0.005, *pb0.05. Infection was
calculated as the percentage of mosquitoes exposed with positive bodies, dissemination rates were calculated as the percentage of mosquitoes with positive bodies containing
positive legs, transmission rates were calculated as the percentage of mosquitoes with positive legs containing positive salivary secretions. B) Body titers of infected mosquitoes,
as estimated by plaque assay in Vero cells (PFU/0.1 ml).
Table 2
Culex quinquefasciatus mosquitoes fed on deletion and full-length virus.
Bloodmeal Day
sampled
nNumber
infected
Mutant
positive
bodies
Number
disseminated
Number
transmitted
DM+FL
FL
DM+FL
FL
14
14
21
21
71
41
56
24
44
33
26
19
7
0
5
0
8
12
8
10
2
6
6
5
14
K.N. Pesko et al. / Virology 427 (2012) 10–17

Page 6

specific antiviral lycorine, and appears to enhance replication by in-
creasing the rate of viral RNA synthesis (Mertens et al., 2010; Zou et
al., 2009a). This substitution has occurred independently in numer-
ous other North American lineages where it is under positive selec-
tion (Armstrong et al., 2011). Whether the presence of 2K-V9M in
the population containing internally deleted genomes was required
for the emergence and propagation of these genomes or evolved in
response to pressure exerted from these genomes is unclear.
In summary, kidney tissues from three lorikeets that died in the
Rio Grande bio park in August of 2005 and 2007 contained as a por-
tion of the virus population genomes with large internal deletions
to the entire envelope coding region and parts of pre-membrane
and nonstructural protein 1 coding regions. Sequencing revealed
deletions of around 2 kb affecting similar regions in each isolate, in
frame, and persisting through multiple passages. Subgenomic RNA
was directly detected by Northern analysis for one isolate, 2774.
This isolate acted as a defective interfering particle when introduced
at high doses to Vero cells. Body titers, infection, and dissemination
rates in Culex pipiens quinquefasciatus fed on stocks containing dele-
tion mutants were lower relative to a full-length only plaque pick
from the same isolate. We found no evidence for attenuation of infec-
tion in a susceptible mouse strain, C3H, inoculated with low or high
dose, or in a resistant strain, C57/B6 mice, inoculated with a low
dose of this virus. Further studies are needed to determine the
prevalence of deletion mutants such as these in vivo and elucidate
their potential roles in infected birds.
Methods
Cells and viruses
Green monkey kidney (Vero), and baby hamster kidney (Bhk-21)
cells were purchased from ATCC. Cells were grown and maintained in
Eagle's MEM supplemented by 10% fetal bovine serum (FBS) at 37 °C
with 5% CO2. Plaque assays to determine titers of isolates were per-
formed as described (Lindsey et al., 1976). WNV positive samples
were provided to our lab from the New Mexico Department of Health.
Three samples were identified as containing deletions to structural
and NS-1 coding regions. Two of these (3337, 3336) were submitted
to the NM-DOH by the Rio Grande Zoo on 8/30/2005, collection
dates are unavailable. The third (2774) was collected and submitted
to the NM-DOH on 8/20/2007. All three samples were taken from
Rainbow Lorikeets (T. haematodus) as confirmed by sequencing of cy-
tochrome oxidase (COI) gene (Wright et al., 2008). Virus stocks were
propagated by addition of 100 μl of sample to T-25 flasks containing
confluent Vero cells. A plaque pick from a plaque assay of 2774, con-
firmed to contain full-length virus only by RT-PCR and Northern anal-
ysis was also propagated in Vero cells, and used as a control in
Fig. 5. (A–C) Survivorship curves for C3H mice inoculated with 102pfu/mouse (DM+FL, red, n=8; FL, green, n=8; WT, blue, n=8) (A), C3H mice inoculated with 105pfu/mouse
(DM+FL, red, n=8; FL, green, n=8; WT, blue, n=5) (B), and C57/Bl6 mice inoculated with 105pfu/mouse (DM+FL, red, n=9; FL, green, n=9; WT, blue, n=8)(C), with
deletion mutant containing isolate (red), plaque purified isolate containing full-length virus only (green), infectious clone derived full-length virus only (blue), or mock inoculated
(black). (D–F) Percent weight change in mice measured daily, shown as the average daily percent difference from starting weight for C3H mice inoculated with 102pfu/mouse
(D), C3H mice inoculated with 105pfu/mouse (E), and C57/Bl6 mice inoculated with 105pfu/mouse (F) with deletion mutant containing isolate (red), plaque purified isolate
containing full-length virus only (green), infectious clone derived full-length virus only (blue), or mock inoculated (black).
15
K.N. Pesko et al. / Virology 427 (2012) 10–17

Page 7

multiple experiments, referred to as full length or FL. Infectious clone
derived control virus was prepared by transfecting BHK-21 cells using
transMessenger kits (Qiagen, Valencia, CA) with RNA transcribed in
vitro from pFLWNV (Shi et al., 2002), using the mMessage mMachine
T7 kit (Ambion, Austin, TX).
Extraction of RNA
RNA was extracted using an RNeasy Protect Mini-Kit, according to
the manufacturer's protocol (Qiagen, Valencia, CA) from 100 μl of
viral stock, mouse or mosquito tissue, or cells homogenized in
350 μl buffer RLT, and eluted into 50 μl of Rnase free H2O.
qRT-PCR, RT-PCR, cloning, and sequencing
Primers used in this study are listed in Table 1. cDNA was reverse
transcribed from RNA using Superscript III First Strand synthesis kit
with a 3′ end primer (Table 1; Invitrogen, Carlsbad, CA). RT-PCR
was conducted using Superscript III One-Step Reverse Transcriptase
Polymerase Chain Reaction (RT-PCR) with Platinum Taq (Invitrogen,
Carlsbad, CA) with the following parameters: 55 °C for 30 min (re-
verse transcription), 95 °C for 15 min (initial denaturation), and 40
cycles of 94 °C for 15 s, 60 °C for 30 s, and 72 °C for 30 s, followed by
final extension at 72 °C for 10 min. PCR was conducted with Promega
2X mastermix (Promega, Madison, WI). PCR products produced for
each isolate using Dh1F/Dh1R primers were cloned directly into
TOPO vector (PCR4.0), following manufacturer's instructions. Positive
colonies were assayed by PCR with M13F/R primers, and products were
purified with Stratagene PCR purification kit (Agilent, Santa Clara, CA).
PCR products were sequenced by UNM sequencing core or Genewiz
(South Plainfield, NJ). Sequences from either end were aligned to the
WNV lineage 1, strain NY99 reference genome, and positions given in
textcorrespondto thisfull-length
NC_009942). Quantitative RT-PCR was performed on an ABI Prism 7000
Sequence Detection System (Applied Biosystems), using a double
quenched probe across the mutant junction (Integrated DNA technolo-
gies, Coralville, IA), and the Brilliant II QRT-PCR kit (Agilent, Santa Clara,
CA). RNA copies were estimated using a full-length infectious clone con-
struct made with fusion PCR at site BamHI and SphI in pFLWNV (Shi et
al., 2002). Plasmid was propagated in HB101 cells, and transcribed in
vitro using the mMessage mMachine T7 kit (Ambion, Austin, TX).
reference genome(NCBI:
Virus dose assay
Twelve-well plates containing confluent Vero cells were infected at
the noted MOIs with either pFLWNV clone derived WNV, or 2774vp2
(isolate 2774 passaged two times in Vero cells). Virus was added to
cells, rocked for 5 min, and allowed to infect at 37 °C for 1 h before
additional media were added. Each MOI was introduced in three
replicates. At 72 h post-infection, supernatants were harvested and
full-length virus was quantified by plaque assay.
Experimental hosts
Five-week old female C3H/HeN and C6 mice were purchased from
Harlan Laboratories (Houston, TX). Mice were housed in groups of
4–6, and allowed one week to acclimate 57BL/6 to BSL-3 conditions.
Mice were provided with food and water ad libitum. At six weeks of
age, mice were inoculated in the left rear footpad with 50 μl of animal
inoculation diluent (cation- and endotoxin-free PBS with 1% FBS)
containing the indicated dose of West Nile virus, or mock inoculated
with inoculation diluent alone. Treatment groups varied by experi-
ment, but all contained between 5 and 9 mice. The exact number
for each is indicated in Fig. 5 legend. Mice were weighed and assessed
daily for clinical signs of disease, and sacrificed at 30% weight loss or
when severe clinical signs were observed. At 28 dpi, serum from
surviving mice was tested by plaque reduction neutralization test to
confirm that infection had occurred. Briefly, serum samples were
spun down for 10 min at 3000 rpm, upper aliquots were removed
and mixed 1:1 with ~100 units of virus (as determined by plaque
assay), incubated at 37 °C for 30 min, and added to confluent Vero
cells in 6 well plates. Samples were overlayed with agar and media,
then agar and media with neutral red at 48 hpi, and plaques were
read at 72 hpi. No more than 1–2 plaques were detected in any
wells containing serum samples, and the majority contained no pla-
ques, except for mock inoculated, indicating neutralizing antibodies
to WNV existed in all surviving mice that had been inoculated with
WNV.
Culex pipiensquinquefasciatus mosquitoes wererearedinthelabora-
toryat27 °C,with16:8photoperiod(Brackneyetal.,2009).At6–7 days
post-emergence, female mosquitoes were provided with defribinated
goose blood (Rockland, Gilbertsville, PA) and 2×107pfu/ml WNV in a
Hemotek membrane feeding apparatus (Accrington, UK). After 1 h of
feeding, mosquitoes were cold anesthetized and engorged mosquitoes
were sorted into a fresh container. Mosquitoes were held at 27 °C for
a 14 or 21 day extrinsic incubation, at which time bodies, legs, and sal-
ivary secretions were collected as described elsewhere (Ebel et al.,
2004). Infection and titers in infected tissues were assessed by plaque
assay.
Northern hybridization
Northern analysis was performed on RNA extracted directly from
cells infected 48 h previously. cDNA clone derived WNV was used as
a control, with virus generated as described elsewhere (Shi et al.,
2002). Probes were designed using the DIG Northern Starter Kit, fol-
lowing the manufacturer's instructions (Roche, Mannheim, Germa-
ny). Briefly, PCR was performed using primers with T7 promoter
sequence on cDNAs derived from WNV infectious clone (pFLWNV).
Probes were generated to target the 3′ end of the NS-5 region
(Table 1, 10296rT7) and the middle of the envelope coding region
(Table 1, 2666rT7), and used in conjunction with 9990f and 2326f, re-
spectively. PCR was run for 30 cycles using 2× PCR mastermix (Pro-
mega, Madison, WI) after which 4 μl of the resultant amplification
product was used as a template for RNA transcription with DIG label-
ing mix and T7 polymerase. RNA samples were diluted 3:1 in North-
ernMax Formaldehyde Load Dye (Ambion, Austin, TX) and run on a
1% agarose gel with 2% formaldehyde, at 60 V for 6 h, in 1× MOPS.
RNA was transferred overnight from gel to positively charged nylon
membranes (Roche, Mannheim, Germany). RNA was fixed to the
membrane by UV-crosslinking on a UV transilluminator for 90 s.
Membranes were hybridized overnight at 65 °C in hybridization buff-
er (5× SSC, 0.1% sarkosyl, 0.02% SDS, 1% block buffer reagent from
Northern Starter kit) with 100 ng/ml probe. Immunodetection pro-
ceeded according to manufacturer's instructions, except BCIP/NBT
phosphatase substrate (KPL, Gaithersburg, MD) was used for detection
after final wash step.
Phylogenetic analysis
Full genome WNV sequences were downloaded from Genbank on
October 14, 2011 (Table 3). Sequences were trimmed to nucleotides
789–10395, because the sequence alignment available for 2774
prior to nt 789 was not clean and aligned manually in BioEdit (Hall,
1999). jModelTest (version 0.1.1) was used to select the best model
for nucleotide substitution using Aikake and Bayesian Information
Criteria, which was determined to be general time reversible plus
gamma distribution of rates (GTR+G) (Posada, 2008). Markov
Chain Monte Carlo (MCMC) tree searches of 5 million generations
each were run in parallel with sampling one in every 1000 trees
using MrBayes version 3.1 (Ronquist and Huelsenbeck, 2003). A 50%
16
K.N. Pesko et al. / Virology 427 (2012) 10–17

Page 8

majority-rule consensus phylogram is shown based on the last 3750
trees, with posterior probabilities given as values at nodes.
Statistics
VirusproductionfromVerocellsafterinfectionatvariousMOIswith
each treatment was compared using a two-way ANOVA in GraphPad
(La Jolla, CA). We compared infection, dissemination, and transmission
rates in mosquitoes fed full length only or deletion mutant containing
virus in a contingency table analysis to χ2distribution or using Fisher's
exact test. Mosquito titers were log transformed and compared by two
tailed t-tests. We compared mouse survival rates by logrank tests on
Kaplan–Meier survivorship curves, and weight change in mice by
group was assessed with repeated measures ANOVA, with Bonferonni
post-teststodeterminecomparisonsresultinginsignificantdifferences.
Acknowledgments
The authors would like to thank Eleanor Deardorff, Doug Brackney,
Kathy Hanley and Lisa Green for technical help and discussions, Adam
Aragon from the New Mexico Department of Health for providing sam-
ples. This project has been funded with Federal funds from theNational
Institute of Allergy and Infectious Diseases, NationalInstitutes of Health
(NIH), Department of Health and Human Services, under grant
AI067380 (GE) and contracts HHSN272200900018C (Broad Inst.) and
HHSN272200900006C (Broad Inst.) from the U.S. NIH. KP was sup-
ported by National Research Service Award Institutional Training
Grant 5T32-AI07538-13.
References
Aaskov, J., Buzacott, K., Thu, H.M., Lowry, K., Holmes, E.C., 2006. Long-term transmission
of defective RNA viruses in humans and Aedes mosquitoes. Science 311, 236.
Anderson, S.L., Richards, S.L., Tabachnick, W.J., Smartt, C.T., 2010. Effects of West Nile
virus dose and extrinsic incubation temperature on temporal progression of
vector competence in Culex Pipiens Quinquefasciatus. J. Am. Mosq. Control
Assoc. 26, 103.
Armstrong, P.M., Vossbrinck, C.R., Andreadis, T.G., Anderson, J.F., Pesko, K.N., Newman,
R.M., et al., 2011. Molecular Evolution of West Nile Virus in a Northern Temperate
Region: Connecticut, USA 1999–2008. Virology 417, 203–210.
Artsob, H., Gubler, D., Enria, D., Morales, M., Pupo, M., Bunning, M., et al., 2009. West
Nile virus in the new world: trends in the spread and proliferation of West Nile
Virus in the western hemisphere. Zoonoses Public Health 56, 357–369.
Barrett, A.D.T., Dimmock, N.J., 1984. Modulation of a systemic Semliki Forest virus in-
fection in mice by defective interfering particles. J. Gen. Virol. 65, 1827–1831.
Brackney, D.E., Beane, J.E., Ebel, G.D., 2009. RNAi targeting of West Nile Virus in mos-
quito midguts promotes virus diversification. PLoS Pathog. 5, e1000502.
Brackney, D.E., Pesko, K.N., Brown, I.K., Deardorff, E.R., Kawatachi, J., Ebel, G.D., 2011.
West Nile virus genetic diversity is maintained during transmission by Culex
pipiens quinquefasciatus mosquitoes. PLoS ONE 6, e24466.
Brinton, M.A., 1983. Analysis of extracellular West Nile virus particles produced by cell
cultures from genetically resistant and susceptible mice indicates enhanced ampli-
fication of defective interfering particles by resistant cultures. J. Virol. 46, 860.
Brinton, M.A., 2001. Host factors involved in West Nile Virus replication. Ann. N. Y.
Acad. Sci. 951, 207–219.
Brown, A.N., Kent, K.A., Bennett, C.J., Bernard, K.A., 2007. Tissue tropism and neuroinva-
sion of West Nile virus do not differ for two mouse strains with different survival
rates. Virology 368, 422–430.
Debnath, N.C., Tiernery, R., Sil, B.K., Wills, M.R., Barrett, A.D.T., 1991. In vitro homotypic
and heterotypic interference by defective interfering particles of West Nile virus. J.
Gen. Virol. 72, 2705–2711.
Ebel, G.D., Carricaburu, J., Young, D., Bernard, K.A., Kramer, L.D., 2004. Genetic and
phenotypic variation of West Nile Virus in New York, 2000–2003. Am. J. Trop.
Med. Hyg. 71, 493–500.
Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and
analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41, 95–98.
Holland, J., Spindler, K., Horodyski, F., Grabau, E., Nichol, S., VandePol, S., 1982. Rapid
evolution of RNA genomes. Science 215, 1577.
Holmes, E.C., 2009. The evolutionary genetics of emerging viruses. Annu. Rev. Ecol.
Evol. Syst. 40, 353.
Jenkins, G.M., Rambaut, A., Pybus, O.G., Holmes, E.C., 2002. Rates of molecular evolution
in RNA viruses: a quantitative phylogenetic analysis. J. Mol. Evol. 54, 156–165.
Jerzak, G., Bernard, K.A., Kramer, L.D., Ebel, G.D., 2005. Genetic variation in West Nile
virus from naturally infected mosquitoes and birds suggests quasispecies structure
and strong purifying selection. J. Gen. Virol. 86, 2175–2183.
Komar, N., Clark, G.G., 2006. West Nile Virus activity in Latin America and the Caribbean.
Rev. Panam. Salud Publica 19, 112–117.
Lancaster, M.U., Hodgetts, S.I., Mackenzie, J.S., Urosevic, N., 1998. Characterization of
defective viral RNA produced during persistent infection of vero cells with Murray
valley encephalitis virus. J. Virol. 72, 2474.
Li, D., Lott, W.B., Lowry, K., Jones, A., Thu, H.M., Aaskov, J., 2011. Defective interfering
viral particles in acute dengue infections. PlosONE 6, e19447.
Lindsey, H.S., Calisher, C.H., Mathews, J.H., 1976. Serum dilution neutralization test for
California group virus identification and serology. J. Clin. Microbiol. 4, 503.
Mertens, E., Kajaste-Rudnitski, A., Torres, S., Funk, A., Frenkiel, M.P., Iteman, I., et al., 2010.
Viral determinants in the NS3 helicase and 2K peptide that promote West Nile virus re-
sistancetoantiviralactionof 2′,5′-oligoadenylate synthetase1b.Virology399,176–185.
Murray, K.O., Mertens, E., Desprès, P., 2010. West Nile virus and its emergence in the
United States of America. Vet. Res. 41, 67.
Posada, D., 2008. JModelTest: phylogenetic model averaging. Mol. Biol. Evol. 25, 1253.
Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inference under
mixed models. Bioinformatics 19, 1572.
Shi, P.Y., Tilgner, M., Lo, M.K., Kent, K.A., Bernard, K.A., 2002. Infectious cDNA clone of
the Epidemic West Nile virus from New York City. J. Virol. 76, 5847.
Thompson, K.A.S., Rempala, G.A., Yin, J., 2009. Multiple-hit inhibition of infection by
defective interfering particles. J. Gen. Virol. 90, 888–899.
Tsai, K.N., Tsang, S.F., Huang, C.H., Chang, R.Y., 2007. Defective interfering RNAs of
Japanese encephalitis virus found in mosquito cells and correlation with persistent
infection. Virus Res. 124, 139–150.
Weaver, S.C., 2006. Evolutionary influences in arboviral disease. Curr. Top. Microbiol.
Immunol. 299, 285–314.
Wright, T.F., Schirtzinger, E.E., Matsumoto, T., Eberhard, J.R., Graves, G.R., Sanchez, J.J.,
et al., 2008. A multilocus molecular phylogeny of the parrots (Psittaciformes):
support for a Gondwanan origin during the Cretaceous. Mol. Biol. Evol. 25, 2141.
Yoon, S.W., Lee, S.Y., Won, S.Y., Park, S.H., Park, S.Y., Jeong, Y.S., 2006. Characterization
of homologous defective interfering RNA during persistent infection of vero cells
with Japanese encephalitis virus. Mol. Cells 21, 112–120.
Zou, G., Puig-Basagoiti, F., Zhang, B., Qing, M., Chen, L., Pankiewicz, K.W., et al., 2009a. A
single-amino acid substitution in West Nile virus 2K peptide between NS4A and
NS4B confers resistance to lycorine, a flavivirus inhibitor. Virology 384, 242–252.
Zou, G., Zhang, B., Lim, P.Y., Yuan, Z., Bernard, K.A., Shi, P.Y., 2009b. Exclusion of West
Nile virus superinfection through RNA replication. J. Virol. 83, 11765–11776.
Table 3
Partial genome sequences used for Bayesian analysis.
DesignationStrain Accession #Host, location, time of sampling
NM053337HM756677Lorikeet kidney, Albuquerque,
August 2005
Lorikeet kidney, Albuquerque,
August 2007
American crow, Staten Island, NY,
October 2000
Culex tarsalis, California 2004
Culex tarsalis, California 2003
Human, Arizona 2004
Blue jay, Texas 2005
Aedes albopictus, Texas 2007
Culex quinquefasciatus, Texas 2009
Human, Colorado 2004
NM072774
NY00WNV3356EF530047
CA04
CA03
AZ04
TX05
TX07
TX09
CO04
COAV689
Impr1075
JF703161
DQ080056
DQ164201
JF415929
JF415919
JF415928
DQ431704
TX5058
M19433
M20122
04-219CO
17
K.N. Pesko et al. / Virology 427 (2012) 10–17