JOURNAL OF CLINICAL MICROBIOLOGY, Apr. 2004, p. 1511–1518
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Vol. 42, No. 4
SYBR Green-Based Real-Time Quantitative PCR Assay for Detection
of West Nile Virus Circumvents False-Negative Results
Due to Strain Variability
James F. Papin, Wolfgang Vahrson, and Dirk P. Dittmer*
Department of Microbiology and Immunology, The University of Oklahoma Health Sciences Center,
Oklahoma City, Oklahoma 73104
Received 11 October 2003/Returned for modification 26 November 2003/Accepted 10 January 2004
Real-time quantitative PCR is used routinely for the high-throughput diagnosis of viral pathogens, such as
West Nile virus (WNV). Rapidly evolving RNA viruses present a challenge for diagnosis because they accu-
mulate mutations that may render them undetectable. To explore the effect of sequence variations on assay
performance, we generated every possible single point mutation within the target region of the widely used
TaqMan assay for WNV and found that the TaqMan assay failed to detect 47% of possible single nucleotide
variations in the probe-binding site and was unable to detect any targets with more than two mutations. In
response, we developed and validated a less expensive assay with the intercalating dye SYBR green. The SYBR
green-based assay was as sensitive as the TaqMan assay for WNV. Importantly, it detected 100% of possible
WNV target region variants. The assay developed here adds an additional layer of protection to guard against
false-negative results that result from natural variations or drug-directed selection and provides a rapid means
to identify such variants for subsequent detailed analysis.
West Nile virus (WNV) belongs to the Flaviviridae, a family
of over 70 related viruses. Specifically, WNV is a member of
the Japanese encephalitis (JE) serocomplex, which also in-
cludes JE virus, St. Louis encephalitis virus, Murray Valley
encephalitis virus, and Kunjin virus (25). Like other flavivi-
ruses, WNV is an enveloped, positive-sense RNA virus with a
ca. 11-kb genome composed of three structural genes and
seven nonstructural genes. The viral sequences for many iso-
lates, such as WN-NY99 and others, have been determined
(4–6, 9, 20, 23). WNV is arthropod borne and is maintained
through an enzootic cycle between mosquitoes and birds. Hu-
mans and other mammals, such as horses, can be incidental
dead-end hosts if bitten by an infected mosquito. In most cases,
WNV causes a self-limited febrile illness; however, infection
may also lead to encephalitis and death.
Originally isolated from the blood of a febrile Ugandan
woman in 1937, WNV is widely distributed throughout Africa,
the Middle East, areas of Europe, and Asia (31). The summer
of 1999 marked the first incidence of WNV on the North
American continent, in the northeastern United States (23).
During this outbreak, more than 60 people became clinically
ill, and seven deaths occurred due to encephalitis (12). WNV
has continued to cause seasonal epidemics and to spread west-
ward. In 2002, the virus was detected in 43 of 50 states, and
more than 3,800 human cases were reported; 225 of these were
fatal (8). Currently, no specific drug or vaccine against WNV is
approved for human use, so that palliative care, nonspecific
antiviral treatment (alpha interferon and ribavirin), surveil-
lance, and screening of potentially contaminated products
(such as blood) are the only available measures against WNV.
At the present time, enzyme-linked immunosorbent assays
and PCR assays are used to monitor the presence of WNV (17,
22, 28). A real-time quantitative reverse transcriptase PCR
(RT-QPCR) assay for WNV has been reported to enable rou-
tine high-throughput screening for WNV (19, 22). This assay is
based on TaqMan technology; in addition to two specific outer
primers, a third, fluorescence-labeled oligonucleotide (Taq-
Man) is used for detection. This assay has been proven to be as
sensitive as gel-based reverse transcriptase PCR (RT-PCR),
with a detection limit of 0.1 PFU/ml.
We have developed an alternative real-time RT-QPCR as-
say with the intercalating dye SYBR green. This assay is less
expensive than TaqMan or beacon-based real-time quantita-
tive PCR but is still faster than gel-based, single, or nested
RT-PCR and should be economical for large-scale routine
testing of clinical samples and blood products. Importantly, we
found that this new assay is insensitive to nucleotide variations
within the amplified region and thus has a substantially lower
false-negative rate than prior assays. Furthermore, through
dissociation curve analysis, the SYBR green-based RT-QPCR
assay allows for the identification of novel WNV strains.
MATERIALS AND METHODS
Virus culture. Vero cells (American Type Culture Collection) were main-
tained in Dulbecco minimal essential medium supplemented with 10% fetal
bovine serum and antibiotics (Cellgro Inc.) at 37°C in 5% CO2. Confluent Vero
cells (six-well plates) were incubated with 100 ?l of a clarified suspension of
WNV-positive tissue for 1 h. WNV-positive bird tissue (from 19 animals) was
generously provided by the Oklahoma Animal Disease Diagnostic Laboratory.
The clarified suspension was prepared by placing tissue samples into 5-ml snap-
top tubes (Falcon 352063) together with 2 ml of homogenization buffer (2?
phosphate-buffered saline, 0.05 M Tris-HCl [pH 7.6], 1% [wt/vol] bovine serum
albumin, 4.2 mM sodium bicarbonate, 0.1 ?g of streptomycin/ml, 1 ?g of am-
photericin B/ml) and four copper-clad steel beads (4.5 mm; Copperhand
[Walmart Inc.]), and the mixture was vortexed for 45 s. The homogenate was
subsequently centrifuged in 2-ml tubes (Sarstedt, Nu ¨mbrecht, Germany) at
13,000 rpm in an Eppendorf centrifuge for 5 min to remove solids from the
supernatant. After incubation with 200 ?l of a tissue-derived suspension, Vero
* Corresponding author. Mailing address: Department of Microbi-
ology and Immunology, The University of Oklahoma Health Sciences
Center, 940 Stanton L. Young Blvd., Oklahoma City, OK 73104.
Phone: (405) 271-2133. Fax: (405) 271-3117. E-mail: dirk-dittmer@ou-
cell monolayers were observed daily for 7 days. For a plaque assay, 0.5% meth-
ylcellulose in 2? Dulbecco minimal essential medium was overlaid 2 h postab-
sorption. At the end of the observation period, the plates were fixed in 100%
ice-cold methanol for 10 min, stained with 1% Giemsa stain (Sigma Inc., St.
Louis, Mo.) for 30 min at room temperature, and rinsed with tap water.
RNA isolation and reverse transcription. RNA was isolated from the homog-
enate by using an Absolutely RNA Micro-Prep kit (Stratagene, La Jolla, Calif.)
according to the manufacturer’s instructions. Reverse transcription was per-
formed as previously described (13). Briefly, 500 ng of RNA was reverse tran-
scribed in a 20-?l reaction mixture with 100 U of SupercriptII reverse transcrip-
tase (Invitrogen Inc., Carlsbad, Calif.), 2 mM deoxyribonucleoside triphosphates,
2.5 mM MgCl2, 1 U of RNasin (the last three from Applied Biosystems, Foster
City, Calif.), and 0.5 ?g of random hexanucleotide primers (Amersham Phar-
macia Biotech, Piscataway, N.J.). The reaction mixture was sequentially incu-
bated at 42°C for 45 min, 52°C for 30 min, and 70°C for 10 min. The reverse
transcription reaction was stopped by heating to 95°C for 5 min. Next, 0.5 U of
RNase H (Invitrogen) was added, and the reaction mixture was incubated at
37°C for an additional 30 min. Afterward, the cDNA pool was diluted 25-fold
with diethyl pyrocarbonate-treated, distilled H2O and stored at ?80°C.
Real-time quantitative PCR. The primers used for WNV in this study were
previously described by Lanciotti et al. (22). They are 5?-TCA CGC ATC TCT
CCA CCA AAG-3? and 5?-GGG TCA GCA CGT TTG TCA TTG-3? and are
specific for the WNV env region. Positive control RNA (strain WN-NY99) was
obtained from the Centers for Disease Control and Prevention Reference Lab-
oratory (Ft. Collins, Colo.). Ninety-six variant 60-mer oligonucleotides were
synthesized at The University of Oklahoma Health Sciences Center genomics
facility. Real-time PCR was conducted with previously established procedures
(11, 13). The final PCR mixture contained 2.5 ?l each of forward and reverse
primers (final concentration of each, 166 nM), 7.5 ?l of 2? SYBR PCR mix
(Applied Biosystems), and 5 ?l of sample. The PCR was set up in a dedicated
room with a CAS-1200 pipetting robot (Pheonix Research, Hayward, Calif.). The
CAS-1200 robot uses filtered carbon-graphite pipette tips (Tecan Inc.) for liquid
level sensing, allowing for a pipetting accuracy of 0.1 ?l and eliminating carryover
contamination. All surfaces were cleaned with 10% bleach and exposed to UV
light overnight on a daily basis. Gowns, gloves, face masks, and equipment were
required for all work. Real-time PCR was performed with an ABI Prism 5700 or
ABI Prism 7700 machine (Applied Biosystems) and universal cycling conditions
(2 min at 50°C, 10 min at 95°C, 40 cycles of 15 s at 95°C, and 1 min at 60°C). Cycle
threshold (CT) values were determined by automated threshold analysis with
ABI Prism version 1.0 software. The amplification efficiencies were determined
by serial dilution and calculated as E ? exp?1/m, where E is the amplification
efficiency and m is the slope of the dilution curve. Dissociation curves were
recorded after each run, and the amplified products were visualized by 2%
agarose gel electrophoresis. The entire procedure from sample receipt to data
analysis was routinely completed in 4 h.
Statistical analysis. Calculations were performed with Excel version 10.1 (Mi-
crosoft Inc., Redwood, Wash.) and SPSS version 12.0 (SPSS, Chicago, Ill.).
Analysis of dissociation profiles was performed with ABI Prism version 1.0
software. The distance between dissociation profiles was calculated as a weighted
sum of squares of differences between pairs of normalized data points. When
necessary, linear interpolation was used between neighboring data points (10).
Initially, dissociation profiles from 96 identical samples were compared to estab-
lish a detection threshold and overall measurement error. In later comparisons,
all distances above the threshold were reported to be different from the refer-
SYBR green is as sensitive as TaqMan detection in a real-
time RT-QPCR assay for WNV. A real-time RT-QPCR assay
has been reported for the detection of WNV (22). We have
adapted this assay by using primers designed on the basis of the
envelope glycoprotein gene (WNV env) of strain WN-NY99.
However, instead of using a specific fluorescence-labeled
probe (TaqMan) for the detection of amplicons, we use SYBR
green dye. Since SYBR green binds only to double-stranded
and not to single stranded DNA molecules, PCR product con-
centrations can be recorded at each cycle, yielding a real-time
amplification curve suitable for automated threshold analysis
and quantification (24).
To compare the sensitivity of the SYBR green-based assay
to that of the TaqMan assay, serial dilutions of strain WN-
NY99 were tested in both assays with the WNV env probe and
primer pairs previously described by Lanciotti et al. (22). Both
assays performed with equal sensitivities and had similar linear
FIG. 1. Linearity and sensitivity comparisons of TaqMan and SYBR green-based real-time quantitative PCRs. Linear regression plots were
generated for TaqMan and SYBR green-based real-time quantitative PCRs of WN-NY99 dilutions with the same WNV env primer set. The sample
dilution on the horizontal axis is plotted against the CTon the vertical axis (n ? 3). (A) Linear regression of TaqMan assay. (B) Linear regression
of SYBR green-based assay.
1512PAPIN ET AL.J. CLIN. MICROBIOL.
ranges (Fig. 1). CTvalues for both assays were similar, with
y-axis intercept points at the greatest dilution of 41.23 ? 3.26
(mean and standard error [SE]) and 40.72 ? 2.99 for TaqMan-
based detection and SYBR green-based detection, respec-
tively. The amplification efficiencies were identical, as evi-
denced by the identical slopes of the regression lines. The
regression coefficients (R2) for the TaqMan assay and the
SYBR green-based assay were 0.9688 and 0.9625, respectively
(Fig. 1). The parallel testing of these two detection methods
establishes that the SYBR green-based assay is as sensitive as
the TaqMan assay with the same (outer) PCR primers.
The sensitivity of SYBR green-based real-time RT-QPCR at
low target concentrations (“real-world” conditions) was mod-
eled by analyzing WNV in mosquito populations. Figure 2A
shows the amplification plot for a representative set of mos-
quito pools (out of a total of 140). The two positive control
reactions were practically superimposed upon each other (CT,
22.06; associated SE, 0.41), attesting to the high reproducibility
of our assay. Positive and negative samples segregated into two
clearly distinguishable clusters, and the threshold was set ac-
cordingly. Conventional agarose gel electrophoresis confirmed
the real-time RT-QPCR readings (Fig. 2B). WNV-positive
samples could easily be distinguished from WNV-negative
pools, which showed no amplification curve and no discernible
band on the gel.
The specificity of SYBR green-based real-time RT-QPCR
was ascertained by comparing the melting temperatures (Tms)
of the amplification products from different samples to that of
the positive control sample (Fig. 2C). The Tmof the 60-bp
WN-NY99 WNV env amplicon was 81.4°C. Cloning and se-
quencing of the PCR product yielded a sequence identical to
the one previously deposited in GenBank under accession
number AF196835 (data not shown). All samples that exhib-
ited a Tmwithin ?1°C of that of the WN-NY99 control were
considered positive for WNV. This threshold is based on ex-
tensive work with other viral assays (11, 13), which established
the reproducibility of Tmprofiles and their use for fingerprint-
ing individual amplification products. Samples that yielded di-
vergent Tmprofiles are discussed below. SYBR green-based
real-time quantitative PCR was as specific and as sensitive as
TaqMan real-time quantitative PCR.
SYBR green-based real-time RT-QPCR can detect polymor-
phic targets, a property which lowers the overall false-negative
rate. In screening natural WNV isolates, we encountered sam-
ples with a substantial viral load (as evidenced by low CT
values) but divergent dissociation profiles. These samples
likely represented naturally occurring WNV variants with se-
quence variations within the amplicon and thus may have es-
caped detection by the TaqMan assay, yielding false-negative
results. These samples yielded products which were indistin-
guishable from those yielded by the WN-NY99 sample (used as
a positive control) in gel electrophoresis, since the readout in
conventional or nested PCR is a band on a 2% agarose gel that
can differentiate only gross variations in size (?5 bp, depend-
ing on the gel system) and not nucleotide substitutions. In
contrast, the SYBR green-based assay provides an amplicon
dissociation profile as an added measure of specificity. With
this method, even single nucleotide changes can yield an al-
tered profile, which should allow for the rapid routine detec-
tion of sequence variations.
To compare directly the ability of the TaqMan assay and the
SYBR green-based assay to detect polymorphisms, we synthe-
sized 60-mer oligonucleotides containing every possible single
point substitution in the central 28-bp amplicon region as well
as selected double and triple mutations (Table 1). These oli-
gonucleotides were each used as templates in both assays. As
expected, all targets yielded a signal with SYBR green, but only
47 (53%) of 88 single-point-substitution amplicons were de-
tected by the probe-based assay (Fig. 3A and B). This finding
represents a false-negative rate of 47% for the TaqMan assay
for single nucleotide polymorphisms across the probe-binding
site. Of the eight double- and triple-substitution amplicons in
our experiment, only five (63%) were detected by the TaqMan
assay, even though a PCR product was amplified in all in-
stances, as determined by agarose gel electrophoresis (Fig.
3C). This outcome indicates that WNV variants with multiple
substitutions will escape detection by TaqMan real-time quan-
FIG. 2. Detection of WN virus in mosquito pools by SYBR green-
based real-time RT-QPCR. (A) Plot of amplification of mosquito
pools with the WNV env primer set. Cycle number on the horizontal
axis is plotted against the relative fluorescence (Rn) on the vertical axis
(log scale). Dotted lines indicate the two positive control reactions.
Positive and negative mosquito pools are indicated by plus and minus
signs, respectively. (B) Agarose (2%) gel electrophoresis of amplifica-
tion products stained with ethidium bromide. Lane M, molecular
weight markers. Lanes ?, positive control (in duplicate). Positive re-
actions showed a single amplification product of approximately 60 bp.
(C) Dissociation plot of amplification products for the positive control
and mosquito pools. Temperature on the x axis is plotted against the
first derivative of the measured fluorescence [d(F)] on the y axis. The
asterisk indicates the peak position of the reference dissociation curve.
VOL. 42, 2004REAL-TIME PCR FOR WNV 1513
TABLE 1. Sequences of 60-mer mutant oligonucleotides that served as targets for amplification with WN env primers
Continued on following page
1514PAPIN ET AL.J. CLIN. MICROBIOL.
titative PCR with a high probability, as would insertion or
To explore the potential for the SYBR green-based assay to
identify WNV mutants, all amplification products from the
mutant oligonucleotide array were subjected to dissociation
profile analysis (Fig. 3E). While every sequence with three or
more substitutions could be easily recognized by the divergent
dissociation profile, not all single or double nucleotide substi-
tutions produced a dissociation profile that differed signifi-
cantly from that of the wild-type sequence. Not surprisingly,
among the single nucleotide substitutions that could be de-
tected best were those with a strong (CG) to weak (AT) change
or vice versa. The best result was obtained for T3G mutations
(which are also transversions), 75% of which could be detected
by melting curve analysis. No T3A or C3G mutation could
be detected, although we are currently optimizing the pattern
recognition software (W. Vahrson and D. P. Dittmer, unpub-
lished data). To assess the potential contributions of variations
within the instrument to differences in Tms, 96 reactions with
WN-NY99 as the template were performed (Fig. 3D). Little
to no variation (mean Tm, 83.9°C; SD, ?0.1°C; n ? 96) was
observed, indicating that differences in Tms are due to differ-
ences in sequences and not to machine variations.
To determine how close the real-time PCR target area in
WNV is to those in other flaviviruses or any other sequence, we
performed a blastn (1) search of GenBank (Table 2). Within
the first 100 hits based on sequence identity, 78 entries be-
longed to WNV sequences. Of these, 76 differed by no more
1181C3A, 1183G3A, 1184C3T
1181C3A, 1189C3T, 1198G3A
1182T3G, 1186T3C, 1195A3G
1200A3C, 1203A3C, 1206T3G
a“1” denotes oligonucleotides which yielded a Tmand a dissociation curve that were significantly different from those of the wild type. “0” denotes oligonucleotides
which yielded a Tmand dissociation curves that did not differ from those of the wild type.
b“0” denotes oligonucleotides which did not yield a detectable signal in the TaqMan assay. “1” denotes oligonucleotides which yielded a detectable signal.
TABLE 2. Flavivirus sequences with significant similarity in the PCR target regiona
No. of identical
entries in GenBank
aMultiple sequence alignment of the top 100 closest matches after a blastn search of GenBank (as of 9 December 2003). Shown are the GenBank identifier for each
sequence class, the sequence alignment (identical residues are indicated by a dot), and the number of identical GenBank entries for individual viral isolates. No other
members of the JE group or other members of the Flaviviridae family exhibit significant sequence homology in this region.
VOL. 42, 2004REAL-TIME PCR FOR WNV1515
than two nucleotides (96%) from the real-time quantitative
PCR target sequence. The remaining two WNV sequences
showed identity for 25 nucleotides (42%) and 27 nucleotides
(45%). The next closest sequences were for three isolates clas-
sified as Kunjin virus, with 46 identical nucleotides (76%). JE
virus isolates exhibited sequence identity in 20 nucleotide po-
sitions (33%) on the right side of this region only and thus
would not have been recognized by the left-hand PCR primer.
All other flaviviruses exhibited no significant sequence similar-
ity at the nucleotide level.
To determine whether the target copy number affected Tm-
based typing, we performed limiting dilution analysis. WNV
(strain OK02) was isolated in Vero cell cultures from the brain
of a dead blue jay (data not shown). The supernatant from this
culture was subjected to fivefold serial dilution, and RNA was
subsequently purified. Even after multiple dilution steps, sam-
FIG. 3. Comparison of the effects of nucleotide variations on the TaqMan and SYBR green-based assays. (A) Plot of amplification of 96 mutant
oligonucleotides in the SYBR green-based assay. The cycle number on the horizontal axis is plotted against the relative measured fluorescence
(Rn) on the vertical axis. (B) CTvalues obtained for individual mutant target oligonucleotides in either the SYBR green-based assay (horizontal
axis) or the TaqMan assay (vertical axis). Samples that did not yield any signal were assigned a CTvalue of 45. (C) Agarose (2%) gel electrophoresis
of the amplification products from the TaqMan assay stained with ethidium bromide. The asterisk indicates an empty lane. (D) Dissociation plot
of 96 identical samples (WN-NY99) after amplification in the SYBR green-based assay with the WNV env primer set. (E) Dissociation plot of the
96 mutant oligonucleotides listed in Table 1 after amplification in the SYBR green based assay with the WNV env primer set. The derivative of
the measured fluorescence [d(F)] on the vertical axis is plotted against the temperature at which it was sampled on the horizontal axis for panels
D and E.
1516PAPIN ET AL. J. CLIN. MICROBIOL.
ples that were amplified above the background had the same
Tmprofile as the positive control sample (Fig. 4). This finding
demonstrated that the Tmand dissociation profiles are inde-
pendent of the initial target concentration.
To improve our dissociation profile analysis, we applied a
mathematical algorithm that takes into account the shape of
the entire dissociation profile rather than the single data point
for Tm(implementation of a standard graph comparison as
described previously ; Vahrson and Dittmer, unpublished).
This algorithm recognized 23 (27%) of all 84 single point
mutations in the target sequence, even though the difference in
Tmfrom the original target sequence was only 1.5°C ? 0.5°C.
These mutations would not have been detected by inspection
of the Tmalone. Interestingly, the algorithm performed much
better with particular types of substitutions; e.g., 75% of all
T3G substitutions were detected, while no T3A substitutions
could be detected. It recognized 50% of double nucleotide
substitutions in the target sequence and 100% of all triple
nucleotide substitutions. It is clear that dissociation profile
analysis does not substitute for single-nucleotide-polymor-
phism analysis for the detection of specific mutants or defined
genotypes (14, 16, 18, 30). However, in the context of routine
screening, samples with an aberrant dissociation profile indi-
cate variant WNV strains, which then may be selected for
detailed characterization by sequence analysis.
We developed and validated an improved real-time RT-QPCR
assay for WNV. Detection with SYBR green is as sensitive as
probe-based real-time PCR without the need for a specialized
probe (Fig. 1). All samples that yielded a fluorescence signal at
least five times that of the nontemplate control reaction in the
SYBR green-based assay also produced an easily discernible
band upon agarose gel electrophoresis. Hence, these samples
would be considered positive by conventional end-point RT-
PCR. Using a pipetting robot, we were able to detect WNV in
field samples with a 15-?l RT-PCR. Using previously pub-
lished WNV env primers, we achieved an SE for CTvalues of
less than 1% in a 96-sample repeat experiment (Fig. 3 and
data not shown). Lanciotti and colleagues introduced the
TaqMan primers and probe for WNV used here and also,
more recently, nucleic acid sequence-based amplification
(NASBA) for WNV (21, 22). Their comparison of TaqMan
real-time RT-QPCR, NASBA, and conventional RT-PCR
revealed a strong concordance for all three methods, except
for operator-based interpretations of threshold settings for
NASBA and real-time RT-QPCR at very low target levels
(?0.1 PFU/ml). We found that real-time quantitative PCR
with SYBR green also has a comparable performance pro-
file. Comparisons at the extreme lower end of the linear
range between different amplification methods are difficult
to interpret, since different primer pairs with associated
different amplification efficiencies (27) and different re-
agents (13) have been used. Previous studies estimated the
limit of detection for the probe-based TaqMan RT-QPCR
assay for WNV to be ?0.1 PFU/ml (2, 22, 26, 28), well below
the detection limit for culture-based assays. Using serial dilu-
tions on Vero cells (Fig. 4), we confirmed these observations
for SYBR green-based real-time quantitative PCR.
The possibility of false-negative test results poses a substan-
tial problem in diagnostic TaqMan assays because mutations
within the probe-binding site can prevent annealing of the
probe and subsequent detection. While this problem has been
documented for herpes simplex virus (3), it is a particular
concern for WNV, severe acute respiratory syndrome virus,
human immunodeficiency virus, hepatitis C virus, and other
RNA viruses, since they exhibit higher sequence variability
than DNA viruses and since it is not always possible to identify
regions of the genome which are highly conserved. By synthe-
sizing and testing every possible point mutation in the target
region for the widely used WNV TaqMan assay, we encoun-
tered a potential false-negative rate of 47% for the TaqMan
assay due to sequence variations in the probe-binding site. In
comparison, the SYBR green-based assay produced a false-
positive rate of 0% (Fig. 3). We did not explore the effects of
mutations within the binding sites for the forward and reverse
primers, since these would affect the TaqMan assay and the
SYBR green-based assay (as well as gel-based PCR assays) in
the same way. The high-stringency annealing temperature used
in real-time quantitative PCR (60°C) for the two outer primers
leaves open the possibility that viral variants with mutations in
the forward or reverse primer-binding sites may escape detec-
tion. Although not tested here, it is likely that the frequency of
FIG. 4. Titration of WNV from an avian sample in a Vero cell
culture. (A) Plot of amplification of serially diluted supernatants from
infected Vero cell cultures. The cycle number on the horizontal axis is
plotted against the relative measured fluorescence (Rn) on the vertical
axis. (B) Dissociation plot of amplification products. The temperature
on the horizontal axis is plotted against the derivative of the measured
fluorescence [d(F)] on the vertical axis.
VOL. 42, 2004 REAL-TIME PCR FOR WNV1517
such false-negatives can be reduced by the use of multiple
primer sets for each virus.
Could a SYBR green-based real-time PCR assay be used to
screen for novel WNV isolates? A number of studies investi-
gated WNV sequence variability (2, 4, 20, 23, 29). For this
flavivirus, the average sequence variability did not exceed 3%
at the nucleotide level, and the majority of mutations did not
lead to amino acid substitutions. For the 1999 outbreak in New
York, Connecticut, and New Jersey, 99% sequence identity
was reported over a 1,278-nucleotide region in 13 isolates from
avian species, humans, and mosquitoes. Many of the same mu-
tations were shared among isolates from mosquitoes, birds, and
humans. Presumably, the need to enter and replicate in cells of
avian as well as arthropod origins places severe constraints on
the variability of this virus. In contrast, human retroviruses,
such as human immunodeficiency virus type 1 (15), exhibit
amino acid variability of up to 20% within the same subtype.
We found that a preliminary measure of natural variability can
be obtained by Tmor dissociation profile analysis (Fig. 3).
SYBR green-based real-time quantitative PCR incorporates
dissociation profile analysis for each sample at no extra cost.
This feature constitutes a considerable improvement over gel-
based or TaqMan-based PCR, neither of which can identify the
presence of single nucleotide mutations or small (?5-bp) in-
sertions or deletions. In the TaqMan assay, these alterations
led to a complete loss of signal (false-negative) or produced
signals that were indiscernible from the wild-type signal (Fig.
3); in the gel-based assay, the signals produced by these mu-
tations were indiscernible from the wild-type signal. Since real-
time quantitative PCR amplicons are small (?100 bp), even a
single base-pair change can result in a distinguishable change
in the Tm(7), although in this study only changes of ?3 nu-
cleotides were identified with 100% accuracy. Currently, we
are trying to improve upon the bioinformatics tools for this
analysis. While not as reliable as sequence analysis, dissocia-
tion profile analysis has the potential to be used for rapid initial
screening for the identification of viral mutants.
We thank Rebecca Hines-Boykin and Arndt Pechstein for technical
help; Johnnie Gilpen, Dan Carr, Joseph Waner, and Gary Saliki for
samples, cell lines, and tissue; and Michelle Staudt and Mike Sakalian
for critical reading of the manuscript.
This work was supported by NIH grants EB00983 and CA97951 to
D.P.D. J.F.P. was supported by NIH training grant T32 AI07364 to the
Department of Microbiology and Immunology, The University of
Oklahoma Health Sciences Center.
1. Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller,
and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation
of protein database search programs. Nucleic Acids Res. 25:3389–3402.
2. Anderson, J. F., T. G. Andreadis, C. R. Vossbrinck, S. Tirrel, E. Wakem, R.
French, A. Garmendia, and H. Van Kruiningen. 1999. Isolation of West Nile
virus from mosquitoes, crows, and a Cooper’s hawk in Connecticut. Science
3. Anderson, T. P., A. M. Werno, K. A. Beynon, and D. R. Murdoch. 2003.
Failure to genotype herpes simplex virus by real-time PCR assay and melting
curve analysis due to sequence variation within probe binding sites. J. Clin.
4. Beasley, D. W., C. T. Davis, H. Guzman, D. L. Vanlandingham, A. P. Travas-
sos da Rosa, R. E. Parsons, S. Higgs, R. B. Tesh, and A. D. Barrett. 2003.
Limited evolution of West Nile virus has occurred during its southwesterly
spread in the United States. Virology 309:190–195.
5. Beasley, D. W., L. Li, M. T. Suderman, and A. D. Barrett. 2002. Mouse
neuroinvasive phenotype of West Nile virus strains varies depending upon
virus genotype. Virology 296:17–23.
6. Brinton, M. A. 2002. The molecular biology of West Nile virus: a new invader
of the western hemisphere. Annu. Rev. Microbiol. 56:371–402.
7. Cantor, C. R., and P. R. Schimmel. 1980. Biophysical chemistry: the behavior
of biological macromolecules. W. H. Freeman & Co., New York, N.Y.
8. Centers for Disease Control and Prevention. 2002. West Nile virus activity—
United States, July 31-August 7, 2002, and Louisiana, January 1-August 7,
2002. Morb. Mortal. Wkly. Rep. 51:681–683.
9. Chambers, T., M. Halevy, A. Nestorowicz, C. Rice, and S. Lustig. 1998. West
Nile virus envelope proteins: nucleotide sequence analysis of strains differing
in mouse neuroinvasiveness. J. Gen. Virol. 79:2375–2380.
10. Chatfield, C., and A. J. Collins. 1980. Introduction to multivariate analysis.
Chapman & Hall, Ltd., London, England.
11. Dittmer, D. P. 2003. Transcription profile of Kaposi’s sarcoma-associated
herpesvirus in primary Kaposi’s sarcoma lesions as determined by real-time
PCR arrays. Cancer Res. 63:2010–2015.
12. Enserink, M. 2002. West Nile’s surprisingly swift continental sweep. Science
13. Fakhari, F. D., and D. P. Dittmer. 2002. Charting latency transcripts in
Kaposi’s sarcoma-associated herpesvirus by whole-genome real-time quan-
titative PCR. J. Virol. 76:6213–6223.
14. Gardner, S. N., T. A. Kuczmarski, E. A. Vitalis, and T. R. Slezak. 2003.
Limitations of TaqMan PCR for detecting divergent viral pathogens illus-
trated by hepatitis A, B, C, and E viruses and human immunodeficiency
virus. J. Clin. Microbiol. 41:2417–2427.
15. Gaschen, B., J. Taylor, K. Yusim, B. Foley, F. Gao, D. Lang, V. Novitsky, B.
Haynes, B. H. Hahn, T. Bhattacharya, and B. Korber. 2002. Diversity con-
siderations in HIV-1 vaccine selection. Science 296:2354–2360.
16. Harris, E., E. Sandoval, A. M. Xet-Mull, M. Johnson, and L. W. Riley. 1999.
Rapid subtyping of dengue viruses by restriction site-specific (RSS)-PCR.
17. Hunt, A. R., R. A. Hall, A. J. Kerst, R. S. Nasci, H. M. Savage, N. A. Panella,
K. Gottfried, K. L. Burkhalter, and J. T. Roehrig. 2002. Detection of West
Nile virus antigen in mosquitoes and avian tissues by a monoclonal antibody-
based capture enzyme immunoassay. J. Clin. Microbiol. 4:2023–2030.
18. Ito, Y., J. C. Grivel, and L. Margolis. 2003. Real-time PCR assay of individ-
ual human immunodeficiency virus type 1 variants in coinfected human
lymphoid tissues. J. Clin. Microbiol. 41:2126–2131.
19. Kauffman, E. B., S. A. Jones, A. P. Dupuis II, K. A. Ngo, K. A. Bernard, and
L. D. Kramer. 2003. Virus detection protocols for West Nile virus in verte-
brate and mosquito specimens. J. Clin. Microbiol. 41:3661–3667.
20. Lanciotti, R. S., G. D. Ebel, V. Deubel, A. J. Kerst, S. Murri, R. Meyer, M.
Bowen, N. McKinney, W. E. Morrill, M. B. Crabtree, L. D. Kramer, and J. T.
Roehrig. 2002. Complete genome sequences and phylogenetic analysis of
West Nile virus strains isolated from the United States, Europe, and the
Middle East. Virology 298:96–105.
21. Lanciotti, R. S., and A. J. Kerst. 2001. Nucleic acid sequence-based ampli-
fication assays for rapid detection of West Nile and St. Louis encephalitis
viruses. J. Clin. Microbiol. 39:4506–4513.
22. Lanciotti, R. S., A. J. Kerst, R. S. Nasci, M. S. Godsey, C. J. Mitchell, H. M.
Savage, N. Komar, N. A. Panella, B. C. Allen, K. E. Volpe, B. S. Davis, and
J. T. Roehrig. 2000. Rapid detection of West Nile virus from human clinical
specimens, field-collected mosquitoes, and avian samples by a TaqMan re-
verse transcriptase PCR assay. J. Clin. Microbiol. 38:4066–4071.
23. Lanciotti, R. S., J. T. Roehrig, V. Deubel, J. Smith, M. Parker, K. Steele, et
al. 1999. Origin of the West Nile virus responsible for an outbreak of
encephalitis in the northeastern United States. Science 286:2333–2337.
24. Morrison, T. B., J. J. Weis, and C. T. Wittwer. 1998. Quantification of
low-copy transcripts by continuous SYBR Green I monitoring during am-
plification. BioTechniques 24:954–958, 960, 962.
25. Murphy, F., C. Fauquet, D. Bishop, S. Ghabrail, A. Jarvis, G. Martelli, M.
Mayo, and M. Summers. 1995. Virus taxonomy, classification and nomen-
clature of viruses. Arch. Virol. 10:1–58.
26. Nasci, R. S., D. J. White, H. J. Stirling, J. Oliver, T. J. Daniels, R. C. Falco,
S. Campbell, W. J. Crans, H. M. Savage, R. S. Lanciotti, C. G. Moore, M. S.
Godsey, K. Gottfried, and C. J. Mitchell. 2001. West Nile virus isolates from
mosquitoes in New York and New Jersey, 1999. Emerg. Infect. Dis. 7:626–630.
27. Pfaffl, M. W. 2001. A new mathematical model for relative quantification in
real-time RT-PCR. Nucleic Acids Res. 29:e45.
28. Shi, P.-Y. 2001. High-throughput detection of West Nile virus RNA. J. Clin.
29. Shi, P.-Y., M. Tilgner, M. K. Lo, K. A. Kent, and K. Bernard. 2002. Infec-
tious cDNA clone of the epidemic West Nile virus from New York City.
J. Virol. 76:5847–5856.
30. Shu, P. Y., S. F. Chang, Y. C. Kuo, Y. Y. Yueh, L. J. Chien, C. L. Sue, T. H.
Lin, and J. H. Huang. 2003. Development of group- and serotype-specific
one-step SYBR green I-based real-time reverse transcription-PCR assay for
dengue virus. J. Clin. Microbiol. 41:2408–2416.
31. Smithburn, K., T. Hughes, A. Burke, and J. Paul. 1940. A neurotropic virus
isolated from the blood of a native of Uganda. Am. J. Trop. Med. Hyg.
1518PAPIN ET AL.J. CLIN. MICROBIOL.