Diagnostic Microbiology and Infectious Disease 77 (2013) 129–132

Contents lists available at ScienceDirect

Diagnostic Microbiology and Infectious Disease

j ourna l homepage: www.e lsev ie r .com/ locate /d iagmicrob io

Virology

Quadraplex qRT-PCR assay for the simultaneous detection of Eastern equineencephalitis virus and West Nile virus

Steven D. Zink a,⁎, Susan A. Jones a, Joseph G. Maffei a, Laura D. Kramer a,b

a Arbovirus Laboratory, Wadsworth Center, New State Department of Health, Slingerlands, NY 12159, USAb School of Public Health, State University of New York at Albany, Albany, 12222, NY, USA

a b s t r a c ta r t i c l e i n f o

⁎ Corresponding author. Tel.: +1-518-485-6864; fax:E-mail addresses: szink@wadsworth.org, sdzink@gm

0732-8893/$ – see front matter © 2013 Elsevier Inc. Alhttp://dx.doi.org/10.1016/j.diagmicrobio.2013.06.019

Article history:Received 18 January 2013Received in revised form 28 May 2013Accepted 14 June 2013Available online 23 July 2013

Keywords:Mosquito surveillanceWest nile virusEastern equine encephalitisqRT-PCRMultiplex PCR

In order to increase testing throughput and reduce cost, we developed a multiplex real-time assay thatidentifies both Eastern equine encephalitis virus and West Nile virus. The assay allows for the screening forthe presence of both the nonstructural and envelope genes of both viruses simultaneously allowing forconfirmatory testing to be done in a single assay. We utilized newly designed primers and probes, eachlabeled with a unique fluorescent label allowing for differentiation using an ABI 7500 real-time PCR machine.The use of Quanta Biosciences qScript XLT One-Step RT-qPCR® Toughmix allowed for a quadraplex assaywithout loss of sensitivity when compared to the previously run singleplex reaction as seen with viral RNAPFU control dilution series. There was no cross reactivity between the viruses within the reaction, andupon utilization of the assay during surveillance, there was no cross reactivity with other historicallyencountered arthropod-borne viruses. The results from the quantitative Reverse Transcriptase - PolymeraseChain Reaction were comparable to those achieved by cell culture which was performed on a subset of thefield mosquito pools screened during the 2012 surveillance season. The multiplex assay resulted in savings inboth time and resources for the lab and faster turn-around of results.

+1-518-485-6669.ail.com (S.D. Zink).

l rights reserved.

© 2013 Elsevier Inc. All rights reserved.

1. Introduction

New York State maintains an active surveillance program inmosquitoes, mammals, and birds for Eastern equine encephalitisvirus (Togaviridae: Alphavirus; EEEV) and West Nile virus (Flaviviridae:Flavivirus; WNV) along with multiple other arboviruses present in thestate. Active monitoring of virus within the natural mosquito vector(s)and/or zoonotic hosts allows for dynamic surveillance and threatdeterminationwithin an ecosystem. EEEV is a known causative agent ofhuman disease (Calisher, 1994; Deibel et al., 1979). The fatality rate ofEEEV is between 30 and 70% in humans and leavesmany survivors withmental deficiency (Deresiewicz et al., 1997). WNV is responsible for arange of symptoms from a benign and often symptomless infection tofever to neuroinvasive disease (Hayes and O'Leary, 2004; Ostlund et al.,2000; Sejvar et al., 2003). The 2012 transmission season was especiallysevere for WNV in the United States with the largest number of humancases seen since 2003. A total of 5674 cases of West Nile disease werereported throughout the country resulting in 286 deaths; over 50% ofthe cases were classified to be neuroinvasive (indicating meningitis orencephalitis) (CDC, 2013). No human cases of Eastern equine encep-halitis were detected in New York State in 2012, but 7 cases werediagnosed in Massachusetts and 2 in Vermont. Both viruses continue topose a threat to public health within the United States. During the 2012

season, New York State screened 5780 pools of mosquitoes from 13counties across the state. Brain tissue was also screened from anyneurologic animal that tested negative for rabies and was submittedfrom a region classified as endemic for EEEV or was of equine originfrom throughout the state.

Rapid and accurate testing is essential in a public health laboratorysetting; quantitative Reverse Transcriptase- Polymerase Chain Reaction(qRT-PCR) provides this through the ability to target-specific viruses.Real-time PCR is able to be multiplexed to allow for confirmatory2-target testing to be done in a single tube. By combining the singleqRT-PCR tests for the envelope (E) and non-structural 5 (NS5) genesof WNV and envelope 1 (E1) and non-structural 1 (NS1), of EEEV, wewere able to perform confirmatory testing for the 2 leading arboviralpublic health threats in New York State. This limited the necessity ofcell culture for viral surveillance and reduced the time to reporting by1 to 2 days as well as reducing the cost of testing by over 50%.

2. Materials and methods

2.1. Mosquito and bird specimens

Mosquitoes were collected by county health departments usingCDC light traps or gravid traps, identified and pooled in groups ofapproximately 50 females by species, and transported on dry ice tothe Arbovirus Laboratory of the Wadsworth Center, New York StateDepartment of Health (NYSDOH). Tissues from neurologic mammals

130 S.D. Zink et al. / Diagnostic Microbiology and Infectious Disease 77 (2013) 129–132

were submitted to NYSDOH Rabies Laboratory and tested for rabies;the negative brain samples were then forwarded to the Arboviruslaboratory for EEEV andWNV testing. All specimens were stored at−70 °C before being processed.

2.2. Primer and probe design

The primers and probes for WNV E and NS5 along with EEEV E1and NS1 were designed utilizing the Clone Manager 9 Professionalsuite (Scientific and Educational Software, Cary, NC, USA). Similarregions of the genome were targeted in the singleplex assayspublished previously (Lanciotti et al. 2000; Linssen et al., 2000; Shiand Kramer, 2003; Shi et al., 2001) (Table 1). All primer and probesites were screened against known and submitted sequences inGenBank using BLAST for specificity toWNV and EEEV (Morgulis et al.,2008; Zhang et al., 2000). The E and E1 probes are Minor GroveBinding sequences labeled at the 5′ end with 2,7′,8′-benzo-5′-fluoro-2′, 4, 7,-trichloro-5-carboxyfluorescein (NED) for WNV and Cyanine5 (CY5) for EEEV, respectively, obtained from Applied Biosystems (CA,USA). The WNV NS5 and EEEV NS1 probes were obtained fromIntegrated DNA Technologies (Coralville, IA) and were labeled atthe 5′ endwith 6-carboxy-4′, 5′-dichloro-2′, 7′- dimethoxyfluorescein(JOE) for WNV and Fluorescein (FAM) for EEEV. A black hole quencherwas attached to the 3′ end of each probe aswell by the relative supplier.

2.3. Sample preparation and nucleic acid extraction

Sample RNA was extracted with Magmax Viral isolation kit(Ambion, CA, USA) on a Tecan Evo 150 liquid handler (Tecan,Morrisvelle, NC, USA). For MagMax extraction, pools of 10 to 50individual mosquitoes were homogenized in diluent containing 20%fetal bovine serum, 50 μg of streptomycin per mL, 50 U of penicillin,and 2.5 μg of amphotericin B per mL in phosphate-buffered saline in aRetsch Mixer Mill set to 24 cycles/s for 30 s. The tubes were thencentrifuged for 3 minutes at 12,000 rpm and the supernatantremoved. The supernatant was centrifuged for another 3 minutes at12,000 rpm to remove any contaminating tissue or mosquito debris,after which it was split, ~500 μL into a clean microfuge tube for viralculture and long-term storage. A total of 50 μL of homogenized sampleRNA was extracted by the robotic workstation according to themanufacturer's instructions. The extracted RNA was eluted in a totalvolume of 50 μL of Elution Buffer for direct use in the multiplexassay. Lysed American Crow kidney tissue with known WNV titerwas added to 2 wells of the 96-well plate robotic extraction as apositive extraction control. Mosquito diluent was added to 7 wellsinterspersed throughout the extraction plate in designated wells asnegative extraction and process controls; any positives in these wells

Table 1Primers and probes designed for the multiplex assay.

Primer/probe Sequence 5′-3′ BP Designation Information

EEEV E1 F CCCTAGTTCGATGTACTTCCG 21 PrimerEEEV E1 FAM FAM-CCGCCGATGCAGTG 14 Probe MGB Applied

BiosystemsEEEV E1 R GCATTATGCACTGCCCTTAG 20 PrimerEEEV NS1 F ATGAAGAGCGCAGAAGACCC 17 PrimerEEEV NS1 CY5 CY5-CAGACTCTACCGCTACG-

CAGACAAG25 Probe Blackhole

Quencher IDTEEEV NS1 R GCGTCGACATTACTGTTAGC 20 PrimerWNV E1 F CTACCGTCAGCGATCTC 17 PrimerWNV E1 NED NED-CATGGGAGAAGCTCAC 16 Probe MGB Applied

BiosystemsWNV E1 R CTGGGTCAGCACGTTTG 17 PrimerWNV NS5 F CGCTGTCCCTGTGAATTG 18 PrimerWNV NS5 JOE JOE-TGGTCCATCCATGCAGG-

AGGAGAG25 Probe Zen Blackhole

quencher IDTWNV NS5 R CCAAACACGGTTCCAGAC 18 Primer

invalidated the run and forced a re-extraction to be done on all originalsamples. All work was performed in a BSL-3 facility, all samples weremaintained in biosafety cabinets, and sealed rotors were utilized toprevent exposure to potential aerosols. The Tecan Evo 150 is locatedinside the BSL-3 laboratory but is not itself sealed; therefore, onlylysed virus was allowed to be extracted on that platform.

2.4. Multiplex qRT-PCR

The qRT-PCR reaction was set up following the protocol of theQuanta Biosciences (Gaithersburg, MD, USA) qScript XLT One-StepqRT-PCR Toughmix kit. The final volume of the reaction was 25 μLconsisting of 10 μL of master mix and 5 μL of template. Each reactioncontained 0.3 μL of 100 μmol/L primer and 0.06 μL of 25 μmol/L probe.The cycling conditions were as follows: 5minutes at 50 °C followed by30 s at 95 °C then 40 cycles of alternating 95 °C for 10 s and 60 °C for 1minute. The 1-minute annealing/extensionwas necessary to allow theABI 7500 standard to fully read the 96-well plate for all 4 dyes.

To prepare the viral RNA standards used in the qRT-PCR, stocks ofWNV and EEEV with known PFUs were extracted by RNeasy method.The extracted RNAs were aliquoted in 60-μL volumes to avoidmultiple rounds of freezing and thawing. This procedure ensuredconsistent standards throughout the surveillance season. All datapresented in this paper utilized these controls. Going forward, aplasmid with the designated target regions will be utilized for EEEVbecause of the select agent status of this virus. Two no-templatecontrols were run on each 96-well plate to control for master mixcontamination and the sample addition process.

3. Results

3.1. Comparison of multiplex to singleplex assay

The primers and probes used in the multiplex were designedspecifically to be used in the combined assay while targeting 2separate regions of the 2 genomes, i.e., the E gene and the NS5 gene ofWNV, and the E1 gene and NS1 gene of EEEV. Table 2 provides thecritical threshold (CT) values for a series of standard curves performedwith both the multiplex and singleplex assays. The CTs for themultiplex are comparable to the singleplex assay throughout thestandard curve. It should be noted that, in initial experiments, therewas a 10-fold decrease (~3 CT) when the primers and probes weremultiplexed using the same reaction parameters as the singleplex. Theloss of detection was eliminated with the replacement of master mixused within the setup from the ABI RT-MM to a Quanta BioscienceMaster mix.With the recovery of the 10-fold sensitivity, themultiplexassay was comparable to the singleplex assays and was able tobe utilized for the simultaneous surveillance of the 2 viruses (datanot shown).

Table 2Real-time PCR CT values for standard curves directly comparing the multiplex assay tothe singleplex assays.

Standard curve dilution series Singleplex

PFU EEE NS1CY5

EEEV E1FAM

WNV ENED

WNV NS5JOE

WNV1160

EEEVEE1881

0.1 36.50 36.82 37.76 38.45 40.00 38.981 33.83 34.01 35.66 35.01 36.08 35.4310 31.68 31.84 33.21 32.15 32.15 31.73100 28.88 28.89 29.64 28.80 28.14 28.241000 25.61 25.63 26.20 25.25 24.81 23.55R2 0.9948 0.9867 0.9817 0.9988 0.9977 0.9833

Ten-fold serial dilution of EEEV and WNV RNA was screened by qRT-PCR in triplicate,and average CT was reported. The linearity of the assay is demonstrated by the R2 valueassociated with each primer/probe combination. The multiplex assay was comparableto the singleplex in both sensitivity and linearity.

Table 4Upon completion of the 2012 surveillance season, the total positive mosquito poolinformation indicated that the multiplex results were comparable to the 2011singleplex results.

No. ofpositivepools

Avg CTE1 or E

Avg CTNS

Min/Max CTPositive E1or E

Min/Max CTPositive NS

AverageabsoluteDifferenceof CT

2012 Surveillance Season (quadraplex)EEEV 24 28.03 27.66 16.80/37.88 15.90/37.73 0.33WNV 828 29.74 28.83 20.69/37.30 21.22/37.33 1.23

2011 Surveillance Season (singleplex)EEEV 66 26.73 28.29 21.06/38.87 16.70/39.24 1.02WNV 317 28.73 28.70 19.32/39.86 18.38/39.05 0.33

The assay looked for sequence within the envelope of both viruses (E for WNV and E1for EEEV) and non-structural proteins (NS5 for WNV and NS1 for EEEV). The topportion shows the average and range of CTs obtained during surveillance while usingthe multiplex primers to screen viral regions. The lower part shows the results for 2011surveillance season utilizing the singleplex reactions for the same regions for eachvirus. This indicates that the sensitivity and range of results were comparable between2 full seasons of mosquito surveillance.

Table 5The supernatant from mosquito pools that presented cytopathic effect in cell culturewas harvested and RNA was extracted.

Sample Flanders Bunyavirus CacheValley virus

Potosivirus

ARB 12MS 0001 22.07 − − −ARB 12MS 0002 23.46 − − −ARB 12MS 0003 15.94 − − −ARB 12MS 0004 17.55 − − −

131S.D. Zink et al. / Diagnostic Microbiology and Infectious Disease 77 (2013) 129–132

Positive samples from the 2011 season were utilized to test thefunctionality of the multiplexed qRT-PCR on field samples prior to itsuse during the surveillance season in 2012. Multiple pools of bloodedmosquitoes were chosen that had confirmed positive for WNV orEEEV. Typically, surveillance is performed on non-blooded mosqui-toes, but the presence of blood, which is an inhibitor of PCR, allowedfor a more robust test of the multiplexed qRT-PCR. The CTs for thepreviously screened mosquito pools re-confirmed as positive by thespecific primer/probe sets (Table 3). The CT values were either loweror equivalent to the singleplex assays for EEEV and WNV providingstrong support that the multiplex assay performs well on environ-mental samples (Table 3).

3.2. Surveillance of WNV and EEEV in mosquito pools

In 2012, the Arbovirus Laboratory screened more than 5000 poolsof mosquitoes from across New York and parts of Vermont. Twenty-four EEEV-infected pools and 828 positive WNV-infected pools weredetected (Table 4). The average CT values that were recorded for EEEVand WNV in 2012 were comparable to the average CT values of thesingleplex results of 2011 for both targeted regions. The variationamong the positive values was slightly less in 2012 than in 2011, butthis is a result of increasing the stringency of the cutoff for the assayfrom 40 to 38. This was done to limit the number of samples that wereshowing a positive CT for only a single target, as it was seen uponsubsequent retesting that these samples failed to repeat the positiveresult. Further analysis of the real-time assays for the single positivepools indicated that they were not true amplification events but slowcreep of signal resulting in a false positive; by lowering the CT cutoff(i.e., increasing the stringency of the assay), these were eliminated.The difference between the 2 primer/probe sites within the samepools was very tight for all assays with a difference of CT of less than1.25 in the assay. The lack of variation indicates that all primers andprobes in the assay are able to work at similar efficiency reducing thechance for false negatives being reported. Further, a subset of sampleswas run on tissue culture, and those results were equivalent to themolecular results for WNV and EEEV.

Mosquito-borne viruses detected in NYS include EEEV, WNV,Flanders-like virus, Potosi virus, Cache Valley virus, California Groupbunyaviruses, Jamestown Canyon virus, and Trivittatus virus. Duringthe 2012 season, there were high titer virus isolations of Flanders

able 3T values from blooded mosquitoes.

Blooded mosquito pool CTs

Pool # EEEV E1 FAM EEEV NS5 CY5 EE1881

1 22.3 22.4 20.622 35.7 38.3 36.473 23.5 23.3 21.614 29.4 31 29.875 31.45 29.8 31.1

WNV NS5 JOE WNV E NED WNV 1160

6 28.16 26.86 28.457 29.88 31.88 32.018 27.48 29.43 29.549 26.03 28.44 28.3710 30.36 30.95 31.2411 29.04 30.87 30.8912 25.58 28.45 27.4513 29.41 30.11 29.98

ools of field captured mosquitoes were processed and tested for EEEV and WNV byngleplex PCR using the EE1881 and WNV 1160 primer/probe sets, respectively; theseools were simultaneously screened by the multiplex assay (EEEV E1 FAM, EEEV NS5Y5, WNV NS5 JOE, WNV E NED) primer/probes to determine the direct sensitivityetween the 2 assays. The EEEV and WNV multiplex values for both regions wereomparable to the singleplex values.

ARB 12MS 0005 17.10 − − −ARB 12MS 0006 − +++ − −ARB 12MS 0007 − +++ − −ARB 12MS 0008 − +++ − −ARB 12MS 0009 − +++ − −ARB 12MS 0010 16.58 − − −ARB 12MS 0011 21.25 − − −ARB 12MS 0012 − +++ − −ARB 12MS 0013 − − − +++ARB 12MS 0014 − +++ − −ARB 12MS 0015 23.06 − −ARB 12MS 0016 − − +++ −ARB 12MS 0017 − − − +++ARB 12MS 0018 − − +++ −ARB 12MS 0019 − − +++ −ARB 12MS 0020 21.35 − −ARB 12MS 0021 − − +++ −ARB 12MS 0022 − − +++ −ARB 12MS 0023 − − − +++ARB 12MS 0024 23.23 − − −ARB 12MS 0025 − − +++ −ARB 12MS 0026 − − +++ −ARB 12MS 0027 − − +++ −

The RNA was screened by qRT-PCR for Flanders virus as well as by RT-PCR for CacheValley virus, Potosi virus, and California serogroup Bunyaviruses. The CT values for theFlanders virus real-time assay are shown; a positive was calculated as having a CT b38.For the standard RT-PCR viral assays, the scaled results of the band intensity ascompared to known standards on a scale of (+) to (+++) are shown. Since thesamples had been isolated following amplification in cell culture, they contained highlevels of viral RNA and were all scored at the maximum scale.

TC

PsipCbc

virus, Bunyavirus, California serogroup, Cache Valley virus, and Potosivirus. (Table 5) The specificity of the assay was maintained in themultiplex assaywith no cross reactivity being seen for any of the poolsthat contained the other viruses. In addition, RNA was extracted fromknown viral stocks of the above viruses to confirm that the primersdid not cross react with historically found viruses. It should be notedthat even though the list of viruses tested is limited to relevant virusesin New York State, upon computer analysis, there is no cross reactivityto other known virus sequences. Furthermore, the pools that weredetermined to be positive by PCR for EEEV were all further confirmed

132 S.D. Zink et al. / Diagnostic Microbiology and Infectious Disease 77 (2013) 129–132

by tissue culture methods validating that there were no false positivesin the course of the surveillance season.

3.3. Detection of EEEV and WNV in vertebrate tissue

Virus was detected in tissues utilizing the same techniques asdescribed above, allowing side-by-side processing of mosquito andvertebrate samples. This greatly reduced testing costs by allowingruns of full assays, diminishing reagent waste. Results from RNAextracted from brain samples from neurologic mammals fromendemic EEEV regions are shown for both EEEV and WNV for the2012 surveillance year (Table 6). The assay performed very wellfor these samples, which often had a large amount of virus present,which can interfere with real-time assays. All tissue samples wereconcurrently placed on Vero cell culture to screen for viral replica-tion. The tissue culture and qRT-PCR results were equivalent for allmammal brain tissue samples tested.

4. Discussion

Surveillance for EEEV and WNV is performed in the United Statestomonitor the prevalence of virus within vector species of mosquitoesand vertebrates. This information is provided to local, regional, andstate health departments as well as the CDC, to allow for public healthdecisions to be made. These pre-human infection data allow for thedecision to institute interventions to protect the public before anoutbreak occurs. For the last decade, molecular methods have becomethe primary means of detection of virus because they are fast,accurate, and relatively cheap as compared to culture methods toidentify virus. The standardmolecular assay utilized in the majority oflabs to detect WNV and EEEV was published over a decade ago andworks very well (Lambert et al., 2003; Shi et al., 2001). However, withthe advances in technology in both equipment and reagents, the assaywas able to be updated to allow for it to be quicker and cheaper.

The current protocol can be completed from the time themosquitopools enter the laboratory for RNA extraction to the time the results areentered into the public health reporting systemwithin the timeframe ofa single 8-hour work day. This allows for the analysis of a maximum of152 pools a day tested by 2 full time lab staff; in addition, it allows for6 vertebrate or repeat samples per assay. By targeting 2 viruses in2 regions in a single assay, it is possible to have confirmatory resultsusing a quarter of the cost of master mix and plastic ware. The averagecost from extraction to reporting to screen a pool of mosquitoes or avertebrate brain for both EEEV and WNV dropped from an average of$17.80 to $6.88 based on a 25 μL qRT-PCR reaction. The multiplexing ofthe assay also reduced the time and equipment needed for turnaround.It previously took 4 runs on a real-time instrument to obtain results on

Table 6WNV and EEEV present in mammal neural tissue tested during 2012 by the multiplexassay.

Species Total received Positive WNV Positive EEEV

Alpaca 5 2Cat 3Cattle 1Coyote 1Chipmunk 1Cow 1Deer 28 1Dog 1Donkey 3Goat 1Grey Fox 1Horse 28 1 2Llama 3Moose 2Squirrel 1

pools being tested for both EEEV and WN, which resulted in only asubset of pools that had been put on culture from the EEEV-endemicregions being tested for both viruses. With the implementation of themultiplex that is reduced to a single run that tests specimens for bothviruses simultaneously, surveillance coverage for the presence of virusis greatly increased. The change in master mix from ABI to Quantaalso decreased the time of the reaction by decreasing the RT step from30 minutes to 5 minutes and eliminating the 95° initial incubationto activate the Taq polymerase due to their Taq/Antibody couplingcomplex, which is unique to the Quanta RT-qPCR ToughMix®.

Recently, it has been published that a negligible percentage ofpositive mosquitoes may be missed due to mutations within thetargeted genomic regions (Brault et al., 2012). This multiplex assayscreens 4 regions simultaneously looking at 2 independent targetregions for each virus, thus virtually eliminating that possibility,without additional man hours for tissue culture work. Additionally,when all samples that were submitted for EEEV testingwere inoculatedon cell culture, no discrepant resultswere seen, indicating themultiplexassay was comparable to tissue culture. We have shown that themultiplexed qRT-PCR assay for EEEV andWNV is efficient, low-cost, andhighly sensitive. Looking forward, the aim is to further multiplex theother viruses that are encountered during surveillance to increaseefficiency, decrease cost, and maintain accuracy and sensitivity.

Acknowledgments

This study was funded by the NYSDOH and National Center ofEmerging and Zoonotic Infectious Disease (Epidemiology and Labo-ratory Capacity for Infectious Disease Program). We would like thankthe State of Vermont Agency of Agriculture and the County HealthDepartments of Erie, Oneida, Onondaga, Suffolk, Westchester, Orange,Rockland, Madison, Chautauqua, Oswego, Lewis, and Nassau countiesfor collecting and submitting mosquitoes.

References

Brault AC, Fang Y, Dannen M, Anishchenko M, Reisen WK. A naturally occurringmutation within the probe-binding region compromises a molecular-based WestNile virus surveillance assay for mosquito pools (Diptera: Culicidae). J MedEntomol 2012;49(4):939–41.

Calisher CH. Medically important arboviruses of the United States and Canada. ClinMicrobiol Rev 1994;7(1):89–116.

Centers for Disease Control and Prevention (CDC). West Nile virus. Found at: http://www.cdc.gov/ncidod/dvbid/westnile/index.htm, 2013.

Deibel R, Srihongse S, Woodall JP. Arboviruses in New York State: an attempt todetermine the role of arboviruses in patients with viral encephalitis andmeningitis.Am J Trop Med Hyg 1979;28(3):577–82.

Deresiewicz RL, Thaler SJ, Hsu L, Zamani AA. Clinical and neuroradiographicmanifestations of eastern equine encephalitis. N Engl J Med 1997;336(26):1867–74.

Hayes EB, O'Leary DR. West Nile virus infection: a pediatric perspective. Pediatrics2004;113(5):1375–81.

Lambert AJ, Martin DA, Lanciotti RS. Detection of North American eastern and westernequine encephalitis viruses by nucleic acid amplification assays. J Clin Microbiol2003;41(1):379–85.

Lanciotti RS, Kerst AJ, Nasci RS, Godsey MS, Mitchell CJ, Savage HM, et al. Rapiddetection of West Nile virus from human clinical specimens, field-collectedmosquitoes, and avian samples by a TaqMan reverse transcriptase-PCR assay. J ClinMicrobiol 2000;38(11):4066–71.

Linssen B, Kinney RM, Aguilar P, Russell KL, Watts DM, Kaaden OR, et al. Development ofreverse transcription-PCR assays specific for detection of equine encephalitisviruses. J Clin Microbiol 2000;38(4):1527–35.

Morgulis A, Coulouris G, Raytselis Y, Madden TL, Agarwala R, Schaffer AA. Databaseindexing for production MegaBLAST searches. Bioinformatics 2008;24(16):1757–64.

Ostlund EN, Andresen JE, AndresenM.West Nile encephalitis. Vet Clin North Am EquinePract 2000;16(3):427–41.

Sejvar JJ, Haddad MB, Tierney BC, Campbell GL, Marfin AA, Van Gerpen JA, et al.Neurologic manifestations and outcome of West Nile virus infection. J Am MedAssoc 2003;290(4):511–5.

Shi P-Y, Kramer LD. Molecular detection of West Nile virus RNA. Expert Rev Mol Diagn2003;3(3):357–66.

Shi P-Y, Kauffman EB, Ren P, Felton A, Tai JH, Dupuis II AP, et al. High throughputdetection of West Nile virus RNA. J Clin Microbiol 2001;39(4):1264–71.

Zhang Z, Schwartz S, Wagner L, Miller W. A greedy algorithm for aligning DNAsequences. J Comput Biol 2000;7(1–2):203–14.