Development of Multiplex Real-Time Reverse Transcriptase PCR Assays for Detecting Eight Medically Important Flaviviruses in Mosquitoes

Abstract

A multiplex real-time reverse transcriptase PCR has been developed for the rapid detection and identification of eight medically important flaviviruses from laboratory-reared, virus-infected mosquito pools. The method used involves the gene-specific amplification of yellow fever virus (YFV), Japanese encephalitis virus (JEV), West Nile virus (WNV), St. Louis encephalitis virus (SLEV), and dengue virus (DENV) serotypes 1 to 4 (DENV-1 to DENV-4, respectively) by use of the flavivirus consensus amplimers located at the RNA-dependent RNA polymerase domain of nonstructural protein 5. Virus-specific amplicons were detected by four newly characterized TaqMan fluorogenic probes (probes specific for YFV, JEV, WNV, and SLEV) and four previously published probes specific for DENV-1 to -4 (L. J. Chien, T. L. Liao, P. Y. Shu, J. H. Huang, D. J. Gubler, and G. J. Chang, J. Clin. Microbiol. 44:1295-1304, 2006). This assay had a specificity of 100% and various sensitivities of at least 3.5 PFU/ml for YFV, 2.0 PFU/ml for JEV, 10.0 PFU/ml for WNV, and 10.0 PFU/ml for SLEV. Additionally, we have developed an in vitro transcription system to generate RNase-resistant RNA templates for each of these eight viruses. These templates can be incorporated into the assay as RNA copy number controls and/or as external controls for RNA-spiked mosquito pools for quality assurance purposes. Although further study with mosquitoes collected in the field is needed, the incorporation of this assay into mosquito surveillance could be used as an early-warning system for the detection of medically important flaviviruses, particularly when the cocirculation of multiple viruses in the same region is suspected.

Flaviviruses are significant causes of disease worldwide and can be classified serologically into several antigenic complexes (12, 20). Medically important members of the Japanese encephalitis virus (JEV) serocomplex include JEV, St. Louis encephalitis virus (SLEV), and West Nile virus (WNV). Each of these viruses causes similar disease syndromes in humans, ranging from an asymptomatic or a mild flu-like illness to clinical encephalitis (18). Until 1999, SLEV was the only mosquito-borne flavivirus causing significant human morbidity and mortality in the United States (22). WNV, traditionally found in Europe, Africa, the Middle East, and Asia, emerged in New York City in the late summer of 1999 (1). By 2005 the distribution of WNV had expanded into areas of known recent SLEV activity, such as the states of Florida, Mississippi, Louisiana, Texas, Arizona, Colorado, and California (3). Mosquito surveillance for detection of virus activity during the transmission season is an essential tool for implementation of an effective mosquito control strategy. As WNV and SLEV occupy similar ecological niches in North America, it is critical to develop a cost-effective assay capable of detecting the presence of either one or both viruses simultaneously in the mosquito pools. In addition, dengue virus (DENV) serotypes 1 to 4 (DENV-1 to DENV-4, respectively), SLEV, and yellow fever virus (YFV) are endemic in Latin America (5). The situation is further complicated by the detection of WNV activity in Mexico and Central America (3). Therefore, the development of a real-time, multiplex reverse transcriptase (RT) PCR platform for the simultaneous detection of viral RNA genomes in the Americas is crucial for supporting mosquito surveillance efforts and clinical diagnosis.

The RNA-dependent RNA polymerase (RdRp) domain, located at the carboxy terminus of nonstructural protein 5 (NS5), is the most conserved coding region in the Flavivirus genomes (15, 16, 23). The consensus sequence primers (primers FU1 and CFD2; YFV genomic positions, nucleotide [nt] 8997 and nt 9233, respectively) have been used in our previous study for genetic characterization of this domain for all registered flaviviruses (16). The nucleotide sequences flanked by these two primers were variable among the 72 aligned virus sequences. The potential to formulate a one-tube, multiplex TaqMan assay for detection of DENV-1 to -4 was demonstrated in our previous publication (4). In the present study, we used this pair of consensus primers; further explored the variability of this region for the design of virus-specific probes; and optimized a multiplex RT-PCR for the simultaneous detection of viral RNA genomes of YFV, JEV, WNV, and SLEV from virus-containing cell culture supernatants. We examined the sensitivities and specificities of these four virus probes based on viral RNAs extracted from viruses of known titer. In addition to these four virus probes, we also included the previously described DENV-1 to -4 multiplex TaqMan assay (4) and developed eight virus-specific, RNase-resistant in vitro-transcribed RNA templates for use as copy number and external RNA controls. These two real-time, one-tube, multiplex TaqMan assays were then used to detect RNAs extracted from the laboratory-reared, virus-inoculated mosquitoes.

TaqMan primers and probes.

The RdRp domain at the carboxy-terminal one-third of NS5 is the most conserved region in the Flavivirus genomes (16, 23). The flavivirus consensus amplification primers (amplimers) and virus-specific probe sequences for WNV, JEV, SLEV, and YFV were located within this region (Fig. 1; Table 1). When these probes were designed, potential mismatched bases were taken into consideration by aligning the sequences of different strains of these four different viruses available from the GenBank databases (data not shown). The DENV serotype-specific probes for the four different serotypes of DENV were included in this study to detect DENV-1 to DENV-4 (4). The 5′ 6-carboxyfluorescein (FAM) reporter dye and 3′ black hole quencher 1 (BHQ1) thermoquench-labeled probes for YFV (YFVP), JEV (JEVP), WNV (WNVP), and SLEV (SLEVP) were used to optimize the assay. Amplimers (Table 1, mFU1 and CFD2) were synthesized by the CDC core facility, and the fluorogenic probes were synthesized by Operon Biotechnologies, Inc. (Huntsville, AL). After the evaluation of the specificities and sensitivities of YFVP, JEVP, WNVP, and SLEVP with a single dye, the probes were changed to four unique reporter dyes (Texas Red, 4,4,7,2′,4′,5′,7′-hexachloro-6-carboxy-fluorescein [HEX], FAM, and Cy5, respectively) for the multiplex application.

FIG. 1.

Multiple-nucleotide-sequence alignment of the YFV group, the DENV group, and the JEV group. The aligned region corresponds to nucleotides 8997 to 9258 of YFV. The 50% majority consensus sequence is listed as the reference to highlight the conservation of this region. In addition to the eight viruses used in this study, viruses in the YFV group (Sepik virus [SEPV], Yokose virus [YOKV] group, and Entebbe bat virus [ENTV] group) and the JEV group (Murray Valley encephalitis virus [MVEV] and Usutu virus [USUV]) are aligned with the respective virus group. The flavivirus consensus amplimers (amplimers mFU1 and CFD2) and eight virus-specific probes are highlighted in boldface and italic.

TABLE 1.

Oligonucleotide amplimers and probes used in the TaqMan RT-PCR

Oligonucleotide	Identification	Specificity	Sequence (5′-3′)	Tma (oC)	5′ Reporter		3′Quencher
					Single	Multiplex
Amplimer	mFU1	Flavivirus	TACAACATGATGGGAAAGCGAGAGAAAAA
	CFD2	Flavivirus	GTGTCCCAGCCGGCGGTGTCATCAGC
Probe	WNVP	WNV	TGCGTGAAGTTGGCACCCGGCCT	62	FAM	FAM	BHQ1
	YFVP	YFV	TCAGAGACCTGGCTGCAATGGATGGT	61	FAM	Texas Red	BHQ2
	SLEVP	SLEV	TGCAAGAAATCTCCCAAATCCCAGGAGGA	60	FAM	CY5	BHQ3
	JEVP	JEV	TCCGTGACATAGCAGGAAAGCAAG	57	FAM	HEX	BHQ1

^aT_m, melting temperature.

Viruses.

Yellow fever virus strain 17D, JEV strain SA14-14-2, WNV strain NY99, SLEV strain MSI-7, DENV-1 strain Hawaii, DENV-2 strain 16681, DENV-3 strain H87, and DENV-4 strain H241 were propagated in C6/36 cells, harvested, aliquoted, and stored at −70°C for later use. Virus titers in the tissue culture fluid were determined by plaque assays with Vero cells in a six-well plate format with a double overlay of 2% agarose and 3% neutral red. The plaques were counted, and the units were converted to the numbers of PFU per milliliter.

Real-time TaqMan assay.

Viral RNA was isolated by using a QIAamp viral RNA kit (QIAGEN, Valencia, CA), and 10 μl of the 50-μl RNA eluate was used for amplification by a one-step RT-PCR. For every assay, homologous viral RNA and water were included as positive and negative controls, respectively. The reactions were performed in a 96-well plate and were subjected to amplification in an iCycler IQ system (Bio-Rad Laboratories, Hercules, CA). The amplification and detection conditions were the same as those described previously (4), and the PCR amplification products were directly detected by monitoring the increase in fluorescence of a dye-labeled oligonucleotide probe.

Briefly, 10 μl of RNA isolated from the tissue culture fluid or mosquito pool was mixed with 100 pmol of the mFU1 and CFD2 amplimers (final concentration, 0.5 μM), 25 pmol each of the virus-specific probes (final concentration, 0.2 μM), and the mixture provided with the iScript one-step RT-PCR ready-mix kit (Bio-Rad Laboratories) in a 50-μl reaction volume. The thermocycler program consisted of a 30-min RT step at 50°C, followed by 1 cycle of 95°C for 15 min to activate the hot-start Taq enzyme and 45 cycles of 95°C for 15 s and 48°C for 3 min with continuous fluorescence data collection. The 5′ nuclease TaqMan assay relies on the 5′ exonuclease activity of the Taq polymerase to free the reporter dye in the quenched probe. The accumulation of the amplified PCR product is proportional to an exponential increase in the fluorescence emitted from the freed reporter dyes (ΔR_n). The detection threshold (C_T) is the cycle number at which the amplification curve crossed the threshold determined by the average of the ΔR_n of the samples between cycles 2 and 10. The C_T value used to estimate the copy number of the initial viral template should remain within the range of linearity of a standard curve with a minimum correlation coefficient of 0.98. Lower C_T values correspond to higher copy numbers of the initial viral template, and a C_T value of 42 or more cycles is considered a negative result.

In vitro-transcribed RNA controls.

In order to develop a reliable, quantitative, safe, and stable standard, we used the in vitro-transcribed RNA as the copy number control. These templates were developed by amplifying YFV, JEV, WNV, SLEV, and DENV-1 to -4 with the flavivirus consensus primers (primers mFU1 and CFD2) by using a Titan one-tube RT-PCR kit (Roche Diagnostics, Indianapolis, IN). The resulting PCR products were cloned into a pCR-II-TOPO vector (Invitrogen Corp., Carlsbad, CA), and the orientation of the correct sequences within the recombinant plasmids was determined by nucleotide sequencing. The positive- or negative-sense RNA (depending upon the orientation of the insert in the pCR-II-TOPO vector) was transcribed with either T7 or SP6 RNA polymerase by using AmpliScribe or DuraScribe T7/SP6 transcription kits (Epicenter, Madison, WI), according to the manufacturer's protocol. The RNase-sensitive or -resistant templates were produced by incorporating ATP, CTP, UTP, and GTP or replacing the CTP and UTP with 2′-fluorine-dCTP (2′-F-dCTP) and 2′-fluorine-dUTP (2′-F-dUTP), respectively. The resulting RNAs were treated with DNase I to remove the DNA templates, purified with an RNA column (QIAGEN), quantified by spectrophotometry, and expressed as the copy numbers/ml. The sizes of the in vitro runoff transcripts were 365 and 371 nucleotides for the transcripts derived from the SP6 and T7 promoters, respectively, both of which contain 266 bases of virus-specific sequence, as shown in Fig. 1.

Mosquito inoculation.

Female Aedes aegypti (Puerto Rico strain) were infected with YFV and DENV-1 to -4 and female Culex pipiens were infected with JEV, WNV, and SLEV by intrathoracic injection. The infected mosquitoes were kept at 28°C for 7 days before they were harvested. The infected mosquitoes were pooled with uninfected mosquitoes at ratios of 1:25 and 1:50, with a maximum of 50 mosquitoes per pool, and were stored at −70°C until they were tested. The pools were homogenized in an MM300 mixer mill (QIAGEN) at 25 Hz for 5 min in the presence of 1 ml BA-1 (Dulbecco minimal essential medium with 3% bovine serum albumin) diluent and one 3-mm sterilized tungsten-carbide bead. The uninfected A. aegypti and C. pipiens mosquito controls and an extraction control consisting of 1 ml cold BA-1 diluent were processed in parallel with each batch of samples. RNA was extracted from 140 μl of the supernatants from each homogenized sample by using a QIAamp viral RNA kit (QIAGEN), eluted in 50 μl of elution buffer according to the manufacturer's protocol, and stored at −70°C before it was tested.

Mosquito-spiked external positive controls.

In order to test whether the in vitro-transcribed RNase-resistant templates could be used as external positive controls, the uninfected mosquito pools of 25 or 50 mosquitoes from both A. aegypti and C. pipiens were homogenized in the presence of 1 ml BA-1 diluent. Mixtures of 140 μl of supernatant from each homogenized sample and 10 μl of 10-fold serial dilutions of the RNase-sensitive or -resistant template with copy numbers ranging from 1 to 10,000,000 for YFV, JEV, WNV, and SLEV were subjected to RNA extraction. The final extracts were eluted in 50 μl for testing.

Sensitivity and specificity evaluation.

The sensitivities and specificities of the real-time RT-PCR assays for YFV, JEV, WNV, and SLEV were evaluated by testing FAM-labeled virus-specific probes against a known virus panel, including WNV (titer, 1.0 × 10⁸ PFU/ml), JEV (titer, 2.0 × 10⁷ PFU/ml), YFV (titer, 3.5 × 10⁷ PFU/ml), and SLEV (titer, 1.0 × 10⁸ PFU/ml). The detection threshold of each assay was determined by using 10-fold serial dilutions of RNA extracted from these stock viruses. The assay sensitivities were 3.5 PFU/ml for YFV (C_T = 40.8), 2.0 PFU/ml for JEV (C_T = 39.8), 10.0 PFU/ml for WNV (C_T = 40.6), and 10.0 PFU/ml for SLEV (C_T = 41.6). We also tested the specificity of the probe by use of the flavivirus study panel, which consisted of YFV, JEV, WNV, SLEV, and DENV-1 to -4. Each probe was tested against eight different viral RNAs, and only homologous viral RNA templates extracted from the stock virus were detected. For instance, the YFV-specific probe detected only YFV RNA and did not detect the other seven flaviviruses in the test panel. We then replaced the single fluorophore (FAM) dye in the four virus-specific probes with Texas Red for the YFV-specific probe, Cy5 for the SLEV-specific probe, HEX for the JEV-specific probe, and FAM for the WNV-specific probe (Table 1). The four-color multiplex RT-PCR was performed under the same conditions as the single-dye assay with FAM. Again, the mixed four-color probes were tested against eight different viral RNAs, and the specificities were 100%, with only the homologous viral RNA detected in the multiplex assay. No cross-talking between different signal channels was observed in the multiplex assay. The detection sensitivities and specificities were compatible for the single-dye and multiplex assays (Fig. 2). The correlation coefficient of the C_T value versus the input virus titer (in PFU) was as high as 0.99 for the four different viruses in duplicate experiments. These results indicate that these newly developed real-time RT-PCR assays are highly effective, with a linear dynamic range of detection of 10¹ to 10⁷ PFU/ml.

FIG. 2.

Performance of the TaqMan RT-PCR assay for detection of four different flaviviruses in a single-dye or four-dye multiplex format. The detection thresholds, expressed as C_T values representative of the two experiments, are plotted against the log of a known infectivity (PFU per reaction) of YFV, JEV, WNV, and SLEV in four-tube, single-dye probe assays (A) or in a one-tube, four-color multiplex probe assay (B).

Comparison of in vitro-transcribed RNase-resistant and RNase-sensitive template as copy number control.

The minimum detection threshold of a TaqMan assay, expressed as the C_T value, correlates with the copy numbers of the nucleic acid templates in the assay specimens, which in this case is the copy number of the reverse-transcribed cDNA from template viral RNA. The quantitative relationship between the infectivity measured by PFU and the true copy number of the viral RNA in the virus seed culture is usually unknown. In addition, reagent preparation and sample handling must be carried out with extreme care to prevent RNase contamination, which may result in RNA degradation. Thus, it is challenging to develop a reliable and reproducible copy number control for real-time RT-PCR assays. In this study, we prepared the RNase-resistant positive- or negative-sense RNA copies by incorporating RNase A-resistant 2′-F-dCTP and 2′-F-dUTP bases into the templates by in vitro transcription with recombinant plasmids containing the viral sequences amplified with primers mFU1 and CFD2. To determine the ability of our multiplex assay to detect the RNase-resistant templates, the RNase-sensitive templates were transcribed and used in the side-by-side study for comparison. Again, the probe detected only the homologous, in vitro-transcribed RNA templates, regardless of their orientations. Our assay had an identical sensitivity of detecting the positive- or negative-sense RNase-resistant templates (10⁴, 10³, 10³, and 10⁴ copies of RNA per reaction mixture for YFV, JEV, WNV, and SLEV, respectively). Interestingly, the assay consistently showed a 1-log-unit difference between detecting RNase-resistant and RNase-sensitive RNA. The minimum detection thresholds were 10³, 10², 10², and 10³ copies of RNase-sensitive RNA per reaction for YFV, JEV, WNV, and SLEV, respectively (Table 2).

TABLE 2.

Comparison of sensitivities and specificities of TaqMan one-step RT-PCR assay with the in vitro transcribed RNA and spiked 25- or 50-mosquito pools

Sample	Quantity (no. of copies/reaction)	Avg CTa
		RNase-sensitive template	RNase-resistant template	RNase-resistant template spiked 25-mosquito pools	RNase-resistant template spiked 50-mosquito pools
YFV	10,000,000	22.59	28.88	29.18	29.02
	1,000,000	27.29	33.08	33.25	33.08
	1,00,000	32.06	37.46	37.59	37.46
	10,000	36.10	40.54	41.04	41.54
	1,000	40.17	NA	ND	ND
	100	NA	NA	ND	ND
	10	NA	NA	ND	ND
	1	NA	NA	ND	ND
JEV	10,000,000	19.37	23.54	24.14	24.00
	1,000,000	24.91	28.00	28.79	28.56
	100,000	28.01	32.46	33.12	33.69
	10,000	32.18	37.01	37.81	38.01
	1,000	36.78	41.12	ND	ND
	100	40.52	NA	ND	ND
	10	NA	NA	ND	ND
	1	NA	NA	ND	ND
WNV	10,000,000	10.92	21.74	22.14	22.52
	1,000,000	15.55	26.03	26.20	27.00
	100,000	20.01	30.22	30.81	31.26
	10,000	24.96	34.92	35.32	35.98
	1,000	29.37	39.21	ND	ND
	100	37.02	NA	ND	ND
	10	NA	NA	ND	ND
	1	NA	NA	ND	ND
SLEV	10,000,000	18.78	26.12	27.00	27.61
	1,000,000	23.44	30.74	31.42	31.74
	100,000	27.97	35.10	35.90	36.10
	10,000	32.54	39.58	40.08	40.02
	1,000	37.07	NA	ND	ND
	100	NA	NA	ND	ND
	10	NA	NA	ND	ND
	1	NA	NA	ND	ND

^aAverage C_T for triplicate reactions. NA, data beyond detection limit; ND, not done.

RNase-resistant template as positive external control.

The RNase-sensitive templates were rapidly degraded by the presence of RNase in the homogenized mosquito preparation (data not shown). Thus, we chose to use the positive-sense RNase-resistant-template-spiked mosquito supernatants as an external control to monitor RNA extraction, RT-PCR, and possible enzyme inhibition. Similar detection sensitivities were observed for the same RNA template spiked with the 25- or the 50-mosquito pool and the RNA-template-spiked water control in the assay without enzyme inhibition (Table 2). These results suggest that the in vitro-transcribed RNase-resistant template can be used as an external control for the detection of flaviviruses in mosquito pools.

Detection of eight different viruses in mosquito pools.

The feasibility of using multiplex probe assays for the detection of eight different flaviviruses was tested via two assays. One assay was designed to detect virus in a pool of mosquitoes infected with any combination of YFV, JEV, WNV, and SLEV; the other assay was designed to detect virus in a pool of mosquitoes experimentally infected with any combination of DENV1, -2, -3, and -4. The assays were able to detect 1 virus-positive mosquito among 25 or 50 uninfected mosquitoes for eight different viruses without any observable false-negative results or cross-contamination.

Discussion.

The emergence of flaviviruses has become an important issue of great public health concern due to the recent incursion and continued transmission of WNV in the Americas and JEV in Oceania (9, 19). Sporadic cases of encephalitis, like those caused by the enzootic flaviviruses, require effective surveillance programs to identify areas of transmission and to enable the immediate implementation of procedures aimed at reducing transmission to humans, including vector control and vaccination. In this study, we successfully developed a multiplex TaqMan RT-PCR assay that simultaneously detected four different flaviviruses, including YFV, JEV, WNV, and SLEV, in a one-tube reaction. This assay has proven specific and sensitive and decreases the risks of laboratory contamination.

Compared to the single-dye real-time assay, this multiplex platform reduces labor and reagent costs. More importantly, this and our previous studies demonstrate the potential for formulating a unified platform, based on the flavivirus consensus amplimers and virus-specific probes located at the RdRp domain of the NS-5 gene, for detection of the presence of any one or a combination of four different viral RNAs in a single-tube reaction. Other published studies have used either four different reactions in four separate reaction tubes (2, 6, 11, 13) or a two-color reaction in one tube (10, 21, 24, 25). The replacement of a fluorescent light quencher, such as 6-carboxy-N,N,N′,N′-tetramethylrhodamine, with a nonfluorescent thermoquencher (BHQ) allows the fluorescent reporters to use greater regions of the spectrum, thus allowing additional reporters to be incorporated into additional probes specific for additional viral RNA types. The same fluorescence combination format has been applied for detection of DENV-1 to -4 in serum specimens from patients with dengue fever (4, 14). The combination of universal primer pairs with four virus-specific probes simplifies the optimization step (4; this study); in contrast, other multiplex designs require four unique primer pairs and four different probes (2, 14).

Another significant improvement is the development of an in vitro-transcribed RNase-resistant RNA template for use as the copy number control. This template control avoids both the hazards of handling infectious viruses and the difficulty in converting the numbers of PFU into reliable RNA copy numbers. Generally, RNA copy numbers were about 1,000 to 3,000 per 1 PFU in our study; however, converting infectivity (PFU) to an RNA copy number contains a level of uncertainty exacerbated by differences among virus strains, tissue culture conditions, and the length of time that the virus was stored. In addition, viral RNA is labile and is subject to rapid degradation by the presence of RNase and by storage duration and/or conditions. Consequently, the use of RNA extracted from virus-infected culture material is often unreliable, in particular among different laboratories. The use of in vitro-transcribed RNase-resistant RNA templates reduces the problems associated with the handling of RNA and allows reliable quantification. The sensitivity of the real-time RT-PCR assay in amplifying an in vitro-transcribed RNase-resistant RNA template was independent of the virus species, although assays with RNase-resistant RNA templates were consistently shown to require a 1-log-unit higher copy number than assays with an in vitro-transcribed RNase-sensitive RNA template. This difference may be due to the reduced enzymatic ability of the RT enzyme to convert the RNase-resistant RNA templates to cDNAs.

Another advantage of our in vitro-transcribed RNase-resistant RNA templates is their ability to be incorporated as positive external controls for viral RNA extraction from mosquito pools. Routine mosquito surveillance necessitates quality control to ensure the validity of the method as well as the detection of false-negative results due to a potential failure during RNA extraction or the presence of residual enzymatic inhibitors in mosquito extracts. Traditional methods of using external quality assurance in PCR detection methods are accomplished by spiking mosquito samples with live viruses of known infectivity. Live viruses are not stable at room temperature and increase the likelihood of false-positive results due to cross-contamination (17). Armored RNA is formed by the assembly of bacteriophage coat proteins around a transcribed RNA target to form pseudoviral particles of defined target sequences, which stabilizes the RNA transcripts and which protects them from nuclease degradation. Although we have not conducted a side-by-side study due to template incompatibility, our in vitro-transcribed RNase-resistant template is expected to perform as well as the “armored RNA” product (Ambion, Inc., Austin, TX) (7). However, the simplicity of preparing an RNase-resistant RNA template should permit the routine application of this methodology in laboratories so that they may develop their own copy number control to monitor the processes of RNA extraction and detection of flaviviruses and other single-stranded RNA viruses as well.

Sequence similarity calculations revealed that the genes encoding NS5 are the most conserved regions in the coding region of the flavivirus genome. The viability of using this region for the development of a real-time, multiplex RT-PCR to detect various flaviviruses has been demonstrated in this study for the detection of YFV, WNV, JEV, SLEV, and DENV-1 to -4 in virus-infected mosquito pools. We have also demonstrated the feasibility of this platform for use for the detection of DENV-1 to -4 in serum specimens from patients with dengue fever (4). The use of flavivirus-consensus amplimers (primers mFU1 and CFD2) and virus-specific probes has great potential to differentiate virus species within a geographic region. For example, flaviviruses, including DENV-1 to -4, SLEV, YFV, Bussuquara virus (BSQV), Cacipacore virus, Iguape virus (IGUV), Ilheus virus, and Rocio virus (ROCV), have been isolated from mosquitoes, animals, and/or humans in Brazil (5, 8). Ideally, in Brazil the relevant probe combinations would include YFV and DENV-1 to -4 in one five-color multiplex assay and WNV, SLEV, ROCV, BSQV, and IGUV in a second five-color multiplex assay. We have shown the specificities of the probes for YFV, DENV-1 to -4, WNV, and SLEV, while in Brazil, ROCV-, BSQV-, and IGUV-specific probes were 100% virus specific in detecting viral RNA extracted from virus seeds (data not shown). Recent improvements in instrument design, such as those found in the Bio-Rad iCycler IQ-5 system, may provide an opportunity for the development of various fiveplex assays based on five-probe combinations for the detection of various flaviviruses. These assays would be well suited for routine mosquito surveillance and clinical diagnosis.