Logo of jcmPermissionsJournals.ASM.orgJournalJCM ArticleJournal InfoAuthorsReviewers
J Clin Microbiol. Apr 2006; 44(4): 1295–1304.
PMCID: PMC1448645

Development of Real-Time Reverse Transcriptase PCR Assays To Detect and Serotype Dengue Viruses


Serotyping dengue virus (DENV) from suspect human specimens is crucial for developing sound epidemiological control measurements early in the transmission season and for effective patient management. We modified DENV consensus D1 (mD1) and serotype-specific TS2 (mTS2) and redesigned serotype-specific TS1 (rTS1) and TS4 (rTS4) as described previously in the conventional capsid and premembrane gene (C-prM) protocol (R. S. Lanciotti, C. H. Calisher, D. J. Gubler, G.-J. Chang, A. V. Vorndam, J. Clin. Microbiol. 30:545-551, 1992). In addition, we designed two new sets of amplimers and probes, located at nonstructural protein 5 (NS5) and the 3′ noncoding region (3′NC) of DENV. The NS5 protocol utilizes two flaviviral consensus outer amplimers (mFU1 and CFD2) and four dengue virus serotype-specific TaqMan fluorogenic probes. The 3′NC protocol uses two DENV consensus amplimers, DC10418 and CDC10564. The conventional gel-based, heminested detection method was adapted for the C-prM protocol for detecting and serotyping dengue viruses. In addition, we developed the real-time SYBR green I and postamplification melting temperature curve analysis for the mD1/TS and 3′NC protocols using identical amplification conditions. The NS5 amplimer/probe set was formulated as a one-tube, multiplex, real-time reverse transcriptase PCR for serotype identification. Three sets of amplimers and probes were verified for their specificity in tests with yellow fever, Japanese encephalitis, St. Louis encephalitis, and West Nile viruses; optimized against 109 DENV strains; and validated for detection of the virus in sera from two different panels of acute-phase human dengue serum specimens and one panel of virus isolates from dengue patients' serum specimens. Clinical evaluation by two separate laboratories indicated that the C-prM was more sensitive (100%) than the NS5 (91%) or the 3′NC (91%) protocol.

Dengue fever and dengue hemorrhagic fever (DHF)/dengue shock syndrome can be caused by any one of the four dengue virus serotypes (DENV-1 to -4). The disease is endemic in Central Africa, the Americas, the Saudi peninsula, Southeast Asia, and the Western Pacific. Prior to 1970, only nine countries had experienced a DHF epidemic; by 1995, the number had increased more than fourfold. In the 1950s, an average of 1,000 DHF cases per year was reported to the World Health Organization (WHO). During the period from 1990 to 1998, the annual average number of cases worldwide increased to half a million. In 1998 alone, a total of 1.2 million cases of dengue and DHF were reported to the WHO, including 15,000 deaths. It is estimated that 51 million infections may occur each year. Factors contributing to increased dengue transmission include the rapid expansion of urbanization, an inadequate supply of drinking water, the increased movement of human populations within and between countries, and the development of insecticide-resistant mosquito populations (http://www.who.int/ctd/dengue/burdens.htm). This trend is expected to continue until concerted, effective mosquito control measures are implemented or effective vaccines are developed.

The majority of diagnostic laboratories employ tissue culture to isolate virus and serological methods to confirm the identity of the DENV isolate (25). This process takes considerable time, during which both clinical and epidemiological information is critical to implement treatment and control measures. The reverse transcriptase PCR (RT-PCR) for amplification of target nucleic acid sequences has provided a rapid and sensitive method for DENV identification and early detection. Conventional methods for detection of PCR-amplified DNA (amplicons) can be grouped into three general categories. The agarose gel electrophoresis-based methods rely upon electrophoresis of the nucleic acids in the presence of ethidium bromide and visual analysis of the resulting bands illuminated by UV light (14), Southern blot methods use labeled oligonucleotide probe hybridization to detect an amplicon (5), and the colorimetric enzyme-linked immunosorbent assay utilizes a biotin-streptavidin interaction to capture and a digoxigenin-specific antiserum to detect an amplicon labeled with a single biotin motif and multiple digoxigenin motifs (2). These methods require multiple handling steps and increase the risk of false-positive results due to amplicon contamination. A recent report reviewed the performance differences and advantages of the four most commonly used conventional RT-PCR assays for detecting dengue viral RNA in clinical specimens (19). Those authors concluded that the heminested protocol using amplimers located at the junction region of the capsid and premembrane genes (C-prM) of DENV was the most sensitive method among them (14). The C-prM protocol utilizes a DENV consensus sequence for outer amplimers D1 and D2 in an initial RT-PCR, followed by a subsequent serotype-specific heminested PCR, combining D1 with one or more of the following serotype-specific internal amplimers: TS1, TS2, TS3, and TS4. In spite of this, several authors have reported false-negative PCR results using this protocol due to a mismatch between the dengue viral RNA sequence and the D1, D2, or TS sequence (8, 20; C. Chin, personal communication).

Advances in the development of fluorophores and nucleotide-labeling chemistries have resolved the need for postamplification manipulations required by conventional RT-PCR and have provided the capability to conduct real-time PCR in a routine diagnostic laboratory (9, 10, 30). However, the fundamental concern regarding false-negative results due to a mismatch in sequences between the amplimer and probe and the continual evolution of the viral RNA or variant viral sequence has not been resolved. Usually, assays are based on a limited number of viral sequences for the construction of the amplimers and probes. The amplimers and probes are then tested against “prototype viruses” to optimize the assay, which is then used to evaluate viruses in clinical specimens.

In this study, we modified amplimers from one existing protocol at the C-prM junction (14) and designed two new sets of amplimers and probes at the nonstructural protein 5 (NS5) and 3′ noncoding (3′NC) regions. Assay protocols were formulated that could be implemented by the conventional gel-based, heminested detection method, the real-time SYBR green I protocol and postamplification melting temperature (Tm) curve analysis, or the multiplex TaqMan. To ensure assay specificity, sensitivity, and detection capabilities, all three amplimer and probe sets were optimized against two viral panels, the flaviviral panel (FP) and the global dengue virus panel (GDP). Protocols were tested concurrently using two panels of acute-phase human dengue serum specimens and one panel of specimens infected with multiple dengue virus serotypes (MDSP).



Three virus panels, the FP, GDP, and MDSP, were assembled for assay development. The FP consists of one representative for each virus, specifically, DENV-1 (Hawaii), DENV-2 (16681), DENV-3 (H87), DENV-4 (H241), Japanese encephalitis virus (SA14-14-2), St. Louis encephalitis virus (MSI-7), West Nile virus (NY99), and yellow fever virus (17D). Dengue virus strains in the GDP were selected from the dengue virus collection maintained at the Division of Vector-Borne Infectious Diseases (DVBID), Centers for Disease Control and Prevention (CDC), Fort Collins, CO. The GDP consists of 32, 38, 14, and 25 unique strains of DENV-1, DENV-2, DENV-3, and DENV-4, respectively, representing the latest isolate of each serotype from each unique country or geographic region in the DVBID collection (Table S1 in the supplemental material). The MDSP consists of 34 viruses available in the DVBID collection (Table S2 in the supplemental material). Viruses comprising the GDP and MDSP were isolated from their original sources (Tables S1 and S2 in the supplemental material) via infection of C6/36 cells and typed using serotype-specific monoclonal antibodies (MAbs) (7; B. Cropp, personal communication). All viruses used in this study were propagated by one additional passage in C6/36 cell culture, aliquoted, and stored at −70°C for future use. The infectivities of the viruses contained in the FP were determined by plaque assay using Vero cells in the six-well plate, double-overlay method and were expressed as numbers of PFU per ml (11).

Clinical specimens.

The human serum specimens used in this study were provided by the Department of Health, Center for Disease Control, Taipei, Taiwan, Republic of China. Dengue virus infection was defined as a febrile illness and confirmed by isolation of dengue virus from acute-phase serum and/or successful amplification of viral RNA from acute-phase serum via RT-PCR or an at least fourfold increase of dengue virus-specific immunoglobulin M or G antibody in paired serum specimens (23, 24). Specimens collected during the period between day 1 and day 7 after the onset of symptoms are referred to as acute-phase specimens. Two human serum collections were used to evaluate the performance of the assays. The collection with a defined infectious viral titer was applied to estimate the sensitivity of the multiplex NS5 TaqMan assay in the clinical laboratory environment.

RNA extraction.

Total RNA was extracted from 140 μl of virus-infected tissue culture fluid or human serum using a QIAmp viral RNA kit (QIAGEN, Inc., Valencia, Calif.), according to the manufacturer's suggested protocol. RNA was eluted in 50 or 70 μl of nuclease-free water. For the conventional gel-based platform, a minimum of two positive-control RNAs from prediluted virus-infected tissue culture fluid with various amounts of target RNA present (i.e., predetermined strong and weak positives) was included as part of the extraction procedure. For the quantitative SYBR green and TaqMan assays, a 10-fold-dilution series containing a known infectivity of target viral RNA was used for RNA extraction.

Oligonucleotide design and test development.

Flavivirus genus sequences in the GenBank database were included in a search for amplimer and probe selection. Three different amplicon detection platforms and three different genomic regions were selected for this study: the conventional heminested RT-PCR using C-prM amplimers; the real-time SYBR green I protocol, followed by a postamplification Tm profile analysis using either C-prM or 3′NC amplimers; and the real-time, multiplex TaqMan fluorogenic assay using NS5 amplimers and four serotype-specific probes. The amplimer and probe sequences are listed in Tables Tables11 and and2.2. The C-prM amplimer was based on those developed by Lanciotti et al. (14). We modified the dengue virus consensus D1 primer (mD1) and DENV-2-specific primer (mTS2) and redesigned DENV-1- and -4-specific primers (rTS1 and rTS4, respectively) but maintained the original primer sequences for D2 and TS3. The RNA-dependent RNA polymerase domain in the NS5 protein is the most conserved coding domain in the Flavivirus genus. A consensus sequence approach was developed for genetic characterization of all registered flaviviruses using the sequence from this domain (4, 13). During this study, we observed that nucleotide sequences flanked by two flavivirus consensus primers (FU1 and CFD2) were varied among the 72 viruses sequenced (13). The real-time, multiplex TaqMan fluorogenic assay is based on this gene domain. We modified FU1 (mFU1), maintained the original CFD2 as described in a previous publication (13), and selected DENV serotype-specific fluorogenic probes (D1P, D2P, D3P, and D4P) flanked by the mFU1 and CFD2 sequences. The amplimers (DC10418 and CDC10564) located at the highly conserved 3′NC region were selected for the real-time SYBR green assay. Unlabeled amplimers were provided by the CDC core facility, and the fluorogenic probes were synthesized by Operon Biotechnologies, Inc. (Huntsville, AL).

Characteristics of dengue virus amplicons generated by C-prM and 3′NC SYBR green I protocols
Characteristics of the NS5 TaqMan assaya

A comprehensive test algorithm was applied for assay optimization. The amplification profile (RT-PCR cycle condition and amplimer/probe concentration) and the amplimer and probe specificities and sensitivities were determined using the viral RNA consisting of eight mosquito-borne flaviviruses in the FP and 109 DENV strains in the GDP. A side-by-side evaluation of the sensitivities and specificities of two C-prM primer sets, one developed by Lanciotti et al. (14) and the other developed from this study, was conducted using RNA extracted from FP specimens and utilizing the conventional heminested and SYBR green protocols. Potential false negatives resulting from amplimer and/or probe mismatch in all three protocols were concurrently evaluated using viral RNA extracted from viruses in the GDP (Table S1 in the supplemental material). The 5′ 6-carboxyfluorescein (FAM) reporter dye and 3′ black hole quencher-1 (BHQ-1) thermoquencher-labeled D1P, D2P, D3P, and D4P probes were used for assay optimization. The original D2P probe produced a false negative, failing to prime the DENV-2 specimen (strain BC124/97, isolated from monkey serum in 1997) in the single-dye assay. This D2P probe was modified to accommodate the observed mismatch, based on the nucleotide sequence of the probe binding region obtained from mFU1/CFD2-amplified DNA (data not shown), and retested against all 38 strains of DENV-2 in the GDP (Table S1 in the supplemental material) to ensure detectability. The assay profiles between the single-dye, four-tube, and multiplex single-tube formats were compared using DENV RNA templates from the FP.

Amplification and detection.

The iCycler IQ system (Bio-Rad Laboratories, Hercules, Calif.) or the Mx400 quantitative PCR system (Stratagene, La Jolla, Calif.) was used throughout this study for the SYBR green and TaqMan assays. An ABI-9600 thermocycler (Applied Biosystems, Foster City, CA) was used for the conventional heminested assay. For every assay, a proper viral RNA extracted from the FP and water was included as a positive and negative control, respectively.

The conventional heminested protocol for serotyping DENV involved two sequential amplifications (14). The initial RT-PCR amplification was assayed in a 50-μl reaction mixture containing 5 μl of RNA, 25 pmol of the mD1 and D2 amplimers, and components of a one-step RT-PCR kit (QIAGEN). The amplification involved the following steps: reverse transcription at 50°C for 30 min; one cycle of initial denaturation of the reverse transcriptase and activation of the HotStartTaq polymerase at 95°C for 15 min, 55°C for 15 s, and 72°C for 30 s; 34 cycles at 95°C for 15 s, 55°C for 15 s, and 72°C for 30 s; and a 10-min, 72°C extension. The heminested PCR was performed using a HotStartTaq master mix kit (QIAGEN) with 5 μl of RT-PCR product from a previous reaction and 25 pmol of each primer (mD1, rTS1, mTS2, TS3, and rTS4) in a 50-μl total reaction mixture. The amplification involved 1 cycle for 15 min at 95°C for polymerase activation and 25 cycles at 95°C for 15 s, 55°C for 15 s, and 72°C for 30 s. After amplification, a 5-μl portion of each product was analyzed by agarose gel electrophoresis utilizing 2.5% NuSieve 3:1 agarose gel, and the serotype was determined by the amplicon size as indicated in Table Table11.

The 3′NC amplimer set was used to detect all four serotypes of dengue viral RNA. Additionally, the C-prM amplimer set using mD1 in combination with rTS1, mTS2, TS3, and rTS4 was applied to identify the specific serotypes of dengue viral RNA in four separate reactions. Both protocols were performed using a QuantiTech SYBR green RT-PCR ready mix (QIAGEN) and utilized identical reaction and postamplification analyses. A one-tube SYBR green I RT-PCR was performed in a 50-μl volume containing 5 μl of RNA, 25 pmol of each primer (for 3′NC, DC10418/CDC10590, and for C-prM, mD1/rTS1 for DENV-1, mD1/mTS2 for DENV-2, mD1/TS3 for DENV-3, and mD1/rTS4 for DENV-4), and reaction mixture. The amplification involved the following steps: reverse transcription at 50°C for 30 min; one initial cycle at 95°C for 15 min, 50°C for 1 min, and 72°C for 1 min; 34 cycles at 95°C for 15 s, 50°C for 15 s, and 72°C for 30 s; and 78.5°C for 10 s. Since SYBR green intercalates nonspecifically with any double-stranded DNA generated during PCR, the Tm curve analysis was performed following amplification to confirm the identity of amplified product by its specific Tm profile (17). The Tm curve analysis included one cycle of denaturation at 94°C for 1 min, followed by 78.5°C for 10 s and a ramp to 94°C at a rate of 0.1°C/10 s with continuous fluorescence measurement.

In the multiplex TaqMan assay, 5 μl of RNA was mixed with 100 pmol each of mFU1 and CFD2, 25 pmol each of serotype-specific probe, and QuantiTech probe RT-PCR ready mix (QIAGEN) in a total of 50 μl reaction mixture. Four serotype-specific probes were labeled at their 5′ ends with a reporter and at their 3′ ends with a thermoquencher as follows: FAM-D1P-BHQ1, Texas Red-D2P-BHQ1, CY5-D3P-BHQ3, and 6-carboxy-2′,4,4′,5′,7,7′-hexachlorofluorescein (HEX)-D4P-BHQ1 for DENV-1, DENV-2, DENV-3, and DENV-4, respectively. The amplification and real-time detection consisted of the following cycle profile: reverse transcription at 50°C for 30 min; 1 initial cycle at 95°C for 15 min, 50°C for 30 s, and 72°C for 1 min; and 45 cycles at 95°C for 15 s and 48°C for 3 min with continuous fluorescence data collection.


One-step RT-PCR with real-time SYBR green I detection.

The SYBR green detection method was adapted for the C-prM and 3′NC amplimer sets. The C-prM region protocol was chosen because it is the most widely used protocol and has been modified and adapted in various laboratories to serotype dengue viruses (8, 14, 19, 20). Mismatching between viral RNA sequences and D1, D2, or TS sequences resulting in false-negative PCR results has previously been observed (8, 20; C. Chin, personal communication). We addressed this observation by modifying the D1 amplimer (mD1) and replacing the DENV-1, DENV-2, and DENV-4 type-specific amplimers with redesigned TS1 (rTS1) and TS4 (rTS4) and modified TS2 (mTS2) after an extensive search using the DENV sequences available in the GenBank database (Table (Table1).1). The conventional heminested and SYBR green protocols were used to compare the sensitivities of serotyping DENV-1, -2, -3, and -4 viruses using 10-fold-serially diluted RNA templates extracted from the FP. No differences were observed in the amplification profiles between the mD1/D2 and D1/D2 pairs (data not shown); therefore, the D1 amplimer was not included in additional studies. The mD1/rTS1, mD1/mTS2, and mD1/rTS4 pairs were at least 10-fold, 10-fold, and 100-fold more sensitive, respectively, at detecting serially diluted dengue viral RNAs than were the mD1/TS1, mD1/TS2, and mD1/TS4 pairs using the SYBR green protocol (Table (Table3).3). The mD1/rTS3 pair demonstrated lower PCR efficiency than the mD1/TS3 pair (76.1% versus 90.1%, respectively) (Table (Table4).4). The conventional heminested protocol, followed by the gel-based detection method, supported the same conclusion (data not shown). Thus, mD1, rTS1, mTS2, TS3, and rTS4 were chosen for the remainder of this study.

Performance comparison for the SYBR green I assay using dengue virus serotype-specific TS amplimers and redesigned or modified TS amplimersa
Correlation coefficients and PCR efficiency values for the SYBR green I assay using the amplimers in Table Table33

Various amplimer combinations (mD1 and D2, mD1 and TS, and mD1, TS, and D2) were tested using the SYBR green protocol (Table (Table5).5). In general, three amplimer combinations for the four DENV serotypes had very similar correlation coefficients, ranging from 0.993 to 1.000. Amplimer pairs mD1/rTS1, mD1/mTS2, mD1/TS3, and mD1/rTS4 had the highest PCR efficiencies within each respective group (Table (Table6).6). Consequently, amplimer pairs mD1/rTS1, mD1/mTS2, mD1/TS3, and mD1/rTS4 were selected and formulated for serotyping DENV in the one-step, four-tube SYBR green assay.

Performance comparison for the C-prM-SYBR green assay using various combinations of dengue virus consensusouter amplimers and dengue virus serotype-specific, redesigned, or modified TS amplimersa
Correlation coefficients and PCR efficiency values for the C-prM-SYBR green assay using the amplimers in Table Table55

The 3′NC amplimer set was developed for detection of all DENV serotypes (Table (Table1).1). This genomic region has been used in several studies (9, 10, 26-28). The 3′NC and the C-prM amplimers assayed via the SYBR green method had similar amplification profiles as well as sensitivities for detecting DENV RNAs (Table (Table1).1). Both assays were optimized using identical RT-PCR cycles and postamplification analyses; thus, both assays could be used concurrently for DENV RNA detection.

One-step RT-PCR with the real-time, multiplex fluorogenic TaqMan assay.

The RNA-dependent RNA polymerase domain of the NS5 protein was selected as the genome region to be used to design the multiplex fluorogenic TaqMan assay. This domain is highly conserved among all known flaviviruses (13). Four DENV serotype-specific probes (D1P, D2P, D3P, and D4P) are flanked by two conserved outer amplimers (mFU1 and CFD2) (Table (Table2),2), giving the assay greater potential to detect and serotype DENV in the multiplex, one-tube format. The RT-PCR and detection profile was developed and optimized using the same reporter (5′ FAM) and thermoquencher (3′BHQ-1) for all four serotypes. This optimized reaction profile featured a unique elongation step at 48°C for 3 min. One concern about the test's use of a lower elongation temperature (48°C compared to established 60°C protocols) is lower PCR efficiency. However, the standard curve obtained by plotting 10-fold-serially diluted DENV RNA against the cycle threshold (CT) value of each dilution indicated that the PCR efficiencies were 94.7, 100, 101.1, and 94.9%, with correlation coefficients of 0.999, 0.997, 0.998, and 1.000, for DENV-1 to -4, respectively. Standard curves obtained by the inclusion of all four serotype probes in the multiplex format demonstrated very similar PCR efficiencies and correlation coefficients for each serotype (data not shown). The detection sensitivities using a single probe, estimated from a standard curve at a CT of 35, were 1.5, 2.2, 1.2, and 0.6 PFU/ml for DENV-1 to -4, respectively (Table (Table2).2). The estimated sensitivities using the multiplex protocol were 2.4, 11.7, 1.5, and 2.3 PFU/ml for DENV-1 to -4, respectively.

Assay performance against the global dengue virus panel.

The GDP consists of 32 unique strains of DENV-1, 38 of DENV-2, 14 of DENV-3, and 25 of DENV-4, selected from the DVBID collection (Table S1 in the supplemental material). Viral RNAs extracted from the GDP viruses were assayed simultaneously for DENV RNA determination by the 3′NC SYBR green, C-prM SYBR green, and multiplex TaqMan NS5 protocols, as described in Materials and Methods. The 3′NC SYBR green assay determines the presence of dengue viral RNA and cannot be used for serotype identification. The C-prM SYBR green protocol has no multiplex capability. Thus, four separate reaction tubes containing mD1 and rTS1, mTS2, TS3, or rTS4 were used for serotype determination. The TaqMan NS5 protocol combines two flaviviral consensus outer amplimers (mFU1 and CFD2) and four different fluorophore-labeled, serotype-specific probes in a single reaction tube. The serotype is determined by the positive CT value associated with the serotype-specific fluorophore. The 3′NC SYBR green assay correctly identified all 109 isolates as DENV (Table S1, column 6, in the supplemental material). All 14 DENV-3 and 25 DENV-4 strains, identified by serotype-specific MAbs following virus isolation in the original data collection, were confirmed by the C-prM SYBR green assay as well as the multiplex NS5 TaqMan assay (Table S1, columns 7 and 8, in the supplemental material). Two DENV-1 isolates (1715 and BC202/97) and one DENV-2 isolate (S-19966), identified by MAbs in the original data collection, appeared to be DENV-2, DENV-3, and DENV-1, respectively, by both C-prM and NS5 assays (Table S1 in the supplemental material). To resolve this conflict, four genomic regions (nucleotides 33 to 763, 1573 to 2167, 2470 to 3700, and 8989 to 9975) of the three viruses were amplified and sequenced with dengue virus consensus primers (G.-J. Chang, unpublished results). The nucleotide sequences of the genomic regions of these four viruses agreed with the C-prM SYBR green and the multiplex NS5 TaqMan assays, and a BLASTN search (GenBank) confirmed the identities of 1715 as DENV-2, BC202/97 as DENV-3, and S-19966 as DENV-1 (data not shown). Given that the region from nucleotides 1573 to 2167, which encompasses domain III of the flaviviral envelope glycoprotein, is the virus type-specific MAb binding site (22), the inconsistent serotyping results between MAb typing in the original data collection and the two RT-PCR assays appear to have resulted from either a technical mistake or a data analysis error. Repeated virus isolation and typing using serotype-specific MAbs of these three strains further confirmed the results of two RT-PCR assays (data not shown).

Assessment of the multiplex capability of the NS5 TaqMan assay.

The MDSP (Table S2 in the supplemental material) consisted of 34 specimens infected with dengue viruses of multiple serotypes in the DVBID collection. All of the viruses in this panel were isolated from their original sources in C6/36 mosquito cell culture and serotyped with serotype-specific monoclonal antibodies (7). This panel was employed to assess the ability of the one-tube TaqMan assay to detect coinfection of multiple serotypes. The performance of this assay was compared to that of the single-dye, four-tube TaqMan assay. Some conflicting specimens in this panel were further tested by the four-tube SYBR green C-prM assay. All 34 specimens tested positive by the 3′NC SYBR green protocol, indicating the presence of dengue viral RNA (Table S2, column 6, in the supplemental material), and none of the three protocols produced false-positive results. Only eight and three specimens in the MDSP were correctly identified by the single-dye and the multiplex NS5 assay, respectively. All three protocols identified BC15/97 as DENV-2 and -4 coinfection and BC113/96 as DENV-3 and -4 coinfection. This result is in good agreement with virus isolation data. Results for 25 specimens (out of a total of 34) tested by the four-tube SYBR green C-prM assay agreed with the virus isolation record, the exception being specimen BC31/96. This specimen was identified as being coinfected with DENV-1, -3, and -4 by virus isolation and by the single-dye NS5 assay; however, only the multiplex NS5 assay confirmed the specimen to be coinfected with DENV-3 and -4. Compared to virus isolation, which produced 73 possibly test-positive scores, the single-dye NS5 assay had 39 (53.4%) and the multiplex NS5 assay had 40 (54.8%) test-positive scores. Statistical analysis based on the score numbers for these two assays indicated that no significant difference existed between the assays (P = 0.677). However, the four-tube SYBR green C-prM assay was much more sensitive than the single or multiplex TaqMan assay for detecting the specimens infected with dengue viruses of multiple serotypes. It is possible that each assay's sensitivity is influenced by the amplimer/probe competition in the specimens infected with viruses of multiple serotypes. Since specimens in the MDSP were propagated by one additional passage in C6/36 cell culture and used for RNA extraction, heterologous virus competition and/or interference among multiple DENV serotypes in the test specimens during one additional passage in C6/36 may have reduced the less-fit or lower-titer serotype virus below the detection threshold of a specific assay. Overall, the four-tube SYBR green C-prM assay was the most sensitive protocol for the detection of specimens coinfected with viruses of multiple serotypes.

Suitability of the multiplex NS5 TaqMan protocol for diagnostics and quantification of virus in DENV patient serum specimens.

The acute-phase serum specimens from dengue virus-infected patients with predetermined serotype and viremic titers were randomly coded and assayed by the multiplex NS5 TaqMan protocol. This collection consisted of 7 DENV-1 (titers ranging from 171 to 1,383,000 PFU/ml), 23 DENV-2 (titers ranging from 100 to 987,900 PFU/ml), 9 DENV-3 (titers ranging from 140 to 8,101 PFU/ml), and 10 DENV-4 (titers ranging from 1,886 to 737,700 PFU/ml) strains. We did not detect false-positive assay results due to contamination or mispriming. The CT value of each specimen was plotted against the log10 virus titer (PFU/ml) of the corresponding specimen (see the figure in the supplemental material). All four serotypes showed an excellent correlation between CT value and infectious-virus titer over a broad dynamic range. The correlation coefficients were −0.99, −0.98, −1.00, and −1.00, with a P of <0.0001, for DENV-1 to -4, respectively. The correlation coefficient of the combined data for all four serotypes was −0.975, with a P of <0.0001, and the 95% confidence interval was −0.986 to −0.956. Similarly, the regression analysis combined with a scatter diagram indicated that only two DENV-2 specimens were outside of the 95% prediction curve (see the figure in the supplemental material).

Evaluation of assay performance for detecting and serotyping dengue viral RNA in the acute-phase serum specimens.

To evaluate the performance of the assay in detecting and serotyping viral RNA in the acute-phase serum specimens, we randomly coded two sets of human serum samples, including five dengue virus RNA-negative samples from patients with fever of unknown origin. The first set of 33 specimens was tested at the DVBID, and the second set of 44 specimens was tested at the Center for Disease Control—Taiwan laboratory. All specimens were collected from dengue patients with infections confirmed by RT-PCR at the Center for Disease Control—Taiwan laboratory (23). Five dengue virus RNA-negative specimens as well as the four reagent controls were negative by three protocols (heminested gel-based, SYBR green, and multiplex TaqMan protocols) using three different amplimer and probe sets (3′NC, C-prM, and NS5). The heminested gel-based protocol generated amplicon fragments of the expected sizes of 208, 119, 288, and 260 bp from positive-control viral RNA for DENV-1 to -4, respectively. The control RNAs were SYBR green positive when the 3′NC and C-prM (mD1/TS) protocols were applied, and each produced a Tm in the expected range (Table (Table1).1). Likewise, the multiplex TaqMan assay produced positive CT values of 27.4 for DENV-1 viral RNA, 21.5 for DENV-2, 21.8 for DENV-3, and 20.2 for DENV-4.

Comparisons were made between the original RT-PCR results reported by the Center for Disease Control—Taiwan laboratory and the three testing protocols described in this study. We did not observe any false-positive results in any of the three testing protocols; thus, all three assay protocols had a 100% test specificity (Table (Table7).7). With a combination of two specimen sets, the 3′NC, C-prM, and NS5 protocols had assay sensitivities of 91%, 100%, and 91%, respectively. In serotyping DENV, the SYBR green C-prM was more sensitive than the multiplex NS5 protocol. However, the multiplex NS5 assay offers the advantage of a one-tube reaction with real-time capability.

Performance comparison of detection and serotyping of dengue viral RNA from serum specimens from patients with acute infectionsa


We have verified three RT-PCR amplimer and probe sets (derived from the C-prM, NS5, and 3′NC regions of the DENV genome) and three detection methodologies (the conventional gel-based, the heminested RT-PCR, the one-tube SYBR green RT-PCR, and the multiplex TaqMan protocols) that can be used independently or concurrently for serotype identification of DENV infections in human serum specimens.

A major concern in the design of a diagnostic RT-PCR for RNA viruses is the genetic variation in the nucleotide sequences of the viruses being studied. One common solution is to synthesize degenerate primers to encompass all possible permutations. However, this approach works properly only when the permutations total 16 or fewer in a single primer (G.-J. Chang, unpublished observation). We searched the GenBank database and retrieved available DENV sequences. The retrieved gene sequences with known phylogenetic clusters were used to design primers that would be compatible with as many sequences as possible. The combination of public-domain sequences with unpublished DENV sequences maintained at the DVBID (G.-J. Chang, unpublished results) contains 61, 83, 17, and 29 sequences available for DENV-1 to -4, respectively, encompassing the C-prM region. The mD1 and D1 (14) amplimers differ by 1 nucleotide at position 19 (Table (Table8).8). This C19A modification increased the mismatch between mD1 and any DENV-1 strain from 0 to 1 nucleotide but decreased the mismatch in any DENV-2 and DENV-4 strain from 3 to 2 nucleotides. Three additional thymines were added at the 3′ terminus of TS2 (Table (Table8).8). This modification (mTS2) is essential for compatibility with the 33 DENV-2 strains with a G21A mutation that may have rendered the TS2 primer less effective. We then verified the capability of these newly designed amplimers to detect the 109 DENV strains in the GDP (Table S1 in the supplemental material). Similar design principles and optimization procedures were employed for developing the multiplex NS5 TaqMan and the 3′NC SYBR green assays.

Mismatch analysis of the C-prM amplimers against available dengue virus sequences in this genomic region

We adapted the conventional heminested, gel-based assay in this study, aware of the possibility of obtaining false-positive results due to amplicon contamination. In spite of this problem, a similar protocol (14) has been used by numerous dengue virus laboratories worldwide to identify virus strains by PCR. The value of our familiarity with this protocol, the availability of the instrumentation, and the reduction of operational costs were the principal reasons for redesigning or modifying the C-prM amplimer set in this study. We noted that the combinations of mD1 and rTS1, mTS2, TS3, and TS4 separately in the single-step, four-tube format were more sensitive than other amplimer combinations used with the SYBR green detection method (Table (Table5).5). Although we did not test various amplimer combinations in the gel-based assay, we can assume that the mD1/TS combinations would perform better than the heminested protocol of the gel-based detection method.

Real-time RT-PCR offers significant improvements in viral-load quantification because of its enormous dynamic range, allowing the accommodation of at least 6 log10 PFU/ml of virus. The use of DNA binding fluorogenic molecules, such as SYBR green I (18), is less expensive than the 5′ nuclease TaqMan assay and has the potential for serotyping DENV based on postamplification melting temperature analysis using a set of DENV consensus amplimers. Shu et al. adapted this approach and developed the group- and type-specific SYBR green-based assays for serotyping DENV (23). A similar concept was applied in this study. The DENV consensus 3′NC amplimers and the DENV consensus C-prM mD1 in combination with rTS1, mTS2, TS3, or rTS4 amplimers were used to detect and serotype dengue viral RNA via the one-step SYBR green assay.

Others have suggested that the Tm profile of the amplicon generated by the 3′NC amplimers could be used for serotyping purposes (3). The melting temperature profile is dependent upon the concentration, the length, and the nucleotide composition of the amplicon (21). The concentration of the amplicon, reflected by the CT value, is dependent upon the initial copy number of the template and the efficiency of the specific PCR protocol. We have determined that the Tm profile alone was not a reliable criterion for serotyping (Table (Table1;1; Fig. Fig.1).1). Composite melting curves of all DENV strains represented in the GDP and MDSP generated average Tm values of 83.8, 82.7, 84.9, and 83.9°C for DENV-1 to -4, respectively (Table (Table1).1). DENV strain 2-HA in the MDSP was identified as containing DENV-2 and -3 serotypes by the C-prM and NS5 protocols (Table S2 in the supplemental material). This strain produced a CT value of 11.3 when subjected to the 3′ NC SYBR green protocol. The relatively similar CT values produced by this strain using mD1/mTD2 (12.2) and mD1/TS3 (10.8) indicated that this specimen might contain similar copy numbers of DENV-2 and -3. Indeed, we observed two distinctive Tm curves obtained from this specimen that were clearly indicated by the postamplification melting temperature analysis under this situation (Fig. (Fig.1,1, upper-left panel).

FIG. 1.
Composite melting temperature profiles of dengue viruses in the GDP and MDSP. The melting temperature curve of each virus was generated from the 3′NC SYBR green I protocol. The upper-left panel displays one isolate (2-H) from the MDSP, identified ...

The 5′ nuclease TaqMan assay is the most popular real-time RT-PCR protocol for serotyping DENV (1, 6, 9, 10, 12, 15, 29, 30). However, none of these published protocols have multiplex capability. The multiplex NS5 TaqMan assay described in this study is unique in that two flaviviral consensus outer amplimers (mFU1 and CFD2) (Table (Table2),2), capable of amplifying any known flaviviruses, with the exception of Tamara bat virus (13), and four DENV serotype-specific probes (D1P, D2P, D3P, and D4P) (Table (Table2)2) were combined in the one-tube, single-step RT-PCR format for detecting and serotyping DENV. Our protocol has also benefited from recent advances in probe chemistry and a new generation of instrumentation (16). In particular, the replacement of a fluorescent light quencher, such as TAMRA (6-carboxy-N,N,N′,N′-tetramethylrhodamine), with a nonfluorescent thermoquencher (such as BHQ) frees up a significant amount of bandwidth that can instead be used to accommodate additional fluorescent reporters (such as FAM, Texas Red, CY5, and HEX, used to label D1P, D2P, D3P, and D4P, respectively) (Table (Table22).

Serotyping DENV early in the transmission season is crucial for developing sound epidemiological control measures. Three sets of amplimers and probes described in this study were thoroughly verified and optimized to identify 109 unique strains of DENV in the GDP. Therefore, we predict that these assays will detect, with a high degree of accuracy, DENVs currently circulating worldwide. The three detection platforms can be adapted to meet individual laboratory needs, taking into account budget and features for high-throughput capability with automated platforms. Three independent amplimer and probe sets provide triple redundancy as an additional safeguard against any potential false-negative result. An unusual DENV-4 strain (8700544), isolated from a Taiwanese patient's serum in 1998, was not detected by the multiplex NS5 TaqMan assay during the course of an independent validation at the Center for Disease Control—Taiwan laboratory. Seven nucleotide substitutions (G2A, G5A, G8C, C14T, A16G, A26G, and C30T) were observed between this 8700544 strain and the D4P. This observation underscores the importance of assay validation. Therefore, a minimum of two amplimer and probe sets are recommended for testing surveillance specimens and suspect human serum specimens early in the transmission season. This procedure is essential to validate the assay performance and to prevent false-negative results due to frequent mutations in the RNA of the flaviviruses.

Supplementary Material

[Supplemental material]


L.-J.C.’s training grant for the development of this study was provided by the Division of Vector-Borne Infectious Diseases, Centers for Disease Control and Prevention.

G.-J.J.C. is indebted to Dennis W. Trent for his mentorship and encouragement in the past and for his critical review of the manuscript.


Supplemental material for this article may be found at http://jcm.asm.org/.


1. Callahan, J. D., S.-J. Wu, A. Dion-Schultz, B. E. Mangold, L. F. Peruski, D. M. Watts, K. R. Porter, G. R. Murphy, W. Suharyono, C.-C. King, C. G. Hayes, and J. J. Temenak. 2001. Development and evaluation of serotype- and group-specific fluorogenic reverse transcriptase PCR (TaqMan) assays for dengue virus. J. Clin. Microbiol. 39:4119-4124. [PMC free article] [PubMed]
2. Chang, G.-J. J., D. W. Trent, A. V. Vorndam, E. Vergne, R. M. Kinney, and C. J. Mitchell. 1994. An integrated target sequence and signal amplification assay, reverse transcriptase-PCR-enzyme-linked immunosorbent assay, to detect and characterize flaviviruses. J. Clin. Microbiol. 32:477-483. [PMC free article] [PubMed]
3. Chutinimitkul, S., S. Payungporn, A. Theamboonlers, and Y. Poovorawan. 2005. Dengue typing assay based on real-time PCR using SYBR Green I. J. Virol. Methods 129:8-15. [PubMed]
4. Crabtree, M. B., R. C. Sang, V. Stollar, L. M. Dunster, and B. R. Miller. 2003. Genetic and phenotypic characterization of the newly described insect flavivirus, Kamiti River virus. Arch. Virol. 148:1095-1118. [PubMed]
5. Deubel, V., M. Laille, J. P. Hugnot, E. Chungue, J. L. Guesdon, M. T. Drouet, S. Bassot, and D. Chevrier. 1990. Identification of dengue sequences by genomic amplification: rapid diagnosis of dengue virus serotypes in peripheral blood. J. Virol. Methods 30:41-54. [PubMed]
6. Drosten, C., S. Göttig, S. Schilling, M. Asper, M. Panning, H. Schmitz, and S. Günther. 2002. Rapid detection and quantification of RNA of Ebola and Marburg viruses, Lassa virus, Crimean-Congo hemorrhagic fever virus, Rift Valley fever virus, dengue virus, and yellow fever virus by real-time reverse transcription-PCR. J. Clin. Microbiol. 40:2323-2330. [PMC free article] [PubMed]
7. Gubler, D. J., G. Kuno, G. E. Sather, M. Velez, and A. Oliver. 1984. Mosquito cell cultures and specific monoclonal antibodies in surveillance for dengue viruses. Am. J. Trop. Med. Hyg. 33:158-165. [PubMed]
8. Harris, E., T. G. Roberts, L. Smith, J. Selle, L. D. Kramer, S. Valle, E. Sandoval, and A. Balmaseda. 1998. Typing of dengue viruses in clinical specimens and mosquitoes by single-tube multiplex reverse transcriptase PCR. J. Clin. Microbiol. 36:2634-2639. [PMC free article] [PubMed]
9. Houng, H. H., D. Hritz, and N. Kanesa-thasan. 2000. Quantitative detection of dengue 2 virus using fluorogenic RT-PCR based on 3′-noncoding sequence. J. Virol. Methods 86:1-11. [PubMed]
10. Houng, H. S., R. C.-M. Chen, D. W. Vaughn, and N. Kanesa-thasan. 2001. Development of a fluorogenic RT-PCR system for quantitative identification of dengue virus serotypes 1-4 using conserved and serotype-specific 3′ noncoding sequences. J. Virol. Methods 95:19-32. [PubMed]
11. Hunt, A. R., C. B. Cropp, and G. J. Chang. 2001. A recombinant particulate antigen of Japanese encephalitis virus produced in stably-transformed cells is an effective noninfectious antigen and subunit immunogen. J. Virol. Methods 97:133-149. [PubMed]
12. Ito, M., T. Takasaki, K.-I. Yamada, R. Nerome, S. Tajima, and I. Kurane. 2004. Development and evaluation of fluorogenic TaqMan reverse transcriptase PCR assays for detection of dengue virus types 1 to 4. J. Clin. Microbiol. 42:5935-5937. [PMC free article] [PubMed]
13. Kuno, G., G.-J. J. Chang, K. R. Tsuchiya, N. Karabatsos, and C. B. Cropp. 1998. Phylogeny of the genus Flavivirus. J. Virol. 72:73-83. [PMC free article] [PubMed]
14. Lanciotti, R. S., C. H. Calisher, D. J. Gubler, G.-J. Chang, and A. V. Vorndam. 1992. Rapid detection and typing of dengue viruses from clinical samples by using reverse transcriptase-polymerase chain reaction. J. Clin. Microbiol. 30:545-551. [PMC free article] [PubMed]
15. Laue, T., P. Emmerich, and H. Schmitz. 1999. Detection of dengue virus RNA in patients after primary or secondary dengue infection by using the TaqMan automated amplification system. J. Clin. Microbiol. 37:2543-2547. [PMC free article] [PubMed]
16. Mackay, I. M., K. E. Arden, and A. Nitsche. 2002. Real-time PCR in virology. Nucleic Acids Res. 30:1292-1305. [PMC free article] [PubMed]
17. Morrison, T. B., Y. Ma, J. H. Weis, and J. J. Weis. 1999. Rapid and sensitive quantification of Borrelia burgdorferi-infected mouse tissues by continuous fluorescent monitoring of PCR. J. Clin. Microbiol. 37:987-992. [PMC free article] [PubMed]
18. Morrison, T. B., J. J. Weis, and C. T. Wittwer. 1998. Quantification of low-copy transcripts by continuous SYBR Green I monitoring during amplification. BioTechniques 24:954-958, 960, 962. [PubMed]
19. Raengsakulrach, B., A. Nisalak, N. Maneekarn, P. T. Yenchitsomanus, C. Limsomwong, A. Jairungsri, V. Thirawuth, S. Green, S. Kalayanarooj, S. Suntayakorn, N. Sittisombut, P. Malasit, and D. Vaughn. 2002. Comparison of four reverse transcription-polymerase chain reaction procedures for the detection of dengue virus in clinical specimens. J. Virol. Methods 105:219-232. [PubMed]
20. Reynes, J.-M., S. Ong, C. Mey, C. Ngan, S. Hoyer, and A. A. Sall. 2003. Improved molecular detection of dengue virus serotype 1 variants. J. Clin. Microbiol. 41:3864-3867. [PMC free article] [PubMed]
21. Ririe, K. M., R. P. Rasmussen, and C. T. Wittwer. 1997. Product differentiation by analysis of DNA melting curves during the polymerase chain reaction. Anal. Biochem. 245:154-160. [PubMed]
22. Roehrig, J. T., R. A. Bolin, and R. G. Kelly. 1998. Monoclonal antibody mapping of the envelope glycoprotein of the dengue 2 virus, Jamaica. Virology 246:317-328. [PubMed]
23. 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. [PMC free article] [PubMed]
24. Shu, P.-Y., L.-K. Chen, S.-F. Chang, Y.-Y. Yueh, L. Chow, L.-J. Chien, C. Chin, T.-H. Lin, and J.-H. Huang. 2003. Comparison of capture immunoglobulin M (IgM) and IgG enzyme-linked immunosorbent assay (ELISA) and nonstructural protein NS1 serotype-specific IgG ELISA for differentiation of primary and secondary dengue virus infections. Clin. Diagn. Lab. Immunol. 10:622-630. [PMC free article] [PubMed]
25. Shu, P.-Y., and J.-H. Huang. 2004. Current advances in dengue diagnosis. Clin. Diagn. Lab. Immunol. 11:642-650. [PMC free article] [PubMed]
26. Sudiro, T. M., H. Ishiko, S. Green, D. W. Vaughn, A. Nisalak, S. Kalayanarooj, A. L. Rothman, B. Raengsakulrach, J. Janus, I. Kurane, and F. A. Ennis. 1997. Rapid diagnosis of dengue viremia by reverse transcriptase-polymerase chain reaction using 3′-noncoding region universal primers. Am. J. Trop. Med. Hyg. 56:424-429. [PubMed]
27. Sudiro, T. M., H. Ishiko, A. L. Rothman, D. E. Kershaw, S. Green, D. W. Vaughn, A. Nisalak, S. Kalayanarooj, and F. A. Ennis. 1998. Microplate-reverse hybridization method to determine dengue virus serotype. J. Virol. Methods 73:229-235. [PubMed]
28. Tanaka, M. 1993. Rapid identification of flavivirus using the polymerase chain reaction. J. Virol. Methods 41:311-322. [PubMed]
29. Wang, W.-K., T.-L. Sung, Y.-C. Tsai, C.-L. Kao, S.-M. Chang, and C.-C. King. 2002. Detection of dengue virus replication in peripheral blood mononuclear cells from dengue virus type 2-infected patients by a reverse transcription-real-time PCR assay. J. Clin. Microbiol. 40:4472-4478. [PMC free article] [PubMed]
30. Warrilow, D., J. A. Northill, A. Pyke, and G. A. Smith. 2002. Single rapid TaqMan fluorogenic probe based PCR assay that detects all four dengue serotypes. J. Med. Virol. 66:524-528. [PubMed]

Articles from Journal of Clinical Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...