Design of Multiplexed Detection Assays for Identification of Avian Influenza A Virus Subtypes Pathogenic to Humans by SmartCycler Real-Time Reverse Transcription-PCR

Clinical samples, viruses, virus HA and M1 genes, and bacteria.Clinical specimens from nasopharyngeal swabs (NPS) were collected from pediatric patients, under 15 years old, experiencing acute respiratory infections (ARI) and who consulted the pediatric department of Shanghai Nanxiang Hospital between October 2006 and January 2008. The swabs were kept in 2 ml viral transport medium (2.95% tryptose phosphate broth, 0.5% gelatin, 500 mg/liter amphotericin B, 4 mg/liter gentamicin, and 0.5 × 10⁻⁶ U/liter penicillin G and streptomycin), transported at 4°C to the Institut Pasteur of Shanghai laboratory, divided into aliquots, and stored at −80°C.

Different IAV and AIV from the collection of the Department of Microbiology, the University of Hong Kong, isolated from embryonating chicken egg allantoic fluids are shown in Table 1. Viruses were not titrated, and RNA was extracted from undiluted fluids.

HK/948/06 (H1N1) and HK/2268/06 (H3N2) viruses from the collection of the Hong Kong Public Health Laboratory were cultivated in MDCK cells and titrated by determining the 50% tissue culture infective dose (TCID₅₀), and RNA was extracted from viruses in cell supernatants (Table 1).

A collection of viruses responsible for ARI was isolated and cultivated in MDCK cells at Institut Pasteur of Shanghai (Table 1).

DNA clones of the subtype H5 HA genes from different H5N1 clades were prepared at Institut Pasteur of Shanghai, either from viral RNA or by gene synthesis (Table 2).

The M1 gene of PR/8/34 (H1N1) virus was amplified by RT-PCR and cloned into the pET14b-M1 plasmid (45), whereas the H1 HA gene of the same virus was cloned into the pPSP-H1 plasmid (16).

Streptococcus hemolyticus (Streptococcus pyogenes) Staphylococcus aureus, Streptococcus pneumoniae, Haemophilus influenzae, Legionella pneumoniae, Chlamydia pneumoniae, and Mycoplasma pneumoniae were isolated from patients with ARI at Institut Pasteur of Cambodia.

RNA extraction and transcript preparation.Total RNA from NPS aliquots obtained from children with ARI in Shanghai Nanxiang Hospital or from virus-infected chicken egg allantoic fluids, virus-infected cell supernatants, or bacterial lysates was extracted using a QIAamp viral RNA minikit (Qiagen, Hilden, Germany) in accordance with the manufacturer's protocol. Purified RNA was frozen at −80°C in aliquots.

In vitro-transcribed RNAs from different HA genes from IAV subtypes and H5 clades was used for internal controls and for the study of sensitivity and specificity of the rRT-PCR (Table 2). The HA genes of IAV were used to construct plasmid DNAs by inserting the full-length genes into the pGEM-T Easy vector (Promega, Madison, WI) or the pCMV vector (Invitrogen, Carlsbad, CA). The plasmidic T7 RNA polymerase transcription initiation site and promoter sequence located upstream of the IAV cloned sequences were utilized to synthesize RNA transcripts in vitro, using a T7 RiboMAX large-scale RNA production system (Promega) in accordance with the manufacturer's instructions. The concentration of RNA transcripts was calculated using a UV spectrophotometer (Bio-Rad, Hercules, CA).

To perform sensitivity tests, the RNAs were serially 10-fold diluted, ranging from 10⁵ RNA copies/reaction to 1 RNA copy/reaction. To compare the detection limits obtained by rRT-PCR with titrated virus particles, RNAs extracted from the H1N1, H3N2, and H5N1 viruses in supernatant fluids (Table 1) were serially diluted 10-fold from 10⁶ to 0.01 TCID₅₀/reaction.

rRT-PCR.Highly conserved regions of each M1 and HA gene were identified using Clustal W sequence alignment algorithms. The primer and probe sequences were designed with highly conserved sequences of IAV genes by using Primer Express (ABI, Foster, CA) software and Beacon Designer (Premier Biosoft) software with selection criteria by default (Table 3). The sequences of the M1 or HA genes of virus subtypes from human, bird, and pig isolates used for alignment are from GenBank (http://www.ncbi.nlm.nih.gov/GenBank/index.html ) and are available by request from the authors. After synthesis (Takara, Dalian, China), primers and probes were divided into aliquots and conserved undiluted at −80°C. H1 and H9 virus probes were labeled at their 5′ ends with Texas Red reporter dye and at their 3′ ends with Black Hole 2 quencher (BHQ-2) dye. The H3 probe was labeled at its 5′ end with TET (6-carboxytetrachlorofluorescein) reporter dye and at its 3′ end with BHQ-1 quencher dye. H2 and H7 virus probes were labeled with 6-carboxyfluorescein reporter dye at their 5′ ends and at their 3′ ends with BHQ-1 quencher dye (Table 3).

We adapted the HA genotyping assay for application to the SmartCycler II system (Cepheid, Sunnyvale, CA), which has four channels and can detect at most four different fluorescent dyes. Reactions were based either on a Sybr technique or on a TaqMan probe technique. Sybr reactions were done with a Sybr green I dye one-step RT-PCR kit (Takara), and the mixtures contained 5 U Takara Ex Taq HS, 5 U Primer Script II RT enzyme mixture, 12 U RNasin (Promega), and 200 pmol of each primer in a 25-μl final volume. TaqMan reactions were performed using a QuantiTect Probe RT-PCR kit (Qiagen). The reaction mixture contained 2.5 U enzyme mixture, 12 U RNasin, and 200 pmol of each primer and probe in a 25-μl final volume. Five microliters of RNA was added to each reaction mixture. All reactions used the same program: reverse transcription at 50°C for 30 min, transcriptase denaturation at 95°C for 5 min and then DNA polymerization for 45 cycles of 94°C for 15 s, 53°C for 30 s, and 72°C for 15 s, ending with (Sybr) or without the melting step. Results were considered positive if the cycle threshold (C_T) values were less than 35 cycles and the ΔS values (representing changes in the magnitude of the fluorescent signal) were higher than 100.

The WHO H5 HA rRT-PCR TaqMan protocol developed by the laboratory at the National Institute of Infectious Diseases (NIID), Tokyo, Japan, the WHO Collaborating Centre for Reference and Research on Influenza, and the WHO H5 Reference Laboratory was used as a reference technique for determining the sensitivity and specificity of H5 HA gene detection assay (51).

rRT-PCR and M1 RT-PCR assays of viruses from NPS responsible for ARI.A total of 202 NPS specimens were prepared for IAV detection using the rRT-PCR described above and for IAV M1 gene detection using the RT-PCR protocol previously described (4). Briefly, 2.5 μl of RNA extracted from NPS specimens was added to 22.5 μl of a master mixture containing PCR buffer (OneStep RT-PCR kit; Qiagen), 0.2 mM of each deoxynucleoside triphosphate, 0.5 μM of each of the M1 primers (5′-CAG AGA CTT GAA GAT GTC TTT GCT AG-3′ and 5′-GCT CTG TCC ATG TTA TTT GGA ATC-3′) previously described (9), and enzymes. Samples were subjected to RT-PCR using the cycle program as follows: 30 min at 50°C, 15 min at 94°C, followed by 40 cycles of 30 s of denaturation at 94°C, 30 s of annealing at 55°C, and 1 min of extension at 72°C. A final extension was programmed for 10 min at 72°C. The amplification products were analyzed in a 0.5 μg/ml ethidium bromide-2% agarose gel. All RT-PCR-positive samples were analyzed by using rRT-PCR for the HA genes of subtypes H1, H3, and H5. Statistical analysis was performed by using SAS software (SAS Institute, Inc., Cary, NC).

rRT-PCR.We have developed a sequential rRT-PCR that contains four reactions (Fig. 1). The AIV M gene-positive samples were further processed for H5 HA gene identification, and negative samples were tested subsequently with either the TaqMan assay for H1 and H3 HA genes or the TaqMan assay for H2, H7, and H9 HA genes and the beacon probe rRT-PCR identification assay. All reactions were programmed with the same PCR program, so they could be run not only with SmartCycler but also with other real-time PCR instruments in case of a large sample quantity.

• Open in new tab
• Download powerpoint

FIG. 1.

Algorithm of the four-tube rRT-PCR assay for the detection of AIV.

Sensitivity.In order to optimize and compare the sensitivities of the four rRT-PCRs, RNA transcripts of each full-length HA gene were prepared and tested by using either monoplex or multiplex reactions (Table 4). The limit of detection was calculated using 10-fold dilutions of the transcripts. The limit of detection of the A/Henan/04 HA RNA transcript found by reaction 2 (Fig. 1) corresponded to 10 RNA copies, and these results were similar to those of the standard rRT-PCR technique recommended by WHO (see Materials and Methods) (data not shown). The limit of detection of reaction 3 for the H1 or H3 HA gene corresponded to 1 RNA copy, whereas the limit was 10 RNA copies of each gene in reaction 3. The limit of detection in reaction 4 was about 10 RNA copies for the H2 and H7 HA genes and 100 copies for the H9 HA gene.

To determine the sensitivity of the detection of live virus particles, serial dilutions of RNAs extracted from titrated H1N1, H3N2, and H5N1 virus stocks (TCID₅₀) were tested with reactions 1, 2, and 3. The fluorescent signals observed for the M1 genes of H1N1 and H3N2 viruses in reaction 1 corresponded to 1 TCID₅₀, and those for the H1 and H3 HA genes in reaction 3 corresponded to 0.1 TCID₅₀ (Table 5). Signals observed for H5N1 virus were 0.1 TCID₅₀ and 0.01 TCID₅₀ for reactions 1 and 2, respectively.

Specificity.The specificity of the H5 rRT-PCR for different clades of viruses, including low-pathogenicity, high-pathogenicity, and avian and human H5N1 viruses, was tested. Eight HA clones of H5N1 virus corresponding to clades 0, 1, 2.1.3, 2.3.1, 2.3.2, and 8 (21) (Table 2) were used in reaction 2 and showed positive results with a melting temperature of 77°C to 78°C for the amplicons (data not shown). All genes also showed positive results when tested using the technique provided by WHO (data not shown).

To assess the integrity of the primers and probes of the six HA subtypes, reactions 1 to 4 were tested for cross-reactivity among H1 to H11 subtypes of AIV (Table 1). The primer set corresponding to the M1 gene of IAV was positive for all subtypes, and only the HA primer-probe set for the homologous virus subtypes showed positive results. The H4, H6, H8, H10, and H11 virus subtypes remained undetected by reactions 2 to 4. None of the several unrelated viruses causing ARI (Table 1) showed any positive results by reactions 1 to 4 (data not shown).

The cross-reactivities of the primers used in the four rRT-PCRs (Table 3) with RNA molecules extracted from seven bacteria commonly found in the respiratory tract were assessed by RT-PCR (see Materials and Methods). None of the samples showed any visible amplified product after agarose gel electrophoresis (data not shown).

Comparative study of RT-PCR and rRT-PCR using NSP samples.As reaction 1 is essential in the detection of IAV, it is important to address its specificity and sensitivity with samples from natural specimens. The specificity of reaction 1 for the IAV M gene was addressed by testing a collection of 202 NPS specimens from children with ARI (Table 6). Calculation of the melting temperature of the amplified products (80°C to 81°C) in reaction 1 helped to discriminate false positives due to internal primer dimers (data not shown). The reaction identified 39 positive samples (P = 19.3%; n = 202; S = 0.028; 95% confidence limits 0.778 to 0.978 [α, <0.05]). To confirm the results obtained by rRT-PCR, samples were tested by RT-PCR (see Materials and Methods). Thirty-eight samples (18.8%) were found positive by RT-PCR, of which two (1%) had previously been found negative by rRT-PCR (Table 6). The specificity between reaction 1 and RT-PCR for detection of the IAV M gene was 99.5%. The sensitivity of reaction 1 relative to that of RT-PCR for the detection of IAV was 94.9%. Statistical analysis showed that the two methods have no significant differences in performance (chi-square = 0.90; continuity adjustment chi-square = 1.0; kappa coefficient = 0.92, with 95% confidence limits of 0.85 to 0.99[α, <0.05]).

Reactions 2 and 3 were performed subsequently with all 202 samples, of which 2 were found positive for H1, 37 were found positive for H3, and all samples were negative for subtype H5. However, two samples found negative by rRT-PCR but positive by RT-PCR were confirmed positive and corresponded to the H3 subtype in reaction 3. These two samples had C_T values of 28 and 29, respectively, values close to the limit of sensitivity for the detection of H3 virus sequences (Table 5), whereas all other positive samples had C_T values between 15 and 19 (data not shown). One sample positive by rRT-PCR but negative by RT-PCR for the M gene was found positive for subtype H3 (Table 6) but with high C_T values of 30 and 34 for the M1 gene and H3 subtype, respectively, indicating a low concentration of viral RNA. Results for two other samples found positive for the M gene in reaction 1 (C_T values of 20.6 and 27.7, respectively) but negative in reactions 2, 3, and 4 should be considered false positives, since the samples were also negative by RT-PCR. The reasons for the positive results in duplicated reaction 1 for the two samples were not evident. Concordance of reaction 3 with reaction 1 was 94.9%.