A one-step, sensitive, specific assay using GeXP for the simultaneous detection of 16 respiratory virus types/subtypes (FluA, FluB, sH1N1,PIV1, PIV2, PIV3, HRV, RSVA, RSVB, HMPV, Adv, CoV OC43, CoV 229E, CoV NL63, CoV HKU1 and HBoV) was described in this study. The GeXP genetic analysis system is a new multitarget, high throughput detection platform, and its application in the differential detection of pandemic Influenza A H1N1 virus and seasonal Influenza A virus was reported recently by our laboratory . To further reduce the assay time and avoid contamination, a one-step multitarget RT-PCR was performed in one tube rather than multiplex two-step RT-PCRs in separate tubes  would be preferred. To develop a one-step GeXP assay, several commercial one-step RT-PCR kits from different companies (Omega, Bio-Rad, Qiagen and Invitrogen) were tested in our preliminary experiments. The one-step RT-PCR kit from Qiagen revealed the best amplification to perform under our current protocol (data not shown).
The temperature switch PCR (TSP)  strategy was adopted to optimize the amplification parameters. The biphasic PCR parameters of the TSP allow a multiplex PCR to be performed under standardized PCR conditions, and therefore do not require optimization of each individual PCR assay. The optimal settings for three different denaturation temperatures and the amplification cycle conditions were determined in the current protocol. The chimeric primers consisted of a specific primer sequence fused to the universal primer sequence. The primers were designed to generate gene fragments with lengths between 140–340 bp. The amount of universal primers included in the RT-PCR was ten times that of the chimeric primers, so in the last 20 cycles of PCR, amplification was carried out predominantly by a single pair of universal primers, of which only the forward universal primer was labeled fluorescently. This should reduce the occurrence of preferential amplification in the reaction. Internal control primers were added to the reaction to ascertain whether the extraction and reverse-transcription steps of the assay were functioning correctly. Only 2 μl of the cy5-labeled PCR amplified fragments were separated by capillary electrophoresis based on size and detected by the GeXP system.
The specificity of the GeXP assay was examined using artificial specimens from the NATRVP-2 panel and clinical specimens confirmed previously to be positive. A specific peak of PCR product was obtained only for the expected viral target using the GeXP system. No cross-reactivity among the 16 respiratory virus types/subtypes was observed. The detection sensitivity for each virus was 20–200 copies per reaction when the assay was performed separately for each virus, and the sensitivity was 1000 copies per reaction when all of the 16 pre-mixed viral targets were present in the multiplex assay. The results indicate that the detection sensitivity of the GeXP assay for mixed virus samples was slightly lower than that for a single type of virus. The detection sensitivity for each virus in this study was similar to that of real-time PCR assays reported in recent years [18–20].
Analyses of 126 specimens using the GeXP assay and RVP Fast assay demonstrated that the GeXP assay had comparable sensitivity and specificity to the commercially available RVP Fast assay (Table 2). Discrepancies in detection results were found for HRV, PIV3, HMPV, Adv, RSV B and HBoV. All of the 9 specimens positive for PIV3 or HRV detected only by the GeXP assay were confirmed by independent PCR and sequencing to be true positives (data not shown), suggesting that the GeXP assay is more sensitive than the RVP Fast assay for the detection of HRV and PIV3. Because the HRV primers used in the RVP Fast assay were able to amplify both HRV and enterovirus, the three specimens positive for HRV detected only by the RVP Fast assay could actually be enteroviruses. All of the specimens positive for HRV detected by the GeXP assay were confirmed by sequencing as true HRV positives. Two of HMPV negative samples, 1 of Adv negative sample, 3 of RSVB negative samples and 4 of HBoV negative samples detected by the GeXP assay were positive by the RVP Fast assay. All of these negative samples had lower median florescence intensity (MFI) values (294–825) in the RVP Fast assay, suggesting that the GeXP assay has a slightly decrease sensitivity for the detection of HMPV, Adv, RSVB and HBoV. However, the difference was not significant (for HMPV, Adv, RSVB and HBoV, p = 0.5, 1, 0.25, 0.125, respectively, using McNemar’s test; Kappa = 0.880, 0.943, 0.949, 0.864, respectively). The detection of coronaviruses (CoV HKU1, CoV NL63, CoV 229E and CoV OC43) by both assays was completely consistent between assays. The overall detection rate of the GeXP assay for each virus was comparable to that of the RVP Fast assay, demonstrating the high sensitivity and specificity of the GeXP assay in the analysis of clinical samples. It should be noted that there are not sufficient positive samples for FluA, sH1N1, FluB, PIV1,PIV2, CoV NL63, CoV 229E, CoV HKU1 and RSVA to determine meaningful statistics in our study. Due to the limited specimens available at this time, only preliminary findings were reported in this study. The detection rate for infections involving two or more viruses was similar for the GeXP assay (68/126) and the RVP Fast assay (66/126). Most occurrences of multiplex infections involved two of three viruses, HRV, RSVB and PIV3 (data not shown).
Two distinct advantages of the GeXP assay are the short assay time and the low cost per test. Though the cost of GeXP equipment is approximately $100.000, the cost of the GeXP assay for simultaneous detection of 16 respiratory virus types/subtypes is approximately $8 per test (the RT-PCR kit and the consumables of detection), versus $120 per test using the RVP Fast kit or $8 per test for each virus using a commercial RT-PCR kit (DaAn, Gene, China). The whole reaction was completed in one tube in a one-step multiplex RT-PCR within 2.5 hours, followed by capillary electrophoresis separation (10 min/12 wells). In addition, two 96-well plates can be placed in parallel in a GeXP machine at the same time to further increase the throughput of the samples.