Development of real-time reverse transcription recombinase polymerase amplification (RPA) for rapid detection of peste des petits ruminants virus in clinical samples and its comparison with real-time PCR test

Peste des petits ruminants (PPR), caused by small ruminant morbillivirus (SRMV), formerly called peste des petits ruminants virus (PPRV), is one of the most important pathogens in small ruminants, and has tremendous negative economic impact on the sheep industry worldwide. Current detection of PPRV in clinical samples mainly relies on real-time RT-PCR. Particularly, samples collected from rural area require highly equipped laboratories for screening. A rapid, real-time reverse-transcription recombinase polymerase amplification assay (RT-RPA), employing primers and exo probe, was thus developed to perform at 42 °C for 20 min, and the detection limit at 95% probability was 14.98 copies per reaction and 0.326 TCID50/mL based on plasmid copy number and tissue culture infectivity titre. All the four lineages of PPRV could be detected with no cross-reaction to other pathogens including measles virus (MeV), goatpox virus (GTPV), canine distemper virus (CDV), foot-and-mouth disease virus (FMDV) and Mycoplasma capricolum subsp. capripneumoniae (Mccp). The performance of real-time RT-RPA assay was validated by testing 138 field samples and compared to real-time RT-PCR. The results indicated an excellent diagnostic agreement between real-time RT-RPA and a reference real-time RT-PCR method with the kappa value of 0.968. Compared to real-time RT-PCR, the sensitivity of real-time RT-RPA was 100%, while the specificity was 97.80%. The developed RT-RPA assay offers a promising platform for simple, rapid, and reliable detection of PPRV, especially in the resource-limited settings.


Results
Optimization of PPRV real-time RT-RPA. In order to optimize the real-time RT-RPA reaction, 4 assemblies of primers and probe were designed and the fluorescence signals were detected. The amplification efficiency of group F4/R was the highest among all candidates during testing RNA of 2 × 10 3 TCID 50 /mL PPRV (Fig. 1A). Based on the fluorescence intensity and its corresponding time, 420 nmol/µL forward and reverse primers, along with 120 nmol/µL probe were ultimately deployed in our assay with the operating temperature of 42 °C (Fig. 1B-D).

Sensitivity and specificity of PPRV real-time RT-RPA.
To determine the sensitivity of PPRV real-time RT-RPA, a dilution ranging from 2 × 10 4 to 0.2 TCID 50 /mL was tested for eight replicates. As shown in Fig. 2A, remarkable increase of fluorescence intensity was observed by the detection from 2 × 10 4 to 2 TCID 50 /mL at 42 °C within 20 min. A 10-fold serial dilution of the in vitro transcribed RNA molecules from 2 × 10 6 to 0.2 copies per reaction was simultaneously detected with the same method (Fig. 2B). Of note, our data on probit regression analysis indicated that the detection limit of PPRV real-time RT-RPA at 95% probability was 0.326 TCID 50 /mL (0.130~11.338 TCID 50 /mL, 95% CI) and 14.99 copies per reaction (4.62~902.77 copies per reaction, 95% CI), respectively (Fig. 2C,D). In order to identify the specificity of RT-RPA assay, PPRV, MeV, GTPV, CDV, FMDV and Mccp were involved in the specificity test. We found positive results by the detection of PPRV from lineage I, II, III and IV, while no cross reaction of the other microbes was shown (Fig. 2E). Performance of PPRV real-time RT-RPA assay on clinical samples and its comparison with real-time RT-PCR testing. In order to evaluate the practical application of PPRV real-time RT-RPA, the correlation analysis was deployed to assess the performance of real-time RT-RPA and real-time RT-PCR on a dilution from 20000 to 0.2 TCID 50 /mL. Higher correlation level between cycle threshold (Ct) value and virus titer (R 2 = 0.99) in real-time RT-PCR was found, compared to that between threshold time and virus titer (R 2 = 0.86) in RT-RPA (Fig. 3A (Tables 2 and 3).     immunochemistry are also used on conjunctival smears and tissue samples collected at necropsy. Particularly, nucleic acid amplification that mainly involves conventional RT-PCR and quantitative real-time RT-PCR significantly improves molecular methods for the detection of PPRV RNA in nasal and mouth discharges, tissues and anticoagulant-treated blood 15,16 . This study describes the development of real-time RPA, a new isothermal nucleic acid amplification targeting PPRV genome. Of note, the whole process of RT-RPA merely requires 20 min at 42 °C, which dramatically shortens the reaction time of real-time RT-PCR (~100 min) and simplifies the thermal cycles. Previous studies showed that RT-RPA assays on the detection of porcine epidemic diarrhea virus (PEDV) 17 , porcine reproductive and respiratory syndrome virus (PRRSV) were performed at 40 °C for 20 min 18 , while within the same method, the rapid diagnosis of Canine distemper virus (CDV) was developed at 40 °C for 3-12 min 19 , which were the less-time consuming and consistent with our test. In addition, the copies of target gene have been often implicated in the evaluation of sensitivity in RPA. For instance, the detection limit of real-time RPA on Capripox virus (CaPV) was 300 copies per reaction within 20 min at 38 °C 20 . The analytical sensitivity of RT-RPA was 31.8 copies in vitro transcribed CDV RNA and 23 copies per reaction in vitro transcribed virus RNA 19 . Particularly, for the detection of PPRV, around 100 copies per reaction with 95% reliability were shown in the sensitivity of the developed real-time RT-RPA 21,22 . However, most of the RPA assays failed to determine the virus quantification by TCID 50 , EID 50 or LD 50 . Notably, in addition to the determine the lowest analyte concentration of the assay (genome copies), our study with real-time RPA method for the first time assessed the minimum detection of virus titer by using TCID 50 and the result showed that the detection limit at 95% probability was 0.326 TCID 50 /mL. As previous evidence on the diagnosis of PPRV revealed that the detection limits of hydrolysis probe (TaqMan) real-time RT-PCR and conventional RT-PCR were approximately 0.1 and 1 TCID 50 , respectively, according to a serial dilution of the live-attenuated PPR vaccine virus 23 , the detection limits in our study were similar to real-time RT-PCR or RT-PCR.
The clinical performance of real-time RT-RPA was determined by the detection of a cohort of 138 clinical samples while real-time RT-PCR testing was conducted in parallel. Our data showed an excellent diagnostic agreement between real-time RT-RPA and a reference real-time RT-PCR method with the kappa value of 0.968, which was similar to the comparison result of RT-RPA DENV 24 . Our data is also consistent with previous data that real-time RT-RPA assay of PPRV was comparable to that of the real-time RT-PCR assay (99.4%, 161/162) 21 . Linear regression analysis on positive samples between threshold time of RT-RPA and Ct value of real-time RT-PCR was further presented and it showed close relationship between the values of two indicators (R 2 = 0.79), suggesting that higher titer of virus corresponds to shorter threshold time and lower Ct value. However, we note that the analysis of field samples gave rise to 2 results positive with RT-RPA and negative with real-time RT-PCR in our study was also found between RT-RPA and RT-qPCR. The sensitivity of laboratory tests may vary due to the methods themselves. For instance, qPCR is generally more sensitive than PCR assay 25 . The difference of sensitivity is probably implicated between real-time RT-RPA and real-time RT-PCR. We cannot exclude these two results were false positive, for neither sequencing nor virus isolation was available for confirmatory validation from the samples with relatively small amount of virus. Preferably, diagnostic agreement between RT-RPA and RT-qPCR was analyzed by semi-log regression analysis and probit regression analysis and we found the sensitivity and the specificity of RT-RPA assay for identification of PPRV were 100% (92.45~100%, 95% CI) and 97.80% (92.29~99.73%, 95% CI), respectively, compared to the reference real-time RT-PCR method. Besides, cutoff based on the detection of a large number of samples is worthy of further consideration regarding the application of RPA in clinical use. Taken together, the real time RT-RPA assay presented a promising tool for PPRV detection, which is simple, rapid and reliable in resource-limited diagnostic laboratories and on-site facilities. Real-time RT-PCR assay. The real-time RT-PCR assay was performed on Light Cycler 480 (Roche, Mannheim, Germany) as previously described 25 . The reactions were prepared as a 25 µL reaction volume containing 12.5 µL 2× reaction mix, 0.5 µL enzyme mix and 3 µL extracted RNA. The following thermal program was: reverse transcription at 50 °C for 15 min, followed by 95 °C for 3 min and 40 cycles of amplification (15 s at 94 °C and 1 min at 60 °C).

Methods
Analytical sensitivity and specificity of real-time RT-RPA. To

Statistical analysis.
For the determination of the PPRV real-time RT-RPA assay analytical sensitivity, a semi-log regression analysis (PRISM, Graphpad Software Inc., San Diego, California) and a probit regression analysis (MedCalc Software bvba, Ostend, Belgium) were performed to calculate the detection limit of the real-time RT-RPA assay at a 95% probability level.  Table 1. RPA primers and probe. a PPRV-RPA F and PPRV-RPA R were defined as forward primer and reverse primer, respectively; PPRV-RPA P was exo probe; b FAM-dT, thymidine nucleotide carrying fluorescein; THF, tetrahydrofuran spacer; BHQ1-dT, thymidine nucleotide carrying black hole quencher 1; C3 spacer to block elongation.