Fecal samples were collected and handled with good animal practices as required by the Chinese Regulations of Laboratory Animals (Ministry of Science and Technology of People's Republic of China, 20110108). The animal study proposal was approved by the Experimental Animal Ethics Committee of Shandong Normal University (approval No. 20160901). Animal owners provided oral consent as per the national ethical regulations.

The reference strain M. paratuberculosis (Table 1) was purchased from the BeNa Culture Collection (BNCC, China) and grown on media following the instructions provided by BNCC. Between September 2016 and September 2017, 320 individual fecal samples and 292 individual serum samples were collected from 10 different dairy farms located in ten distinct geographic regions of Shandong province, China. The herds were selected at random and herd size varied between 200 and 800 animals. There was no history of paratuberculosis in these farms and no results of previous serological tests. Approximately 10% of the animals (adult cows over 24 months old showing diarrhea and emaciation) were selected for sampling. All fecal samples were collected by local field clinical veterinarians via per-rectal drags and were kept at 4℃ for up to 48 h for DNA extraction. Serum samples were collected from cattle tail vein by using a 10 mL sterile syringe and 5 mL were kept in a coagulant tube. The genomic DNA of the reference strain and clinical samples was extracted by using a bacterial genome DNA extraction kit and a stool DNA kit (Tiangen Biotech, China). Specific extraction steps were carried out in accordance with the manufacturer's instructions. DNA extracted from field fecal samples was tested by RPA-LFD and qPCR assays, serum samples were tested by ELISA. During analysis, the results from two different assays were compared.

The extracted DNA of M. paratuberculosis was amplified using the 2× EasyTaq PCR Supermix kit (Beijing TransGen Biotech, China). Forward primer, 5′-ATCAGCGCGGCAC GGCTCTTG-3′, and reverse primer, 5′-CGGGTAGTTACCG CGGCGAAG-3′, were used to amplify 632 nucleotides of the IS900 gene of M. paratuberculosis (691–1322 of GenBank accession No. S74401.1; National Center for Biotechnology Information, USA). The PCR temperature profile was as follows: initial activation at 94℃ for 3 min, 35 cycles of 94℃ for 30 sec, 58℃ for 30 sec, and 72℃ for 30 sec, and a final extension step of 72℃ for 10 min. The amplified fragment was ligated into the plasmid pEASY-T3 cloning vector by using the pEASY-T3 cloning kit (Beijing TransGen Biotech) and designated as pEASY-T3-MAP. The pEASY-T3-MAP standard DNA was extracted by using the TIANpure mini plasmid kit (Tiangen Biotech) and measured by using a NanoPhotometer P-Class P360 (Implen, Germany). The DNA copy number was calculated as described previously [19]. The DNA standard was then stored at 20℃ until needed.

The primer and lateral flow probe for the RPA-lateral flow (RPA-nfo) reaction were designed on the basis of the content of the appendix to the manual for the TwistAmp reaction kit (TwistDx, UK). Molecular biology detection genes are generally based on the IS900 gene sequence for M. paratuberculosis; therefore, six combinations of candidate primers (6 forward and 2 reverse) and two probes were designed according to this gene sequence (Table 2). A specificity analysis was performed by using BLAST and the amplification locus of a 632 nucleotide region indicated 100% identity (query coverage 100%) with that of the M. paratuberculosis IS900 gene. There were no matches to other bacteria.

The detail sequence and the modification information are displayed in Table 2. Synthesis and modifications of the oligonucleotides were accomplished by Sangon Biological Engineering, China. Initially, the RPA-nfo reaction system was optimized. The 50 µL reaction volume was prepared according to the instructions for the TwistAmp nfo kit (TwistDX); the volumes of primer and probe (10 µM) were adjusted in accordance with the manufacturer's recommendations. Amounts of rehydration buffer (29.5 µL), DNA template (2 µL), and 280 nM magnesium acetate (2.5 µL) were kept constant. The beginning of the reaction was controlled by the magnesium acetate. All test samples were incubated for a prototypical 30 min at 37℃; the heating equipment was a thermostatic water tank (Shanghai Jing Hong Laboratory, China) unless otherwise noted. The RPA products were purified by using MiniBEST DNA fragment purification kit (ver. 4.0; TaKaRa, China). The size of the RPA oligonucleotides was detected by performing 2% agarose gel electrophoresis (Agarose Broad Range; Beijing TransGen Biotech) with ethidium bromide staining. Because specific amplification was essential for the primer and probe properties, the specificity of primer and probe was first resolved by performing agarose-gel electrophoresis. When the lateral flow probe is cut by the nfo nuclease at the dSpacer position, the probe is transformed into a primer and acts by priming the polymerase extension. Therefore, there were two bands visible in the electrophoresis results.

Visualization of RPA-nfo amplicons was carried out by using LFD assays (HybriDetect; MileniaBiotec, Germany), with the specific assay steps performed according to HybriDetect's instructions. Briefly, 2 µL of the RPA-nfo product and 98 µL of the HybriDetect assay buffer were blended adequately in a 200 µL centrifuge tube. Subsequently, the LFD was directly dipped into the product-buffer mixture. The visual result should be observed within 5 min.

In order to obtain optimal primer and probe conditions, various combinations were assessed and tested by examining the detected specificity and sensitivity. Next, the incubation temperature and duration of the RPA-nfo reaction were evaluated according to the supplier's (TwistDX) instructions and as reported previously [11]. The tested temperature range was 20℃ to 50℃, and time range tested was 1 to 35 min.

The specificity of the RPA-LFD assay was determined by using DNA from the bacterial strains listed in Table 1. The genomic DNA of M. paratuberculosis and nuclease-free water were used as the positive and negative controls, respectively, in every run. All bacteria species were provided by BNCC and grown on media by following the instructions given by the BNCC. The analytical sensitivities of the RPA-LFD and qPCR assays were tested in a template with 10-fold serial dilutions of plasmid standard DNA from 4 × 10⁶ to 4 genome copies per microliter.

The qPCR assay for M. paratuberculosis was performed as described previously [9]. Forward primers M-F (5′-CCACAACCACCTCCGTAACC-3′) and reverse primers M-R (5′-CGCTAATTGAGAGATGCGATTG-3′) were used to amplify a 100 base pair sequences in the 1223–1322 regions of the IS900 gene. The reaction was prepared as a 20 µL reaction volume containing 2× SYBR Green Premix, the forward and reverse primers (10 mM, 0.7 µL each), and 2 µL of DNA template. The following thermal cycling parameters were used: initial denaturation at 95℃ for 30 sec, followed by 40 cycles of 95℃ for 5 sec, and 60℃ for 20 sec. Melting curves parameters were: 95℃ for 5 sec, then 1 cycle of 60℃ for 1 min, and 95℃ for 5 sec followed by a final cooling step at 50℃ for 30 sec. The results were analyzed by using gene scanning software version 1.5 (Roche, Germany). After qPCR, melting curves were analyzed to confirm the specificity of the amplified product.

In order to compare the performance of the RPA-LFD and ELISA assays, 292 serum samples were tested for the presence of antibodies to M. paratuberculosis by using a commercial antibody test kit (IDEXX, USA). Specific operation steps followed the manufacturer's instructions.

The experimental data were stored in a Microsoft Excel 2007 spreadsheet (Microsoft, USA). Statistical analysis was performed by using SPSS software (ver. 16.0; SPSS, USA). The independent-samples t-test was used to evaluate the results. For all analyses, p < 0.05 was considered to indicate significance. The diagnostic performance of the RPA-LFD, qPCR, and ELISA assays were assessed by calculating the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), 95% confidence interval, and coincidence rate as described previously [15].

The specificities and sensitivities of the RPA-nfo primer and probe sets were assessed by agarose gel electrophoresis. The various primer pairs and LF (TwistAmp reaction kit; TwistDx, UK)-probes combinations were classified according to specificity and product yield. The performance of the primer and probe in electrophoresis is demonstrated by non-specific amplification (panel A in Fig. 1). In same-primer pairs and LF-probe sets, the second group strip was the brightest in the LFD, and it took the shortest time in the test zone position (panel B in Fig. 1). Due to resource constraints, one set (IS900-F2-IS900-R1 with IS900-LF probe 1) was chosen for subsequent evaluations (Table 2).

In order to determine the optimum reaction temperature for the M. paratuberculosis RPA-LFD assay, a temperature range of 20℃ to 50℃ was assessed through the reaction of 30 min. The results indicated that the reaction could be effectively completed over a wide interval of temperatures from 30℃ to 45℃. However, the reaction band was brightest at a reaction temperature between 37℃ and 42℃ (panel A in Fig. 2), indicating the optimal temperature range for amplification efficiency. Therefore, in subsequent RPA-LFD assays, the selected reaction temperature was 39℃. The best reaction time was evaluated over a duration scale of 1 to 35 min. The results showed that a distinct band in the test zone position was visible with 10 to 35 min reaction durations. However, the stripe was very weak at 10 min (panel B in Fig. 2). Based on the results, the time of incubation was set at 30 min for subsequent RPA-LFD assay testing.

The specificity of the RPA-LFD assay was tested by using DNA extracted from a number of other pathogens that present similarly in the clinic (Table 2). The RPA-LFD assay did not detect the genomic DNA of those 10 bacteria; only the M. paratuberculosis genomic DNA produced a positive result on the LFD strip (Fig. 3). A gradient dilution of M. paratuberculosis plasmid standard DNA (4 × 10⁶ to 4 genome copies per microliter) was used to evaluate the sensitivities of the RPA-LFD and qPCR assays. The results showed that they all were capable of detecting 8 copies per reaction standard DNA (Fig. 4); thus, they had the same threshold detection level.

In order to compare the performance of RPA-LFD and qPCR assays, 320 field fecal samples were examined in the same sample set. The RPA-LFD assay had a slightly lower sensitivity than that of the qPCR assay (Table 3). There were 114 positives in the 320 samples in the RPA-LFD assay, and the positive rate was 35.63%. The positive rate of the qPCR assay was 34.06% (109/320) in the same sample set. The results from both assays were not significantly different based on independent-samples t-test results (t [18] = 0.031, p = 0.976; p > 0.05). The RPA-LFD assay yielded 100% sensitivity, 97.63% specificity, and a 98.44% concordance rate with the qPCR assay (Table 4). The results indicate that RPA-LFD and qPCR assays have very almost the same detection performance. Finally, the serum samples of 292 cattle were screened by the RPA-LFD assay and ELISA, respectively. The positive rate of the ELISA assay was 36.99% (108/292); however, the RPA-LFD assay only detected 23.29% (68/292) positive rate in the same sample set (Table 5). The performances of the RPA-LFD assay and ELISA assay are summarized in Table 6, it shown the RPA-LFD assay yielded 88.24% sensitivity, 78.57% specificity, and a 95.65% NPV with the ELISA assay. Based on the results, the RPA-LFD assay should be used to test field samples in the future.