Rapid isothermal duplex real-time recombinase polymerase amplification (RPA) assay for the diagnosis of equine piroplasmosis

Equine piroplasmosis (EP) is a severe disease of horses caused by the tick-borne protozoa Theileria equi (T. equi) and Babesia caballi (B. caballi). Infectious carriers are not always symptomatic, meaning there is a risk to non-enzootic areas. Regulatory tests for EP include sero-epidemiological methods for equine babesiosis, but these lack specificity due to cross-reactivity with other Babesia species. In this study, we present a real-time quantitative recombinase polymerase amplification (qRPA) method for fast simultaneous detection of both T. equi and B. caballi. In this method, primers and probes targeting the 18S rRNA gene of both T. equi and B. caballi, the ema-1 gene of T. equi and the bc48 gene of B. caballi were designed and evaluated. The sensitivity of qRPA was evaluated using the pUC57 plasmid DNA containing the target gene. For the pUC57-bc48 gene DNA, the R2 value was 0.983 for the concentration range 0.2 ng (4.1 × 107 DNA copies) to 2.0 fg (4.1 × 101 DNA copies). For the pUC57-ema gene DNA, the R2 value was 0.993 for the concentration range 0.2 ng (5.26 × 107 DNA copies) to 2.0 fg (5.26 × 102 DNA copies). For the pUC57-Bc18S gene DNA the R2 value was 0.976 for the concentration range 2.0 ng (4.21 × 108 DNA copies) to 2.0 fg (4.21 × 102 DNA copies). For the pUC57-Te18S gene DNA, the R2 value was 0.952 (Fig. S3b) for the concentration range 2.0 ng (4.16 × 108 DNA copies) to 2.0 fg (4.16 × 102 DNA copies). Furthermore, a duplex qRPA analysis was developed and optimized and the results showed that primers and probes targeting for the bc48 gene of B. caballi and the 18S rRNA gene of T. equi is the best combination for a duplex qRPA analysis in one reaction. The developed duplex qRPA assay has good specificity, and had negative amplification for several similar parasite. For DNA extracted from real horse blood specimens, this qRPA method has comparable sensitivity to traditional qPCR, but a simpler and more rapid operating process to obtain positive amplification. The qRPA, including the duplex strategy described here, could allow fast identification of the EP-causing T. equi and B. caballi, showing great potential for on-site EP screening of horses.


Results and Discussion
Duplex real-time RPA strategy for the diagnosis of equine piroplasmosis. A molecular diagnostic method based on duplex real-time RPA for the diagnosis of both B. caballi and T. equi from infected horse blood was developed (Fig. 1b). To achieve rapid diagnosis, specimen DNA extraction and separation were performed using magnetic beads, as an alternative to traditional high-speed centrifugation. To simplify the steps, all reagents (including extracted DNA, primers, probes, freeze-dried enzymes and rehydration buffer) were one-time mixed and the amplification reaction was initiated by MgAc 2 reagent at 37-42 °C. During RPA amplification, the fluorescent signal was created using an oligonucleotide probe flanked by a dT-fluorophore and a corresponding dT-quencher group, and was observed with a portable fluorescence detection device. Therefore, this strategy allows rapid diagnosis of piroplasmosis on-site. Primer design. Four primers targeting the bc48 gene of B. caballi and the ema-1 gene of T. equi were tested for sensitivity and specificity. The B. caballi-bc48-F1/R2 primer could amplify pUC57-bc48 (Fig. 2a, line 1), but not T. equi genomic DNA (Fig. 2a, line 1-Te). The B. caballi-bc48-F2/R1 primer was able to amplify pUC57-bc48 (Fig. a, line 2) and T. equi genomic DNA to some extent (Fig. 2a, line 2-Te). The B. caballi-bc48-F1/R1 primer produced an amplification curve with only water as the template (Fig. 2a, line 3). The primer set B. caballi-bc48-F2/R2 has a low amplification efficiency for T. equi genomic DNA (Fig. 2a, line 4). Therefore, B. caballi-bc48-F1/R2 was chosen to test the sensitivity for pUC57-bc48 and B. caballi genomic DNA.
Duplex RPA assay. To develop the duplex RPA assay for B. caballi and T. equi, we first investigated the primers and probes targeting the bc48 gene of B. caballi and the ema-1 gene of T. equi using pUC57-bc48 and pUC57-ema plasmid DNA. A carboxy fluorescein (FAM)-signal was only detected from pUC57-bc48 plasmid DNA (Fig. 3a, lines b& d); and a ROX-signal was only detected from pUC57-ema plasmid DNA (Fig. 3a, lines a&c). This indicated good specificity of these primers and probes. Compared with the primers and probes targeting the ema-1 gene, those targeting the 18S rRNA gene of T. equi had greater amplification (Fig. 3b), possibly due to ema-1 gene sequence heterogeneity in a subset of the original isolates 27 . Therefore, we chose T. equi-18S-F2/R2 with a T. equi-18S-probe to amplify T. equi genomic DNA.
When the primers and probes targeting the 18S rRNA genes of B. caballi and T. equi were mixed in one reaction tube, the amplification signals showed cross interference (Fig. 3c-e), which meant that developing a duplex assay based on these primers and probes was inappropriate. Therefore, primers B. caballi-bc48-F1/R2 and T. equi-18S-F2/R2 were chosen to establish a duplex RPA strategy for B. caballi and T. equi. These primers and probes were added into one reaction tube to detect water (Fig. 3f, blue lines), B. caballi genomic DNA (Fig. 3f, black lines) and T. equi genomic DNA (Fig. 3f, red lines). Results showed specific amplification of B. caballi or T. equi genomic DNA and no cross interference. Therefore, a valid duplex RPA strategy was developed by employing a B. Sensitivity and specificity of real-time RPA assay. To evaluate the sensitivity of the primers and probes for the bc48 gene of B. caballi and the ema-1 gene of T. equi, we chose B. caballi-bc48-F2/R2 and T. equi-ema-F1/R1 to evaluate the sensitivity and linearity of the real-time RPA assay. The amplification curves are shown in Fig. 4a,b. For the pUC57-bc48 gene DNA, the R 2 value was 0.983 ( Fig. S2a) for the concentration range 0.2 ng (4.1 × 10 7 DNA copies) to 2.0 fg (4.1 × 10 1 DNA copies). For the pUC57-ema gene DNA, the R 2 value was 0.993 ( Fig. S2b) for the concentration range 0.2 ng (5.26 × 10 7 DNA copies) to 2.0 fg (5.26 × 10 2 DNA copies).
Real-time PCR is the widely-recognized gold standard of genomic DNA identification and quantification due to its good specificity and high sensitivity. In this study, we employed the reported PCR primers and probes to assay pUC57-Bc18S and pUC57-ema DNA 26,28 and the results showed that 2.0 fg pUC57-Bc18S DNA (Figs. 4e) and 2.0 fg of pUC57-ema DNA (Fig. 4f) could be detected. The onset of real-time PCR was linearly fitted against the logarithm of initial pUC57-Bc18S concentration from 2.0 fg (4.21 × 10 2 DNA copies) to 2.0 ng (4.21 × 10 8 DNA copies), with an R 2 value of 0.998 (Fig. S4a); and for pUC57-ema concentration from 2.0 fg (5.26 × 10 2 DNA copies) to 2.0 ng (5.26 × 10 8 DNA copies), with a R 2 value of 0.997 (Fig. S4b). This shows that the sensitivity of the developed real-time RPA is comparable to real-time PCR assay.
The specificity of the developed qRPA was evaluated using several other parasite (Sample No. 16-22 in Table 1), and the results (Fig. 5f) showed that none of these parasites generated FAM or ROX fluorescence signals in the duplex RPA reaction containing the primers and probes targeting for bc48 gene of B. caballi and the 18S gene of T. equi. But the B. caballi genomic DNA from sample HQ1-4 could generate remarkable fluorescence signals of FAM Dye and T. equi genomic DNA from sample DQ1 and DQ2 produced obvious fluorescence signals of ROX Dye (Fig. 5f), indicating that the developed primers and probes have good specificity.
Applicability of RPA assay for horse blood samples. To investigate the efficacy of the developed primers and method in a real sample, genomic DNA extracted from horse blood containing B. caballi or T. equi was www.nature.com/scientificreports www.nature.com/scientificreports/ used as a template for an RPA assay. Horse blood samples collected from the Hongshan Military Horse Farm of the Beijing Military Region during a period of high incidence of equine piroplasmosis in April, 2018 were detected. Before detection, we evaluated the amplification efficiency of primer sets (the 18S rRNA, bc48 and ema-1 genes). For B. caballi, the bc48 gene primers had better amplification efficiency than the 18S gene primers (Fig. 5a). For T.  equi, the 18S gene primers had better amplification efficiency than the ema-1 gene primers (Fig. 3b). Therefore, B. caballi-bc48-F1/R2 and T. equi-18S-F2/R2 were applied to assay the real horse blood sample.
The horse blood samples containing B. caballi or T. equi were assayed using real-time RPA (Fig. 5d,e), and the results compared to real-time PCR (Fig. 5b,c). Results of all the samples were consistent according to the onset time of the two methods. Furthermore, the chosen specimens containing B. caballi were assayed using the developed duplex real-time RPA method. The results showed that only FAM signals were generated from B. caballi samples (Fig. 5f), confirming the validity of the developed method for diagnosing EP. The new method has the advantages of being more rapid, whilst maintaining high sensitivity and specificity.
Repeatability of RPA assay. During RPA amplification, fluorescence intensity can easily be affected by the concentration of the probe and DNA template, even by the pipette due to the small volume of DNA template 40 . The repeatability of the RPA assay was therefore evaluated by measuring 0.02 ng (4.1 × 10 6 DNA copies) and 20 fg (4.1 × 10 3 copies) of pUC57-bc48 plasmid DNA four times (Fig. 6a). Since quantification was based on the onset time of amplification, we investigated this across the four experiments. The results showed that the deviations of onset time were 8.53% for 0.02 ng and 6.03% for 20 fg DNA (Fig. 6b), an acceptable level of repeatability.

conclusion
In view of the harm caused by EP and the inadequacy of existing rapid diagnosis methods, we developed a new real-time qRPA system for the rapid diagnosis of the typical pathogens of EP: B. caballi and T. equi. The presented qRPA assay, by amplifying the bc48 or 18S rRNA gene of B. caballi and the ema-1 gene or 18S rRNA gene of T. equi, showed strict specificity and comparable sensitivities. The developed qRPA strategy has a simpler operating process with a lower constant reaction temperature (39~42 °C) and was more suitable to assay blood specimens in the field as a rapid diagnosis method compared with traditional qPCR. Furthermore, a duplex analysis strategy based on the bc48 gene of B. caballi and the 18S rRNA gene of T. equi was established, and needed only 20 min to amplify the pathogens in horse blood specimens. The qRPA, including the duplex strategy described here, shows great potential for on-site EP screening. DNA extraction procedure. Total DNA was extracted from 200 μl of whole blood using the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. Briefly, 20 μl Qiagen Protease was added into the blood sample, then 200 μl Buffer AL was added, followed by a brief vortexing. This mixture was incubated at 56 °C for 10 min with four times up-and-down mixing. After a brief vortexing, 200 μl ethanol was added to the sample, and mixed thoroughly by vortexing. Then the DNA released from the sample was purified with the DNeasy mini spin column with AW1 and AW2 washes, respectively. At last, 100 μl of elution buffer was used to elute the DNA retained on the membrane of the spin column. The eluted DNA was quantified using a Qubit ® 2.0 Fluorometer and a Qubit ® dsDNA HS Assay kit.
Primer design. Two genes from B. caballi were chosen as the target of the RPA assay: bc48 (GenBank No.  Tables 2 and 3. Primers and probes for bc48 and ema-1 were listed in Table 2, and for 18S rRNA in Table 3. The forward and reverse primers had 30-35 bp. The probes were labelled with a FAM/ Black Hole Quencher 1 or ROX/Black Hole Quencher 1. All primers and probes were evaluated for biophysical properties and dimer formation by Oligoanalyzer 3.1 (IDT, Leuven, Belgium) and blasted (https://www.ncbi.nlm.nih.gov/tools/primer-blast) against the NCBI nucleotide database to make sure that there was no homology with sequences from another organism.
Construction of recombinant plasmid carrying targeting genes of B. caballi or and T. equi. Sequence of the bc48 or 18S rRNA genes from B. caballi, and the ema-1 or 18S rRNA genes from T. equi, were commercially synthesized and cloned into plasmid pUC57 (Genscript, Jiangsu, China), respectively. The prepared recombinant plasmids termed pUC57-bc48, pUC57-Bc18S, pUC57-ema and pUC57-Te18S, respectively. Then, plasmids were replicated in E. coli DH5α cells, extracted and purified using an endotoxin-free Plasmid Maxiprep Kit. After that, plasmid DNA was quantified using a Qubit ® 2.0 Fluorometer and a Qubit ® dsDNA HS Assay kit.
Development of RPA assay. The RPA assay was performed based on the manual instructions. In brief, each assay was performed in a test tube filled with TwistAmp TM exonuclease and 50 μl reaction mixture. In process, the enzyme pellet was rehydrated by 46.5 μL of a master-mix containing rehydration buffer (29.5 μL), www.nature.com/scientificreports www.nature.com/scientificreports/ nuclease-free water (12.2 μL), 10 μM forward and reverse primer (each 2.1 μL) and 10 μM probe (0.6 μL). Next, 2 μL of plasmid DNA or DNA specimen extracted from parasites was added to each reaction, with a final concentration of 420 nM for each primer and 120 nM for probe. Last, 2.5 μL of Mg(Ac) 2 (280 mM) was added to the lid of each tube. All reactions were simultaneously initiated by centrifuging the magnesium acetate into the reaction mixture and transferring the tubes to a 39 °C heat block for 30 min. Fluorometric data were collected every 10 s. Referring to the previous works, the onset time of amplification was used to calculate the linearity and limit of detection, which was more practical than fluorescence intensity and helpful to improve the repeatability of the real-time RPA assay 55,56 . Evaluation of specificity and sensitivity. Sensitivity of the RPA assay were tested using pure recombinant plasmid genomic DNA, which allowed accurate determination of the detection limit in terms of copy numbers. A ten-fold serial dilution of recombinant plasmid DNA, ranging from 2.0 ng to 0.02 fg, was used as standard to determine sensitivity level, establish amplification efficiency and resolve the limit of detection. The DNA copy number of 2.0 ng of the pUC57-bc48 plasmid is equal to 4.1 × 10 8 copies; 2.0 ng of pUC57-ema equates to 5.26 × 10 8 copies; 2.0 ng of pUC57-Bc18S equates to 4.21 × 10 8 copies; and 2.0 ng of pUC57-Te18S is 4.16 × 10 8 copies. For specificity, the interfering DNA samples from trypanosoma evansi (strain STIB 805), trypanosome brucei (brucei TREU927), toxoplasma gondii (RH), plasmodium falciparum (3D7), theileria hirci (Dschunkowsky), theileria sinensis (Blood extract), trichinella spiralis inorganic (Blood extract), were evaluated using the developed duplex RPA assay. The onset time of amplification was plotted against the concentration of total DNA.
As a comparison, real-time PCR assay was also performed. Primers and probes reported in a previous duplex qPCR assay of EP 26,28 (Table 4) were synthesized by Sangon Biotech (Shanghai, China). The PCR reactions were run on a Light Cycler 480 (Roche, USA) using a TaqMan Gene Expression Master Mix (Thermo Fisher Scientific, USA). The cycling protocol consisted of the initial activation cycle at 95 °C for 10 min, followed by 45 cycles of 30 s at 95 °C, 30 s at 57 °C and 20 s at 72 °C with fluorescence data acquisition after the elongation step. Negative controls consisted of nuclease-free water and DNA extracted from the blood of an EP-free horse.     Table 4. Nucleotide sequences of primers and probes used in the duplex qPCR assay. Abbreviations: MGB, minor groove binder; NFQ, non-fluorescent quencher