The present study was conducted in April during a period of high incidence of equine piroplasmosis in 2019 at Ulan Butong Grassland of Chifeng City, Inner Mongolia Autonomous Region, China. Thirty-six tick-bitten malnourished Mongolia horses (4 stallions, 14 geldings and 18 mares) aged 6.7 ± 1.5 years and with mean live weight 286.5 ± 21.3 kg were used for the study. The general condition of horses were fully monitored by routine clinical examinations. The body condition score of horses ranged between 3 and 4, scored on a 9-point scale (1–9, where 1 represents lean minus and 9 represents fat plus), according to the guidelines described by Henneke et al. (27).

Blood samples were obtained from the jugular vein of the horses into 10 mL vacuum tubes containing ethylenediaminetetraacetic acid as an anticoagulant (BD Vacutainer, USA). All blood samples were stored in 4 °C and centrifuged for 10 min at 3,000 rpm at 4 °C to separate the red blood cells, and then kept frozen at −80 °C for further analyses.

DNA from 200 uL of red blood cells was extracted. The extraction was carried out using TaKaRa MiniBEST Universal Genomic DNA Extraction Kit Version 5.0 (Takara Biomedical Technology Co., Ltd, Dalian, China) according to the instructions. Standard negative blood samples were obtained from EP-free horses which were detected and confirmed negative using the method described in a previous study (28). Theileria sinensis DNA samples were donated by Heilongjiang Bayi Agricultural University.

18S rRNA coding sequences (Z15104.1 and Z15105.1) of B. caballi and T. equi (according to the sequence information published by GenBank) were synthesized using Nanjing Kingsley Biotechnology Co., Ltd., and cloned them into pUC57 vector, which were named pUC57-T. equi and pUC57-B. caballi, respectively.

Based on the 18S rRNA gene sequence of T. equi (Z15105.1, DQ287951.1, EU642511.1, KM046918.1, KF559357.1, AY150063.2) and B. caballi (KY952238.1, KY952236.1, KY952233.1, AY309955.1, AY534883.1, Z15104.1) from GenBank, a set of primers were designed from the conserved regions shared by all gene sequences (Supplementary Figure 1) using Primer 5.0. Table 1 lists the designed internal and external nPCR primers.

The first round of nPCR was conducted in a 20 μL reaction system containing 10 μL of 2 × PCR Master Mix (Beyotime Biotechnology, Shanghai, China), 0.5 μL of each primer (18S Out – F, 10 μM; 18S Out – R, 10 μM), and 1 ng of a standard positive plasmid. A pre-denaturation step at 95 °C for 5 min was conducted and a total of 35 amplification cycles was performed. Each cycle consisted of denaturation at 95 °C for 15 s, annealing at 56 °C for 15 s, and extension at 72 °C for 30 s. A final extension was also included at 72 °C for 7 min in the PCR profile. In total 0.5 μL of the amplified products were used in the second round of the nPCR, also in a 20 μL reaction system, changing the primers to 18S In – F (10 μM) and 18S In – R (10 μM). The PCR reaction conditions and programs were same as in round one. The amplified products were electrophoresed in 4% agarose gels and the results were documented using the gel doc-IT² imager system (UVP Inc, CA, USA).

To determine detection specificity of nPCR, known standard positive plasmids were used for amplification, negative blood DNA samples and genomic DNA of T. sinensis were used as negative controls, and ddH₂O was used as a blank control. To determine detection sensitivity, the standard positive plasmid was diluted into different concentrations (1 × 10⁻¹ to 1 × 10⁻⁸ ng/μL) and nPCR amplification were performed. The lowest detectable concentration was determined by observing agarose gel electrophoresis results of second round PCR products.

Using nPCR, 36 horse blood samples were tested. The detection method was performed in triplicate, and second round PCR products were sequenced at Sangon Biotech Co. Ltd., Shanghai. Further, the same set of samples was also detected by a multinested PCR (mnPCR) targeting the EMA-1 gene of T. equi and the RAP-1 gene of B. caballi as previously described (15) and the primers were shown in Table 2.

Table 3 lists the primer sequences and probes used in this study. Primers and probes were designed from the species-specific region of the 18S rDNA (Supplementary Figure 1), using Primer Express (version 3.0) and Primer (version 5.0) software. TaqMan probes with reporter dye 6-Carboxyfluorescein (FAM) or 5-Hexachloro-fluorescein (HEX) at the 5' end, and Black Hole Quencher-1(BHQ1) at the 3' end were labeled.

To determine detection specificity of duplex fluorescence qPCR, a 50 μL reaction volume was used, containing 1 μL (20 pmol) of each forward and reverse primer, 0.5 μL (10 pmol) of TaqMan probes, 0.5 μL (1 × 10⁻¹ ng/μL) of standard positive plasmid, 25 μL of TaqMan™ Real-Time PCR Master Mixes (Thermo Fisher Scientific, MA, USA), and ddH₂O to 50 μL. The following thermocycling program was used: 40–45 cycles of 95 °C for 15 s, 60 °C for 15 s, and 72 °C for 30 s with an initial cycle of 95 °C for 30 s, using the Roche LightCycler 480 real time PCR system (Hoffmann-La Roche Ltd, Basel, Schweiz). All amplifications and detections were conducted using a 96-well reaction plate with optical caps (Sangon Biotech, Shanghai, China). At each cycle, accumulation of PCR products were detected by monitoring the increase in reporter dye fluorescence of the TaqMan probes. To test sensitivity, the standard positive plasmid was diluted into different concentrations (1 × 10⁻¹ to 1 × 10⁻⁸ ng/μL) and the lowest detection efficiency and number of copies were determined.

The 36 horse blood samples consisted with those detected by nPCR were tested using duplex fluorescence qPCR, and were performed in triplicate.

The nPCR products were analyzed of standard positive plasmids using agarose gel electrophoresis. First round amplification products of pUC57-T. equi and pUC57-B. caballi (using 18S Out primers) were about 1,000 bp in length, demonstrated by clear and complete electrophoretic bands of the amplified products (Figure 1A). Then, first-round amplification products were amplified a second time (with 18S In primers), revealing clear and bright electrophoretic bands at about 400 bp (Figure 1B).

Fragment lengths obtained from the amplification were consistent with the target product, indicating successful primer design. For further verification, the amplified products were sent to Shanghai Biotechnology Co., Ltd. for sequencing, which confirmed correct sequences. Therefore, the designed primers could be effectively used in subsequent experiments.

Figure 1 shows a single band at the expected sizes following nPCR, indicating successful primer design. Moreover, Figure 2A shows the absence of a target band when primers were used with T. sinensis, suggesting primers are specific for EP detection. nPCR results, however, failed to easily distinguish the two T. equi and B. caballi pathogens by electrophoresis alone. To determine nPCR sensitivity, a 10-fold serial dilution of the standard positive plasmid (from an initial concentration of 1 ng/μL) was used. Using this method, standard positive plasmids at 1 × 10⁻¹ to 1 × 10⁻⁷ ng/μL could be detected (Figure 2B), indicating that the nPCR method has high sensitivity and can amplify DNA concentrations as low as 1 × 10⁻⁷ ng/μL.

Using nPCR, analysis of the 36 blood samples revealed 29 positive samples, demonstrating a positivity rate of 80.56%. Of the 29 positive samples, 21 were infected with T. equi, six with B. caballi, and two were mixed (Supplementary Figure 2, Table 4). This analysis was performed in triplicate, and showed the same results. Sequencing analysis of the nPCR products confirmed these results which showed that six were positive for B. caballi and 21 for T. equi, suggesting that nPCR can be an effective tool in detecting EP infection in clinical disease. However, to separate the 27 bp difference in PCR product of DNA from horse blood samples simultaneous infected with T. equi and B. caballi, electrophoresis with 4% agarose gels was needed. The mnPCR for the 36 blood samples showed that 22.22% of the blood samples (8/36) were positive for equine piroplasmids. The prevalence of infection for T. equi was 2.78% (1/36), and 19.44% (7/36) for B. caballi (Supplementary Figure 3, Table 4). The results mean that nPCR method is more sensitive when the parasitemia was low.

An array of fluorescent probes and templates (Figures 3A,B) were initially tried, before deciding on HEX to label T. equi (using a detection wavelength range of 533–580 nm) and FAM to label B. caballi (using a detection wavelength range of 465–510 nm). Only a FAM-B. caballi amplification curve in the FAM panel (Figure 3A), and a HEX-T. equi amplification curve in the HEX panel were detected (Figure 3B).

The two different types of primers and probes were then added into one reaction vessel, changing only the template (Figures 3C,D). An amplification curve in the FAM panel was observed when B. caballi template was used (Figure 3C), and an amplification curve in the HEX panel when T. equi template was used (Figure 3D). These results indicate that the duplex fluorescence qPCR method demonstrates good specificity.

The standard positive plasmid template was diluted into different concentrations (ng/μL): 1 × 10⁻¹, 1 × 10⁻², 1 × 10⁻³,1 × 10⁻⁴, 1 × 10⁻⁵, 1 × 10⁻⁶, 1 × 10⁻⁷, and 1 × 10⁻⁸, respectively, using 1 μL of template in each reaction system and sterile ddH₂O as a blank control, and assayed the samples using duplex fluorescence qPCR. In the absence of a template, we failed to observe an amplification curve in either FAM or HEX panel (Figures 4A,B). When 1 × 10⁻⁷ and 1 × 10⁻⁸ ng/μL of template was used, we observed a weak amplification curve (Figures 4A,B). Figure 5 shows standard curves generated from combined data of these runs. The correlation coefficient of B. caballi was R² = 0.996, and the standard curve equation was y = 16.39–3.1 log(x) (Figure 5A). The correlation coefficient of T. equi was R² = 0.999, and the standard curve equation was y = 14.75–3.5 log(x) (Figure 5B). These results demonstrate that duplex fluorescence qPCR detected a minimum of 4.47 × 10² copies/μL, indicating that it can identify pathogenic EP species with high efficiency.

Using duplex fluorescence qPCR, analysis of the 36 blood samples revealed 14 infected samples (shown in Supplementary Figure 4, demonstrating a positivity rate of 38.89%. Of the 14 positive samples, 12 were infected with T. equi, one with B. caballi, and one was mixed (Table 4). Then, 36 samples were tested in triplicate using this method and the results were the same. Compared with the nPCR results, duplex fluorescence qPCR demonstrates lower sensitivity.