Development and Evaluation of Multiple Droplet Digital PCR Method for Specific Detection and Differentiation of Brucella melitensis and Brucella abortus

TL;DR

A new digital PCR test was developed to accurately detect and distinguish two types of Brucella bacteria, improving early diagnosis and disease control.

Contribution

A novel multiplex ddPCR assay for specific detection and differentiation of Brucella melitensis and Brucella abortus with high sensitivity and specificity.

Findings

The ddPCR assay detected as low as 2.54–3.11 copies of Brucella per reaction with high reproducibility.

The assay showed superior sensitivity compared to qPCR, detecting 17 positives versus 13 in clinical samples.

The method achieved 100% sensitivity and 95.96% specificity in clinical validation.

Abstract

Brucellosis is a widespread and highly contagious zoonotic disease with serious implications for both animal and human health. In China, Brucella melitensis and Brucella abortus are the predominant species, accounting for 80–90% of human brucellosis cases. This study presents the development and validation of novel multiplex droplet digital PCR (ddPCR) assay designed for the highly sensitive and specific detection and differentiation of B. melitensis and B. abortus. The assay achieved exceptional detection limits of 3.11 copies/reaction (95% CI: 2.68–3.61) for B. melitensis and 2.54 copies/reaction (95%CI: 2.32–2.79) for B. abortus, with strong linearity (R2 > 0.99) and high reproducibility (CV < 9%). In a comparative evaluation using a panel of clinical nucleic acid samples, ddPCR detected 17 positives versus 13 by qPCR, demonstrating superior sensitivity, particularly in samples with low bacterial load. This robust and precise multiplex ddPCR assay provides a reliable tool for early diagnosis and epidemiological surveillance, enabling improved brucellosis control.

Brucellosis remains a major global public health concern, with B. melitensis and B. abortus being the primary causative agents in China. This study describes the development of a multiplex droplet digital PCR (ddPCR) assay for the simultaneous detection and differentiation of B. melitensis and B. abortus. The assay employs TaqMan probes targeting the bcsp31 gene (genus-specific), a transposase gene (B. melitensis-specific) and an autotransporter-associated beta strand repeat-containing protein gene (B. abortus-specific), each labeled with distinct fluorophores (FAM, HEX, ROX). The optimized assay exhibited no cross-reactivity with other pathogens and exhibited significantly higher sensitivity than both qPCR and conventional PCR, with detection limits as low as 2.54–3.11 copies/reaction. Repeatability was excellent, with intra- and inter-assay coefficients of variation below 9%. When validated on a panel of clinical nucleic acid samples, the ddPCR assay showed strong agreement with qPCR (kappa = 0.85), with a sensitivity of 100% (79.42%~100%, 95%CI) and specificity of 95.96% (92.08%~99.84%, 95%CI). These findings establish the multiplex ddPCR as a rapid, sensitive, and highly specific diagnostic platform that improves brucellosis detection accuracy and supports targeted control strategies.

Figures

Full text

1. Introduction

Brucellosis is a globally significant zoonosis caused by bacteria of the genus Brucella [1,2], with B. melitensis and B. abortus being the most prevalent and pathogenic species affecting both humans and livestock [3]. In animals, infection often leads to reproductive failures such as abortion and orchitis, while in humans, it causes debilitating conditions including undulant fever and chronic inflammation [4,5]. Conventional detection methods, such as bacterial culture, are time-consuming and pose biosafety risks [6]. Although molecular techniques like quantitative PCR (qPCR), loop-mediated isothermal amplification (LAMP), recombinase aided isothermal nucleic acid amplification (RAA) have improved detection speed, they suffer from limitations including the need for standard curves, subjective interpretation, and non-specific amplification [7,8,9,10,11].

Droplet Digital PCR (ddPCR) offers a transformative approach by enabling absolute quantification of nucleic acids without external calibration, thereby enhancing precision, reproducibility and sensitivity [12]. In human medicine, ddPCR has been widely applied for the diagnosis of conditions such as ovarian cancer, lower respiratory tract infections, bloodstream pathogens, and tumor biomarkers [13,14,15,16]. In veterinary contexts, ddPCR has been successfully used for detecting B. abortus and B. suis. For instance, Fusco et al. [7] developed a ddPCR assay for B. abortus with a detection limit of 225 CFU/mL. When tested on 599 tissue samples, the assay showed 100% sensitivity, 93.4% specificity, and 94.15% accuracy compared to qPCR. Similarly, Wang et al. designed a ddPCR method targeting SNP sites to distinguish B. suis S2, improving detection sensitivity by optimizing probe design and achieving a detection limit of 10 copies/µL for both wild-type and vaccine strains [17]. Despite these advances, the application of ddPCR for multiplex discrimination of B. melitensis and B. abortus remains underexplored.

In this study, we report the development and validation of a novel multiplex ddPCR assay for the simultaneous detection and differentiation of B. melitensis and B. abortus. The assay’s performance was rigorously evaluated using a well-characterized panel of clinical nucleic acid samples to assess its diagnostic reliability and potential for improving brucellosis surveillance and control.

2. Materials and Methods

2.1. Bacterial Strains

All bacteria strains used in this study, including B. melitensis M5, B. abortus A19, B. suis S2, Brucella canis CQ3, Mycobacterium tuberculosis, Escherichia coli, Salmonella choleraesuis, Staphylococcus aureus, Ochrobactrum anthropi, Pasteurella multocida and Haemophilus parasuis, were obtained as standard reference strains (frozen stocks) from the Zoonosis Monitoring Center of the China Animal Health and Epidemiology Center (Qingdao, China). All bacterial culture procedures, particularly for Brucella spp., were conducted under appropriate biosafety containment in accordance with the recommendations of the WOAH Manual of Diagnostic Tests and Vaccines for Terrestrial Animals [18,19]. The purity of extracted bacterial DNA was confirmed by an A260/A280 ratio between 1.8 and 2.0.

2.2. A Panel of Nucleic Acids from Clinical Samples

A validation panel of nucleic acids extracted from clinical samples was obtained from the National Brucellosis Para-reference Laboratory at the China Animal Health and Epidemiology Center. The panel consisted of 112 characterized samples, which comprised the following: 31 vaginal swab samples (including 5 positive for B. melitensis and 1 for B. abortus), 30 environmental swab samples (including 5 positive for B. abortus), 25 milk samples (including 1 positive for B. abortus), and 26 abortion specimens (including 1 positive for B. melitensis), The characterization (positive/negative status and species identification) was determined by a referenced qPCR assay [20], which served as the reference standard for this study. A cycle threshold (CT) value of 35 served as the cut-off, with samples of CT ≤ 35 defined as positive. All 13 samples identified as positive by this reference qPCR assay were successfully confirmed by bacterial isolation culture.

2.3. Primers and Probes

According to previous studies, specific genomic regions were selected from the NCBI GenBank database, including the bcsp31 gene of Brucella spp. (Sequence ID: AM040264.1, B. abortus 2308 strain), the transposase gene of B. melitensis (Sequence ID: CP001851.1, B. melitensis M5-90 strain), and the autotransporter-associated beta strand repeat-containing protein gene of B. abortus (Sequence ID: CP000888.1, B. abortus S19 strain) [21]. All primers and TaqMan probes used in this study are listed in Table 1. These oligonucleotides were designed using the online PrimerQuest tool (https://sg.idtdna.com/PrimerQuest/Home/Index, accessed on 17 July 2024). The specificity of each primer and probe set was verified using NCBI Primer-BLAST (accessed on 17 July 2024), and potential primer-dimer formation and secondary structures were assessed with Primer Premier 5.0. All primers and probes were synthesized by Sangon Biotech Co., Ltd. (Shanghai, China).

2.4. PCR Assay

The Multiplex PCR assays were conducted in a final volume of 25 µL, following a previously described protocol with minor modifications [22]. Each reaction mixture consisted of 12.5 µL of Taq PCR Master Mix, 1 µL of each primer (BAA13334_I00002f/r and BMEA_A1673f/r, final concentration 400 nM), 1 µL of DNA template, and 7.5 µL of nuclease-free water. Amplification was performed under the following conditions: initial denaturation at 94 °C for 10 min; 30 cycles of 94 °C for 1 min, 56 °C for 1 min and 72 °C for 90 s; and a final elongation step at 72 °C for 8 min. The PCR products were analyzed by electrophoresis on a 1.5% agarose gel stained with a nucleic acid dye at 100 V for 30 min. The expected sizes of the amplification products are 1154 bp for B. abortus, and 745 bp for B. melitensis, which can be resolved as distinct bands.

2.5. Quantitative PCR Assay

The triplex qPCR assay developed in this study was performed in a final volume of 20 µL using Premix Ex Tag^TM^ Probe qPCR Mix (Vazyme, Nanjing, China). Each reaction contained 10 µL of 2× Premix Ex Tag^TM^ Probe qPCR Mix, 0.8 µL of each primer (400 nM final concentration), 0.4 µL of each probe (200 nM final concentration), 1 µL of DNA template, and 3 µL of nuclease-free water. Amplification was carried out under the following conditions: initial denaturation at 95 °C for 2 min, followed by 40 cycles of denaturation at 95 °C for 10 s and annealing/extension at 60 °C for 30 s. Each run included a negative control (Sterile ultrapure water) and a positive control (genomic DNA from B. melitensis or B. abortus). When applied to the characterized clinical panel, this in-house assay yielded results in complete agreement with the reference standard, detecting the same 13 positive and 99 negative samples.

2.6. Droplet Digital PCR Assay

Probe concentrations and annealing temperature were optimized to establish the multiplex ddPCR assay. The reaction mixture (22 µL total volume) contained 11 µL of 2× ddPCR probe Master mix (Sniper, Suzhou, China), 1 µL each of Bru, BruM, and BruA primer stock solutions (final concentration of 455 nM per primer), 0.44 µL of Bru-P probe (200 nM final concentration), 0.66 µL of BruM-P probe (300 nM final concentration), 1.1 µL of BruA-P probe (500 nM final concentration), 1 µL of DNA template. Nuclease-free water was added to adjust the final volume to 22 µL. Droplet generation and fluorescence reading were performed using the Sniper DQ24pro^TM^ dPCR system (Sniper, Suzhou, China). The thermal cycling protocol consisted of an initial step at 60 °C for 5 min (ramp rate: 2 °C/s), followed by 40 cycles of denaturation at 95 °C for 20 s and annealing/extension at 54 °C for 30 s, with a final hold at 12 °C [12]. Data were analyzed using SightPro V1.3.1 software. Each run included a negative control (Sterile ultrapure water), and a positive control (genomic DNA from B. melitensis or B. abortus).

2.7. Analytical Sensitivity and Repeatability

Serial dilutions of B. melitensis M5 and B. abortus A19 ranging from 1 ng/µL to 5 × 10^−6^ ng/µL were subjected to PCR, qPCR and ddPCR to evaluate the sensitivity of each assay. To relate DNA mass to biological units, genome equivalents (GE) were calculated. Based on an average genome size of 3.3 Mb for Brucella spp., the mass of a single genome is approximately 3.4 fg. Consequently, 1 ng of pure genomic DNA corresponds to approximately 2.95 × 10^5^ GE.

The limit of blank (LOB) for each detection channel was determined by analyzing 24 nuclease-free water samples. The LOB was calculated as the result at the position [N_B_ × (p/100) + 0.5], where N_B_ represents the total number of blank samples and p denotes the proportion of experimental data assigned in descending order (p = 95% in this study). Accordingly,

$$[eqn]$$

Samples with copy numbers exceeding the LOB were classified as positive [23].

The limit of quantification (LOQ) was defined as the lowest concentration of the standard that could be quantitatively measured with a coefficient of variation (CV) below 25% [24]. Subsequently, further 2-, 4-, 10- and 20-fold serial dilutions were prepared from 100 fg/µL DNA solutions of B. melitensis M5 and B. abortus A19, with eight replicates for each dilution. The 95% limit of detection (LOD) was determined by probit regression analysis using data from all eight replicates for each sample [25].

Serially diluted samples of B. melitensis M5 and B. abortus A19 DNA were prepared across a ten-fold dilution series (5 × 10^−1^, 5 × 10^−2^, and 5 × 10^−3^ ng/µL). Repeatability was assessed at two levels: intra-assay and inter-assay. Intra-assay repeatability was determined by testing each dilution in triplicate within the same run. Inter-assay repeatability was evaluated by performing three independent experiments on different days. The mean, standard deviation, and coefficient of variation (CV) were calculated from the detection results to evaluate the repeatability of the method [26].

2.8. Analytical Specificity

The analytical specificity of the assay was evaluated using a panel of 11 pathogenic bacterial strains, including B. melitensis M5, B. abortus A19, B. suis S2, B. canis CQ3, M. tuberculosis, E. coli, S. choleraesuis, S. aureus, O. anthropi, P. multocida and H. parasuis.

2.9. Assay Performance on the Nucleic Acid Panel

The clinical utility of the ddPCR assay was evaluated using a panel of 112 characterized clinical nucleic acid samples. A well-established qPCR assay [20] was employed as the reference standard. This reference method was used to characterize the panel, and all 13 qPCR-positive samples were confirmed by bacterial isolation culture. The detection results of ddPCR assay were compared against this reference standard. For this comparison, samples were categorized as follows: true positive (TP, positive by both methods), false positive (FP, positive by ddPCR but negative by qPCR), true negative (TN, negative by both), and false negative (FN, negative by ddPCR but positive by qPCR). The sensitivity and specificity of the ddPCR assay were calculated as follows: Sensitivity = [TP/(TP + FN)] × 100%; Specificity = [TN/(TN + FP)] × 100%. Agreement between the two methods was assessed by calculating the kappa statistic using MedCalc V23.4.8 software (MedCalc Software bvba, Ostend, Belgium).

3. Results

3.1. Optimization of Multiplex ddPCR

The presence of multiple target sequences in the multiplex ddPCR assay necessitated careful optimization of probe concentration and annealing temperature to ensure efficient and balanced amplification of all targets. Probe concentrations were optimized using 0.5 ng/µL DNA from B. melitensis M5 and B. abortus A19. Various probe concentration combinations were evaluated, and their performance was analyzed using SightPro software (Sniper Technologies, China). The optimal probe concentrations were selected based on the combination that yielded the clearest separation between positive (colored) and negative (gray) droplet clusters, with well-defined boundaries and distinct fluorescence amplitude intervals. As shown in Figure 1A–C, the optimal concentrations were determined to be 200 nM for Bru-P, 300 nM for BruM-P, and 500 nM for BruA-P. To optimize the annealing temperature, reactions containing 0.5 ng/µL DNA from each strain were tested across a temperature gradient from 54 to 64 °C. The highest number of positive droplets was observed at 54 °C, indicating optimal amplification efficiency and effective target separation (Figure 1D–F).

3.2. Evaluation of the Multiplex ddPCR with DNA Samples

For the ddPCR assay, three fluorescence channels (FAM, HEX and ROX) were utilized. The LOB values determined from blank samples were 0, 0.53 and 0 copies/reaction for the FAM, HEX and ROX channels, respectively (Figure 2). Assay sensitivity was evaluated using serial dilutions of B. melitensis M5 and B. abortus A19 DNA, ranging from 1 ng/µL to 5 fg/µL. A sample was identified as B. melitensis if both the bcsp31 and transposase gene targets tested positive, and as B. abortus if both the bcsp31 and autotransporter-associated beta strand repeat-containing protein gene targets were positive. Importantly, a sample positive for the bcsp31 target alone was interpreted as indicating the presence of Brucella DNA from a species other than B. melitensis and B. abortus (e.g., B. suis or B. canis). Conversely, samples negative for all three targets were considered negative for Brucella DNA. The LOD, defined as the lowest template concentration detectable with 95% probability, was determined from eight replicate measurements. Probit regression analysis established LOD values of 3.11 copies/reaction (95%CI: 2.68–3.61) for B. melitensis M5 and 2.54 copies/reaction (95%CI: 2.32–2.79) for B. abortus A19 (Table 2). Assuming the target genes are present in single copy per genome, these correspond to approximately 3.11 GE/reaction and 2.54 GE/reaction, respectively. Quantitative analysis demonstrated a strong linear relationship (R^2^ > 0.99) across the concentration range of 1 to 5 × 10^−5^ ng/µL, with log_10_ (concentration fg/µL) plotted against log_10_ (copies/reaction) (Figure 3). The LOQ was determined to be 5.0 × 10^1^ copies/reaction for B. melitensis and 2.6 × 10^1^ copies/reaction for B. abortus (Table 3).

Repeatability was assessed by determining intra- and inter-assay coefficients of variation (CVs). As summarized in Table 4, the CV values for ddPCR were inversely related to template concentration. At DNA concentrations ranging from 5 × 10^−1^ to 5 × 10^−3^ ng/µL, the intra-assay CVs were 3.57–5.29% for B. melitensis M5 and 1.02–8.98% for B. abortus A19, while the inter-assay CVs ranged from 0.13% to 3.86% for B. melitensis M5 and from 3.85% to 5.42% for B. abortus A19. These low CV values indicate excellent repeatability for the detection of both Brucella species.

The specificity of the ddPCR assay was evaluated using genomic DNA from a panel of bacterial strains, including B. melitensis M5, B. abortus A19, B. suis S2, B. canis CQ3, M. tuberculosis, E. coli, S. choleraesuis, S. aureus, O. anthropi, P. multocida, and H. parasuis. As shown in Figure 4, positive droplets were exclusively observed for one or more target genes in the corresponding Brucella strains, while all non-target species showed no positive signals, demonstrating high analytical specificity of the assay.

3.3. Comparison Analysis of the Sensitivity and Standard Curves Between the Multiplex ddPCR and Multiplex qPCR

The sensitivity of multiplex PCR, qPCR and ddPCR was evaluated using serial dilutions of B. melitensis and B. abortus DNA, ranging from 1 × 10^6^ to 5 × 10^0^ fg/µL. For B. melitensis, the detection limits of PCR, qPCR and ddPCR were 5 × 10^3^, 5.1 × 10^0^, and 4.6 × 10^0^ copies/reaction, respectively. Similarly, for B. abortus, the detection limits were 5 × 10^3^, 2.3 × 10^1^ and 4.1 × 10^0^ copies/reaction, respectively (Table 5). These results demonstrate that multiplex ddPCR exhibits superior sensitivity compared to both multiplex qPCR and conventional PCR. Moreover, a strong linear correlation (R^2^ = 0.9986) was observed between the quantitative results generated by multiplex ddPCR and multiplex qPCR across both bacteria species (Figure 3C), reflecting excellent agreement in measurement accuracy.

3.4. Performance of Multiplex ddPCR on the Nucleic Acid Panel and Its Comparison with qPCR

The clinical performance of the multiplex ddPCR assay was evaluated using a panel of 112 nucleic acid samples and compared with a reference qPCR assay. According to qPCR, 13 samples were positive for Brucella DNA (6 for B. melitensis and 7 for B. abortus), and 99 were negative. In comparison, multiplex ddPCR detected 17 positive samples (9 for B. melitensis and 8 for B. abortus) and 95 negatives. To resolve the four discordant samples (ddPCR-positive/qPCR-negative), bacterial isolation culture was performed. Brucella was successfully isolated from two of these samples (one from an abortion specimen and one from a swab), confirming them as true positives that were below the detection limit of the qPCR assay. Considering the qPCR results as the initial reference standard, the multiplex ddPCR assay demonstrated a sensitivity of 100% (79.42%~100%, 95%CI) and a specificity of 95.96% (92.08%99.84%, 95%CI). The agreement between the two methods was substantial, with a kappa value of 0.85 (0.531.00, 95%CI) and an overall concordance rate of 96.43% (108/112) (Table 6). These results indicate that the multiplex ddPCR assay offers superior detection capability for low-level Brucella infections compared to the qPCR method used in this study.

4. Discussion

Brucellosis posed a significant threat to both public health and livestock production safety. B. melitensis, B. abortus and B. suis can infect various livestock species—including cattle, sheep, goats, pigs, sika deer, and camels—and are responsible for severe human infections [27,28]. In China, efforts are currently underway to establish brucellosis-free zones in animals, particularly in pastoral and agro-pastoral regions where mixed farming of cattle and sheep is common. Several studies have reported that ruminants, including cattle and sheep, are susceptible hosts for both B. melitensis and B. abortus [29,30]. Given that B. melitensis poses a greater threat to human health, the early identification of risk factors is crucial for safeguarding public health. Currently, qPCR is the most widely used molecular method for detecting and characterizing Brucella spp. in diverse sample types such as swabs, tissues, and milk [31,32,33]. However, qPCR has several limitations, including the requirement for a standard curve [8] and susceptibility to inhibition by endogenous matrix components—such as proteinases, heavy metals, immunoglobulin G, and lipids—which can lead to inaccurate results [34,35,36]. In this context, we evaluated the diagnostic performance of ddPCR in comparison with qPCR. Using a panel of 112 clinical nucleic acid samples, this study aimed to assess whether ddPCR can overcome the limitations of qPCR and improve the accuracy and timeliness of differentiating B. melitensis and B. abortus infections.

In this study, we developed and validated a multiplex ddPCR assay based on a probe ratio strategy. Three target genes-bcsp31, a transposase gene and an autotransporter-associated beta strand repeat-containing protein gene were labeled with FAM, HEX, and ROX fluorescent probes, respectively. We optimized probe concentrations and annealing temperature due to their critical influence on amplification efficiency and droplet separation. As shown in Figure 1A–C, the optimal probe concentrations were determined to be 200 nM for Bru-P, 300 nM for BruM-P, and 500 nM for BruA-P. Higher annealing temperatures reduced the number of positive droplets, indicating lower amplification efficiency, whereas lower temperatures increased intermediate droplets, reflecting incomplete amplification [37]. Based on the results in Figure 1D–F, an annealing temperature of 54 °C was selected, as it yielded higher fluorescence amplitudes and clearer cluster separation.

For B. melitensis detection, Marcia Ashmi et al. [38] developed a LAMP assay combined with lateral flow immunoassay (LFIA), achieving a sensitivity of 12.1 fg of genomic DNA, comparable to that of qPCR. Shen et al. [39] integrated recombinase polymerase amplification (RPA) with the CRISPR/Cas12a system, enabling detection via fluorescence (FL) or lateral flow strip (LFS). Their RPA-CRISPR/Cas12a-FL and RPA-CRISPR/Cas12a-LFS assays achieved detection limits of 1 and 10 copy/µL, respectively. In comparison, our multiplex ddPCR assay demonstrated a LOD of 3.11 copies/reaction for B. melitensis, indicating superior sensitivity to both LAMP-LFIA and RPA-CRISPR/Cas12a-LFS. For the B. abortus detection, Sung-Il Kang et al. [40] established an LAMP assay with a LOD of 20 fg/µL, while Yang et al. [41] combined LAMP with a nanoparticle-based LFIA, achieving an LOD of 100 fg/reaction for pure genomic DNA. In contrast, our multiplex ddPCR assay achieved an LOD of 2.54 copies/reaction for B. abortus, demonstrating higher sensitivity than both LAMP and LAMP-LFIA. Overall, the multiplex ddPCR assay developed here showed enhanced sensitivity and the ability to detect nucleic acids at lower concentrations.

Analysis of the nucleic acid panel revealed that the multiplex ddPCR assay detected four additional positive samples compared to the reference qPCR. Critically, follow-up bacterial culture confirmed brucella infection in two of these four discordant samples. This microbiological validation demonstrates that these ddPCR signals represented true low-level infections that were missed by the qPCR assay, rather than false positives. Consistent with this finding, all four discordant samples had an estimated DNA load below 50 copies/reaction, highlighting the superior capability of ddPCR to detect minute amounts of target DNA. These findings demonstrate that the multiplex ddPCR assay developed in this study provides a highly sensitive and reliable diagnostic approach for detecting and differentiating B. melitensis and B. abortus, even at low bacterial DNA concentrations, offering substantial practical value for brucellosis surveillance and control programs (Figure 5).

5. Conclusions

A multiplex ddPCR method for the specific detection and differentiation of B. melitensis and B. abortus was successfully established in this study. The assay demonstrated low detection limits (2.54–3.11 copies/reaction) and high specificity for both target species. Although ddPCR remains more costly than qPCR and may have limited applicability in routine diagnostic settings, it represents a valuable tool for clarifying equivocal cases or testing clinical samples with low bacterial loads—particularly for early detection of brucella spp. in suspected herds located in non-free areas, thereby aiding in containment of disease spread. These findings indicate that the multiplex ddPCR assay is a sensitive, specific and reproducible technique for species-specific detection of brucellosis.