Reference genomic DNA (gDNA) samples (n = 19; Supplementary Table S1) including 13 strains of targeted bacteria (Campylobacter coli, C. jejuni, Clostridium perfringens, Escherichia coli, Enterococcus faecalis (2 strains), Enterococcus faecium, L. monocytogenes (2 strains), Salmonella enterica (2 strains), Shigella spp., and Staphylococcus aureus), non-targeted bacteria (Listeria marthii), protozoa (Crithidia fasciculata and Leishmania martiniquensis), severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), human (Applied Biosystems, California, United States), and plant (teak leave), were utilized as DNA templates for developing and optimizing the MassARRAY-based assay. A collection of 85 laboratory bacterial strains, sourced from diverse environments, animals, and humans (Supplementary Table S2) were used for assay evaluations. These bacterial strains had undergone rigorous species identification, using various methods, including conventional bacterial culture with biochemical tests, conventional PCR, and whole-genome sequencing.

Escherichia coli, Shigella spp., Klebsiella spp., Listeria spp., Salmonella spp., and S. aureus were cultured on tryptic soy agar (Difco Laboratories Inc., Franklin Lakes, New Jersy, United States) at 37°C for 18–24 h. Enterococcus spp. was grown on De Man–Rogosa–Sharpe (MRS) (Difco Laboratories Inc., Franklin Lakes, New Jersy, United States) at 37°C for 24–48 h. Campylobacter spp. was grown on sheep blood agar (Thermo Fisher Scientific Inc. Massachusetts, United States) at 42°C for 24 h under micro-aerophilic condition using AnaeroPack™-MicroAero (Mitsubishi Gas Chemical Company, Inc., Japan). Streptococcus spp. was grown on Columbia blood agar (Thermo Fisher Scientific Inc. Massachusetts, United States) at 37°C, in 5% CO₂ incubator for 18–24 h. Vibrio parahaemolyticus was cultured overnight on tryptic soy agar (TSA, Difco Laboratories Inc., Franklin Lakes, New Jersy, United States) with 2% NaCl at 37°C in 5% CO₂ incubator for 16–18 h. Clostridium spp. was grown on Reinforced Clostridial Medium (RCM) (Difco Laboratories Inc., Franklin Lakes, New Jersy, United States) under anaerobic condition (10% H₂/10% CO₂/80% N₂) at 37°C for 48 h. Following the culturing process, a single colony of each bacterial species was selected for the bacterial culture to prepare gDNA subsequently.

Three distinct DNA extraction kits were employed to ensure comprehensive and reliable DNA isolation for specific purposes in this study. The ZymoBIOMICS DNA Miniprep Kit (Zymo Research Corp, California, United States) was used for gDNA extraction following the manufacturer’s protocol, yielding highly pure genomic DNA suitable for primer specificity testing and method development experiments. For MassARRAY-based bacterial detection in field samples, the MagPurix^® Bacterial DNA Extraction Kit (ZP02006) (Zinexts Life Science Crop., New Taipei, Taiwan) in combined with MagPurix^® 12EVO automatic instruments was utilized to obtain high-quality gDNA templates. Heat-lysis extraction using QuickExtract DNA Extraction Solution (LGC Biosearch Technologies, Hoddesdon, United Kingdom) was conducted to obtain gDNA samples for the LOD determination experiment. The quality of extracted gDNA was evaluated using a NanoDrop spectrophotometer (Thermo Fisher Scientific Inc., Massachusetts, United States). Samples with an OD260/280 ratio ranging from 1.8 to 2.0 and an OD260/230 ratio within the range of 2.0 to 2.2 were chosen for subsequent experiments.

Meat samples from various sources (n = 103) preserved in culture transport medium, along with bacterial culture plates, including both selective and enrichment plates, were generously provided by the Bureau of Quality Control of Livestock Products (BQCLP), Department of Livestock Development, Ministry of Agriculture and Cooperatives, Thailand. The preparation of meat samples and bacterial cell culture was conducted at the BQCLP laboratory (ISO/IEC 17025), adhering to FDA’s Bacteriological Analytical Manual (BAM) guidelines, including those for aerobic plate count (FDA’s BAM 2001, Chapter 3), Coliform and E. coli (FDA’s BAM 2020 Chapter 4), and Clostridium spp. (FDA’s BAM 2001, Chapter 16 and API). Additionally, protocols outlined in the International Organization for Standardization (ISO) Food Production Guideline (ISO 6759:2017/Amd. 1:2020 for Salmonella spp. and ISO 6888-3:2003 for S. aureus), and Thailand’s Integrated Antimicrobial Resistance Surveillance with One Health Approach Guideline were followed (National Antimicrobial Resistant Surveillance Center, Thailand, 2023; The United States Food and Drug Administration, 2024). Briefly, 25–50 g of meat samples were transferred to culture transport mediums, such as DF (0.1% Peptone normal saline solution), BPB (Butterfield’s Phosphate-Buffered), and BPW (Buffered Peptone Water), in accordance with standard protocols. Subsequently, approximately 5 mL of culture samples were processed for gDNA extraction. For bacterial culture plate samples, a few colonies were selected for gDNA extraction. The DNA extraction procedure employed the MagPurix^® Bacterial DNA Extraction Kit (ZP02006), following the manufacturer’s instructions. The high-quality gDNA samples obtained were subsequently used as DNA templates for bacterial identification using the developed MassARRAY-based assay.

Species-specific genes were chosen based on primer sequences provided in previously reported PCR-based methods (Liu, 2011). The primer design process involved retrieving a substantial number of full coding sequences from the NCBI genome database. For each targeted gene, a substantial number of sequences, ranging from approximately 100–2,000 sequences, were obtained. These sequences were then subjected to multiple sequence alignment using an alignment tool, CLUSTAL-W (Thompson et al., 1994) to identify highly conserved regions within the sequences across different strains or isolates of the same species. The highly conserved regions of the species-specific targeted genes were subsequently employed for primer design using Assay Design 4.0 software (Agena Bioscience, Inc., California, United States).

All targeted sequences were inputted into Assay Design 4.0 software (Agena Bioscience, Inc., California, United States) for primer and assay design, following the manufacturer’s recommendations. The primer design process adhered to stringent parameters to minimize primer-dimer formation, hairpin loops, and false priming, ensuring the production of high-quality and specific primers. A set of primers specific for the conserved region of the 16S rRNA gene was included as an internal control, enabling the detection of any species of bacteria, if presence in the sample. The specificity of designed primers was evaluated using the Basic Local Alignment Search Tool (BLAST) to ensure the highest level of primer specificity was achieved. Subsequently, each primer pair (forward and reverse primers) and different set of multiplex primers were rigorously evaluated using conventional PCR reaction with gDNA samples from targeted bacteria, non-targeted bacteria, and other organisms, as well as negative control (no DNA template) to ensure primer specificity and absence of cross-PCR reactions. Only primers exhibiting the highest specificity were selected for use in the MassARRAY-based assay system. It is important to highlight that manual adjustments, such as altering bases and adjusting primer length, were implemented as necessary to achieve uniformity of melting temperature (Tm) and guanine-cytosine content (%GC) values among multiplex primers, facilitating PCR optimization. All primers were synthesized by Macrogen, Inc. (Macrogen, Inc., Seoul, Korea).

The initial step of the MassARRAY-based assay involved optimizing a multiplex PCR reaction, using gDNA of known bacterial strains as reference DNA templates (n = 14; Supplementary Table S1). L. marthii (non-targeted bacterium) and the non-bacterial gDNA samples, including C. fasciculata and L. martiniquensis, SARS-CoV-2, human, and plant, were used as negative controls. The MassARRAY-based assay panel was optimized and modified based on the manufacturer’s recommended protocol. In brief, the 5 μL of PCR reaction consisted of 2.5 μL of EconoTaq^® PLUS 2X Master Mix (LGC Biosearch Technologies, Hoddesdon, United Kingdom), 0.5 μL of the primer mixture (resulting in a final concentration of 500 nM for each forward and reverse primer), 1.0 μL of gDNA template, and 1 μL of DNase-free distilled water. The PCR reaction includes: an initial denaturation at 94°C for 2 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 55°C for 30 s, extension at 72°C for 1 min, and a final extension step at 72°C for 5 min. The reaction was then cooled to 4°C to stop the reaction. Following the PCR reaction, the excess dNTPs in the reaction were eliminated by shrimp alkaline phosphatase (SAP) (Agena Bioscience, California, United States). The dephosphorylation reaction was performed at 37°C for 40 min, and then inactivated at 85°C for 5 min. Following the SAP reaction, the single-base extension (SBE) reaction was conducted using a mixture of extension reaction cocktail that was prepared according to the iPLEX^® Pro and Gold Reagents User Guide. The SBE reaction was performed with an initial step at 95°C for 30 s, followed by 40 cycles consisting of denaturation at 95°C for 5 s, and 5 cycles of an annealing step at 52°C for 5 s and a denaturation step at 80°C for 5 s. Finally, the extension was carried out at 72°C for 3 min, and the reaction was then held at 4°C. Subsequently, the SBE products were dispensed onto the SpectroCHIP through a nano-dispenser and then loaded into the mass spectrometer (MS) for the measurement of the molecular mass of the SBE products (Agena Bioscience, California, United States). The molecular mass and base calling data corresponding to specific SBE products acquired from the MS were analyzed and interpreted by MassARRAY^® Typer Viewer v4.0 software (Agena Bioscience, California, United States).

The developed MassRRAY-based assay was validated using purified gDNA from 85 bacterial strains, including 75 strains of the targeted bacterial species and 10 strains of non-targeted bacterial species (Supplementary Table S2). For each assay reaction, 10 ng/μL of purified gDNA were used as a sample, and the experiments were done independently in triplicate. The correlation between the developed MassARRAY-based assay results and known bacterial isolates was determined.

The efficiency of the developed MassARRAY-based assay’s reaction was evaluated using a mixture of gDNA from all nine targeted species, including C. coli, C. jejuni, Cl. perfringens, E. coli/Shigella spp., E. faecalis, E. faecium, L. monocytogenes (variant-1), L. monocytogenes (variant-2), Salmonella spp., and S. aureus. The gDNA from each targeted species was combined, with concentrations varying from 10 to 50 ng/μL for each species. Subsequently, 1 μL of the gDNA mixture was utilized as the template, resulting in a final concentration of 1–5 ng/μL for each targeted species. The experiments were conducted independently and in duplicate.

The limit of detection of the developed MassARRAY-based assay was determined using bacterial cells. The bacterial culture was adjusted to an optical density (OD) of 1.85 at 600 nm, equivalent to approximately 10⁹ colony forming units (CFU)/mL. The bacterial cell count was determined through a plate count assay using appropriate culture conditions for the species. The bacterial suspension was 10-fold serially diluted to extinction. One hundred microliters of each bacterial dilution were centrifuged to obtain bacterial cell pellets which were subjected to gDNA extraction using 50 μL of QuickExtract DNA Extraction Solution (LGC Biosearch Technologies, Hoddesdon, United Kingdom), following the manufacturing protocol. Briefly, bacterial cell suspensions were subjected to heat-lysis extraction at 65°C for 6 min, followed by thorough vortexing for 15 s. The mixture solution was then incubated at 98°C for 2 min. Subsequently, 1 μL of the extracted gDNA solution was analyzed using the developed MassARRAY-based assay for the LOD assessment. The results were reported as the average values with their corresponding standard deviations. The experiments were conducted independently, with each experiment performed at least in duplicate.

The developed MassARRAY-based assay for bacterial detection was further validated with field samples (n = 103). The results obtained from the developed assay were subsequently compared to those results obtained from standard microbiological testing methods routinely performed by the BQCLP laboratory. This comparison ensured the accuracy and reliability of the developed MassARRAY-based assay in bacterial identification.

In this study, a MassARRAY-based assay panel was developed for the identification of 10 enteropathogenic bacterial species. The assay panel utilized eleven targeted amplicons to design 11 pairs of multiplex-PCR primers and 12 extension primers. The assay panel comprised eight single-targeted amplicons, including C. jejuni (Camp002), Cl. perfringens (Clos001), E. coli/Shigella spp. (Eco001N), E. faecalis (Ent001), E. faecium (Ent003), L. monocytogenes (Lis001), Salmonella spp. (Sal002), and S. aureus (Stap001), while double-targeted amplicons were designed for C. coli (Camp005 and Camp006). Additionally, one target amplicon was designed for the universal detection of bacteria (Bac16 1–1). As a result, 11 pairs of multiplex-PCR primers were specified for each target. An individual extension primer was designed for each amplicon, except for L. monocytogenes, which required two extension primers. Since a novel single nucleotide polymorphism (SNP) was discovered in the targeted gene of L. monocytogenes strains isolated from samples in Thailand, two extension primers, Lis001 and LisG, were designed to target both wild type (variant-1) and SNP (variant-2), respectively. Therefore, a total of 12 extension primers were utilized in the SBE reaction (Table 1).

According to Assay Design Suite software V4.0, the multiplex-PCR primer lengths were 30–33 bp with a melting temperature (Tm) range of 58.6–66.9°C, and a GC content percentage varying from 38.7 to 54.5%. These multiplex-PCR primers enabled the generation of PCR products with lengths, ranging from approximately 94–150 bp. According to the species-specific sequence selected, the software designed a set of extension primers (EPs) with a Tm range of 45.7–52.1°C, a GC content percentage varying from 21.7 to 62.5%, and molecular mass ranging from 4955.20 to 7670.00 Da, enabling the generation of non-overlapping SBE products with molecular masses ranging from 5282.30 to 7957.20 Da. Detailed information regarding the designed primers was summarized in Table 1.

The MassARRAY-based assay was initially optimized using reference gDNA samples from 10 bacterial species, including C. jejuni, C. coli, Cl. perfringens, E. coli, Shigella spp., E. faecalis, E. faecium, L. monocytogenes (both variant-1 and variant-2), S. enterica, and S. aureus. The resulting chromatogram (Figure 1) yielded the molecular mass of the SBE products for each designed target as expected, suggesting that the assay functioned effectively. All reactions with bacterial gDNA also demonstrated a mass spectrum of Bac16 1–1 SBE product (5816.80 Da). The complete chromatograms fully displayed the specific molecular mass of EPs and SBE products (Supplementary Figure S1). The negative control without any DNA template yielded only the peaks corresponding to the molecular mass of EPs without any of the SBE products or unexpected peaks, indicating a high specificity of the designed primers without any cross-reactivity resulting from the mixed primers.

When a non-targeted bacterium, L. marthii, was used as a template, the expected universal SBE product for bacteria was obtained, indicating suitability of the Bac16 1–1 primer set as an internal control to confirm the successful occurrence of all reactions and the validity of the results. Furthermore, when gDNA samples of non-bacterial samples, C. fasciculata, L. martiniquensis, SARS-CoV-2, human, and plant, were used as templates in the assay, the results showed no SBE products present in the reactions (Supplementary Figure S2).

The evaluation of the developed MassARRAY-based assay’s reaction using a mixture of gDNA from all nine targeted species revealed the capability of a single MassARRAY-based assay to simultaneously detect all targeted species. The distinct molecular mass of the SBE products were clearly observed in the lowest concentration of gDNA as 1 ng/μL (Figure 2). When the mixture of gDNA from targeted bacteria and non-targeted bacteria (e.g., E. coli, E. faecium, Salmonella spp., S. agalactiae, S. suis, C. fasciculata, L. martiniquensis) was used as DNA templates, the complete chromatograms clearly displayed the specific molecular mass of specific SBE products of targeted bacterial species (data not shown). These results suggested the precision of the developed MassARRAY-based assay in accurately identifying the targeted bacteria species in a single run.

The developed MassARRAY-based assay underwent validation using gDNA samples obtained from pure cultures of well-identified bacterial species (n = 85; Supplementary Table S2). The MassARRAY results demonstrated 100% correlation between the molecular mass of SBE products and all bacterial species, with no occurrences of false-negative or false-positive results (Supplementary Table S3), highlighting the high accuracy of the assay. When non-targeted bacteria, including C. lari, K. pneumoniae, L. innocua, S. agalactiae, S. suis, and V. parahaemolyticus, sourced from distinct origins were used as templates, the complete chromatograms demonstrated a mass spectrum of Bac16 1–1 SBE product (5816.80 Da), which was the expected universal SBE product for bacteria (data not shown), indicating the validity of the results.

The limit of detection (LOD) for the MassARRAY-based assay was assessed using serially diluted bacterial cell suspension, from which the gDNA was extracted using QuickExtract DNA Extraction Solution, as described in Section 2.7. The results revealed that the developed method achieved the lowest LOD in the reaction for Cl. perfringens at 357 ± 101 cells and E. coli/Shigella spp. at 445 ± 56 cells. The detection limit for C. jejuni (2,920 ± 651 cells) was comparable to that of Salmonella spp. (5,040 ± 825 cells). The LOD value for E. faecalis and L. monocytogenes was 36,000 ± 13,294 cells and 90,500 ± 48,790 cells, respectively. The highest LOD value was observed for S. aureus (282,000 ± 79,196 cells), C. coli (244,000 ± 10,392 cells), and E. faecium (221,000 ± 26,870 cells) (Table 2).

The performance of the developed MassARRAY-based assay in detecting targeted bacteria in livestock products was assessed using 103 meat samples collected from various sources. The meat samples underwent bacterial culture processing, which involved culture transport medium, selective plates, and enrichment plates. The assay validation was conducted across all stages of sample processing. The results obtained from the developed assay were subsequently compared to analyses conducted at the BQCLP laboratory. This comparison ensured the accuracy and reliability of the developed MassARRAY-based assay for bacterial identification in real-world application. Out of the 103 tested samples, 12 samples were positive for bacterial contamination, while the remaining samples (n = 91) tested negative for targeted bacterial species. These results were consistent with those obtained from bacterial culture-based method routinely conducted by the BQCLP laboratory. Notably, the negative control reaction without DNA template exhibited no SBE products, confirming the absence of contamination. All samples tested positive for Bac16 1–1, confirming the presence of bacteria in all field samples as expected. Furthermore, the specific molecular masses of the SBE products observed in the twelve positive samples matched the respective targeted bacteria, confirming the presence of specific bacterial species. The result demonstrated a 100% correlation between the data obtained from the standard bacterial-culture methods and the MassARRAY-based assay (Table 3). These findings indicated a high level of accuracy and reliability of the developed MassARRAY-based assay in bacterial identification, suggesting its potential for application within real-world field samples.