The standard strains used in this study are listed in Table 1. These strains were employed to assess the specificity of the dual RPA-LFD assay. For this purpose, S. typhimurium and S. enteritidis were utilized as reference strains. Using these strains, the reaction system of the dual RPA-LFD assay was optimized, along with its detection limit analysis. The above strains were cultured in the Luria-Bertani (LB) medium at 37 °C for 24 h. Bacterial cultures were used for nucleic acid extraction or routine plate counting. To determine the detection limit, S. typhimurium and S. enteritidis solution were serially diluted 10 fold, then 1 mL of these diluted solutions were added into LB medium and incubated at 37 °C for 24 h. To avoid laboratory personnel infection and sample contamination, the whole operation process must be conducted in the biosafety cabinet, and the LB medium culture dishes were inverted during culturing. According to the Bacterial Analysis Manual (BAM), the number of bacterial colonies per plate (repeated 3 times) multiplied by the dilution ratio was the number of bacteria per milliliter (CFU/mL). Notably, only plates containing 30 to 300 colonies could be used for plate counting.

The DNA of the bacteria listed in Table 1 was extracted according to the instructions of the bacterial DNA extraction kit (TIANGEN, Beijing). After centrifugation of the bacterial solution, the cell wall was removed by lysozyme digestion. The DNA was released from the cell through lysate lysis and protease K digestion. The DNA was added to the binding solution to adjust the optimal binding conditions. The DNA solution was transferred to the purification column and centrifuged. The DNA was selectively bound to the filter membrane, while impurities such as proteins entered the filtrate. The residual contaminants and enzyme inhibitors were removed through two washing steps. Finally, the DNA was eluted with a small amount of buffer and stored at −80 °C.

Nucleic acid sequences of the STM4497 gene of S. typhimurium and the safA gene of S. enteritidis were obtained from NCBI. The RPA primers were designed using Primer Premier 5.0 software according to the guidelines provided by the TwistAmp™Basic Kit (TwistDx, US). Specific parameters were set as follows: Primer pair length was between 32–35 bp, GC content was between 30% and 70%, amplicon length was between 100–300 bp, Tm value was between 50% and 100%, and a maximum allowable length of a mononucleotide repeat was “5". Additionally, primer specificity was initially confirmed by homologous comparison with available sequences on NCBI. Subsequently, experimental analysis was conducted to select the primers with the highest amplification efficiency. Furthermore, their specificity was verified by amplifying DNA from various bacterial strains. The best-selected primer pairs (Table 2) were labeled with biotin and digoxin, carboxyfluorescein (FAM), and digoxin. All primers were synthesized by Shanghai Shenggong Bioengineering Technology Service Corporation.

The dual RPA amplification system consisted of 29.5 μL primerless rehydration buffer, 1.2 μL forward and 1.2 μL reverse primer of S. typhimurium, 1.2 μL forward and 1.2 μL reverse primer of S. enteritidis, 2.5 μL magnesium acetate, 13.2 μL DNA template, and water. The steps of the dual RPA amplification system were as follows: the components without template DNA were fully mixed; then the mixture was added to a reaction tube containing lyophilized enzyme powder and mixed gently with a pipette; two kinds of template DNA were added to the test tube, and 2.5 μL magnesium acetate was added to the test tube lid. After centrifugation, the reaction was performed under a constant temperature for 20 min. Finally, the amplification products of RPA were detected by LFD and gel electrophoresis.

The LFD was mainly composed of a sample pad, binding pad, absorption pad, liner, and nitrocellulofiltration membrane, with two test lines and one control line (Figure 1). The Au-labeled anti-digoxin monoclonal antibody was sprayed on the binding mat, and then the two test lines T1 and T2 were prepared with anti-biotin monoclonal antibody and anti-FAM monoclonal antibody respectively. The fixed anti-digoxin anti-antibody on the control line was used as the assay control. Before pipetting the liquid onto the test strip, the amplification products were diluted 50–100 times with ultra-pure water. The result of the test line would be read after 5–10 min.

The principle of dual RPA-LFD assay is shown in Figure 1. Firstly, the DNA of S. typhimurium and S. enteritidis was extracted using a DNA extraction kit. Then dual RPA amplification was performed using labeled upstream and downstream primers to produce biotin and digoxin, or FAM and digoxin labeled double-stranded DNA amplification products. Finally, the amplification products were diluted 50 times and added to the sample pad, and the results were observed after 5–10 min.

When the amplified solution migrated to the other end, the labeled double-stranded DNA products were bound to the Au-labeled anti-digoxin monoclonal antibody on the binding pad. When the amplification solution continued to migrate to the test line, biotin or FAM-labeled double-stranded DNA products were captured on the T1 or T2 test line containing corresponding anti-biotin monoclonal antibodies or anti-FAM monoclonal antibodies, respectively. The excess gold nanoparticles were captured by control lines immobilized with a secondary antibody. In the absence of target DNA, red bands did not appear on the T1 or T2 test line. If amplification was successful, red bands would be produced on the test and control lines. The entire detection process, including amplification and reading of results, needed about 30 min.

To establish a dual RPA assay, preliminary single-tube experiments were performed to screen the optimal primer pair and primer ratio and to test different reaction conditions. The single-tube amplification system for S. typhimurium and S. enteritidis was optimized. The amplification efficiency of different targets was inconsistent in the experiment. Different concentration ratios of the two primers were investigated to determine the optimal primer ratio to achieve similar amplification efficiency of the two fragments. To determine the optimal ratio of STM4497 primers and safA primers, we set the ST-F4/R4: SE-F4/R4 ratio to 10 μM: 2 μM, 10 μM: 4 μM, 10 μM: 6 μM, 10 μM: 8 μM, 10 μM: 10 μM, 2 μM: 10 μM, 4 μM: 10 μM, 6 μM: 10 μM, and 8 μM: 10 μM, then added equal amounts of primers with different concentrations in reaction system, finally performed the reaction at 38 °C for 20 min. Gel electrophoresis analysis was used to detect the product bands and the optimal primer ratio was determined by observing the brightness of the product bands. Through the primer optimization, the optimal reaction temperature, reaction time, and DNA concentration of dual RPA were also investigated. The dual RPA reaction was performed at five different reaction temperatures (35 °C, 37 °C, 38 °C, 39 °C, and 41 °C), five different reaction times (20 min, 25 min, 30 min, 35 min, and 40 min), and three different template DNA concentrations (1 μL: 1 μL, 2 μL: 2 μL, and 3 μL: 3 μL). Gel electrophoresis analysis was used to detect the product bands and the optimal reaction temperature, reaction time, and DNA concentration were determined by observing the brightness of the product bands. Genomic DNA (10⁷ CFU/mL) of S. typhimurium and S. enteritidis were extracted as target templates according to the above protocol. All optimization experiments were carried out in a metal bath.

In detection limit experiments, the detection limits were determined using a tenfold series dilution of two reference strains. Genomic DNA was extracted from bacterial cultures of S. typhimurium and S. enteritidis at concentrations ranging from 10⁰ to 10⁷ CFU/mL. Each concentration level was amplified three times, and the resulting products were subjected to gel electrophoresis and tested with strips for detection purposes. The genomic DNA (10⁷ CFU/mL) extracted from the five bacterial strains listed in Table 1 was incorporated into the reaction to ascertain the specificity of the dual RPA-LFD assay.

Both S. typhimurium and S. enteritidis were cultured in LB medium at 37°C for 18 h, and the cultures were utilized to prepare template DNA solution at various concentrations. Stool samples were collected from the Affiliated Hospital of Youjiang Medical University for Nationalities, all of which tested negative for S. typhimurium and S. enteritidis using the BAM method established by the US Food and Drug Administration. Subsequently, these collected samples were artificially contaminated with different concentrations of S. typhimurium and S. enteritidis ranging from 10⁰–10⁷ CFU/mL. Each sample was then subjected to DNA extraction.

To extract bacterial DNA from the artificially contaminated samples, 20 mg/mL protease K was added to digest the protein in the sample and destroy the bacterial cell wall. The DNA extracted from each sample was used in a dual RPA amplification reaction. Contaminated samples were also tested according to BAM. Uncontaminated samples were used as negative controls.

To evaluate the clinical applicability, the dual RPA-LFD assay was employed for the analysis of 25 stool samples from the Affiliated Hospital of Youjiang Medical University for Nationalities. Among these, 8 stool samples tested positive for S. typhimurium, 4 were positive for S. enteritidis, and 13 were negative for both S. typhimurium and S. enteritidis. Untreated stool samples or extracted DNA were stored at −80°C. The samples underwent testing using two different methods: BAM and the dual RPA-LFD assay.

LFD and gel electrophoresis results were visually read. The optimum conditions were selected according to the brightness of the strips. Data collected from the dual RPA-LFD assay and culture methods were analyzed, including the positive detection rates of S. typhimurium and S. enteritidis in clinical samples with BAM and the dual RPA-LFD assay.

A primer in RPA required longer oligonucleotides, typically 30–35 base pairs. For the STM4497 gene of S. typhimurium and the safA gene of S. enteritidis, a total of 5 primer pairs were designed (Figure 2). Among these, F4/R4 primers exhibited the highest amplification efficiency with single bands. Consequently, this primer pair was selected for subsequent dual RPA-LFD assay experiments.

The presence of two primers in the dual RPA amplification reaction might result in mutual interference between them. To determine the optimal ratio of STM4497 primers and safA primers. We assessed gel electrophoresis bands after adding equal amounts of primers with different concentrations at 38 °C for 20 min (Figures 3A,B). When the ratio reached 1:1, minimal interference was observed and the amplification efficiency was maximized. Additionally, we investigated the impact of varying template amounts on amplification efficiency by adding different quantities of templates under identical conditions (Figure 3C). The highest amplification efficiency was achieved when adding 2 μL each of genomic DNA from S. typhimurium and S. enteritidis.

To evaluate the optimal amplification temperature, we incubated the dual RPA reaction system at different temperatures (35 °C, 37 °C, 38 °C, 39 °C, and 41 °C) in a metal bath for 30 min (Figure 3D). Gel electrophoresis analysis was used to detect the product bands. The brightest band was obtained at a temperature of 38 °C.

In order to determine the optimal reaction time, we performed the dual RPA amplification reaction at a constant temperature of 38°C for various durations (20 min, 25 min, 30 min, 35 min, and 40 min). Gel electrophoresis analysis was used to detect the product bands and the optimum reaction temperature was determined by observing the brightness of the product bands. As depicted in Figure 3E, the product band became visible after only 20 min. However, as time extended to 40 min, the product band did not become brighter. Therefore, we selected 20 min as the most suitable duration for performing the dual RPA amplification reaction.

To assess the detection limits of the dual RPA-LFD assay, a 10-fold serial dilution of the pure bacterial solution was prepared using purified bacterial genomic DNA as a template. The amplified products were analyzed by LFD and gel electrophoresis. When the bacterial concentrations of S. typhimurium and S. enteritidis were below 10¹ CFU/mL, the color of the test line was not visible to the naked eye (Figure 4A). Therefore, the detection limits of the dual RPA-LFD assay for simultaneous detection of S. typhimurium and S. enteritidis were 5.23×10¹ and 3.59 × 10¹ CFU/mL, respectively.

To investigate the specificity of the dual RPA-LFD assay, DNA samples from a panel of S. typhimurium, S. enteritidis, and other common pathogens were extracted separately and used for the specificity determination of the dual RPA-LFD assay (Figure 5A strip 1, 2, 3, 4, 5, 6). Only S. typhimurium and S. enteritidis strains (Figure 5A stripe 1, 2, 3) tested positive, while the other bacteria including Klebsiella pneumoniae (stripe 4), Escherichia coli (stripe 5), Enterococcus faecalis (stripe 6) were tested as negative. The dual RPA-LFD assay was able to correctly differentiate S. typhimurium, S. enteritidis, and several other common pathogens, meaning that the dual RPA-LFD assay had a high degree of specificity. The experiment was repeated three times under the same conditions.

Different concentrations of S. typhimurium and S. enteritidis were added to the negative stool sample to investigate the usefulness of the dual RPA-LFD assay. As shown in Figure 6, the detection limits of the dual LDF-RPA assay for S. typhimurium and S. enteritidis were 8.30×10¹ and 2.70 × 10² CFU/mL, respectively. The detection limits of the dual RPA-LFD assay for the detection of S. typhimurium in artificially contaminated samples were identical to those of the pure bacterial solution, but the detection limits for the detection of S. enteritidis in artificially contaminated samples were higher than those of the pure bacterial solution.

The specificity of the dual RPA-LFD assay in artificial pollution samples were also investigated. In this experiment, a double-blind design was adopted. 10 unknown artificial contaminated samples were tested by dual RPA-LFD assay and the results were shown in Figure 7. 3 samples tested positive for S. typhimurium and S. enteritidis, 3 samples tested positive for S. typhimurium, 3 samples were tested positive for S. enteritidis and 1 control was tested as negative. The results were completely consistent with the actual contaminated samples.

S. typhimurium and S. enteritidis were detected in 25 clinical samples using the dual RPA-LFD assay (Figure 8) and BAM. The results of the two tests were identical, with sensitivity and specificity of 100% (8/8) and 100% (17/17) for S. typhimurium, and 100% (4/4) and 100% (21/21) for S. enteritidis (Table 3).