Phage-based magnetic capture method as an aid for real-time recombinase polymerase amplification detection of Salmonella spp. in milk

Salmonella is a major cause of foodborne diseases worldwide. Conventional rapid assays for detecting Salmonella in real samples often encounter severe matrix interference or detect a limited number of species of a genus, resulting in inaccurate detection. In this study, we developed a method that combined phage-based magnetic capture with real-time recombinase polymerase amplification (RPA) for the rapid, highly sensitive, and specific detection of Salmonella in milk with an ultra-low detection limit. The Felix O-1 phage-conjugated magnetic beads (O-1 pMBs) synthesized in this method showed excellent capture ability for Salmonella spp. and ideal specificity for non-Salmonella strains. After O-1 pMBs-based magnetic separation, the limit of detection of the real-time RPA assay was 50 cfu/mL in milk samples, which was significantly increased by a magnitude of 3 to 4 orders. The method exhibited a high sensitivity (compatibility) of 100% (14/14) for all tested Salmonella serotype strains and an ideal specificity (exclusivity) of 100% (7/7) for the tested non-Salmonella strains. The entire detection process, including Salmonella capture, DNA extraction, and real-time RPA detection, was completed within 1.5 h. Furthermore, milk samples spiked with 10 cfu/25 mL of Salmonella were detected positive after being cultured in buffered peptone water for only 3 h. Therefore, the proposed method could be an alternative for the rapid and accurate detection of Salmonella .


INTRODUCTION
Salmonella infection, one of the major foodborne diseases, is an important public health issue particularly in impoverished areas, posing a substantial threat to human health and food safety and increasing the economic burden on society (Chen et al., 2016;Hadler et al., 2020).Humans infected with Salmonella strains may experience a range of symptoms, including fever, vomiting, abdominal pain, nausea, and hemorrhagic enteritis (Chiu et al., 2004;Ferrari et al., 2019;Fu et al., 2021).Milk is a staple food consumed daily in many countries (de Klerk and Robinson, 2022), and various bacteria easily grow in milk because of its high nutritional content (Han et al., 2018).It has been reported that contaminated milk was the primary source of Salmonella infection (Brewster and Paul, 2016;Bahroun et al., 2018;Ding et al., 2022).Therefore, the development of a rapid and highly sensitive method for the detection Salmonella contamination in milk is crucial.
The primary detection methods for Salmonella are conventional culture-based methods owing to their accuracy, sensitivity, and reproducibility (Wang et al., 2022).However, these conventional methods generally require 4 to 7 d to obtain results, making them unsuitable for rapid detection of Salmonella (Liu et al., 2022).Consequently, several technologies and methods, including protein-based methods, biochemical metabolism-based methods, and nucleic acid sequence-based assays, have been developed for the rapid detection of Salmonella (Liu et al., 2022;Li et al., 2023).Nucleic acid-based methods represented by PCR are widely recognized for their high sensitivity, rapidity, and specificity in detecting pathogens (Wang et al., 2022;Chen et al., 2023a).However, PCR requires complicated equipment to precisely regulate the reaction temperature and skilled personnel to conduct the analysis (Lin et al., 2021).Subsequently, isothermal amplification technologies, such as recombinase polymerase amplification (RPA), loop-mediated isothermal amplification, rolling circle amplification, helicase dependent amplification, and strand displacement amplification have emerged as alternative technologies in recent years (Guan et al., 2023).These technologies eliminate the dependence on sophisticated equipment and have demonstrated potential applications in pointof-care testing (Bodulev and Sakharov, 2020).Among these, RPA has gained increasing attention because of its powerful amplification ability to obtain results within 20 min.In addition, It is noteworthy that RPA can be performed without precise temperature control at an accessible and broad reaction temperature range (37-42°C), which is closest to room temperature among all the aforementioned isothermal amplification technologies (Liu et al., 2017;Chen et al., 2023b).However, the detection of pathogens using RPA is often challenged because the multienzyme system of RPA can be easily affected by the natural inhibitors such as the sample matrix, background microflora and other inhibitory natural substances (Ding et al., 2023).
Immunomagnetic separation (IMS) is commonly employed in the enrichment of targets, nucleic acid extraction, pretreatment of samples, and immobilization of enzymes (He et al., 2014).Immunomagnetic separation typically relied on high-quality antibodies as bio-recognition elements for specific recognition and capture of targets (Wang et al., 2020).However, unlike other bacteria, Salmonella has more than 2,650 serotypes (Xu et al., 2021).General antibodies against Salmonella can only recognize limited kinds of serotypes (Lu et al., 2023), which undoubtedly leads to the difficulty on IMS-based detection methods of the Salmonella genus.Additionally, the process of screening and identifying high-quality antibodies against Salmonella is time consuming and labor intensive (Ertürk and Lood, 2018;Huang et al., 2021).Therefore, an improved bio-recognition element is urgently required for IMS-based detection methods of Salmonella.
Bacteriophages (phages) are kinds of naturally parasitic viruses which can recognize host bacteria with high specificity.Phages exhibit greater stability toward fluctuations in pH, temperature, and ionic strength than antibodies and have lower production costs (Huang et al., 2021).Furthermore, compared with other bio-recognition elements, phages can potentially discriminate between live and dead bacteria (Xu et al., 2023).The Felix O-1 phage, a member of the A1 group of the Myoviridae, is fairly specific for Salmonella (Whichard et al., 2010).The O-1 phage can bind to Salmonella via its tail fiber protein, which can identify lipopolysaccharides on cytomembrane of Salmonella.This phage can identify and lyse 98.2% of Salmonella strains (Welkos et al., 1974).Therefore, the O-1 phage is potentially an ideal biorecognition element for the development of phage-based magnetic separation technology for Salmonella spp.
In this study, we successfully developed a real-time RPA assay for the rapid, highly specific, and sensitive detection of Salmonella in milk, which employs phagebased magnetic capture method as an aid in the detection process.As illustrated in Figure 1, Felix O-1 phageconjugated magnetic beads (O-1 pMBs) were prepared to separate and enrich Salmonella from milk samples.Nucleic acids were extracted from the captured Salmonella via simple boiling without complex extraction procedures.Finally, the detection results were quickly obtained using a real-time RPA process.

Strains and Culture Conditions
A total of 100 μL of strain seeds stored at −20°C were inoculated into 10 mL of nutritional broth medium and then cultured at 37°C and 200 rpm for 14 h.The plate counting method was employed for bacterial quantification.A 10-fold serial dilution of bacteria ranging from 10 8 to 10 1 cfu/mL was prepared for subsequent magnetic capture experiments.
To verify the O-1 phage, the lytic spectrum of the phage was tested by plaque experiments against 6 dif-ferent serotypes of Salmonella and 6 non-Salmonella strains before use.Cultivation and determination of the O-1 phage titer were conducted as previously described (Duyvejonck et al., 2019), and the results were presented in Supplemental Figure S1 (see Notes).

Phage-Based Magnetic Capture of Salmonella
Synthesis of O-1 pMBs.Two types of commonly used carboxylated magnetic beads (MBs; submicron scale and micron scale) were compared with coupled O-1 phages to achieve a higher capture efficiency.The O-1 pMBs were synthesized using the EDC/NHS method as follows: A total of 100 μL of MBs were well-dispersed via ultrasound and washed thrice with 3 mL of MES (50 mM, pH 6.0).Next, 200 μL of EDC (20 mg/mL) and 200 μL of NHS (20 mg/mL) were added and mixed vertically on a tube rotator for 1 h at 25°C to activate the carboxylic groups.The activated beads were washed thrice with PBS (0.01 M, pH 7.4) to remove excess NHS and EDC before being re-suspended by PBS (100 mM,pH 4,5,6,7,8,9,and 10). Thereafter,various volumes (50,100,125,150,and 200 μL) of Felix O-1 phage (10 11 pfu/mL) were added into the suspension and vertically mixed at 25°C to couple on magnetic beads for different time durations (1, 1.5, 2, 2.5, 3, 3.5, and 4 h).After magnetic separation, the O-1 pMBs were re-suspended with 1.5 mL of PBS (0.01 M, pH 7.4) containing 3% (wt/vol) BSA, followed by an incubation for 1 h at 37°C to block residual sites.The obtained O-1 pMBs were washed thrice with PBS (0.01 M, pH 7.4) and stored in 1 mL of PBS (0.01 M, pH 7.4) containing 0.1% (wt/vol) BSA at 4°C.

Characterization of O-1 pMBs.
To demonstrate the biological activity of phage on the O-1 pMBs, MBs, O-1 phage, and O-1 pMBs were compared on the formation of plaques through plaque experiments.A total of 200 μL of Salmonella bacterial culture with a concentration of 10 8 cfu/mL was evenly spread onto nutrient agar plates.Then 3 μL of MBs, O-1 phage, and O-1 pMBs were added dropwise.The plates were cultured overnight at 37°C and plaque formation was then assessed.Further, the O-1 phage, MBs, and the binding complex of O-1 pMBs with Salmonella were also observed using trans-mission electron microscopy (TEM).The TEM samples were prepared using the previously described negative staining method (Hooton et al., 2011).
Magnetic Capture of O-1 pMBs.The capture of bacteria cells was performed using 1 mL of bacteria growth medium in centrifuge tubes to optimize the conditions.In the normal detection procedure, capture was accomplished in 50 mL of bacterial growth medium to achieve a higher detection sensitivity.The process of magnetic capture is illustrated in Figure 1B.Initially, the O-1 pMBs were incubated with 1 mL of bacteria, followed by a slight shake at a certain temperature for a period of time.The bacteria-O-1 pMBs complexes were separated by magnetic separation rack for 3 min.Capture efficiency (CE) was calculated by counting the cells before (N 0 , cfu/mL) and after (N b , cfu/mL) magnetic separation using the following equation: To improve the capture efficiency of O-1 pMBs, the crucial capture conditions, including temperature of capture (4, 25, 37, 45, and 55°C), time of capture (from 10 to 50 min), and usage amount of O-1 pMBs (from 50 to 400 μg) were optimized (the concentration of Salmonella in the optimization process was 10 4 cfu/mL).The capture efficiency of O-1 pMBs under different conditions was discussed.

Evaluation of Capture Properties
Magnetic Capture Ability of Salmonella in Different Liquids Matrixes.The Salmonella Paratyphi B bacterial cell suspension was used to generate a 10-fold serial dilution ranging from 10 8 to 10 3 cfu/mL in 3 different liquid matrixes: H 2 O, NB medium, and milk.Next, 1 mL of Salmonella Paratyphi B was mixed with 200 μg of O-1 pMBs and then rotated at 37°C for 30 min.Following magnetic separation for 3 min, the supernatant was used for plate count to calculate capture efficiency.
Specificity and Stability of O-1 pMBs.Capture efficiency was tested on various strains, including different serotypes of Salmonella and various non-Salmonella, to evaluate specificity.In addition, the stability was evaluated by measuring the changes in capture efficiency of O-1 pMBs stored at 4°C according to different storage days (0, 7, 14, 21, 28, and 35 d).

Detection of Salmonella by Real-Time RPA Assay
Design and Screening of Primers and Probes for RPA Assay.The invA gene (GenBank accession number M90846) is a common target gene for Salmonella detection (Ma et al., 2023).Primers and probes were designed using the Primer and Probe Design Tool for RPA/RAA (https: / / ezassay .com/primer) and are listed in Supplemental Tables S1 and S2 (see Notes).First, the screening of capillary electrophoresis diagrams for RPA products was conducted to select the primer pairs that amplified target fragment.Then another screening of fluorescent RPA was performed to distinguish these combinations on obvious degree between positive and negative results.
DNA Extraction.Through the previous capture steps, the O-1 pMBs-Salmonella mixture was re-suspended in 200 μL of PBS (0.01 M, pH 7.4).The DNA was extracted using a simple heat treatment method at 100°C for 10 min.The suspension was centrifuged at 13,201 × g for 10 min.Subsequently, 2 μL of supernatant was used as the DNA template for real-time RPA.
Establishment of Fluorescent Probe-Based Real-Time RPA.The DNA template was used for the fluorescent probe-based real-time RPA.The limit of detection (LOD), sensitivity, and specificity of the method were also examined.Each 50-μL RPA response unit included 10 μM of forward primer (2 μL) and reverse primers (2 μL), 10 μM of probe (0.6 μL), reaction buffer (liquid A, 29.4 μL), and sterile water (11.5 μL).These materials were mixed before pipetting into the freeze-dried powder tubes.Next, 2 μL of extracted genomic DNA and 280 mM magnesium acetate solution (liquid B, 2.5 μL) were sequentially added into the reaction tubes.Conversely, a negative control contained 2 μL of ddH 2 O. Finally, the test tube was quickly placed into the QuantStudio 12K Flex at 39°C for 30 min to collect fluorescent signals every 30 s.The threshold time (TT) was determined by calculating the increase in fluorescence above a set threshold (550,000).A higher TT value indicated that a longer time was taken for amplification to reach the threshold (Gao et al., 2021).

Performance Evaluation of Real-Time RPA Method
Sensitivity and Specificity.We selected 14 Salmonella serotype strains, 6 non-Salmonella strains, and Felix O-1 phage to test the sensitivity (compatibility) and specificity (exclusivity) of this method.
Detection of Salmonella in Artificially Contaminated Milk.Milk samples were purchased from the local market.A total of 25 mL of negative milk samples were spiked with Salmonella at a final concentration of approximately 10 cfu/25 mL.Thereafter, the 25 mL of spiked milk samples were placed into 225 mL of BPW and incubated at 36 ± 1°C with shaking at 200 rpm.Next, 51 mL of the culture solution was removed every hour from 1 to 8 h.Finally, 50 mL of the culture solution was tested by fluorescent probe-based real-time RPA with treatment of O-1 pMBs, and meanwhile another 1 mL of the culture solution was used for colony count in Salmonella chromogenic medium.
Moreover, milk samples spiked with various concentrations of Salmonella (10 0 , 10 2 , and 10 3 cfu/25 mL) were also tested for the required pre-enrichment time.For comparison, the milk samples spiked with different concentrations of Salmonella (10 0 , 10 1 , 10 2 , and 10 3 cfu/25 mL) testing with RPA were also performed by real-time PCR partly according to the Chinese entry-exit inspection and quarantine industry standard SN/T1059.7-2010.The primers and TaqMan probe used for PCR were listed in Supplemental Table S3 (see Notes).

Optimization of Synthesis Conditions of O-1 pMBs
To verify the O-1 phage, the lytic spectrum of the phage seeds was tested through the plaque experiments before use.As shown in Supplemental Figure S2 (see Notes), plaques were obviously appeared on all the tested Salmonella standard strains plates.And for all the non-Salmonella strains, no plaques was found.This indicated that O-1 phage can develop an IMS method for Salmonella.
The O-1 pMBs were a crucial auxiliary elements in the RPA assay for the enrichment and separation of Salmonella from contaminated samples.To obtain the best quality O-1 pMBs for the capture experiment, the key conditions including the size of MBs, reaction ratio of phage to MBs, coupling time, and pH of coupling buffer were optimized.Compared with micron-scale MBs (2 μm), submicron-scale MBs (200 nm) coupled with O-1 phage exhibited a higher capture efficiency for different concentrations of Salmonella (Figure 2A).Consequently, the submicron scale MBs were selected to prepare the O-1 pMBs.As shown in Figure 2B, the capture efficiency of O-1 pMBs increased with the reaction ratio of phage to MBs increasing from 0.5 × 10 11 pfu/mg to 1.25 × 10 11 pfu/mg, but slightly decreased with further increase of the reaction ratio (1.25 × 10 11 − 2 × 10 11 pfu/mg).Due to excessive phage conjugation on MBs, steric hindrance formed to reduce the capture efficiency.The capture efficiency increased with an increase in the coupling time but remained almost unchanged after 3 h of coupling (Figure 2C).The results exhibited in Figure 2D demon-strated that O-1 pMBs kept a steady capture efficiency in pH of the coupling buffer (0.01 M PBS) from 5 to 9 due to the tolerance of phages to acid as previous reported (Huang et al., 2021), which meant the synthetic O-1 pMBs showed good matrix tolerance.In general, the synthesis of O-1 pMBs was realized in the test by using the optimized conditions, including optimized size of the MBs (200 nm), reaction ratio of the phages to MBs (1.25 × 10 11 pfu/mg), coupling time (3 h), and pH of coupling buffer (pH = 7).

Characterization of MBs, O-1 Phage, and O-1 pMBs
As can be seen in Figure 3A, the O-1 phage exhibited a white shadow shape because of varying degrees of aggregation.As shown in Figure 3B, the MBs were of a uniform size of approximately 200 nm and wrapped by a ring of coatings.Notably, some shadow-shaped phages were clearly observed around the MBs (Figure 3C) after coupling.In Figure 3D, MBs modified with O-1 phages could be found adhering onto the cell wall of Salmonella, indicating that the O-1 pMBs had favorable adhesion ability on Salmonella.In the top right corner of Figure 3A, 3B, and 3C, plaques were only observed in the plates

Optimization of O-1 pMBs Capture Conditions for Salmonella
Several key conditions were optimized to maximize the capture efficiency.The results in Figure 4A showed that both excessively low and high temperatures could affect the biological activity of phage, so a capture temperature of 37°C was regarded as optimal.As shown in Figure 4B, the capture efficiency of O-1 pMBs for Salmonella increased with prolonged capture time but remained unchanged or decreased slightly after 30 min of capture.The capture efficiency increased with an increasing of the dosage of O-1 pMBs (50-200 μg).However, we found no significant differences in the capture efficiency between 200 and 400 μg of O-1 pMBs (Figure 4C).Accordingly, 200 μg of O-1 pMBs was considered a suitable dosage for effective capture of Salmonella.The optimized conditions for maximizing capture efficiency were 30 min at 37°C with 200 μg of O-1 pMBs.

Capture Ability of O-1 pMBs
Capture Ability in Different Liquids Matrixes.The O-1 pMBs were tested under optimized conditions in various liquids to evaluate their tolerance of the matrix.
In Figure 5A, the capture efficiency in the low concentration of Salmonella could reach up to 85%.However, the capture efficiency decreased slightly at high concentrations of Salmonella (10 5 -10 8 cfu/mL).This result was also inconsistent with the experimental conclusions of previous IMS-based detection methods (Zengin et al., 2014;Wei et al., 2016;Zheng et al., 2016;Su et al., 2024).Compared with the other recognition elements, such as antibodies, phages are larger in size, have more complex structures unrelated to recognition and more notable steric hindrance effects.Thus, compared with antibody-immunomagnetic beads, O-1 pMBs have a weaker capturing ability, especially in systems with high target concentrations.In addition, compared with H 2 O, the capture efficiency of O-1 pMBs in NB medium and milk was somewhat reduced, but it still satisfied the requirements for capture of Salmonella in the milk samples.These results indicated that O-1 pMBs could overcome matrix interference to rapidly capture Salmonella from complex food samples.
Capture Ability of O-1 pMBs for Different Bacterial Strains.Generally, phages exhibit extremely high specificity for their host strains.In this study, the specificity of the O-1 pMBs was evaluated.The O-1 pMBs were able to capture all the tested serotypes of Salmonella (6/6), with the capture efficiency ranging from 65% to 90% (Figure 5B).However, the capture efficiency of the O-1 pMBs was considerably low for other bacterial strains, including Pseudomonas aeruginosa (PA), Escherichia  Screening of Primers and Probes for RPA Assay.The electrophoretic results depicted in Supplemental Figure S3 (see Notes) indicated that invA #3, invA#4, invA#5, invA#6, invA#8, invA#9, invA#10, invA#11, and invA#12 possessed the anticipated positive bands (prod-uct lengths) listed in Supplemental Table S1.Thus, these primers were chosen to design the corresponding probes (Supplemental Table S2) used in the fluorescence-based screening of the primer-probe combinations.During the fluorescent probe-based RPA screening (Supplemental Figure S4, see Notes), a notable fluorescence signal gap was observed particularly between the positive control and negative controls when using the primer invA#5 and probe 1 combination.Therefore, this combination was identified as the optimal primer-probe combination and was presented in Table 2.
To accurately determine the positive and negative results of the RPA fluorescence curves, the following criteria were applied: samples with a TT value less than 15 and a fluorescence signal exceeding 1,000,000 at 30 min were classified as positive, and those failing to meet one of parameters above were classified as negative.Details of the fluorescence curves are provided in Figures  6, 7, and 8, including the TT value, and the fluorescence intensity at 30 min is listed in Supplemental Table S4 (see Notes).
LOD for RPA Detection Method.The LOD of the RPA detection method was assessed in 3 distinct liquid matrixes: H 2 O, NB, and milk.The fluorescence amplification curves were presented in Figure 6.Based on the analysis of the TT value and fluorescence intensity at 30 min, the LOD for RPA detection in H 2 O, NB, and milk without treatment of O-1 pMBs were determined to be 2 × 10 4 cfu/mL, 5 × 10 4 cfu/mL, and 10 6 cfu/mL, respectively (Figure 6A, 6B, and 6C).In contrast, treatment with O-1 pMBs highly increased the LOD to 10 cfu/mL, 30 cfu/mL, and 50 cfu/mL, respectively (Figure 6D, 6E,  and 6F).This highlighted the crucial role of O-1 pMBs in increasing LOD by 3 to 4 orders of magnitude.Because matrix interference was greatly reduced and tested bacteria were significantly concentrated, the sensitivity of this method was greatly enhanced.
The proposed phage-based IMS combined with realtime RPA detection method was compared with other methods for detecting Salmonella (Supplemental Table S5, see Notes).Among the molecular biological detection methods, compared with isothermal amplification-based detection methods, the detection time of PCR-based detection methods is relatively longer.In comparison to such molecular biological detection methods, our detection method demonstrated better detection sensitivity.In addition, immunological recognition-based detection methods represented by lateral flow immunochromatography assay are simply performed, with a short detection time.However, for such immunological recognition-based detection methods, the number of serotypes that can be detected is always limited, and the detection sensitivity of these methods is usually not high.Therefore, the proposed phage-based IMS combined with a real-time RPA detection method can be used for the rapid, highly specific, sensitive, and accurate detection of Salmonella spp.. Sensitivity and Specificity.As shown in Figure 7, the fluorescence curves of all the tested 14 Salmonella strains (10 6 cfu/mL) were manifested as obvious positive, confirming that the method established in this study had high sensitivity (compatibility) of 100%.Meanwhile, all 7 non-Salmonella strains (10 6 cfu/mL; PA, EC1, SA, EC2, SF, and Fliex O-1 phage) were detected as negative, indicating that the detection method had high specificity (exclusivity) of 100%.
Detection of Salmonella in Milk.Furthermore, the applicability of the proposed assay to actual samples was evaluated by detecting milk samples spiked with an ultra-low amount of Salmonella (10 cfu/25 mL).The milk samples after different culture times (2, 3, 4, 6, and 8 h) were pre-enriched by O-1 pMBs and then detected using RPA.As shown in Figure 8, no discernible changes were visually observed in the pre-enrichment culture medium.However, the Salmonella-spiked milk sample cultured at 3 h could be detected as positive, with a TT value of 10.059 and a fluorescence intensity of more than 1,000,000.Simultaneously, the concentration of Salmonella in the milk samples was 45 cfu/mL after being cultured in BPW for 3 h.
Moreover, the pre-enrichment times of various Salmonella-spiked concentrations (10 0 , 10 2 , and 10 3 cfu/25 mL) in milk were also explored.As shown in Supplemental Table S6 (see Notes), the milk samples spiked with Salmonella of 10 0 , 10 2 , and 10 3 cfu/25 mL could be detected as positive after pre-enriched in BPW for 6 h, 2 h, and 0 h, respectively.In addition, the results of the real-time RPA were highly consistent with those of the real-time PCR, indicating the accuracy of the real-time RPA established in this study.
Thus, the RPA detection method, combined with the phage-based magnetic capture established in this study, greatly simplified detection procedure, reduced detection time, and showed high applicability in real samples.

CONCLUSIONS
In this study, we developed an O-1 pMBs-based capture strategy coupled with a real-time RPA method for the rapid, highly specific, and sensitive detection of Salmonella spp. in milk with an ultra-low detection limit.
The LOD of the RPA detection method was significantly increased by a magnitude of 3 to 4 orders with the participation of O-1 pMBs.Furthermore, the detection method showed a high sensitivity (compatibility, 14/14) and an ideal specificity (exclusivity, 7/7) for the tested Salmonella strains and non-Salmonella strains.In addition, the milk samples spiked with 10 cfu/25 mL of Salmonella were detected positive after being cultured in BPW for only 3 h.The proposed method provided a new choice as a universal platform for Salmonella detection, especially in complicated sample matrixes.

NOTES
This study was supported by the National Natural Science Foundation of China (82260644, 82003467; Beijing, China), General Project of Jiangxi Natural Science Foundation (20202BAB206066; Nanchang, China), Science and Technology Fund Plan of Jiangxi Provincial Health Commission (202130977; Nanchang, China), Department of Science and Technology of Jiangxi Province (20232BAB21605; Nanchang, China), and the Jiangxi Province High-Level and High-Skill Leading Personnel Training Project (2021;Nanchang, China).Special thanks to Xiaoqing He, Jian He, and their families for donating phage seeds and providing valuable guidance.Supplemental material for this article is available at https: / / data .mendeley.com/datasets/ trvt5fxyrm/ 1.No human or animal subjects were used, so this analysis did not require approval by an Institutional Animal Care and Use Committee or Institutional Review Board.The authors have not stated any conflicts of interest.

Figure 1 .
Figure 1.Schematic illustration of the O-1 phage-based magnetic capture method as an aid for real-time RPA detection of Salmonella in milk.(A) Preparation of O-1 pMBs, (B) magnetic capture and DNA extraction, (C) RPA detection, and (D) names of the elements.

Figure 2 .
Figure 2. Optimization of preparation conditions of O-1 pMBs.(A) Salmonella concentration and size of MBs.(B) Reaction ratio of phage to MBs. (C) Coupling time.(D) pH of the coupling buffer.Error bars indicate SD.
coli (EC1), Staphylococcus aureus (SA), Enterobacter cloacae (EC2), and Shigella flexneri (SF).The results showed the O-1 pMBs to be a Salmonella spp.-specific tool for the separation and enrichment of rapid and accurate assays of Salmonella.Stability of O-1 pMBs.The stability of O-1 pMBs was determined by monitoring changes in the capture efficiency of O-1 pMBs on different storage days.As shown in Figure 5C, the capture efficiency of O-1 pMBs decreased gradually as storage time increased.Although the capture efficiency decreased over time, the capture efficiency of O-1 pMBs stored after 28 d was still greater than 50%.

Figure 4 .
Figure 4. Optimization of conditions for the capture of Salmonella by O-1 pMBs.(A) Temperature of capture.(B) Time of capture.(C) Dosage of O-1 pMBs.Error bars indicate SD.Red boxes indicate optimal values.

Figure 5 .
Figure 5. Capture ability of O-1 pMBs.(A) Capture of Salmonella in ddH 2 O, NB medium, and milk.(B) Capture ability of O-1 pMBs for different bacterial strains (as listed in Table 1).(C) Stability of O-1 pMBs.Error bars indicate SD.

Figure 6 .
Figure 6.Limit of detection of the RPA detection method.Fluorescence amplification curve of real-time RPA detection for 10-fold serial dilution of Salmonella from 10 8 to 10 cfu/mL per RPA without treatment by O-1 pMBs in (A) H 2 O, (B) NB medium, and (C) milk, and with treatment by O-1 pMBs in (D) H 2 O, (E) NB medium, and (F) milk.
Figure 7. Sensitivity (A) and specificity (B) of real-time RPA detection method.Fluorescence amplification curves of the real-time RPA were tested by 10 6 cfu/mL of Salmonella and non-Salmonella.a.u.= arbitrary units.

Figure 8 .
Figure 8. Fluorescence amplification curve of the real-time RPA detection for the spiked milk samples after cultured at different times.a.u.= arbitrary units.

Table 1 .
Liu et al.: REAL-TIME SALMONELLA DETECTION IN MILK Sources and serial numbers of Salmonella strains