Development of a lateral flow dipstick test for the detection of 4 strains of Salmonella spp. in animal products and animal production environmental samples based on loop-mediated isothermal amplification

Objective This study aimed to develop loop-mediated isothermal amplification (LAMP) combined with lateral flow dipstick (LFD) and compare it with LAMP-AGE, polymerase chain reaction (PCR), and standard Salmonella culture as reference methods for detecting Salmonella contamination in animal products and animal production environmental samples. Methods The SalInvA01 primer, derived from the InvA gene and designed as a new probe for LFD detection, was used in developing this study. Adjusting for optimal conditions by temperature, time, and reagent concentration includes evaluating the specificity and limit of detection. The sampling of 120 animal product samples and 350 animal production environmental samples was determined by LAMP-LFD, comparing LAMP-AGE, PCR, and the culture method. Results Salmonella was amplified using optimal conditions for the LAMP reaction and a DNA probe for LFD at 63°C for 60 minutes. The specificity test revealed no cross-reactivity with other microorganisms. The limit of detection of LAMP-LFD in pure culture was 3×102 CFU/mL (6 CFU/reaction) and 9.01 pg/μL in genomic DNA. The limit of detection of the LAMP-LFD using artificially inoculated in minced chicken samples with 5 hours of pre-enrichment was 3.4×104 CFU/mL (680 CFU/reaction). For 120 animal product samples, Salmonella was detected by the culture method, LAMP-LFD, LAMP-AGE, and PCR in 10/120 (8.3%). In three hundred fifty animal production environmental samples, Salmonella was detected in 91/350 (26%) by the culture method, equivalent to the detection rates of LAMP-LFD and LAMP-AGE, while PCR achieved 86/350 (24.6%). When comparing sensitivity, specificity, positive predictive value, and accuracy, LAMP-LFD showed the best results at 100%, 95.7%, 86.3%, and 96.6%, respectively. For Kappa index of LAMP-LFD, indicated nearly perfect agreement with culture method. Conclusion The LAMP-LFD Salmonella detection, which used InvA gene, was highly specific, sensitive, and convenient for identifying Salmonella. Furthermore, this method could be used for Salmonella monitoring and primary screening in animal products and animal production environmental samples.


INTRODUCTION
Foodborne pathogens cause a wide range of diseases with major consequences for human health and the economy. A foodborne disease outbreak occurs when two or more cases of a similar illness occur as a result of the consumption common food. The World Health Organization reported that, poisoned foods and water affect over 600 million people globally, collected into a microcentrifuge tube and centrifuged for 3 min at 10,000 rpm/min. The pellet was washed twice in 1 mL of sterile deionized water after the supernatant was discarded. Subsequently, the pellet was suspended in 100 μL of deionized water and heated at 95°C for 15 min. The supernatant was centrifuged at 12,000 rpm/min for 3 min and transferred to a new tube for the DNA template in LAMP reaction.

Primers and DNA probes of LAMP-LFD
The InvA gene was chosen as the target gene for detecting Salmonella in this study, and it was used to design all of the primers. The SalInvA01 primer set shown in the LAMP and PCR tests was derived from Masphol et al [13]. Meanwhile, a pair of PCR primers were used to compare the LAMP assay. The LAMP assay was carried out in 25 μL of reaction mixture containing 1× ThermoPol Buffer, 0. , distilled water, and 2 μL of DNA template, as previously described in Masphol et al [13]. The LAMP reaction mixture was incubated for 60 min at 61°C before being terminated at 80°C for 5 min. The PCR assay was performed in accordance with Masphol et al [13]. In 20 μL of reaction mixture contained 1× PCR Buffer, 2.5 μM of each primer (SalInvA03-F3 and SalInvA03-B3), 2.5 mM dNTPs, 0.3 U of Taq DNA polymerase, 4 mM MgCl 2 , distilled water and 2 μL of DNA template. The PCR mixture was pre-denatured at 95°C for 5 min followed by 35 cycles of denaturation at 95°C for 30 seconds, annealing at 61°C for 30 seconds, and extension at 72°C for 30 seconds. The final-extension step was completed at 72°C for 5 min. LAMP and PCR products were electrophoresed in 2% agarose gel, stained with ethidium bromide and viewed under ultraviolet light.
For detection with the LAMP-LFD, the DNA probe was created by combining sequences from the SalInvA01-FIP and SalInvA01-BIP primers. The DIG was labeled using the 5′ end of the SalInvA01-FIP primer, giving rise to the new SalInvA01-DIG primer. Biotin was used to label the 5′ end of the DNA probe (Petty patent submission number 2003 002289). The InvA nucleotide sequences in Salmonella spp. (GenBank Accession No: DQ644633.1) were aligned with the DNA probe using the CLUSTALW software (https://www. genome.jp/tools-bin/clustalw). All of the LAMP primers or PCR primers and the probe of DNA were synthesized and labeled by (Ward Medic IDT, Bangkok, Thailand).

LAMP-LFD assay conditions
The LAMP condition used in conjunction with the hybrid-ization processes was modified in line with Masphol et al [13]. This reaction combination was put to the test by varying the amplification temperature, reaction time, and probe concentration. The amplification temperature was adjusted by amplifying the reaction at 61°C, 63°C, and 65°C for 60 min. The reaction time was varied between 30, 40, 50, and 60 min. The probe concentrations of 0.8, 1.6, 3.2, and 4.0 μM were added to the reaction mixture. SalInvA01-DIG and other components were also involved in this reaction. The SalInvA01-DIG for LAMP assay was carried out in 0.8 μM of DIG-labeled SalInvA01-FIP primer, 0.8 μM of SalInvA01-BIP primer, 0.4 μM of SalInvA01-F3 primer, 0.4 μM of SalInvA01-B3 primer, 0.8 M betaine, 1.4 mM each dNTPs, 4 mM MgSO 4 , 8 U of the Bst DNA polymerase, 1× ThermoPol Buffer and DNA template 2 μL by the boiling method. After hybridization, LAMP amplicons 8 μL of the LAMP product hybridized with the probe was moved into a new tube containing 100 μL of the Hybri Detect buffer, and the DNA LFD (Serve Science Co., Ltd., Bangkok, Thailand) was dipped in the mixture. The detection result for the test and control lines was detected on the LFD strip within 1 min. For the positive result, two bands were visible on the LFD (one band on the test line and one band on the control line). The sample showed only one band on the control line for the negative result. The test is not functioning correctly if no band is visible on the control line.

Specificity tests of the LAMP-LFD assays in pure culture
The SP of the LAMP assay was established using 22 isolated bacteria (Table 1). Salmonella and non-Salmonella strains were tested under optimum conditions using the InvA gene. The DNA template of each microbe was amplified by LAMP using SalInvA01-DIG primers. Then, the LAMP product was analyzed with 2% agarose gel electrophoresis and the DNA LFD.
Limits of detection of the LAMP-LFD and LAMP-AGE assays in pure culture S. Enteritidis was cultured in tryptic soy broth for 24 hours at 37°C, then diluted ten-fold in sterile water at concentrations ranging from 3×10 1 CFU/mL to 3×10 8 CFU/mL (confirmed by viable cells on XLD agar). The DNA template was prepared as described earlier and each of the LAMP products was examined using both agarose gel electrophoresis and the DNA LFD.

Limits of detection of the LAMP-LFD and LAMP-AGE assays in genomic DNA
The detection limits of the LAMP-LFD and LAMP-AGE assays were determined using genomic DNA from an overnight Salmonella pure culture in Tryptic Soy Broth. The genomic DNA was obtained by using the previously described simple boiling method after overnight growth. The Nanodrop 800 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) was used to measure the concentration of DNA at A260/280. And serially diluted 10-fold, each genomic DNA concentration has been used as a DNA template in this LAMP-LFD assay.
Artificial contamination detection limits in minced chicken with and without pre-enrichment S. Enteritidis was cultured in Tryptic Soy Broth at 37°C for 24 hours before being diluted in sterile water at a concentration range of 3×10 1 CFU/mL to 3×10 8 CFU/mL (confirmed by viable cells on XLD agar). The minced chicken was sterilized by autoclaving for use as decontaminated chicken meat. The artificially contaminated food was inoculated with 1 mL of S. Enteritidis at various concentrations into the samples. Then 25 g of artificially contaminated minced chicken sample was homogenized in 225 mL buffer peptone water and incubated at 37°C for 5 hours. After incubation, samples were analyzed at 0 hours (without pre-enrichment) and 5 hours (with pre-enrichment). The DNA of each sample was tested and analyzed using electrophoresis on a 2% agarose gel and the dipstick for DNA lateral flow.
Detection of Salmonella spp. in animal products and animal production environmental samples Sample collection: All the samples were collected within Nakhon Pathom province in Thailand. One hundred twenty animal product samples were sampled and purchased in retail packs from various local markets and supermarkets. Three hundred fifty animal production environmental samples were randomly collected from different farms and slaughterhouses. The samples were collected for determination between June 2020 and July 2021.
Sample collection in animal product samples: A total of 120 animal product samples were collected in separate clean plastic bags from 10 samples each of pork, minced pork, chicken meat, minced chicken, beef, chicken balls, pork balls, beef balls. And 20 samples each of chicken cut drumsticks and chicken cut fillets. All samples were tested immediately after purchase and transferred in an icebox to a laboratory.
Sample collection in animal production environmental samples: A total of 350 samples were obtained from different areas. In swine housing, 60 samples were collected from 20 swab samples each of the floor in a dry area, feed bunks, and the floor in a wet area. In poultry housing, 60 samples were collected from 20 swab samples each of the floor, feeder, and cage wire mesh. In milking parlor, 60 samples were collected from 20 swab samples each of the teat wipe towels, floor, and milking unit. In feed bulk truck, 60 samples were collected from 20 swab samples each of the feed conveyor, feed tank, and wheel. In beef slaughterhouse, 60 samples were collected from 20 swab samples each of the butcher table, knife, and floor. In poultry slaughterhouse, 50 samples were collected from 10 swab samples each of the chiller water, floor, conveyor belt, cutting board and knife. Sterilized cotton swabs (1 swab/sample) dipped in 10 mL of phosphate-buffered saline in sterile conical tubes with caps were used to swab and collect samples from the exposed surface area of 100 cm 2 . All samples were sent to the laboratory on ice and were analyzed.
Sample preparation and analysis: In animal product samples, 25 g of animal product samples were weighed and sliced into small pieces by using a sterile scalpel and forceps on a sterile petri dish. Then, it was placed in a sterile bag with 225 mL of buffer peptone water. For environmental samples from animal production, each animal production environmental sample swab was vortex mixed and homogenized to aid the release of organisms into the diluent. The diluent sample swab was added into a sterile bag containing 225 mL of buffer peptone water. Then, each animal product sample and the animal production environmental samples in a sterile bag with 225 mL of buffer peptone water were homogenized by using a food stomacher for 2 min before pre-enrichment step. For 5 hours; the DNA template of each sample was prepared as described earlier and analyzed using LAMP-AGE, LAMP-LFD, and PCR assays. For the culture method, it was continuously incubated at 37°C for 24 hours. Salmonella spp. was isolated according to an international standard culture method (Including WHO 2010 for detecting Salmonella) [14]. Following pre-enrichment, 1 mL aliquots of samples were added to 9 mL of Mueller Kauffman Tetrathionate Novobiocin Broth Base (HiMedia Laboratories, Mumbai, India), and incubated at 37°C for 24 hours, while 0.1 mL aliquots of samples were added to 9.9 mL Rappaport Vassiliadis Soya Broth (HiMedia Laboratories, India), and incubated at 42°C for 24 hours. A 10-μL loop full of selective enrichment was spread and incubated overnight at 37°C for each selective plating on Xylose Lysine Deoxycholate Agar (HiMedia Laboratories, India) and Brilliant Green Agar (Hi-Media Laboratories, India). Salmonella colonies were selected and transferred to tubes for biochemical identification from each selective agar.

Kappa statistics and accuracy analysis
The performance indicators for qualitative microbiological methods were calculated accuracy (AC), sensitivity (SE), specificity (SP), positive predictive value (PPV). The degree of agreement between the methods tested, as described by using Cohen's kappa (k), according to previously study [15] that described the kappa index was used: perfect agreement (1.00), nearly perfect agreement (0.81 to 0.99), substantial agreement (0.61 to 0.80), moderate agreement (0.41 to 0.60), fair agreement (0.21 to 0.40), slight agreement (0.01 to 0.20),

Optimization of LAMP-LFD reaction conditions
The optimized temperature of the SalInvA01-DIG primer was between 63°C and 65°C. The agarose gel electrophoresis results revealed that 63°C showed the cleanest bands with a ladder-like pattern of DNA. The optimal temperature for the LAMP-LFD reaction was chosen as 63°C in Figure 1A. In the LAMP-LFD study, LFD analysis revealed cherry-red bands in both the test and control lines at 63°C and 65°C that 63°C and both the test and control lines had red bands at 65°C, as well as a cherry-red band in the control line at  Figure 1B). The SalInvA01-DIG primer amplified the DNA ladder band at 40 min but dense and clear bands appeared at 60 min ( Figures 1C, 1D). The optimized LAMP reaction time for the SalInvA01-DIG primer was 60 min. For concentrations of the probe, the result showed that 0.8 μM of DNA probe provided the lowest concentration while still producing a cherry-red color on the test line of the LFD (Figures 1E, 1F).

The specificity of the LAMP-LFD reaction
The LAMP-AGE tests are used to determine the SP in Figure  2A, while the LAMP-LFD assays are shown in Figure 2B for

The detection limit of the LAMP-LFD and LAMP-AGE assays in genomic DNA
The initial genomic DNA concentration was measured at 901 ng/μL. To establish the lowest genomic DNA concentration required for the SalInvA01-DIG primer, the detection limit of genomic DNA was 9.01 pg/μL based on the InvA gene. The results are shown by LAMP-AGE in Figure 4A and LAMP-LFD in Figure 4B.

Limits of detection in artificially contaminated minced chicken with and without pre-enrichment
All autoclaved chicken meat samples containing 25 g in 225 mL of buffer peptone water were inoculated with S. Enteritidis, giving a positive result. The sample of non-inoculated chicken meat tested negative. The Salmonella detection limit in LAMP-LFD without pre-enrichment testing of artificially contaminated chicken meat samples was 3.4×10 7 CFU/mL ( Figure 5). After incubating the chicken meat samples at 37°C for 5 hours with pre-enrichment, the limit of detection increased to 3.4×10 4 CFU/mL or 680 CFU/reaction ( Figure 6).   The total of 350 animal production environmental samples were examined by 60 samples of swine housing, poultry housing, milking parlor, feed bulk truck, beef slaughterhouse, and 50 samples of poultry slaughterhouse. The results shown in Table 3 reveal that 91/350 (26%), 91/350 (26%), 91/350  three methods.

Kappa index and accuracy
The SE, SP, PPV, AC, and Kappa Index of LAMP-LFD, LAMP-AGE, PCR, compared with culture method were determined and showed in Table 4. The SE, SP, PPV, and AC for LAMP-LFD and LAMP-AGE were 100%, 95.9%, 87.2%, and 96.8%, respectively. Kappa Index was 0.90, indicating nearly perfect agreement with culture method. While, PCR method, The SE, SP, PPV, and AC were 95.0%, 95.7%, 85.7%, and 95.5%, respectively. Kappa Index was 0.88, indicating nearly perfect agreement with culture method.

DISCUSSION
The InvA gene is the most commonly utilized and successful gene for generating LAMP primers to detect Salmonella [12]. LAMP was originally presented for Salmonella detection in 2005. Later, modified Salmonella diagnosis methods based on various target genes were developed, such as stn, hilA, bcfD, and fimY [13]. LAMP products have been detected using a variety of approaches. The LFD assay provides good SP testing and visual result evaluation [12,16]. In the previous study, the researchers developed the LAMP technique in conjunction with the LFD technique to examine and report the results. In this study, the LAMP technique in combination with the LFD technique (LAMP-LFD) was selected based on the InvA gene and achieved the best primer set from previous studies (Masphol et al [13]). A modified LAMP product analysis protocol from gel electrophoresis and the addition of SYBR Green I to the LFD for detecting Salmonella spp. Also, new probe was designed to target the InvA. labeled DIG-with FIP primers and Biotin-labeled DNA probe, primer set and DNA probe labeled was used for the LAMP technique in conjunction with the LFD technique (LAMP-LFD) and focused on the LAMP-LFD technique to examine Salmonella contamination compared to other methods such as LAMP-AGE, PCR, and culture methods. For LAMP-LFD, a reaction temperature of 63°C for 60 min and an increased probe concentration of 0.8 μM within the LAMP reaction are required. Hybridization of the probe is used in most of the studies that follow the LAMP reaction [1,5]. Probes were hybridized along with the LAMP method, which was used in this study. It suggests that hybridizing a probe with the LAMP reaction shortens the step of the LAMP combination in the LFD reaction.
The LAMP-LFD method for Salmonella targeting the InvA gene was effectively developed in this study. The SalInvA01-DIG primer set has no cross-reactivity with target genes in non-Salmonella strains. The result shows that the LAMP-LFD assays based on the SalInvA01-DIG primer set of InvA gene are very efficient and specific for detecting Salmonella spp. According to the same protocol as in the previous study [13], the genomic DNA of each of the 19 microorganisms was tested using the fluorescent dyes (LAMP-SYBR green I) and agarose gel electrophoresis (LAMP-AGE  In this study, the limit of detection based on LAMP-LFD for Salmonella pure culture was 3×10 2 CFU/mL while genomic DNA was 9.01 pg/μL. In previous studies, the Salmonella genomic DNA detection limit targeting similar gene with our Table 2. Detection of Salmonella spp. in animal product using by culture method, LAMP-AGE, LAMP-AGE, and PCR 1) Positive  TP  FP  FN  TP  FP  FN  TP  FP   work in the InvA gene was 89 fg/μL [12]. For Salmonella targeting different gene with this work, in the hilA gene was 6.7 CFU/mL in pure culture and 13.5 fg/μL in genomic DNA [17]. And Salmonella cells targeting the siiA gene was 3.7 CFU/mL [18]. This result indicated that the efficiency of amplification may be affected by the primers used for different genes, or even the same target gene. It may lead to a diverse range of detection limits [17]. Moreover, Liu et al [12] reported the LAMP-LFD was repeatable and obtained identical findings at 8.9 pg/μL. This study found that LAMP with a hybridized probe has a better limit of detection than LAMP because the probe may increase the LAMP reaction by hybridizing to the LAMP product during the LAMP reaction. Similarly Kumvongpin et al [19] and Augkarawaritsawong et al [20] demonstrated increasing of the probe and limit of detection.

Culture method LAMP-LFD LAMP-AGE PCR
In the case of artificial contamination of the chicken sample, the Salmonella detection limits used by LAMP-LFD with pre-enrichment were 3.4×10 4 CFU/mL (680 CFU/reaction), and 3.4×10 7 CFU/mL without pre-enrichment. The detection limit of Salmonella spp. in pure culture was higher than Salmonella spp. in minced chicken samples. This is because the chicken samples may contain inhibitors found in the food sample such as fats, proteins, enzymes or other compounds that affect the detection limit of Salmonella spp. [13]. The enrichment step is a procedure to increase the SE of pathogen analysis by increasing the number of target microorganisms in low or susceptible conditions. The sample is then processed for further analysis using standard methods such as culture, PCR, and LAMP. This demonstrates that the enrichment procedure ensures enough SE to be correct and accurate [21]. Previous studies have discovered that LAMP-LFD has a detection limit equal to or lower than this study, based on the InvA gene: 8×10 4 CFU/25 g in artificially con-taminated chicken feed [12] and 1×10 4 CFU/mL in milk [22]. Another study discovered various genes at, 2.2 CFU/mL in powdered infant formula based on the siiA gene [18], and 1.44 CFU/mL in the hilA gene-based food matrix (raw milk, pork, beef, and chicken meat) [17]. According to several governmental regulations, live Salmonella must not be detectable in 25 g of food. The limit of detection (corresponds to 3×10 4 CFU/mL) does not meet the requirement for quantification of low concentration, but is sufficient for qualitative detection. Despite the effectiveness of the LAMP-LFD method, it still needs to be developed, popularized, and used in public health and food detection [23].
The LAMP-LFD, LAMP-AGE, PCR, and culture method were evaluated and compared for their AC, SE, SP, PPV, and Kappa index. LAMP-LFD and LAMP-AGE showed excellent SE, SP, PPV, AC, and Kappa Index more than PCR at 100%, 95.9%, 87.2%, and 96.8%, respectively. For Kappa index of LAMP-LFD was 0.90, indicated nearly perfect agreement with culture method. It was found that the result obtained from this study were better than previous [24]. In another study, Hyeon et al [25] reported that the label-based LFD with multiplex LAMP detected Salmonella spp. and Cronobacter spp. in powdered infant formula (n = 80) and showed 100% SP in this method. Among the animal products, the most positive for Salmonella was found to be minced pork (3/10) and chicken cut drumsticks (3/20) by using the four methods. Minced meat contains fat, which can protect Salmonella from the environment. Other food samples lack these qualities and their surface dries rapidly making them unsuitable for Salmonella [26]. For chicken cut drumsticks, which are bone-in products, a previous study suggested that internalized Salmonella through bone could be the cause of disease in these products [27]. Almost all detection methods are consistent, including LAMP-LFD, LAMP-AGE, PCR, and the culture method. As with LAMP-LFD, closer DNAbased detection methods like PCR and other comparable assays produce Salmonella detection results which are not always totally compatible with the culture method. Furthermore, PCR and LAMP can identify non-viable Salmonella as well as viable but non-culturable Salmonella as a reaction to poor environmental conditions [17]. This was also shown in this study, particularly for swabs acquired from the floor and cage wire mesh, as the samples were swabbed after being cleaned with powerful chemical detergents. The characteristics of methods based on nucleic acid amplification may also lead to false-positive examination results for the culture method. For example, knives and cutting boards produced false positive results. Because it is possible that DNA from Salmonella cells that have died will be amplified, which would explain the increased Salmonella molecular positive in culturenegative samples [28,29]. The enrichment has been incubated for 5 hours rather than overnight in the culture method, the Table 4. Kappa values and accuracy of agreement between alternative methods and culture method in animal products (n = 120) and animal production environmental samples (n = 350)  incubation time may be insufficient to ensure sub-lethally dead cells are revived or resurrected. Despite most reports of improved molecular testing methods and enhanced sample processing efficiency, enrichment is still required [30]. This is because low concentrations are possible and cause sublethal wounds to the target, as well as a large number of other bacteria. Non-culturable Salmonella or incubation times that are too short may reduce diagnosis SE, resulting in false negative results [31]. This is especially true for chilled water, which may be contaminated with background microbiota that would inhibit Salmonella growth. This LAMP-LFD assay has several benefits that make it valuable in conditions where time and resources are limited. First, the LAMP-LFD reaction does not necessitate the use of specialist equipment includes PCR machine and gel electrophoresis systems. Second, LAMP-LFD reactions could be carried out at an isothermal temperature of 60°C to 65°C. Alternatively, the PCR necessitates strict temperature control during pre-denaturing, denaturing, annealing, extension, and final-extension. Third, the LAMP-LFD assay is a diagnostic tool that saves time. It only takes 1.30 hours to detect LAMP and LFD using the two-step assay. The LAMP-AGE reaction is also a twostep assay that takes 2 hours to amplify by LAMP and analyze by agarose gel electrophoresis. Meanwhile, PCR reactions are usually doubled by LAMP-LFD. Fourth, the fluorometer is not required to monitor the fluorescent signal in the LAMP-LFD assay. LAMP results were seen with the naked eye on the dipsticks, making this platform easier to use than others.

Number of sample LAMP-LFD LAMP-AGE PCR
In conclusion, The LAMP-LFD assay was developed to detect Salmonella with accuracy, simplicity, and rapidity. The InvA gene was used to determine the specific primer combination. The LAMP-LFD assay could also be completed in 1.30 hours at a temperature of 63°C. This assay offers notable benefits. Only the primer set and a DNA probe are utilized to identify the amounts of target DNA. Meanwhile, the LAMP amplicons could produce visible lines on LFDs without the need for gel electrophoresis. Furthermore, SE studies found that the LAMP-LFD assay was capable of detecting as low a level as 9.01 pg/μL of genomic DNA of Salmonella. Furthermore, the LAMP-LFD assay detected Salmonella at higher rates in animal products and animal production environmental samples than the LAMP-AGE and PCR. This LAMP-LFD assay has high potential for use as a primary Salmonella screening assay in animal products and animal production environmental samples.