Multiplex touchdown Polymerase Chain Reaction for rapid detection of Salmonella enterica subsp. enterica serovars Enteritidis and Typhimurium in food

The Salmonella outbreak is one of the leading foodborne diseases in the world with increasing cases being reported annually. However, the current methods for Salmonella detection in foods are outdated, laborious and time - consuming. This necessitated developing a technique that is rapid for Salmonella detection in foods. Thus, the current study aimed to develop a multiplex touchdown PCR (m - TdPCR) protocol for rapid and simultaneous detection of Salmonella enterica subsp. enterica serovars Enteritidis and Typhimurium in foods. A two - phase m - TdPCR protocol was developed and optimized with primer pairs targeting the Salmonella enterica subsp. enterica (ST11/ST15 - 0.15 µM), serovars Enteritidis ( sdfI gene - 1.2 µM), Typhimurium ( fliC gene - 1.5 µM) and an internal amplification control (16S rRNA - 0.08 µM). It was found that the m - TdPCR protocol is highly sensitive detecting up to 1 ng of Salmonella DNA and its specificity was verified using the in - silico method. Furthermore, the developed m - TdPCR shows no non - specific PCR amplicons and is able to detect both S. enterica ser. Enteritidis and S. enterica ser. Typhimurium in real - time when tested against the artificially contaminated food samples at up to 10 - 3 dilutions. Therefore, the validated m - TdPCR protocol in this study can be used as a tool for rapid detection of S. enterica ser. Enteritidis and S. enterica ser. Typhimurium in food samples and this may significantly reduce any related foodborne incidences in future.


Introduction
The Salmonella enterica subsp.enterica is a clinically relevant bacteria that have caused a significant threat to public health and animal welfare in recent years.This subspecies is generally divided into typhoidal and nontyphoidal salmonellae, 99% of which can cause disease in humans and animals (Issenhuth-Jeanjean et al., 2014).The typhoid fever causing Salmonella enterica serovar Typhi and S. enterica ser.Paratyphi are largely restricted to the human host, non-zoonotic and are commonly witnessed in many underdeveloped countries due to poor sanitation practices (Havelaar et al., 2015).Contrarily, the non-typhoidal salmonellae (NTS) are zoonotic, foodborne, and found to be widespread in many industrialized countries, primarily due to the consumption of poultry and food products contaminated by these bacteria (Eng et al., 2015;Bangera et al., 2019).The two most prevalent NTS serotypes reported worldwide were S. enterica ser.Enteritidis (65%) and S. enterica ser.Typhimurium (12%) (Eng et al., 2015).In general, NTS contributes to a low mortality rate in humans, but due to its high morbidity and environmental widespread distributions, the prevalence of these bacteria in the global food chain has substantially increased the clinical and economic Enteritidis and S. enterica ser.Typhimurium was found not only in raw carcasses but also in many cooked and pre-cooked foods (Anukampa et al., 2017;Siala et al., 2017).Consequently, the food poisoning incidents documented due to the NTS were seen to be on the rise and an estimated 93.8 million cases occur globally each year, 80.3 million cases of which were linked to the foodborne spread resulting in 155 000 deaths annually (Majowicz et al., 2010).The infection caused by NTS primarily elicits gastroenteritis and bacteraemia; besides, it may also cause other extra-intestinal complications such as pneumonia, meningitis, urinary tract infections and endocarditis (Arii et al., 2002).As such, the median cost for treatments (2006 -2008) was estimated at US$ 468 for gastroenteritis, US$ 1465 for septicaemia and US$ 1918 for infections at other sites (Chen et al., 2012).
The medical cost for treating salmonellosis is ballooning and could implicate severe financial distress to both industrialised and developing countries.However, the current practice for Salmonella detection in foods is rather outdated, laborious, and time-consuming.Today, the gold standards in Salmonella identification in many laboratories are the conventional culture method using selective media, followed by biochemical characterization and molecular serotyping.These methods typically take more than five days to confirm the serovar, and at the critical outbreak, the infection would have deeply spread in the community and sustained further damage by the time the results were made available (Lee et al., 2015).Furthermore, the detection and identification of Salmonella in foods using the culture methods demand highly trained biomedical personnel to avoid reporting false negative or false positive (Paniel and Noguer, 2019).Conversely, many of the commercially available Salmonella detection kits such as serological assays and miniaturized biochemical assays are mostly automated and rapid but at the same time, they are too costly (Lee et al., 2015).Thus, this necessitated the need to develop a rapid, sensitive, and cost-effective method to detect and identify the NTS.Such criteria were fairly fulfilled by the polymerase chain reaction (PCR) techniques in pathogen detection and identification, particularly in the food samples (Lee et al., 2015).However, the method may still require further optimization, especially in the simultaneous detection of several bacteria and their targets in food samples.
Foods sold in the outlets and premises can be easily contaminated by various foodborne pathogens (Latchumaya et al., 2021).Since the PCR assay is highly sensitive, it can produce misleading non-specific amplification for the non-targeted bacteria even when the DNA is presented at a trace amount (Garibyan, and Avashia, 2013).On top of that, the development of multiplex PCR for simultaneous detection of several targets can lead to low amplification efficiency, low reproducibility, and succumb to self-inhibition due to the interference among the multiple primer pairs used in the PCR master mix (Xu and Shang, 2016).To prevent this, hot-start PCR and touchdown PCR can be used to improve the target specificity and avoid any non-specific PCR amplifications.Thus, this study aimed to develop a multiplex-touchdown PCR (m-TdPCR) which would be rapid and capable of simultaneously detecting S. enterica serovars Enteritidis and Typhimurium in food samples.The development of this protocol will be cost-effective and will add simplicity besides reducing the time for NTS detection in food samples.This may substantially prevent any significant impact of foodborne NTS outbreaks in future.

Bacterial strains and DNA isolation
Salmonella enterica subsp.enterica serovar Typhimurium (ATCC® 14028™) and Salmonella enterica subsp.enterica serovar Enteritidis (ATCC® 13076™) were obtained from the Department of Food Science Laboratory, Universiti Putra Malaysia.These strains were individually inoculated overnight in nutrient broth (Chem Soln, India) at 37.0±1.0°Cfor DNA isolation using the cell boiling method.The DNA isolation was conducted according to the method defined by Pui et al. (2011).

PCR primers and in-silico primers specificity
A total of four primer pairs (Vivantis, Malaysia) developed from previous studies were used to optimize the PCR amplification of targeted sequences as shown in Table 1; ST11/ ST15 primers (429 bp), sdfI gene (304 bp), fliC gene (620 bp), bacterial 16S rRNA universal primers (180 bp) (Soumet et al., 1999;Alvarez et al., 2004;Kang et al., 2017).Besides, these primer pairs were also verified for bacterial specificity using the insilico PCR amplification program (insilico.ehu.es)against the complete genome (including their plasmids) of Salmonella genus (n = 45) and other common and related Enterobacteriaceae families such as Escherichia spp.(n = 65), Yersinia spp.(n = 19) and Shigella spp.(n = 10).Up to two mismatches between the template DNA and primers were allowed.

Multiplex-touchdown PCR development
Both singleplex and multiplex PCR (mPCR) were optimized with the 2× Taq master mix (Vivantis, Malaysia) prior to developing the m-TdPCR protocol.Malaysia) was also optimized to the final concentration of 3.5 mM and 2.0 U, respectively.Additionally, 5% DMSO (Duchefa, Netherlands) and nuclease-free H 2 O were added to prepare a final mixture of 25 µL/reaction tube.The annealing temperature was optimized at 58°C.The DNA concentration was measured by Nanodrop one TM (Thermo Fisher, USA) and added to the master mix with at least 1 ng concentration per reaction mix.Two-phase m-TdPCR protocol (Table 2) for the rapid detection of S. enterica ser.Enteritidis and S. enterica ser.Typhimurium was adopted from a previous study (Korbie and Mattick, 2008) where the annealing temperature was further optimized as depicted in Table 2. Escherichia coli DH5α strain was used as the control.
The PCR mix was amplified in a thermocycler (Bio-Rad, USA), and the amplicons were analyzed in the gel documentation system (UVP, UK) after the gel electrophoresis [1.5%] step.

Bacterial artificial contamination and multiplextouchdown PCR bacterial detection in food sample
Bacterial cultures (S. enterica ser.Typhimurium and S. enterica ser.Enteritidis) were prepared at 10 8 CFU/ mL concentration and artificially contaminated with a protocol modified by Rohde et al. (2015).The contamination of food sample protocol was adapted from a previous study (Rohde et al., 2015).This study used cooked and autoclaved chicken nuggets (25 g) to represent food samples and it was inoculated with 10 8 CFU/mL S. enterica ser.Typhimurium and S. enterica ser.Enteritidis culture.The inoculum was incubated at room temperature for 4 hrs and then added to 225 mL of Buffered Peptone Water (BPW) (Chemsoln, India) for homogenization.Serial dilution (10 -2 and 10 -3 ) were prepared in triplicates, and these tubes were then subjected to DNA isolation using the cell boiling method.The resulting DNA was used as a template in the m-TdPCR amplification.

Results and discussion
In this study, four primers pairs were utilized to develop an m-TdPCR protocol for simultaneous and rapid detection of Salmonella enterica subspecies enterica (subsp.I), S. enterica ser.Enteritidis and S. enterica ser.Typhimurium from food samples.The universal total bacterial 16S rRNA primers were used as The incorporation of the IAC is important to avoid falsenegative results in the enumeration of bacteria from the food samples using the PCR method (Copin et al., 2012).
Table 3 shows the in-silico PCR outcome against ST11/ ST15 primer pair, three out of forty-five Salmonella strains [S.bongori (n = 2) and S. enterica subsp.arizonae (n = 1)] did not show any PCR band.No PCR band was also observed when the same primer pair was tested against the Escherichia spp.(n = 65), Yersinia spp.(n = 19), and Shigella spp.(n = 10).The failure of the ST11/ST15 primers to detect S. bongori and S. enterica subsp.arizonae is much anticipated as these oligo primers has been shown to lack specificity for Salmonella enterica subspecies II, III, IV, V and VI earlier (Aabo et al., 1993).The primers to amplify sdfI was only visible for the S. enterica ser.Enteritidis (n = 1) and was not present in any other Salmonella strains, Escherichia spp., Yersinia spp.and Shigella spp.A similar outcome was also observed for the primers targeted fliC with the band only visible for the S. enterica ser.Typhimurium (n = 13), but no crossreactivity was noticed with closely related bacterial clades.This theoretically warrants that the primers ST11/ ST15, primers for sdfI and the primers for fliC are specific to the S. enterica subspecies enterica (subsp.I), S. enterica ser.Enteritidis and S. enterica ser.Typhimurium, respectively.The in-silico PCR and the primers validation methods are highly specific, and it is a validated technique.A study conducted by Goay et al. (2016), looked into the S. enterica ser.Typhi gene amplification compared the study's outcome using the in -silico and in vitro PCR methods.When a whole genome of bacteria was made available (n = 5), the predicted insilico PCR showed 100% specificity, and a similar outcome was obtained using in vitro PCR method.However, this outcome was achieved if only the whole bacterial genome were available for the in-silico test (Goay et al., 2016).Thus, the amplification specificity testing using the in-silico PCR technique is equally valid compared to the in vitro PCR technique.On top of that, the in-silico PCR technique is also cheap, labour effortless and safe as it prevents any risk of contracting harmful pathogens (van Weezep et al., 2019).Assessing the cross-reactivity among the primers used in this study against other important Enterobacteriaceae is vital.Bacteria such as E. coli, Yersinia spp.and Shigella spp.are closely related to the Salmonella genus, and they not only possess similar clinical symptoms but are also found to have high similarity in their genome and their G + C content (Forsdyke, 1996).As such, careful scrutinization for the cross-reactivity within the primer pairs in the developed m-TdPCR protocol is necessary for simultaneous and rapid detection of S. enterica., S. enterica ser.Enteritidis and S. enterica ser.Typhimurium to prevent any false-negative or falsepositive outcomes.
In the singleplex PCR amplification, the Salmonella bacteria were amplified against all four primer pairs (Figure 1a).For the ST11/ ST15 and the Universal primers 16S rRNA, both S. enterica ser.Enteritidis and S. enterica ser.Typhimurium amplicons were observed.Alternatively, for the primers targeting the sdfI and fliC, the amplicons were only noticed for the S. enterica ser.Enteritidis and S. enterica ser.Typhimurium, respectively, with no cross-reactivity.Conversely, the control E. coli DH5α amplification occurred only for the universal bacterial 16S rRNA primers.Besides that, when a multiplex PCR protocol was developed with the same four primer pairs, all amplicons with respective size bands were observed in a single PCR reaction (Figure 1b).The minimal DNA concentration required was 1 ng in this protocol.This was optimized in the sensitivity experiment where mPCR amplification with 0.1 ng of DNA concentration failed to obtain results for all four amplicons.A study conducted by Wang et al. (2019) to detect six pathogenic bacteria in poultry using mPCR showed that the developed protocol's sensitivity FULL PAPER (detection limit) was mostly up to 0.5 -1 ng (Wang et al., 2019).Moreover, a typical PCR may require a 3×10 5 DNA template molecule which is reasonable to start with 1 ng of bacterial genomic DNA (Møller, 2016).
However, the protocol developed in mPCR presented a non-specific amplicon at approximately 100 bp even after the annealing temperature and other PCR reagents in the master mix were optimized.The non-specific amplicon in the mPCR is expected as impurities or inhibitors in the PCR reaction mixture due to the cell boiling method during the DNA extraction in this study.Nevertheless, the cell-boiling DNA extraction method is more cost-effective and less laborious compared to the phenol/chloroform or DNA extraction kits (Oliveira et al., 2014).Salmonella genus generally has high G + C regions in their genome, contributing to many inter and intra hairpin loops.This can generate many complex secondary structure formations, which may inhibit target amplification in the PCR reactions (Musso et al., 2006).Higher G + C rich regions could also resist template denaturation and lead to improper annealing conditions, thus preventing primers annealing at the desired site (Hubé et al., 2005).This may subsequently add to nonspecific target amplification.This issue was addressed when the two-phase m-TdPCR was introduced where the non-specific band was not observed, therefore increasing the accuracy and specificity of the target.This could be due to the wide annealing temperature range employed in the m-TdPCR that enhances the specificity of the amplicon formation (Korbie and Mattick, 2008).
Besides that, all three dilution tubes (10 -1 , 10 -2 and 10 ) in the artificially contaminated food experiment were subjected to the m-TdPCR amplification, respective amplicon bands for both S. enterica ser.Enteritidis (ST11/ ST15, sdfI gene and universal total bacteria 16S rRNA) and S. enterica ser.Typhimurium (ST11/ ST15, fliC gene and universal total bacteria 16S rRNA) were evident (Figure 2).This has demonstrated that the m-TdPCR protocol can detect and identify the presence of S. enterica ser.Enteritidis and S. enterica ser.Typhimurium in the food samples with no unspecific amplicons observed.Several studies in the past had successfully developed the m-TdPCR protocol for various pathogen detections in food products.For instance, a study conducted by Moezi et al. (2019) showed that the m-TdPCR has better sensitivity and specificity for E. coli O157:H7, L. monocytogenes, Staphylococcus aureus, and S. enterica ser.Enteritidis detection in the raw milk samples.The same study also had shown that the m-TdPCR protocol had successfully  FULL PAPER eliminated any non-specific bands that appeared in the mPCR besides delivering comparatively similar pathogen detection outcomes to the conventional culture method (Moezi et al., 2019).As the pathogen detection using the conventional method may take up to 2 -5 days, the currently developed m-TdPCR protocol was shown rapid (detection within 24 hrs), sensitive and specific in detecting S. enterica ser.Enteritidis and S. enterica ser.Typhimurium in the food samples.

Conclusion
The developed multiplex touchdown PCR protocol was shown to have high sensitivity (1 ng of DNA template added to 25 µl PCR mix) in detecting S. enterica ser.Enteritidis and S. enterica ser.Typhimurium.Besides, the in-silico results had proved that the PCR primers used in this protocol were highly specific as no cross-reactivity was observed within the Salmonella genus and other closely related bacterial clades.On top of that, when the food samples were artificially contaminated with S. enterica ser.Enteritidis and S. enterica ser.Typhimurium, the m-TdPCR was able to detect the bacteria up to 10 -3 dilutions.Additionally, the inclusion of the internal amplification control (IAC) in the protocol can rule out any possible false-negative result using this m-TdPCR protocol.Thus, the developed m-TdPCR protocol is suitable for detecting NTS such as S. enterica ser.Enteritidis and S. enterica ser.Typhimurium in the food samples as it is rapid, sensitive, and specific.

Figure 1 .
Figure 1.Detection of Salmonella enterica subsp., S. enterica ser.Enteritidis and S. enterica ser.Typhimurium using PCR.An unspecific band (~100 bp) was noticed in the mPCR but not in the Td-mPCR protocol.

Figure 2 .
Figure 2. Gel electrophoresis analysis of the m-TdPCR for artificially contaminated food experiment.Artificially contaminated home cooked nuggets with S. enterica ser.Enteritidis (a) and S. enterica ser.Typhimurium (b) able to be detected in all triplet dilution tubes (10 -1 , 10 -2 and 10 -3 ) during the m-TdPCR amplification.The figure shown is a representation of PCR results in triplicates.M: 100 bp DNA Ladder (Thermo Scientific, US); N: Negative Control (dH 2 O as template).

Table 2 .
Two phase m-TdPCR protocol for a rapid detection of Salmonella enterica subsp., S. enterica ser.Enteritidis and S. enterica ser.Typhimurium in food.