International Journal of Food Microbiology 166 (2013) 213–218

Contents lists available at ScienceDirect

International Journal of Food Microbiology

j ourna l homepage: www.e lsev ie r .com/ locate / i j foodmicro

Multiplex TaqMan® detection of pathogenic and multi-drugresistant Salmonella

Prashant Singh, Azlin Mustapha ⁎Food Science Program, University of Missouri, Columbia, MO, USA

⁎ Corresponding author at: Food Science Program, 256Wof Missouri, Columbia, MO 65211, USA. Tel.: +1 573 882 2

E-mail address: MustaphaA@missouri.edu (A. Mustap

0168-1605/$ – see front matter © 2013 Elsevier B.V. All rhttp://dx.doi.org/10.1016/j.ijfoodmicro.2013.07.023

a b s t r a c t

a r t i c l e i n f o

Article history:Received 24 February 2013Received in revised form 26 June 2013Accepted 25 July 2013Available online 2 August 2013

Keywords:Antibiotic resistanceSalmonellaMultiplex PCR

Overuse of antibiotics in the medical and animal industries is one of the major causes for the development ofmulti-drug-resistant (MDR) food pathogens that are often difficult to treat. In the past few years, higher inci-dences of outbreaks caused by MDR Salmonella have been increasingly documented. The objective of thisstudy was to develop a rapidmultiplex real-time polymerase chain reaction (PCR) assay for simultaneous detec-tion of pathogenic and MDR Salmonella spp. A multiplex TaqMan®real-time PCR was designed by targeting theinvasin virulence gene (invA), and four commonly found antibiotic resistance genes, viz. ampicillin, chloram-phenicol, streptomycin and tetracycline. To avoid false negative results and to increase the reliability of theassay, an internal amplification control (IAC) was added which was detected using a locked nucleic acid (LNA)probe. In serially diluted (5 ng–50 fg) DNA samples, the assay was able to detect 100 genomic equivalents ofSalmonella, while in a multiplex format, the sensitivity was 1000 genomic equivalents. The assay performedequally well on artificially contaminated samples of beef trim, ground beef of different fat contents (73:27,80:20, 85:15 and 93:7), chicken rinse, ground chicken, ground turkey, egg, spinach and tomato.While the detectionlimit for un-enriched inoculated food samples was 104 CFU/g, this was improved to 10 CFU/g after a 12-henrichment in buffered peptone water, with 100% reproducibility. The multiplex real-time assay developedin this study can be used as a valuable tool to detect MDR virulent Salmonella, thus enhancing the safety offood.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Salmonella outbreaks in the U.S. are generally associated withcontaminated products of animal origin, such as ground beef, turkeyand eggs (CDC, 2010). Salmonella leads to the second highest numberof foodborne infections in the United States, making it a pathogen ofmajor concern. On top of that, the emergence of antibiotic resistantstrains of Salmonella is an even greater challenge. In the last fewdecades, widespread use of antibiotics in the medical field and animalhusbandry has been thought to lead to an emergence of antibiotic resis-tant strains of Salmonella that were responsible for some of the majorSalmonella outbreaks worldwide. Antimicrobials administered to dairyor meat cattle frequently belong to the same class as those used inhumans (Aarestrup et al., 2008). Alarmingly, Salmonella Enteritidisand Salmonella Typhimurium isolated from animal feed have beenreported to be highly resistant to antimicrobials (Li et al., 2012). Inaddition to resistance towards commonly used antibiotics, such asampicillin, streptomycin, sulfonamides, trimethoprim and tetracycline,there have been several reports of resistance to the cephalosporinclass of antibiotics which were recently approved for the treatment of

CSWing, Eckles Hall, University649; fax: +1 573 884 7964.ha).

ights reserved.

severe cases of Salmonella infection in humans (Sjölund-Karlssonet al., 2010). Antibiotic resistant Salmonella can cause severe infectionsthat could lead to prolonged hospitalizations or even death. Themultipledrug resistant (MDR) S. Typhimurium DT104 is known to show resis-tance to ampicillin, chloramphenicol, streptomycin, sulfonamide, andtetracycline. In the U.S., human isolates of S. Typhimurium DT104(R-type ACSSuT) increased from 9% in 1990 to 32% in 1996 whenthis strain led to the first major outbreak of human salmonellosis(Akkina et al., 1999). Apart from an increasing rate of antibiotic resis-tance among Salmonella, there has also been an increase in the rate ofantibiotic resistance in the other members of the Enterobacteriaceaefamily. These MDR members of Enterobacteriaceae have becomeone of the biggest challenges among hospital-acquired infections(Cohen, 2013). The Study for Monitoring Antimicrobial ResistanceTrends (SMART) has also shown an increase in cephalosporin resis-tance in Escherichia coli and Klebsiella pneumoniae (Hoban et al.,2012), making this a critical problem for vulnerable patients.

Multidrug resistance in bacteria is frequently linked to genetic ele-ments called integrons. These genetic elements play an important rolein the dissemination of a number of antibiotic resistance genes. Thereare three major classes of integrons, of which the class I integrons aremost commonly found and extensively studied in Salmonella serovarsof animal origin (Molla et al., 2007; Zhao et al., 2007). Integrons arenatural expression systems that allow the insertion and expression of

214 P. Singh, A. Mustapha / International Journal of Food Microbiology 166 (2013) 213–218

various antibiotic resistance genes by their ownpromoters, thusmakingthe inserted gene functional. Integron 1 has been shown to harbor up to10 different antibiotic resistance gene cassettes. These gene cassetteswere reported to confer resistance towards aminoglycosides, ampicillin,rifampicin, spectinomycin, streptothricin, sulfasulfonamides, tetracy-cline and trimethoprim (White et al., 2001). Apart from integrons, allpathogenic strains of Salmonella have also been reported to possessthe invA virulence gene. This gene is located on the Salmonella path-ogenicity island 1 (SPI-1) of Salmonella spp., which encodes Type IIIsecretion system proteins. This, along with other genes regulatesthe expression of other virulence genes present on SPI-1 involvedin pathogenicity (Boyd et al., 1997).

Real-time PCR methods have been increasingly used as a rapid andsensitive technique for the detection of food pathogens. With the useof TaqMan® probes, real-time PCR offers an advantage of rapid, sen-sitive and specific detection of food pathogens, while avoiding cross-contamination from other closely related bacteria (Malorny et al.,2007). Incorporation of locked nucleic acids (LNA) molecules in theprobe has further helped to increase the sensitivity of these assays.LNA are nucleic acid analogs containing a locked bicyclic furanoseunit in an RNA-mimicking sugar conformation. These moleculesallow for better base stacking and, therefore, show a higher stabilityand affinity towards LNA, DNA and RNA targets. LNA residues confera relative degree of resistance towards exo- and endonucleases(Kumar et al., 1998; Whitcombe et al., 1999). LNA modification in a5′nuclease probe leads to a significant increase in the melting tempera-ture (Tm) of the probe, allowing for designing of shorter fluorescentprobes with higher specificity (Tolstrup et al., 2003).

To date, limited studies have been done on rapid detection ofantibiotic resistant Salmonella. Chiu et al. (2006) described a multi-plex PCR method for the detection of antibiotic resistant strains ofS. Typhimurium. Khan et al. (2011) developed intercalating dye-basedmultiplex real-time PCR for the detection of bacteria showing resistancetowards quaternary ammonium compounds, tetracycline, gentamicinand chloramphenicol. Because of the importance of MDR Salmonella infood safety, a rapid and specific method for simultaneous detectionof pathogenic and antibiotic resistant strains of Salmonella is muchneeded. The aim of this study, therefore, was to develop a novel mul-tiplex TaqMan® real-time PCR based method for the detection ofpathogenic as well as antibiotic resistant strains of Salmonella usinga LNA dual labeled probe for increased reliability.

2. Material and methods

2.1. Bacterial strains

Standard cultures of S. Typhimurium 14028, S. Typhimurium 13311and S. Enteritidis 13076 were procured from ATCC (Rockville, MD),S. Typhimurium LJH 666, Salmonella Newport 692, Salmonella Agona LJH1132, S. Agona LJH 1122 and S. Typhimurium var Copenhagen LJH 788were generously provided by Dr. Linda Harris, University of California-Davis. Cultures were grown at 37 °C in Tryptic Soy broth supplementedwith 0.5% yeast extract (TSBY) (Difco Labs., MD). Twenty-two S.Enteritidis isolates were from the University of Missouri Food Micro-biology Lab culture collection.

2.2. Phenotypic and genotypic determination of antibiotic resistance

All Salmonella cultures were screened for antibiotic resistance by thestandard agar disk diffusion assay according to the criteria suggested bythe Clinical Laboratory Standards Institute (Watt et al., 2008), with thefollowing quality control strains: E. coli ATCC 25922 and Staphylococcusaureus 25923. Isolates were also tested for the presence of ampicillin,chloramphenicol, streptomycin and tetracycline resistance genes byconventional PCR, as described by Chiu et al. (2006).

2.3. Bacterial DNA isolation

GenomicDNAwas isolated fromactively growing cells in TSBY broth.One milliliter of an overnight culture was centrifuged at 6000 ×g andthe obtained cell pellet was resuspended in 100 μL of sterile MilliQwater. The DNA from the obtained cell pellet was isolated using aGeneration Capture Column Kit (Qiagen, Hilden, Germany). Concentra-tions andpurity of thus obtainedDNAwere determinedusing aNanodrop2000 spectrophotometer (Thermo Fisher Scientific, Wilmington).

2.4. Primers and probes

Primers for the antibiotic resistance genes, chloramphenicol (floR),ampicillin (pse), streptomycin (aadA2) and tetracycline (tetG) weredesigned on GenBank accession No. AY339985.2 and GU987052, as de-scribed by Chiu et al. (2006). The invA gene primer was modified usingGenBank accession No U43239.1. Primers and TaqMan® probes for theamplification of antibiotic resistance and virulence geneswere designedusing Primer Express software version 3.0 (Applied Biosystems, FosterCity, CA) and Primer3 software (http://primer3.wi.mit.edu) (Rozenand Skaletsky, 2000). The specificity of the designed primerswas testedusing NCBI's Primer BLAST tool. The melting temperature (Tm) of theLNA probe was estimated using the Exiqon Tm prediction program(Tolstrup et al., 2003). Fluorophores for labeling the probeswere selectedbased on two criteria described by Molenkamp et al. (2007). First, eachfluorophore was detected by separate detectors and secondly, excitationand emission wavelengths of each fluorophore were separated by30 nm. This approach allowedminimal cross-talk between the differentdetectors and provided a platform for quantitative detection ofmultipletargets in the sample.

Dual-labeled probes, minor groove binding (MGB) TaqMan® probesand LNA-dual labeled probe were commercially synthesized by IDT(Coraville, IA), Applied Biosystems and Sigma Life Science (SigmaAldrich, TX), respectively. PCR amplification was performed in a 20 μLreaction volume using 2× TaqMan® Universal PCR Master Mix withno AmpErase UNG, 20 ng of genomic DNA, and the respective primerconcentrations, as shown in Table 1, on a 7500 Real-Time PCR System(Applied Biosystems). A three-step protocol was used for DNA amplifi-cation: initial denaturation at 95 °C for 10 min, and 40 cycles of 95 °Cfor 30 s, 58 °C for 45 s and 72 °C for 45 s. An amplification plot wasconstructed between ΔRn versus cycles. Because of the limited numberof channels available on the ABI 7500 instrument, this multiplex assaywas designed in two sets: the first set amplifying the invA virulencegene, aadA2 streptomycin resistance gene, blaPSE ampicillin resistancegene and an internal amplification control (IAC); and the secondtargeting floR chloramphenicol resistance gene, tetG tetracyclineresistance gene and an IAC.

2.5. Specificity of the invasin primer

The specificity of the modified invasin primer set was vali-dated using 73 bacterial strains spread across 20 genera and31 species of Gram-negative and Gram-positive bacteria, includ-ing Acinetobacter, Alcaligenes, Bacillus, Bifidobacterium, Citrobacter,Clostridium, Enterobacter, Escherichia, Hafnia, Klebsiella, Lactobacillus,Leuconostoc, Listeria, Micrococcus, Pseudomonas, Salmonella, Serratia,Shigella, Staphylococcus and Yersinia.

2.6. Sensitivity of the real-time PCR assay

To determine the sensitivity of the developed multiplex assay,genomic DNA from a MDR isolate, S. Typhimurium DT104 strain LJH788, was isolated using Generation Capture Column Kit (Qiagen). DNAstock solution (50 ng/μL) was ten-fold serially diluted from 5 ng/reaction to 50 fg/reaction of DNA. Salmonella genomic copy numberequivalents were calculated based on the S. Typhimurium genome

Table 1Primers and probes for real-time PCR assays.

Gene Primer sequence (5′–3′) Product size (bp) Primer conc. Reference

Str F CAGCCA(T/c)GATCGACATTGATCT 66 0.16 μMStr R CCAAGGCAACGCTATGTTCTC This studyStr-Probe VIC-CTGCTTACAAAAGC-NFQ 0.17 μMTet F CGGTACTTCTGGCTTCTCTT 148 0.22 μMTet R GAATCGGCAATGGTTGAG This studyTet-Probe FAM-CAGGAGCCG/ZEN/CAGTCGATTACACG-Iowa Black® FQ 0.1 μMChl F GGCAGGCGATATTCATTACT 197 0.12 μMChl R CGAGAAGAAGACGAAGAAGG This studyChl-Probe VIC-CTAAAGCCGACAGTGTA-NFQ 0.12 μMAmp F GATTTGGTGCTCGGAGTATT 92 0.12 μMAmp R CATTGAAGCCTGTGTTTGAG This studyAmp-probe NED-CTTGATGCTCACTCCA-NFQ 0.1 μMInv F CCAGTTTATCGTTATTACCAAAGG 200 0.15 μM This studyInv R ATCGCACCGTCAAAGGA(A/g)CSal-invA-SO-WH probe FAM-CTCTGGATGGTATGCCCGGTAAACA-BHQ1 0.2 μM Anderson et al. (2011)IAC-probe Cy5-CTC[+T]TGCTC[+T][+C]T[+T]C[+C][+T]-BHQ3 0.1 μM This studyIAC TCGCCTGTATTTCCAGTTTATCGTTATTACCAAAGGTTTGATCTCTT

GCTCTCTTCCTACCGTTCAAGYTCCTTTGACGGTGCGATGATTAACGGCAG0.1 pg This study

215P. Singh, A. Mustapha / International Journal of Food Microbiology 166 (2013) 213–218

size. One S. Typhimurium double stranded genomic DNA weighs about5.0 fg. Thus, one genomic equivalent of Salmonella DNA is equal to5.0 fg of Salmonella genomicDNA (McClelland et al., 2001). Each seriallydiluted DNA was used in duplicates in the real-time PCR for the con-struction of standard curves and determination of the assay's amplifica-tion performance. The reaction was performed in duplex format withthe specific gene targets and a specially designed IAC (Table 1). Becausethe amount of PCR product is doubled at the end of each cycle in a PCRreaction, the PCR efficiency should ideally be 100%. PCR efficiency is cal-culated from the slope of the standard curve using the formula:

E ¼ 10 −1=slopeð Þ−1:

A PCR reaction is considered good if the reaction efficiency rangesbetween 90% and 110%, which corresponds to a standard curve slopebetween −3.58 and −3.10. In this study, the PCR efficiency of eachgene was calculated using StepOne® Software V2.2.2. Microsoft Excelsoftware was used to compute the regression coefficients and standarddeviations for the standard curves.

2.7. Preparation of artificially spiked samples

Ground beef of different fat contents (73:27, 80:20, 85:15 and 93:7),whole chicken, ground chicken, ground turkey, raw egg, spinach andtomato were purchased from local supermarkets. Beef trimwas obtainedfrom the University of Missouri Meat Laboratory. These food productswere tested for the presence of Salmonella using a standard culture-based protocol (Andrews et al., 2007). Chicken rinse was prepared asdescribed by Wang and Mustapha (2010). Aerobic plate count of eachfood sample was estimated using standard serial dilution and the pourplate method on plate count agar (Difco Labs).

Twenty-five grams of ground beefwith the highest fat content (27%)was mixed with 225 mL of buffered peptone water. The homogenatewas inoculated with 10 to 107 CFU/g of MDR isolate S. Typhimuriumvar Copenhagen LJH 788 and stomached for 2 min. Samples were usedfor construction of a standard curve and determination of the sensitivityof the assay. Upon inoculation of cells at 0 h, 2 mL of broth was with-drawn from each stomacher bag and briefly centrifuged (1120 ×g for1 min) to separate out beef particles from the broth. The supernatantwas transferred to a fresh centrifuge tube and used for isolatingDNA using PrepMan® Ultra Sample Preparation Reagent (AppliedBiosystems) according to the manufacturer's instructions. The isolatedDNA was used for real-time PCR to construct a standard curve for eachtarget in duplex format with the IAC.

Twenty-five grams each of beef trim, ground beef of different fatcontents (73:27, 80:20, 85:15 and 93:7), chicken, ground chicken,

ground turkey, raw egg, spinach and tomato were placed in filteredstomacher bags (Fisherbrand, Houston, TX) and mixed with 225 mLbuffered peptone water (BPW). BPW has been reported to be betterthan universal pre-enrichment broth for the enrichment of food samplescontaminated with Salmonella (Hein et al., 2006). The samples wereinoculated with S. Typhimurium LJH 788 at different concentrationsranging from 1 to 105 CFU/mL. One negative (un-inoculated) controlfor each sample was also included in the study. To determine the enrich-ment time required for detection of food samples contaminatedwith lowlevels of Salmonella, 1 mL of spiked samples was withdrawn after inter-vals of 0, 12 and 24 h. DNAwas isolated from the enriched sample usingPrepMan® Ultra Sample Preparation Reagent, and used for detection ofSalmonella by TaqMan® real-time PCR.

3. Results

3.1. Antibiotic resistance of tested Salmonella strains

Standard agar disk diffusion assay results showed that S. Typhimurium14028, S. Typhimurium 13311, S. Typhimurium LJH 666, S. Newport 692,S. Agona LJH 1132 and S. Agona LJH 1122 were resistant to streptomycinand tetracycline. S. Typhimurium var. Copenhagen LJH 788 was found tobe resistant to all four antibiotics tested, viz streptomycin, ampicillin,chloramphenicol and tetracycline (Table 2). This strain was, thus, consid-ered to be an MDR isolate. On the other hand S. Enteritidis 13076 wasfound to be sensitive to all fourmentioned antibiotics. The antibiotic resis-tance profiles of other Salmonella isolates are shown in Table 2. Theseisolates also testedpositive for thepresence of the respective antibiotic re-sistance genes by PCR, whichmostly paralleled their respective antibioticresistance phenotypic patterns. Based on these results, S. Typhimuriumvar Copenhagen LJH 788 was subsequently used for the standardizationof amultiplex 5′ nuclease assay for the detection of virulent and antibioticresistant Salmonella.

3.2. Specificity of the assay

The PCR cycling conditions of the primers designed for this assaywere standardized using a gradient thermal cycler (Veriti, AppliedBiosystems). The primers were found to perform efficiently at anannealing temperature of 58 °C. A multiplex real-time PCR assaywas standardized for the multiplex detection of virulent as well asantibiotic resistant Salmonella. The specificity of the modified invAgene primers was tested with genomic DNA of a wide range ofGram-negative and Gram-positive bacteria. No cross-reactivity wasobserved in any of the other 73 bacterial strains evaluated while all

Table 2Phenotypic1 and genotypic2 antibiotic resistance profile of Salmonella isolates.

Isolate Antimicrobial susceptibility1 and real-time PCR2 results

STR1 aadA22 TET1 tetG2 AMP1 Pse2 CHL1 floR2

S. Enteritidis F3-1 R + R + R − S −S. Enteritidis F3-2 R + R + R − S −S. Enteritidis F3-6 R + R + R − S −S. Enteritidis F3-7 R + R + S − S −S. Enteritidis F3-8 R + R + R − S −S. Enteritidis F3-9 R + R + S − S −S. Enteritidis F3-10 R + R + S − S −S. Enteritidis G5-02 R + R + R + R +S. Enteritidis G5-08 R + R + R − S −S. Enteritidis G5-09 R + R + S − S −S. Enteritidis G5-10 R + R + S − S −S. Enteritidis I4-1 R + R + S − S −S. Enteritidis I4-2 R + R − S − R −S. Enteritidis I4-3 R + R − S − S −S. Enteritidis I4-4 R + R − R + R −S. Enteritidis I4-5 R + R + R + S −S. Enteritidis I4-6 R + R + S − S −S. Enteritidis I4-7 R + R + S − S −S. Enteritidis I4-8 R + R + S − S −S. Enteritidis I4-9 R + R + S − S −S. Enteritidis I4-10 R − R − S − S −S. Enteritidis I4-11 R + R + S − S −S. Typhimurium 13311 R + R + S − S −S. Typhimurium 14028 R + R + S − S −S. Enteritidis 13076 S − S − S − S −S. Typhimurium LJH 666 R + S − S − S −S. Newport 692 R + R + S − S −S. Agona LJH 1132 R + R + S − S −S. Agona LJH 1122 R + R + S − S −S. Typhimurium LJH 788 R + R + R + R +

15

20

25

30

35

40

0 1 2 3 4 5 6 7

Ct

Log genomic equivalents

Amp

Chl

Inv

Str

Tet

Fig. 1. Standard curve for all five targets: ampicillin (Amp), chloramphenicol (Chl), invasin(Inv), streptomycin (Str) and tetracycline (Tet) constructed using serially diluted DNA. 1Genomic equivalent = 5 fg of Salmonella genomic DNA.

15

20

25

30

35

40

0 1 2 3 4 5 6 7 8

Ct

Bacterial count (Log CFU)

Amp

Chl

Inv

Str

Tet

Fig. 2. Standard curve for all five targets: ampicillin (Amp), chloramphenicol (Chl), invasin(Inv), streptomycin (Str) and tetracycline (Tet) constructed using a spiked beef sample.

216 P. Singh, A. Mustapha / International Journal of Food Microbiology 166 (2013) 213–218

30 Salmonella strains tested were positively detected by the assay(data not shown).

3.3. Sensitivity of the assay

The duplex assay for the detection of virulent strains of Salmonella,using the invA gene primer/probe with IAC was able to successfullydetect as low as 100 genomic equivalents of Salmonella from a 10-foldserially diluted DNA sample. The standard curve for this assay wasfound to be linear from 2 to 6 log genomic equivalents of Salmonellawith a regression coefficient of 0.98 and a reaction efficiency of97.948% (Fig. 1). For other antibiotic resistant genes being targeted inthis assay, the detection limit was found to be 500 fg/reaction (100genomic equivalents). The obtained regression coefficient and reactionefficiency of these targets, as shownon Fig. 1,were as follows: ampicillin(R2 = 0.97, PCR efficiency = 97.51%), chloramphenicol (R2 = 0.97,PCR efficiency = 96.03%), streptomycin (R2 = 0.99, PCR efficiency =91%) and tetracycline (R2 = 0.93, PCR efficiency = 110%). The sensi-tivity of the multiplex assay was found to be 5 pg/reaction (1000genomic equivalents).

A standard curve constructed from spiked beef sample for each ofthe five targets in duplex format with IAC was found to be linear fromlog 2 to log 7 CFU/g (Fig. 2). The obtained regression coefficient andreaction efficiency of these targets were as follows: invasin (R2 = 0.97,PCR efficiency = 72.34%), ampicillin (R2 = 0.94, PCR efficiency =66.81%), chloramphenicol (R2 = 0.97, PCR efficiency = 75.59%),streptomycin (R2 = 0.99, PCR efficiency = 106%) and tetracycline(R2 = 0.93, PCR efficiency = 114.89%).

3.4. Detection of antibiotic resistant Salmonella in spiked samples

Except for ground turkey, all other food samples obtained fromlocal supermarkets were found to be negative for Salmonella. Stan-dard aerobic plate count of food samples used in this study was as

follows: beef trim (5.2 × 104 CFU/g), ground beef 73:27 (4.5 ×102 CFU/g), ground beef 80:20 (8.2 × 103 CFU/g), ground beef85:15 (8.3 × 103 CFU/g), ground beef 93:7 (7.0 × 102 CFU/g),ground chicken (6.2 × 107 CFU/g), ground turkey (1.1 × 103 CFU/g), egg(N10 CFU/g), spinach (3.1 × 108 CFU/g) and tomato (1.1 × 107 CFU/g).All Salmonella-negative food samples were artificially inoculated with theMDR isolate, S. Typhimurium var. Copenhagen LJH 788 at five differentconcentrations, starting from 10 CFU/mL to 105 CFU/mL. The multiplexassay for all five targets was able to detect Salmonella, in samples spikedwith104 CFU/mL or higher without any enrichment (Table 3). Foodsamples spiked with 10 to 103 CFU/mL were detected after an enrich-ment period of 12 h. Different fat percentages in ground beef sampleswere not an inhibitory factor for this assay (Table 3).

4. Discussion

Multiplex real-time PCR has the ability to detect more than one tar-get in a single reaction, thus saving time and cost. The invA primer setdesigned by Rahn et al. (1992) is generally considered as a standardset of primers for the detection of pathogenic strains of Salmonellaspp. by the FOOD-PCR project (Malorny et al., 2003). The primer sethas been validated to show 99.4% specificity with 630 Salmonella strainsand 142 non-Salmonella strains (Rahn et al., 1992) and has also beentested with 100 other Salmonella strains from both Salmonella entericaand Salmonella bongori species (Anderson et al., 2011). This publishedinvA primer-probe set was initially tested for its compatibility withother primers and probes being used in our multiplex assay. However,the results with this invA primer and probe set were not reproduciblein either uniplex or multiplex format. This might have been due totwo reasons: first, the Tm difference in the forward and reverse primers

Table 3Multiplex real-time PCR results of artificially contaminated food samples.

Sample Spiked Salmonellaconcentration (CFU/mL)

Real-time PCR resultafter enrichment

0 h 12 h 24 h

Beef trim 10 − + +102 − + +103 − + +104 + + +105 + + +

Ground beef (73:27) 10 − + +102 − + +103 − + +104 + + +105 + + +

Ground beef (80:20) 10 − + +102 − + +103 − + +104 + + +105 + + +

Ground beef (85:15) 10 − + +102 − + +103 − + +104 + + +105 + + +

Ground beef (93:7) 10 − + +102 − + +103 − + +104 + + +105 + + +

Ground chicken 10 − + +102 − + +103 − + +104 + + +105 + + +

Egg 10 − + +102 − + +103 − + +104 + + +105 + + +

Spinach 10 − + +102 − + +103 − + +104 + + +105 + + +

Tomato 10 − + +102 − + +103 − + +104 + + +105 + + +

217P. Singh, A. Mustapha / International Journal of Food Microbiology 166 (2013) 213–218

is very high, and secondly, the annealing temperature of the probe waslower than that of the forward primers. This Tm mismatch between theprimers and probe is not ideal for real-time PCR amplification. Even theamplicon size of this invA gene primer set (284 bp) was on the higherside of the generally recommended amplicon size for real-time PCR.As these primer and probe sets were already validated by various labsfor their specificity for Salmonella, some minor changes in the primerset were made in our study, to allow for the real-time assay to bemore thermodynamically stable. The annealing temperature of the re-verse primer was adjusted by reducing it by one base from the 3′-endand a new forward primer was designed using Primer3 software(Rozen and Skaletsky, 2000). This new primer set was validated with73 bacterial strains and found to be highly specific for Salmonella (datanot shown).

Malorny et al. (2007) reported a 5′ nuclease uniplex assay for thedetection of Salmonella with a sensitivity of 100 genomic equivalents.Wang and Mustapha (2010) reported a duplex TaqMan® assay for thedetection of viable Salmonella with a sensitivity of 100 CFU/mL and1000 CFU/mL in uniplex and multiplex assays, respectively. Hein et al.(2006) and Anderson et al. (2011) reported a lower detection limit of10 CFU/mL. In the present study, a detection limit of 100 and 1000 ge-nomic equivalents was achieved in uniplex and multiplex assay format,

respectively. The detection limit of the real-time PCR assay developed inthis study for the detection of antibiotic resistant Salmonella wascomparable to results obtained by Malorny et al. (2007) and WangandMustapha (2010). However, the assay demonstrated a lower sensi-tivity when compared to real-time multiplex PCR assays described byHein et al. (2006) and Anderson et al. (2011). This might be due to ahigher number of primers and probes used in this multiplex assay.Food samples artificially contaminated with Salmonella concentra-tions lower than 104 CFU/mL could not be detected without an en-richment step. A detection limit of 10 CFU/g was achieved after anenrichment period of 12 h, for all the artificially contaminated foodsamples tested (Table 3). The use of a sterile filter stomacher bagfor the enrichment of artificially contaminated food samples allowedfor a good separation of fat and other debris from the enriched foodsample. Thus, ground beef samples with a higher fat content showedno inhibitory effects on the assay. When the sensitivity of the spikedsamples was compared, a sample artificially contaminated with10 CFU/g of Salmonella could be detected after a 12-h enrichment.This enrichment time is much shorter than the 16–20 h enrichment pe-riod reported by previous researchers (Anderson et al., 2011; Hein et al.,2006; Malorny et al., 2007).

Food samples contain many inhibitory factors that include organicand inorganic substances, such as phenolic compounds, fats, enzymes,polysaccharides, proteins, and salts. These inhibitory substances caneither inhibit PCR amplification, leading to false-negative results, ortheymay cause a reduction in the amplification efficiency of the reaction,leading to a reduction in the detection limit of the assay (Španová et al.,2000). These inhibitory compounds could be one of the reasons for thelow PCR efficiency observed in the standard curve constructed usingground beef (Fig. 2).

An IAC is a non-target DNA sequence present in the sample reactionmixture that is co-amplified with the target sequence in the reaction.The IAC is used to overrule chances of false negative results in a PCRassay. In this assay, a 98-base long DNA sequence was uniquelydesigned to serve as the IAC, which has a primer-binding site for theinvA gene primers and a different probe binding sequence. Thus, theIAC target DNA could be co-amplified with the same primer set butdetected using a different dual-labeled probe. This approach helped usto reduce the number of primers used in this multiplex assay. Becausethe detection of an IAC is an important part of any real-time PCRassay, a LNA dual-labeled probe was used for the detection of the IACtarget in this study. Probes with LNA bases have the property of bindingto complementary target molecules (DNA, RNA or LNA) with higheraffinity (Kumar et al., 1998;Whitcombe et al., 1999). The probeworkedefficiently even at a starting target DNA concentra7tion of 0.1 pg/reaction.The LNA-dual labeled probe used for the detection of the IAC target waslabeled with Cy5 fluorophore and BHQ-3 quencher. BHQ-2 quencher isgenerally recommended for used in a dual-labeled probe with Cy5fluorophore (Marras, 2006). In a study done by Reynisson et al. (2006),Cy5–BHQ3 dual-labeled probes showed a lower cycle threshold (Ct)value and stronger fluorescence signal than Cy5–BHQ2 labeled probe.This might be because of a better overlap in the fluorescence spectra ofCy5 dye and BHQ-3 quencher than other quenchers.

Other real-time assays only detected the presence of Salmonella in afood sample but, to our knowledge, the assay described in this study isthe first to successfully identify both virulent and antibiotic resistantstrains of Salmonella in a single reaction. The primer set developed inthis study, which uses the same set of primers for amplification of theIAC and was detected with a LNA dual-labeled probe, demonstratedthat such combinations do work robustly in multiplex PCR. Aslamet al. (2012) reported an unconditional association between antibioticresistance genes, which can be co-selected and transferred throughmobile genetic elements to other members of Enterobacteriaceaefamily by horizontal gene transfer mechanisms. Hence, the assay de-veloped in this study might also be used to detect antibiotic resistantstrains of other members of Enterobacteriaceae. The assay can be

218 P. Singh, A. Mustapha / International Journal of Food Microbiology 166 (2013) 213–218

performed in less than 16 h and, with the use of an IAC, has the po-tential to be applied for pathogenic and antibiotic resistant pathogentesting in the food industry to ensure food safety and minimize theassociated health hazards.

Acknowledgments

This researchwas supported by the Beef Checkoff dollars. The authorssincerely thank theMissouri Beef Industry Council for funding this project.

References

Aarestrup, F.M., Wegener, H.C., Collignon, P., 2008. Resistance in bacteria of the foodchain: epidemiology and control strategies. Expert Review of Anti-Infective Therapy6, 733–750.

Akkina, J.E., Hogue, A.T., Angulo, F.J., Johnson, R., Petersen, K.E., Saini, P.K., Fedorka-Cray,P.J., Schlosser, W.D., 1999. Epidemiologic aspects, control, and importance ofmultiple-drug resistant Salmonella Typhimurium DT104 in the United States. Journalof the American Veterinary Medical Association 214, 790–798.

Anderson, A., Pietsch, K., Zucker, R., Mayr, A., Müller-Hohe, E., Messelhäusser, U., Singh, A.,Busch, U., Huber, I., 2011. Validation of a duplex real-time PCR for the detection ofSalmonella spp. in different food products. Food Analytical Methods 4, 259–267.

Andrews,W.H., Jacobson, A., Hammack, T., 2007. Salmonella. In: Jackson, G.J., et al. (Ed.), Bac-teriological Analytical Manual Online. U.S. Food and Drug Administration, Gaithersburg,MD (Nov 2011 ed., http://www.fda.gov/Food/ScienceResearch/LaboratoryMethods/BacteriologicalAnalyticalManualBAM/ucm070149.htm).

Aslam, M., Checkley, S., Avery, B., Chalmers, G., Bohaychuk, V., Gensler, G., Reid-Smith,R., Boerlin, P., 2012. Phenotypic andgenetic characterization of antimicrobial resistancein Salmonella serovars isolated from retail meats in Alberta, Canada. Food Microbiology32, 110–117.

Boyd, E.F., Li, J., Ochman, H., Selander, R.K., 1997. Comparative genetics of the inv-spainvasion gene complex of Salmonellaenterica. Journal of Bacteriology 179, 1985–1991.

Centers for Disease Control and Prevention, 2010. Salmonella. http://www.cdc.gov/Salmonella/general/index.html.

Chiu, C.H., Su, L.H., Chu, C.H., Wang, M.H., Yeh, C.M., Weill, F.X., Chu, C., 2006. Detection ofmultidrug-resistant Salmonella enterica serovar Typhimurium phage types DT102,DT104, and U302 by multiplex PCR. Journal of Clinical Microbiology 44, 2354–2358.

Cohen, J., 2013. Confronting the threat of multidrug-resistant Gram-negative bacteria incritically ill patients. Journal of Antimicrobial Chemotherapy 68, 490–491.

Hein, I., Flekna, G., Krassnig, M., Wagner, M., 2006. Real-time PCR for the detection ofSalmonella spp. in food: an alternative approach to a conventional PCR system sug-gested by the FOOD-PCR project. Journal of Microbiological Methods 66, 538–547.

Hoban, D.J., Lascols, C., Nicolle, L.E., Badal, R., Bouchillon, S., Hackel, M., Hawser, S., 2012.Antimicrobial susceptibility of Enterobacteriaceae, including molecular characteriza-tion of extended-spectrumbeta-lactamase-producing species, in urinary tract isolatesfrom hospitalized patients in North America and Europe: results from the SMARTstudy 2009-2010. Diagnostic Microbiology and Infectious Disease 74, 62–67.

Khan, S.A., Sung, K., Nawaz, M.S., 2011. Detection of aacA-aphD, qacE [delta] 1, marA, floR,and tetA genes from multidrug-resistant bacteria: comparative analysis of real-timemultiplex PCR assays using EvaGreenо and SYBRо Green I dyes. Molecular andCellular Probes 25, 78–86.

Kumar, R., Singh, S.K., Koshkin, A.A., Rajwanshi, V.K., Meldgaard, M., Wengel, J., 1998. Thefirst analogues of LNA (locked nucleic acids): phosphorothioate-LNA and 2′-thio-LNA. Bioorganic & Medicinal Chemistry Letters 8, 2219–2222.

Li, X., Bethune, L., Jia, Y., Lovell, R., Proescholdt, T., Benz, S., Schell, T., Kaplan, G.,McChesney, D., 2012. Surveillance of Salmonella prevalence in animal feeds andcharacterization of the Salmonella isolates by serotyping and antimicrobial suscepti-bility. Foodborne Pathogens and Disease 9, 692–698.

Malorny, B., Hoorfar, J., Bunge, C., Helmuth, R., 2003. Multicenter validation of the analyticalaccuracy of Salmonella PCR: towards an international standard. Applied and Environ-mental Microbiology 69, 290–296.

Malorny, B., Bunge, C., Helmuth, R., 2007. A real-time PCR for the detection of SalmonellaEnteritidis in poultry meat and consumption eggs. Journal of MicrobiologicalMethods 70, 245–251.

Marras, S.A.E., 2006. Selection of fluorophore and quencher pairs for fluorescent nucleicacid hybridization probes. Methods in Molecular Biology 335, 3–16.

McClelland,M., Sanderson, K.E., Spieth, J., Clifton, S.W., Latreille, P., Courtney, L., Porwollik, S.,Ali, J., Dante,M., Du, F., 2001. Complete genome sequence of Salmonella enterica serovarTyphimurium LT2. Nature 413, 852–856.

Molenkamp, R., Van Der Ham, A., Schinkel, J., Beld, M., 2007. Simultaneous detectionof five different DNA targets by real-time TaqMan PCR using the RocheLightCycler480: application in viral molecular diagnostics. Journal of VirologicalMethods 141, 205–211.

Molla, B., Miko, A., Pries, K., Hildebrandt, G., Kleer, J., Schroeter, A., Helmuth, R., 2007. Class1 integrons and resistance gene cassettes among multidrug resistant Salmonellaserovars isolated from slaughter animals and foods of animal origin in Ethiopia.Acta Tropica 103, 142–149.

Rahn, K., De Grandis, S., Clarke, R., McEwen, S., Galan, J., Ginocchio, C., Curtiss, R., Gyles, C.,1992. Amplification of an invA gene sequence of Salmonella typhimurium by poly-merase chain reaction as a specific method of detection of Salmonella. Molecularand Cellular Probes 6, 271–279.

Reynisson, E., Josefsen, M.H., Krause, M., Hoorfar, J., 2006. Evaluation of probe chemistriesand platforms to improve the detection limit of real-time PCR. Journal of MicrobiologicalMethods 66, 206–216.

Rozen, S., Skaletsky, H.J., 2000. Primer3 on the WWW for general users and for biologistprogrammers. In: Krawetz, S., Misener, S. (Eds.), Bioinformatics Methods and Proto-cols: Methods in Molecular Biology. Humana Press, Totowa, NJ, pp. 365–386.

Sjölund-Karlsson, M., Rickert, R., Matar, C., Pecic, G., Howie, R.L., Joyce, K., Medalla, F.,Barzilay, E.J., Whichard, J.M., 2010. Salmonella isolates with decreased susceptibilityto extended-spectrum cephalosporins in the United States. Foodborne Pathogensand Disease 7, 1503–1509.

Španová, A., Rittich, B., Karpíšková, R., Čechová, L., Škapova, D., 2000. PCR identification ofSalmonella cells in food and stool samples after immunomagnetic separation.Bioseparation 9, 379–384.

Tolstrup, N., Nielsen, P.S., Kolberg, J.G., Frankel, A.M., Vissing, H., Kauppinen, S., 2003.OligoDesign: optimal design of LNA (locked nucleic acid) oligonucleotide captureprobes for gene expression profiling. Nucleic Acids Research 31, 3758–3762.

Wang, L., Mustapha, A., 2010. EMA real time PCR as a reliable method for detection ofviable Salmonella in chicken and eggs. Journal of Food Science 75, M134–M139.

Watt, J.L., Shryock, T.R., Apley, M., Bade, D.J., Brown, S.D., Gray, J.T., Heine, H., Hunter,R.P., Mevius, D.J., Papich, M.G., Silley, P., Zurenko, G.E., 2008. Performance Standardsfor Antimicrobial Disk and Dilution Susceptibility Tests for Bacteria Isolated fromAnimals: Approved Standard, CLSI DocumentM31–A3, 3rd (Edn). Clinical LaboratoryStandards Institute, Wayne, PA (http://www.clsi.org/source/orders/free/m31-a3.pdf).

Whitcombe, D., Theaker, J., Guy, S.P., Brown, T., Little, S., 1999. Detection of PCR prod-ucts using self-probing amplicons and fluorescence. Nature Biotechnology 17,804–807.

White, P.A., McIver, C.J., Rawlinson, W.D., 2001. Integrons and gene cassettes in theEnterobacteriaceae. Antimicrobial Agents and Chemotherapy 45, 2658–2661.

Zhao, S., McDermott, P.F., White, D.G., Qaiyumi, S., Friedman, S.L., Abbott, J.W., Glenn, A.,Ayers, S.L., Post, K.W., Fales, W.H., Wilson, R.B., Reggiardo, C., Walker, R.D., 2007.Characterization of multidrug resistant Salmonella recovered from diseased animals.Veterinary Microbiology 123, 122–132.