International Journal of Food Microbiology 148 (2011) 8–14

Contents lists available at ScienceDirect

International Journal of Food Microbiology

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

Multiple-locus variable number of tandem repeat analysis (MLVA) ofListeria monocytogenes directly in food samples

Shu Chen a,⁎, Jiping Li a, Saleema Saleh-Lakha a, Vanessa Allen b, Joseph Odumeru c

a Laboratory Services Division, University of Guelph, Guelph, Ontario, Canadab Public Health Laboratory, The Ontario Agency for Health Protection and Promotion, 81 Resources Road, Etobicoke, Ontario, Canada M9P 3T1c Laboraotory Services Branch, The Ontario Ministry of Environment, 125 Resources Road, Etobicoke, Ontario, Canada M9P 3V6

⁎ Corresponding author at: University of Guelph,P.O. Box 3650, 95 Stone Road West, Guelph, Ontario, C823 1268x57319; fax: +1 519 767 6240.

E-mail addresses: schen@uoguelph.ca (S. Chen), jipissaleh@uoguelph.ca (S. Saleh-Lakha), Vanessa.Allen@oaJoseph.Odumeru@ontario.ca (J. Odumeru).

0168-1605/$ – see front matter © 2011 Elsevier B.V. Adoi:10.1016/j.ijfoodmicro.2011.04.014

a b s t r a c t

a r t i c l e i n f o

Article history:Received 3 January 2011Received in revised form 4 April 2011Accepted 12 April 2011Available online 22 April 2011

Keywords:Listeria monocytogenesMLVAFood-borne pathogenSubtyping Listeria monocytogenes

Listeria monocytogenes is the etiologic agent of listeriosis responsible for severe and fatal infections in humans.Listeria contamination occurs quite often in a wide range of foods due to its ubiquitous nature. Isolates need tobe characterized to a strain level for accurate diagnosis of Listeria infection, epidemiological studies,investigation of outbreaks and effective prevention and control of food-borne listeriosis. The purpose of thisresearchwas to evaluate themultiple-locus variable number of tandem repeat analysis (MLVA) for sub-typingL. monocytogenes isolates in pure cultures and in food matrices. Two multiplex PCR assays were formulated toamplify six specific loci using fluorescently-labeled primers; and the amplicons were analyzed by capillaryelectrophoresis. The MLVA method resulted in 34 unique DNA fingerprint patterns from 46 L. monocytogenesisolates of 10 serotypes which had 29 or 30 PFGE patterns with a single restriction enzyme and 34 AFLPpatterns. The MLVA patterns of the 46 isolates remained unchanged in the presence of pre-enrichedfood matrices including sausage, ham, chicken, milk and lettuce. The MLVA method successfully typedL. monocytogenes strains spiked in cheese, roast beef, egg salad and vegetable samples after 48 h enrichment atthe initial inoculation levels of 1–5 CFU per 25 g of food or higher. The limits of detection (typing) of theMLVAmethod were 103–104 CFU/mL of pre-enriched food broth when evaluated using post-spiked sausage, ham,chicken, milk and lettuce samples. The MLVA method was simple, highly discriminatory, and easy to performwith portable (numerical) results. To our knowledge, this is the first report that describes the application ofthe MLVA method directly to food samples and demonstrates the possibility to obtain rapid and accuratesubtyping results before an isolate is obtained.

Laboratory Services Division,anada N1H 8J7. Tel.: +1 519

ngli@uoguelph.ca (J. Li),hpp.ca (V. Allen),

ll rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Listeria monocytogenes, a ubiquitous food-borne bacterium capableof growing under refrigeration conditions, is responsible for severeand fatal infections in humans. The pathogen can cause invasivehuman listeriosis, particularly in high-risk populations such as theelderly, pregnant women and neonates. The symptoms of listeriosisinclude meningitis, encephalitis, septicaemia and abortion (Jefferset al., 2001; Nightingale et al., 2005). It is often difficult to diagnoseand identify exact sources of this pathogen due to the long incubationperiod (up to 70 days) associated with listeriosis (Swaminathan andGerner-Smidt, 2007; Wiedmann, 2002). The annual cases of humanlisteriosis in Canada have increased from approximately 47–79between 1995 and 2000 to 100–140 in recent years (Pagotto et al.,

2006; PHAC, 2009), perhaps due to enhanced reporting systems, andincreased aging population. Among Listeria outbreaks, one of themosttragic occurrences was the 2008 listeriosis outbreak in Ontario,Canada caused by consumption of contaminated ready-to-eat meats.The outbreak resulted in 57 serious illnesses and loss of 22 lives, andprompted the country's largest food recall in history (PHAC, 2009).Food-borne listeriosis remains a public health threat.

Themain mode of transmission of Listeria has been considered to bethrough environmental contamination of ready-to-eat food productsduring their production and processing (Evans et al., 2004; Lappi et al.,2004; Vazquez-Boland et al., 2001). The pathogen can be isolated fromvarious raw and ready-to-eat foods and also from environmentalsamples such as water, equipment swabs and soil (Gombas et al.,2003; Miettinen and Wirtanen, 2006; Wiedmann, 2003). Food proces-sing plants following “GMP” operations with a sound HACCP system inplace cannot always ensure that contaminated foods do not reachconsumers. Food recalls due to possible contamination of the productsby L. monocytogenes continue to occur (CFIA, 2009, 2010). Production offoods free from Listeria contamination remains a challenge for the foodindustry. Proactive approaches beyond “plant floor inspection” and “testandhold”protocols are needed in order to effectively protect consumers.

9S. Chen et al. / International Journal of Food Microbiology 148 (2011) 8–14

Isolates of L. monocytogenes need to be characterized in order toaccurately diagnose their infection, understand the epidemiology ofinfection, investigate outbreaks, and effectively prevent and minimizethe transmission of Listeria through the food chain and other routes.Among the 13 recognized serotypes of L. monocytogenes, serotypes 1/2a,1/2b and 4b, are themost prevalent in cases of human listeriosis (Evanset al., 2004; Farber and Peterkin, 1991; Hofer et al., 2000) and differentstrains of a same serotype also have varying degrees of virulence (Evanset al., 2004; Farber and Peterkin, 1991; Hofer et al., 2000). Molecularmethods such as pulsed-field gel electrophoresis (PFGE) (Lindstedt et al.,2008; Revazishvili et al., 2004), ribotyping (Fonnesbech et al., 2004; Grifet al., 2006), multi-virulence-locus sequence typing (MVLST) (Zhanget al., 2004; Chen et al., 2007) and amplified fragment length polymor-phism (AFLP) (Aarts et al., 1999; Riikka et al., 2003) have been used tosubtype strains of L. monocytogenes. These methods can differ in theirability to differentiate strains, their reproducibility, ease of use, cost andtime to obtain a result. PFGE has been considered the “gold standard” forsubtyping common food-borne pathogens and have been successfullyapplied to subtyping of many food-borne pathogens, includingL. monocytogenes (Lindstedt et al., 2008; Revazishvili et al., 2004).However, PFGE is time-consuming, labor-intensive, has low throughputand limited inter-person or inter-lab reproducibility and does not workon every strain (Heir et al., 2000). An AFLP method has beendemonstrated to have similar discriminatory power to PFGE, higherreproducibility, automation capability and higher throughput (Aartset al., 1999; Riikka et al., 2003). However, the method is technicallydemanding, and data interpretation of AFLP fingerprints can be com-plicated, especially when there is no AFLP database for the pathogens.

Recently, multiple-locus variable number of tandem repeatanalysis (MLVA) has been developed as a new generation method tosubtype food-borne pathogens including L. monocytogenes (Lindstedtet al., 2008; Miya et al., 2008; Murphy et al., 2007). This approach isbased on the detection of the number of tandem repeats (TRs) at aspecific locus in the genome of a microorganism. MLVA is a PCR-basedmethod that is simple and reproducible and has been demonstrated tobe comparable or more discriminatory than PFGE (Lindstedt et al.,2008; Murphy et al., 2007), and has been successfully applied tosubtyping other food-borne pathogens such as Salmonella (Lindstedtet al., 2003a; Yichun et al., 2003) and E. coli O157:H7 (Lindstedt et al.,2003b, 2004). In this study, we have evaluated recently publishedMLVAprimers, established twomultiplex PCR assays basedon selectedloci, and evaluated the MLVA procedure to subtype a collection ofL. monocytogenes strains of various serotypes, as well as its utility as afast-typing tool directly on positive food sampleswithout the isolationof a pure culture. To our knowledge this is thefirst report of applicationof the MLVA method directly to food samples.

2. Materials and methods

2.1. Bacterial strains

The strains of L. monocytogenes were from a collection ofLaboratory Services, University of Guelph, which were originallyisolated from various food and other samples, including meat, milk,and vegetables. All isolates were previously identified as L. mono-cytogenes by a culture method (Pagotto et al., 2001), and furthercharacterized by serotyping. These isolates were maintained at−80 °C on cryostat beads. L. monocytogenes ATCC 19115 (serotype4b) was included as a reference strain in this study.

2.2. Food sample preparation

Food samples were collected from local retail stores in Ontario,Canada. Samples were stored on ice during shipping and processedimmediately upon arrival. Food samples (25 g or 25 mL) werehomogenized with 225 mL of Listeria enrichment broth (LEB) (Becton,

Dickinson and Company, Sparks, MD) in a stomacher bag andincubated for 48 h at 30 °C. The pre-enriched samples were testedfor L. monocytogenes using the culture method (Pagotto et al., 2001)and were also used in the MLVA analysis. In the artificial inoculationexperiments, only those food samples that were confirmedL. monocytogenes negative by the standard method were used. Foodsamples were inoculated in two ways: (1) Food samples (250 g or250 mL) were inoculated with 10–50 and 50–100 CFU of L. mono-cytogenes; two to five sub-samples of 25 g (or 25 mL) each were takenand then homogenized with 225 mL LEB broth in a stomacher bagusing a Smasher (AES Chemunex, Inc. Cranbury, NJ) at a normal speedfor 1 min and incubated at 30 °C for 48 h, and these food samplesweredesignated as pre-enrichment-spiked samples. (2) Food homogenate(250 mL of 10% food in LEB) was pre-enriched at 30 °C for 48 h andthen inoculated with 102–105 CFU/mL of L. monocytogenes prior toDNA extraction, and these food samples were designated as post-enrichment-spiked samples.

2.3. DNA isolation

Bacterial cells were cultured on Tryptone Soy Agar (Oxoid, Nepean,ON) at 30 °C for 24 h and a single isolated colony was inoculated into3 mL Tryptone Soy Broth (TSB) (Becton, Dickinson and Company) andincubated at 30 °C for 24 h. The bacterial cells were recovered from1 mL culture by centrifugation at 14,000×g for 5 min. Supernatantwas discarded and total genomic DNA was isolated using theDNeasy™ tissue kit (Qiagen, Mississauga, ON) according to themanufacturer's protocol for Gram-positive bacteria. For DNA extrac-tion from spiked food samples, cells were recovered from 1 mL of foodbroth by centrifugation at 14,000×g for 5 min. The total genomic DNAwas extracted from the cell pellet using the same method as the purecultures. The purified DNA was eluted in 100 μL AE Buffer (Qiagen)and stored at −20 °C until use.

2.4. Multiple-locus variable number of tandem repeat analysis (MLVA)

The VNTR loci, targeting the main serotypes of L. monocytogenes,1/2a, 1/2b, 1/2c and 4b were as described previously (Miya et al.,2008; Murphy et al., 2007) and two triplex PCR assays wereformulated and optimized in this study. A summary of the six lociselected in this study is provided in Table 1. Primers were synthesizedat the Laboratory Services, University of Guelph (Guelph, ON) using anABI 3900 HT DNA synthesizer (Applied Biosystems, Foster City, CA)and used to amplify the selected genomic regions according to thecombinations as indicated in Table 1. Fluorescent TR sequences wereamplified from bacterial DNA or total genomic DNA of spiked foodsamples using multiplex PCR. The PCR reaction mixtures (25 μL)contained 1× HotStarTaq Master Mix (Qiagen), 100–600 nM corre-sponding primers according to the ratios listed in Table 1, 2 mMadditional MgCl2, and 1–5 μL of template DNA (approximately 10–100 ng). Thermal cycling conditionswere one cycle of 10 min at 95 °C;7 cycles of 15 s at 95 °C, 40 s at 65 to 59 °C (1 °C touching down fromthe previous cycle), 1 min and 10 s at 72 °C; 38 more cycles of 15 s at95 °C, 40 s at 58 °C, 1 min and10 s at 72 °C; and a final extension timeat 72 °C for 7 min. Each amplification run included a positive controland a no-DNA template control. All amplifications were conductedusing a thermal cycler GeneAmp PCR System 9700 (AppliedBiosystems). Amplified products were resolved by automated capil-lary electrophoresis using an ABI 3730 Genetic Analyzer with POP-7polymer and 3730 Buffer with EDTA (Applied Biosystems). The GSLIZ1200 (Applied Biosystems) was used as an internal size standard.The PCR fragments were sized using the GeneMapper v4.0 software(Applied Biosystems), and the number of tandem repeats (TRs) wascalculated, and rounded down to form whole numbers. An allelenumber string, based on the number of TRs at each locus, was assignedto the amplified DNA fragments from each isolate.

Table 1Summary of the MLVA primers used to subtype L. monocytogenes in this study.

Primer ID Sequence(5′–3′)

PCR grouping(and primer ratio)

Amplicon sizea

(bp)Reference

LM-TR 1F ggcggaaaatgggaagc I (1) 695 Murphy et al., 2007LM-TR 1R tgcgatggtttggactgttg I (1)LM-TR 4F tccgaaaaagacgaagaagtagca I (0.8) 478 Murphy et al., 2007LM-TR 4R tggaacgacggacgaaataataat I (0.8)LM4b-TR1f acatgggaagggttgcaa I (1) 299 Miya et al., 2008LM4b-TR1r ggatttacttgatttgacgggt I (1)LM-TR 3F gcgtgtattagatgcggttgag II (4) 548 Murphy et al., 2007LM-TR 3R gcattccactatcccctgtttt II (4)LM4b-TR2f ccatggaagactactgtttgta II (2) 468 Miya et al., 2008LM4b-TR2r gacggtactgttatcggaaa II (2)LM4b-TR3f gaaggtaaaaacggcgaaaaa II (1) 383 Miya et al., 2008LM4b-TR3r attgcttctccgtatccctca II (1)

Note: All forward primers were labeled with the fluorescent dye FAM. The primers were grouped into two multiplex assays I and II. The numbers in the brackets following the PCRGroup numbers indicate the optimized relative ratios of the primers used. The amplicon sizes were determined based on the genome sequence of L. monocytogenes strains 1/2a EGD-e (accession no. AL591824) and 4b F2365 (accession no. AE017262).

10 S. Chen et al. / International Journal of Food Microbiology 148 (2011) 8–14

2.5. Pulsed-field gel electrophoresis (PFGE) analysis

PFGE was conducted on the strains of L. monocytogenes using thestandardized laboratory protocol for sub-typing L. monocytogenes(PulseNet, Centres of Disease Control and Prevention, Atlanta, Georgia,http://www.cdc.gov/pulsenet/protocols.htm). Two restriction enzymes,AscI and ApaI were used separately to generate two sets of PFGE patterns.

2.6. Amplified fragment length polymorphism (AFLP) analysis

The basic procedures used for Amplified Fragment LengthPolymorphism (AFLP) analysis were as described previously (Aartset al., 1999; Guan et al., 2002), and described as follows. Approxi-mately 100 ng genomic DNA was digested with EcoRI and MseI (NewEngland Biolabs, Pickering, ON), and the resulting DNA fragmentswere ligated to EcoRI and MseI adapters of the AFLP microbialfingerprinting kit (Applied Biosystems) by following the specifica-tions of the manufacturer. The EcoRI–MseI fragments tagged withspecific adapters were then selectively amplified by using theGeneAmp 9700 PCR system (Applied Biosystems). Primers, EcoRI-A(labeled with the fluorescent dye FAM) and MseI-C were used inthe selective amplification. The following thermo-cycling conditionswere used: one cycle of 2 min at 96 °C, 30 s at 65 °C, and 2 min at72 °C; eight cycles of 1 s at 94 °C, 30 s at 64 to 57 °C (1 °C touchingdown from the previous cycle), and 2 min at 72 °C; 26 cycles of 1 s at94 °C, 30 s at 56 °C, and 2 min at 72 °C; and a final incubation at 60 °Cfor 30 min. The amplified DNA fragments were separated usingABI 3730 Genetic Analyzer with POP-7 polymer and 3730 Buffer

500030001000

250 290 330 370 410 450

Flu

ores

cenc

e In

tens

ity

500030001000

500030001000

500030001000

Fig. 1. Representative MLVA patterns of L. monocytogenes generated in two multiplex P

with EDTA (Applied Biosystems). The GS LIZ 600 size standard(Applied Biosystems) was included as an internal standard. The strainof L. monocytogenes ATCC 19115 was used as a positive control andPCR water was used as a negative control. The fragment data wereanalyzed for size calling and intensity normalization by using theGeneMapper v4.0 software (Applied Biosystems). All electrophero-grams were also visually inspected to ensure data quality. Fragmentsin the range from 60 to 600 bp were included in data analysis.

2.7. Data analysis

The DNA fingerprinting data were analyzed using the BioNumericsv5.1 software program (Applied Maths, Austin, Texas). The relation-ship among isolates of L. monocytogenes was determined using themodules of Fingerprint Types and Comparison and Cluster Analysis ofthe BioNumerics v5.1 software. The files generated using theGeneMapper software or from the PFGE analysis were imported intothe BioNumerics software, and the fragments were compared usingthe Band Matching function of the software. Uncertain bands wereexcluded in the analysis by using the software's Uncertain Bandsfunction. Cluster analysis was conducted as follows: the similaritiesamong the fingerprints of each sample were calculated using theband-based Dice coefficient; dendrograms were constructed by usingUnweighted Pair Group Method using Arithmetic averages (UPGMA).The cluster grouping was based on a 95% similarity of the fingerprints.The fingerprints of representative groups were visually examined toensure that the software data analysis outcomes are matchedwith thevisual pattern determination.

A, I

A, II

B, I

B, II

460 530 570 610 650 730 (bp)

CR assays I and II: A, strain 1047-218, serotype 1/2b; B, strain LI0521, serotype 4b.

11S. Chen et al. / International Journal of Food Microbiology 148 (2011) 8–14

3. Results

3.1. Establishment and optimization of the multiplex MLVA PCR assays

Fourteen published primer pairs (Lindstedt et al., 2008; Miya et al.,2008; Murphy et al., 2007) were evaluated using a collection of 46isolates of L. monocytogenes. Six of the primer pairs (Table 1) were

Strain name Serotype Source

LI0529 1/2a Raw Milk

LI0530 1/2a Raw Milk

LI0575 3a Beef

L341 1/2a Cheese

LI0576 1/2a Meat

7163 1/2a Meat

L90 1/2c Faeces

LI0539 1/2c Salami

ATCC 7644 1/2c Human

LI0548 1/2c Cheese

7148 1/2a Meat

H12 4a Cucumber

LI0578 1/2a Ice Cream

LI0513 3a Beef

LI0527 1/2a Turkey

ATCC 19115 4b Human

LI0540 1/2b Salami

LI0511 3b Sausage

LI0521 4b Outbreak II

LI0522 4b Outbreak II

LI0524 4b Outbreak II

LI0525 4b Outbreak II

LI0549 4b Outbreak II

310 4b Outbreak II

LI0512 1/2b Beef

LI0514 1/2b Beef

LI0533 4b Outbreak I

1029-250 4b Outbreak I

1042-221 4b Outbreak I

1088-193 4b Outbreak I

1087-194 4b Outbreak I

LI0507 4e Chicken

V9 4b Beef

L487 1/2b Cheese

1047-218 1/2b Sporadic, C

LI0523 4b Outbreak II

1089-192 3b Sporadic, U

L77 4b Liquor

LI0506 4d Sheep

V37CE 4b Raw Milk

L47 4b Liquor

ATCC 19113 3a Human

409 1/2b Cheese

V84 1/2b Pork

L46 4a Placenta

LI0505 4c Chicken

% Similarity PFGE-ApaI

0 20 40 60 80 100

Fig. 2. Cluster analysis of 46 strains of L. monocytogenes based on PFGE-ApaI type using the DiAFLP methods. Subtype characteristics of the L. monocytogenes strains, including strain napatterns are indicated. MLVA allele string is indicated in the order of LM-TR 1F/R, LM-TR 3F/PCR product. The PFGE and AFLP pattern numbers are arbitrary.

selected for further study based on their discrimination capacity,amplification success rate and compatibility for multiplexing in PCR.PCR conditionswith the six selected primer pairs were then optimizedfor balanced amplification of the targets under one condition and forefficient amplification of low copy numbers of target DNA sequencesin food samples. The touchdown annealing temperatures used foramplification, from 65 to 59 °C, resulted in good yields of all the PCR

MLVA allele string

PFGE-ApaI

pattern

PFGE-AscI

patternAFLP

pattern

06-np(--)-03-np-13-08 1 1 1

06-13(--)-03-np-13-08 1 1 2

08-17(19)-03-np-np-08 2 2 3

np-12(15)-03-np-np-08 3 3 4

03-23(19)-03-np-np-08 4 4 5

03-23(--)-03-np-np-08 4 4 5

08-15(21)-03-np-np-08 5 5 6

08-21(--)-03-np-np-08 6 5 7

09-21(25)-03-np-np-08 7 6 8

08-21(15)-03-np-np-08 8 7 9

np-22(21)-03-np-np-08 9 8 10

np-np(--)-02-np-15-np 10 9 11

07-20(15)-03-np-np-08 11 10 12

09-32(19)-03-np-np-08 11 11 13

np-21(15)-03-np-np-08 12 12 14

00-25(--)-02-25-12-06 13 13 n/a

02-21(--)-02-21-17-04 14 14 15

02-13(--)-02-13-17-04 15 14 16

00-24(--)-02-24-12-06 16 15 17

(food) 00-24(--)-02-24-12-06 16 15 17

00-24(--)-02-24-12-06 16 15 17

00-24(--)-02-24-12-06 16 15 17

(cheese) 00-24(--)-02-24-12-06 16 15 17

(cheese) 00-24(--)-02-24-12-06 16 15 17

02-20(--)-02-20-17-04 17 16 18

02-20(--)-02-20-17-04 17 16 18

(coleslaw) 00-19(--)-03-19-17-08 18 17 19

00-19(--)-03-19-17-08 18 17 19

00-19(--)-03-19-17-08 18 17 19

00-19(--)-03-19-17-08 18 18 20

00-19(--)-03-19-17-08 18 17 19

00-21(--)-03-21-17-08 19 19 21

00-18(--)-03-18-17-08 20 20 22

06-15(--)-03-15-15-05 21 21 23

anada 03-15(--)-03-15-15-05 21 22 24

(unrelated) 01-14(--)-05-14-14-04 22 23 25

SA 05-15(--)-03-15-15-05 23 22 26

00-22(26)-02-22-12-06 24 24 27

00-25(--)-02-25-12-06 24 13 28

01-20(--)-05-20-14-04 25 25 29

00-26(--)-02-26-12-06 26 26 30

np-21(18)-03-np-17-08 27 27 31

02-22(--)-02-22-17-04 28 28 32

02-22(--)-02-22-17-04 28 28 32

np-15(22)-02-np-21-04 29 29 33

np-np(--)-02-19-15-np 29 30 34

ce coefficient and UPGMA, and comparative results obtained using the PFGE, MLVA, andme, serotype, source of isolation, MLVA allele string, PFGE-Apal, PFGE-AscI and AFLPR, LM-TR 4F/R, LM4b-TR1f/r, LM4b-TR2f/r, and LM4b-TR3f/r. np denotes the absence of

Flu

ores

cenc

e

240 280 320 360 400 440 480 520 560 600 640 680 (bp)

90007000500030001000

90007000500030001000

90007000500030001000

A

B

C

Fig. 3. RepresentativeMLVA results of L. monocytogenes obtained in presence of background DNA prepared from a foodmix, generated in twomultiplex PCR assays I (red peaks) and II(blue peaks). The food mix was an equal mixture of pre-enriched food samples, including sausage (beef and pork), homogenized milk, cooked ham, smoked chicken and lettuce. A,MLVA pattern generatedwith 5% DNA of L. monocytogenes strain LI0523 and 95% DNA of the foodmix; B,MLVA pattern generatedwith 100% DNA of L. monocytogenes strain LI0523; C,MLVA pattern generated with 100% DNA of the food mix.

12 S. Chen et al. / International Journal of Food Microbiology 148 (2011) 8–14

products with minimal nonspecific amplification. It was necessary toincrease the Mg2+ concentration to 4.0 mM and to perform 45 cycles(7 touchdown cycles+38 regular cycles) of PCR in order to detect lowtarget copies in food samples using the multiplex PCR assays. TheMLVA assays produced electropherograms which were simple andclear with well-defined peaks. Fig. 1 illustrates the results of tworepresentative strains of L. monocytogenes of serotypes 1/2b and 4b.

3.2. MLVA analysis of L. monocytogenes strains

A total of 46 strains of L. monocytogenes were analyzed in twomultiplex PCR assays based on the selected six loci, and were alsoanalyzed using PFGE and AFLP analyses. The MLVA and AFLP methodsresulted in 34 unique DNA fingerprint patterns from the 46 L.monocytogenes isolates of 10 serotypes (Fig. 2). The PFGE methodresulted in 29 and 30 unique patterns with restriction enzymes Apaland AscI respectively (Fig. 2). The MLVA results (Fig. 2) successfullydiscriminated between the outbreak isolates and the unrelated food,animal or environmental isolates, and displayed identical MLVApatterns for known outbreak-related isolates (outbreak I isolates:LI0533, 1029-250, 1042-221, 1088-193 and 1087-194; outbreak IIisolates: LI0521, LI0522, LI0524, LI0525, LI0549 and 310). The AFLPmethod showed similar discrimination capacity as the MLVA. ThePFGE analysis with either of the single restriction enzyme was unableto differentiate some of the unrelated strains; and the combined PFGEresults from both restriction enzymes resulted in similar discrimina-tion or clustering capacity to the MLVA or AFLP. One outbreak isolate(1088-193) produced a slightly different AFLP and PFGE-AscI profilesfrom the remaining outbreak strains while the MLVA and the PFGE-ApaI results of the isolate showed the same pattern as the otherstrains from the same outbreak. The results demonstrate that the

Table 2Detection/typing of L. monocytogenes in pre-enrichment-spiked foods. Cells of L.monocytogenes were added into 250 g of food at two levels; two to five sub-samples of25 g each were taken and incubated in 225 mL LEB at 30 °C for 48 h; total genomic DNAwas extracted from the enriched broth for the MLVA analyses.

Food type Strain serotypeand source

No. of samples typedsuccessfully/total no. of samplestested (CFU/25 g food)

0 1–5 10–50

Egg salad 4b, coleslaw 0/2 5/5 2/2Vegetable mix 4b, coleslaw 0/2 5/5 2/2Soft cheese 4b, cheese 0/2 5/5 2/2Beef roast 4b, cheese 0/2 5/5 2/2Smoked salmon 1/2b, salami 0/2 0/5 2/5

MLVA method was highly discriminatory, and able to cluster relatedisolates together and discriminate between the unrelated isolates.

3.3. MLVA analysis of L. monocytogenes in spiked food samples

To test if the MLVA method could be performed in the presenceof other background DNA, DNA preparations (5 ng) from the 46L. monocytogenes isolates were re-analyzed with addition of back-ground DNA (95 ng) which was prepared from 1 mL pellet of anequal-volume mixture of pre-enriched food samples includingsausage (beef and pork), homogenized milk, cooked ham, smokedchicken, and lettuce using the same method as for the individualspiked food samples. TheMLVA fingerprints of the isolates obtained inthe presence of the 95% background DNA were same as those of theirpure cultures (e.g. Fig. 3), suggesting that the MLVA method could beperformed in the presence of other DNA.

To evaluate the limit of detection/typing of the MLVA methodfor L. monocytogenes in foods, food samples were pre-spiked orpost-spiked with L. monocytogenes cells. Pre-spiked samples wereprepared by adding two levels of target cells into 250 g of food; asub-sample of an analytical size (25 g) was then incubated in 225 mLLEB at 30 °C for 48 hours; the resulting broth was used for the MLVAanalysis. The detection/typing limit was 1–5 CFU/25 g for egg salad,vegetable mix, soft cheese and beef roast when evaluated using thepre-enrichment-spiked samples (Table 2). For the smoked salmonsample, successful typing results were obtained from two of the fivereplicates at a spiking level of 10–50 CFU/mL. Tables 3 and 4 show theresults obtained from the post-spiked samples that were prepared byincubating 25 g (or mL) of food in 225 mL LEB at 30 °C for 48 h andthen adding a known number of the target cells at 3 levels to the pre-enriched broth prior to the MLVA analysis. The limits of detection/typing of theMLVAmethod for post-enrichment-spiked sampleswere

Table 3Detection/typing of L. monocytogenes in post-enrichment-spiked foods. 25 g or mL offood was incubated in 225 mL LEB at 30 °C for 48 hours; cells of L. monocytogenes (ATCC19115) were added to the pre-enriched broth; total genomic DNA was extracted fromthe broth for the MLVA analyses.

Food type No. of samples typed successfully/total numberof samples tested (CFU/mL pre-enriched food)

0 102 103 104

Sausage (beef and pork) 0/5 0/6 4/6 6/6Homogenized milk 0/5 0/6 6/6 6/6Cooked ham 0/5 0/6 6/6 6/6Smoked chicken 0/5 0/6 4/6 6/6Lettuce 0/5 0/6 5/6 6/6

Table 4Limit of detection/typing of L. monocytogenes in post-enrichment-spiked foods withdifferent strains of serotypes 4b (ATCC 19115, human), 1/2a (LI0530, raw milk) and 1/2b (LI0514, beef). Samples were prepared and analyzed in the sameway as described inTable 3.

Food type LOD (CFU/mL) of different serotypes

4b 1/2a 1/2b

Sausage (beef and pork) 104 104 104

Homogenized milk 103 104 103

Cooked ham 103 104 103

Smoked chicken 104 104 103

Lettuce 104 104 103

13S. Chen et al. / International Journal of Food Microbiology 148 (2011) 8–14

103 to 104 CFU/mL when evaluated using sausage, homogenized milk,cooked ham, smoked chicken and lettuce samples and the 4b strainATCC 19115 (Table 3). The results were similar when the 1/2a or 1/2bstrains of food sources were used for spiking (Table 4).

3.4. Reproducibility of the MLVA method

The reproducibility of the MLVA method was evaluated byconducting three experiments on three different days and also bytwo lab technicians. The analyses were conducted in two to fivereplicates within a run. The results shown in Table 2 were repeatedthree times as described above; the results presented in Table 3 werecombinations of three experiments with a duplicate for each of thespiked samples in each of the experiments. The results summarized inTable 4 were obtained in a same way as those in Table 3. The intra-and inter-assays were reproducible at or above the limit of detectionlevels.

Sample storage effect was investigated using post-enrichment-spiked food samples stored at 4 °C for 2 days or at −80 °C for twoweeks prior to DNA extraction. Each of the food samples, includingsausage, homogenizedmilk, cooked ham, smoked chicken and lettuce,was prepared in duplicate and spiked with cells of L. monocytogenes(1/2a strain LI0530) at a concentration of 104 CFU/mL. Similar results(not shown) were observed between freshly-prepared samples andthe stored samples except that samples stored at 4 °C for 2 daysresulted in reduced signal with the multiplex PCR II.

4. Discussions

In this study, two multiplex PCR assays were formulated based onsix VNTR loci and were used to subtype a collection of 46 strains of L.monocytogenes of 10 serotypes. Three of the loci (LM-TR1, LM-TR3 andLM-TR4) were developed by Murphy et al. (2007) to subtype foodisolates withmost of them representing serotype1/2a. The other threeloci (LM4b-TR1, LM4b-TR2 and LM4b-TR3) were developed by Miyaet al. (2008) specifically for subtyping 4b strains. The PCR assaysformulated in this study using a combination of the six loci provide asingle MLVA protocol to facilitate subtyping of L. monocytogenesstrains of an unknown serotype. As expected, the MLVA protocolclustered the related isolates together and differentiated the non-related isolates of various serotypes, demonstrating distinctionsamong these isolates based on their origin. The MLVA resultscorrelated well with the AFLP sub-typing data, and the combinedPFGE results from both restriction enzymes AscI and ApaI. The PFGEwith a single restriction enzyme was less discriminatory than eitherAFLP or MLVA, which is consistent with the results of Miya et al.(2008) who reported a higher discrimination power of MLVA thanApaI PFGE for sub-typing L. monocytogens. The MLVA method wassimple, rapid and reproducible with easy-to-interpret and portableresults as compared with PFGE or AFLP. The accuracy of the MLVAresults can be comparable to direct sequencing of the amplicons assuggested by Lindstedt et al. (2003a,b, 2004) since the MLVA assay is

based on specific amplification of a target region, and accurate sizingof the amplicons or number of tandem repeats in an isolate. The PCR-based method is also relatively easy to be standardized, facilitatingdata exchange, in a manner similar to PulseNet. Our results suggestthat the MLVA approach provides an attractive alternative to existingtyping methods, particularly PFGE, for this organism.

To our knowledge this is the first report of MLVA application tosubtyping of L. monocytogenes directly in food samples. The MLVAprotocol used in this study was able to generate accurate, reproduc-ible and high quality results from different types of artificiallycontaminated food samples with an exception of smoked salmon,which appeared to have inhibited the growth of L. monocytogenesduring pre-enrichment. Inhibition of smoked salmon on growth andrecovery of L. monocytogenes was observed by other researchers(Nyachuba et al., 2007). The limits of detection of the MLVA methodwere 1–5 CFU/25 g of food for pre-enrichment-spiked samples and103–104 CFU/mL for post-enrichment-spiked samples, which weresimilar to many of the commercial or reported PCR assays (Badosa etal., 2009; Ennaji et al., 2009; Omiccioli et al., 2009). The application ofan MLVA method directly to food samples is possible because theMLVA method, in principle, is same as a multiplex PCR assay fordetection of a specific pathogen in which the primers are specific andPCR conditions are optimized to produce clear and well-defined PCRproducts. Each of the loci in the multiplexed MLVA method can alsoserve as an internal control for the PCR assay, providing moreconfidence on null-allele results and safe-guarding against false-negative results caused by PCR inhibitors often inherent in foodsamples. The MLVA results can be obtained within 8 h followingstandard pre-enrichment of food samples.

Direct typing of a pathogen in food samples represents a newpossibility for future food safety diagnostics. A PCR method can beused for screening of a particular pathogen in a food sample followingpre-enrichment; a PCR-positive sample is then subject to MLVAanalysis using the same pre-enrichment broth, or even the same DNAextract that is used in the initial screening. The strain type informationcan be obtained by using the MLVA analysis without the need of apure culture, which is not possible by using either PFGE or AFLP. Theaccuracy of the MLVA results obtained directly from food samples canbe compromised when a food sample is contaminated by more thanone strain of L. monocytogenes. This can be overcome by conventionalculture isolation, or by comparison to existing MLVA patterns if thepolymorphism between two strains is limited to single locus and if acomprehensive MLVA database is established. Food contamination bymultiple L. monocytogenes strains is expected to happen rarely. AnMLVA result from a food sample can be compared with the patterns inan active database to determine relatedness of isolates, allowing “real-time” tracking of potential sources of contamination and provision ofearly warning of an outbreak, and supporting control measures toeffectively protect consumers.

Acknowledgments

We are grateful to Honghong Li, Laboratory Services, University ofGuelph for conducting the AFLP analysis, Carlos Leon-Velarde andShailaWadud, Laboratory Services, University of Guelph for preparingthe Listeria monocytogenes cultures and spiked food samples, andMarina Lombos, Ontario Agency for Health Protection and Promotionfor conducting the PFGE analysis.

References

Aarts, H.J., Hakemulder, L.E., Van Hoef, A.M., 1999. Genomic typing of Listeriamonocytogenes strains by automated laser fluorescence analysis of amplifiedfragment length polymorphism fingerprint patterns. International Journal of FoodMicrobiology 49, 95–102.

Badosa, E., Chico, N., Pla, M., Parés, D., Montesinos, E., 2009. Evaluation of ISOenrichment real-time PCR methods with internal amplification control for

14 S. Chen et al. / International Journal of Food Microbiology 148 (2011) 8–14

detection of Listeria monocytogenes and Salmonella enterica in fresh fruit andvegetables. Letters in Applied Microbiology 49 (1), 105–111.

Canadian Food Inspection Agency (CFIA), 2009. Health Hazard Alert. Certain maple leaf,shopsy's and hygrade brand wieners may contain Listeria monocytogenes. http://www.inspection.gc.ca/english/corpaffr/recarapp/2009/20090803e.shtml.

Canadian Food Inspection Agency (CFIA), 2010. Health Hazard Alert. Ready-to-eatcookedmeats produced by Establishment 294 (Smith's Quality Meats) may containListeria monocytogenes. http://www.inspection.gc.ca/english/corpaffr/recarapp/2010/20100611be.shtml.

Chen, Y., Zhang, W., Knabel, S.J., 2007. Multi-virulence-locus sequence typing identifiessingle nucleotide polymorphisms which differentiate epidemic clones and out-break strains of Listeria monocytogenes. Journal of Clinical Microbiology 45 (3),835–846.

Ennaji, H., Timinouni, M., Ennaji, M.M., Ait m'hand, R., Hassar, M., Cohen, N., 2009. Rapiddetection of Listeria monocytogenes in food by polymerase chain reaction. Cell andMolecular Biology (Noisy-le-grand) (55 Suppl.), OL1104–OL1110.

Evans, M.R., Swaminathan, B., Graves, L.M., Altermann, E., Klaenhammer, T.R., Fink, R.C.,Kernodle, S., Kathariou, S., 2004. Genetic markers unique to Listeria monocytogenesserotype 4b differentiate epidemic clone II (hot dog outbreak strains) from otherlineages. Applied and Environmental Microbiology 70, 2383–2390.

Farber, J.M., Peterkin, P.I., 1991. Listeria monocytogenes, a food-borne pathogen.Microbiological Reviews 55 (3), 476–511.

Fonnesbech, V.B., Fussing, V., Ojeniyi, B., Gram, L., Ahrens, P., 2004. High-resolutiongenotyping of Listeria monocytogenes by fluorescent amplified fragment lengthpolymorphism analysis compared to pulsed-field gel electrophoresis, randomamplified polymorphic DNA analysis, ribotyping, and PCR-restriction fragmentlength polymorphism analysis. Journal of Food Protection 67 (8), 1656–1665.

Gombas, D.E., Chen, Y., Clavero, R.S., Scott, V.N., 2003. Survey of Listeria monocytogenesin ready-to-eat foods. Journal of Food Protection 66 (4), 559–569.

Grif, K., Heller, I., Wagner, M., Dierich, M., Würzner, R., 2006. A comparison of Listeriamonocytogenes serovar 4b isolates of clinical and food origin in Austria byautomated ribotyping and pulsed-field gel electrophoresis. Foodborne Pathogensand Disease 3 (1), 138–141.

Guan, S., Xu, R., Chen, S., Odumeru, J., Gyles, C., 2002. Development of a procedure fordiscriminating among Escherichia coli isolates from animal and human sources.Applied and Environmental Microbiology 68 (6), 2690–2698.

Heir, E., Lindtstedt, B.-A., Vardund, T., Wasteson, Y., Kapperud, G., 2000. Genomicfingerprinting of shiga-toxin producing Escherichia coli (STEC) strains: comparisonof pulsed-field gel electrophoresis (PFGE) and fluorescent amplified-fragmentlength polymorphism (FAFLP). Epidemiology and Infection 125, 527–548.

Hofer, E., Ribeiro, R., Feitosa, D.P., 2000. Species and serovars of the genus Listeriaisolated from different sources in Brazil from 1971 to 1997. Memorias do InstitutoOswaldo Cruz 95, 615–620.

Jeffers, G.T., Bruce, J.L., McDonough, P.L., Scarlett, J., Boor, K.J., Wiedmann, M., 2001.Comparative genetic characterization of Listeria monocytogenes isolates fromhuman and animal listeriosis cases. Microbiology 147, 1095–1104.

Lappi, V.R., Thimothe, J., Nightingale, K.K., Gall, K., Scott, V.N., Wiedmann, M., 2004.Longitudinal studies on Listeria in smoked fish plants: impact of interventionstrategies on contamination patterns. Journal of Food Protection 67, 2500–2514.

Lindstedt, B.A., Heir, E., Gjernes, E., Kapperud, G., 2003a. DNA fingerprinting ofSalmonella enterica subsp. enterica serovar Typhimurium with emphasis on phagetype DT104 based on variable number of tandem repeat loci. Journal of ClinicalMicrobiology 41, 1469–1479.

Lindstedt, B.A., Heir, E., Gjernes, E., Vardund, T., Kapperud, G., 2003b. DNAfingerprinting of Shiga-toxin producing Escherichia coli O157 based on Multiple-Locus Variable-Number Tandem-Repeats Analysis (MLVA). Annals of ClinicalMicrobiology and Antimicrobials 2, 12.

Lindstedt, B.A., Vardund, T., Kapperud, G., 2004. Multiple-locus variable-numbertandem-repeats analysis of Escherichia coli O157 using PCRmultiplexing andmulti-colored capillary electrophoresis. Journal of Microbiological Methods 58, 213–222.

Lindstedt, B.A., Tham, W., Danielsson-Tham, M.L., Vardund, T., Helmersson, S.,Kapperud, G., 2008. Multiple-locus variable-number tandem-repeats analysis ofListeria monocytogenes using multicolour capillary electrophoresis and comparisonwith pulsed-field gel electrophoresis typing. Journal of Microbiological Methods 72(2), 141–148.

Miettinen, H., Wirtanen, G., 2006. Ecology of Listeria spp. in a fish farm and moleculartyping of Listeria monocytogenes from fish farming and processing companies.International Journal of Food Microbiology 112, 138–146.

Miya, S., Kimura, B., Sato, M., Takahashi, H., Ishikawa, T., Suda, T., Takakura, C., Fujii, T.,Wiedmann, M., 2008. Development of a multilocus variable-number of tandemrepeat typing method for Listeria monocytogenes serotype 4b strains. InternationalJournal of Food Microbiology 124 (3), 239–249.

Murphy, M., Corcoran, D., Buckley, J.F., O'Mahony, M., Whyte, P., Fanning, S., 2007.Development and application of Multiple-Locus Variable Number of tandem repeatAnalysis (MLVA) to subtype a collection of Listeria monocytogenes. InternationalJournal of Food Microbiology 115 (2), 187–194.

Nightingale, K.K., Windham, K., Wiedmann, M., 2005. Evolution and molecularphylogeny of Listeria monocytogenes isolated from human and animal listeriosiscases and foods. Journal of Bacteriology 187, 5537–5551.

Nyachuba, D.G., Donnelly, C.W., Howard, A.B., 2007. Impact of nitrite on detection ofListeria monocyotgenes in selected ready-to-eat (RTE) meat and seafood products.Journal of Food Science 72, 267–275.

Omiccioli, E., Amagliani, G., Brandi, G., Magnani, M., 2009. A new platform for Real-TimePCR detection of Salmonella spp., Listeria monocytogenes and Escherichia coli O157in milk. Food Microbiology 26, 615–622.

Pagotto, F., Daley, E., Farber, J., Warburton, D., 2001. Isolation of Listeria monocytogenesfrom all food and environmental samples. MFHPB-30. Compendium of AnalyticalMethods, Volume 2. Bureau of Microbiological Hazards, Food Directorate, HealthProducts and Food Branch, Health Canada.

Pagotto, F., Ng, L.K., Clark, C., Farber, J., 2006. Canadian listeriosis reference service.Foodborne Pathogens and Disease 3 (1), 132–137.

Public Health Agency of Canada (PHAC), 2009. Listeria monocytogenes outbreak. FinalUpdate: http://www.phac-aspc.gc.ca/alert-alerte/listeria/listeria_2009-eng.php.

Revazishvili, T., Kotetishvili, M., Stine, O.C., Kreger, A.S., Morris Jr., J.G., Sulakvelidze, A.,2004. Comparative analysis of multilocus sequence typing and pulsed-field gelelectrophoresis for characterizing Listeria monocytogenes strains isolated fromenvironmental and clinical sources. Journal of Clinical Microbiology 42 (1),276–285.

Riikka, O., Keto-Timonen, T., Autio, J., Korkeala, H.J., 2003. An improved amplifiedfragment length polymorphism (AFLP) protocol for discrimination of Listeriaisolates. Systematic and Applied Microbiology 26, 236–244.

Swaminathan, B., Gerner-Smidt, P., 2007. The epidemiology of human listeriosis.Microbes and Infection 9, 1236–1243.

Vazquez-Boland, J.A., Kuhn, M., Berche, P., Chakraborty, T., Dominguez-Bernal, G.,Goebel, W., Gonzalez-Zorn, B., Wehland, J., Kreft, J., 2001. Listeria pathogenesis andmolecular virulence determinants. Clinical Microbiology Reviews 14, 584–640.

Wiedmann, M., 2002. Molecular subtyping methods for Listeria monocytogenes. Journalof AOAC International 85, 524–531.

Wiedmann, M., 2003. ADSA Foundation Scholar Award—an integrated science-basedapproach to dairy food safety: Listeria monocytogenes as a model system. Journal ofDairy Science 86 (6), 1865–1875.

Yichun, L., Lee, M.A., Ooi, E.E., Mavis, Y., Tan, A.L., Quek, H.H., 2003. Molecular typing ofSalmonella enterica serovar typhi isolates from various countries in Asia by amultiplex PCR assay on variable-number tandem repeats. Journal of ClinicalMicrobiology 41, 4388–4394.

Zhang, W., Jayarao, B.M., Knabel, S.J., 2004. Multi-virulence-locus sequence typing ofListeria monocytogenes. Applied and Environmental Microbiology 70 (2), 913–920.