JOURNAL OF CLINICAL MICROBIOLOGY, May 1994, P. 1280-12870095-1 137/94/$04.00+0Copyright © 1994, American Society for Microbiology

Verification of Causal Relationships between Listeria monocytogenesIsolates Implicated in Food-Borne Outbreaks of Listeriosis by

Randomly Amplified Polymorphic DNA PatternsJOHN CZAJKA AND CARL A. BATT*

Department of Food Science, Comell University, Ithaca, New York 14853

Received 20 September 1993/Returned for modification 7 December 1993/Accepted 8 February 1994

Food and clinical isolates of Listeria monocytogenes recovered from four different outbreaks of listeriosis wereanalyzed by their PCR-based randomly amplified polymorphic DNA (RAPD) patterns to verify their causalrelationships. The generation of DNA fingerprints by PCR-based RAPD analysis is a fast and sensitive methodfor the epidemiological tracking and identification of bacteria implicated in food poisoning outbreaks. The L.monocytogenes strains used in the study were obtained from the following four outbreaks: California, 1985,Mexican-style cheese; Canadian Maritime Provinces, 1981, coleslaw; Canada, 1989, brie cheese; and Canada,1989, alfalfa tablets. RAPD profiles were generated by using random 10-mer primers for at least one food andone clinical isolate recovered from each outbreak. Identical profiles for 20 different primers were observed foreach pair of food and clinical isolates from two of the four outbreaks. Isolates from the outbreak involving alfalfatablets exhibited identical patterns for 19 primers; however, primer OPA-1 produced one additional 1.8-kbfragment, designated OPA-1-1.8, that was found in the food isolate but not in the corresponding clinical isolate.Hybridization analysis revealed that the absence of the OPA-1-1.8 polymorphic fragment in the clinical isolatewas due to a deletion of at least 1.8 kb. Loss of the OPA-1-1.8 polymorphic fragment could not be induced byinfective passage of the L. monocytogenes isolate from the alfalfa tablet through a mouse or by growth of thisisolate under selective conditions. This suggests that the isolate recovered from the food was not identical to theisolate recovered from the patient. The ability to produce unique RAPD patterns allows for the discriminationbetween isolates even if they are of the same serotype and multilocus enzyme electrophoretic type.

Listeria monocytogenes is a food-borne pathogen that isubiquitous in nature. As a result, Listeria contamination occurs

quite often in a wide range of foods, such as raw andpasteurized milk, soft cheeses, ice cream, coleslaw, raw meat,fermented sausage, and raw fish (10, 13, 14, 32). Consumptionof contaminated food can cause severe and often fatal infec-tions in susceptible people. Several outbreaks of listeriosis inthe past few years have increased the need for rapid, more

sensitive methods for the detection and identification of L.monocytogenes strains associated with food-borne illness (13,14, 17, 32, 33).The length of time for the clinical symptoms of listeriosis to

become apparent may take from a few days to several weeksafter the ingestion of the contaminated food product. Thisfactor, in combination with the lack of sensitive methods ofidentification, increases the difficulty in tracking and confirm-ing the source of a food-borne infection. The typing methodsused in the past have lacked the ability to consistently verifythat two different bacterial isolates are actually the same strain.Newer techniques have a greater ability to differentiate twoclosely related isolates. Current methods include serotyping,phage typing, multilocus enzyme electrophoresis, randomlyamplified polymorphic DNA (RAPD) analysis, and wholegenome restriction digests (1, 3, 4, 7, 9, 10, 15, 18, 20, 21).RAPD analysis provides a fast and simple method for differ-entiating bacterial strains (6, 7, 12, 19. 20, 35).RAPD polymorphisms can be used to discriminate between

two otherwise indistinguishable isolates of the same serotype(7, 19). The establishment of identical RAPD patterns allows

* Corresponding author. Mailing address: 413 Stocking Hall, Cor-nell University, Ithaca, NY 14853. Phone: (607) 255-2896. Fax: (607)255-8741.

for the confirmation of a bacterium recovered from a foodsource that is suspected of being the causal agent in a foodpoisoning outbreak. Various genetic events give rise to poly-morphisms, including primer-template mismatches, deletionsor insertions within the primer-binding site, and deletions or

insertions between two primer-binding sites, which can eitherincrease the fragment size or prevent amplification (if thedistance between two primer-binding sites is greater than 4.0kb).One of the largest documented listeriosis outbreaks oc-

curred in 1985 in the state of California. Jalesco cheese, a

Mexican-style soft cheese, was implicated in the epidemic. L.monocytogenes was recovered from a sample of a patient'srefrigerated Mexican-style cheese as well as from unopenedpackages of cheese recovered from the factory. It was specu-lated that the source of L. monocytogenes was raw milk, whichwas mixed with pasteurized milk during the cheese-makingprocess. The L. moniocytogenes isolates recovered from thecheese and the patients were serotype 4b and electrophoretictype (ET) 1 (17). Another outbreak of listeriosis, whichoccurred in 1981 in Nova Scotia, Canada, implicated coleslawas the vehicle of infection. L. monocytogenes was recoveredfrom a sample of a patient's refrigerated coleslaw and fromunopened packages obtained from the factory. Further inves-tigations revealed that the cabbage used for processing was

contaminated with manure fertilizer obtained from a flock ofsheep with listeriosis. The L. monocytogenes isolates recoveredfrom the coleslaw and the patient were again serotype 4b andET 1 (31). One isolated case of listeriosis that occurred inCanada involved the consumption of brie cheese (10). L.monocytogenes isolates, serotype 1/2b, ET 22 (28), were recov-

ered from the patient and unopened packages of the brand ofsoft cheese that he had consumed (11). A second isolated case

1280

Vol. 32, No. S

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

RAPD PATTERNS OF L. MONOCYTOGENES 1281

TABLE 1. L. monocytogenes strains used for RAPD analysis

Outbreak Food isolates Clinical isolates

Year Source No. Sero- ET No. Sero- ETtype type

1981 Coleslaw PP-2 4b 1 PP-37 4b 1

1985 Mexican-style PP-22 4b 1 PP-23 4b Icheese

1989 Brie cheese PP-395 1/2b 22 PP-447 1/2b 22

1989 Alfalfa tablets PP-295 4b 1 PP-448 4b IPP-296 4b 1PP-297 4b 1PP-298 4b IPP-299 4b 1PP-397 4b 1PP-861 4b 6

of listeriosis involved the consumption of alfalfa tablets con-taminated with L. monocytogenes. The isolates recovered fromalfalfa tablets and the patient were serotype 4b, ET 1 (10). Oneadditional isolate of L. monocytogenes, serotype 4b, ET 6, wasrecovered from the alfalfa tablets. No other isolates of L.monocytogenes were recovered from food samples present inthe patient's home. The food and clinical isolate, serotype 4b,ET 1, were thought to be identical to the isolates recovered inthe Nova Scotia outbreak (11).The present study was carried out to further confirm the

previous identification of causal sources implicated in listerio-sis outbreaks. The RAPD profiles of the food and clinicalisolates of L. monocytogenes from four different listeriosisoutbreaks were analyzed. The results support the conclusionthat the causative strain was recovered in three of the fourlisteriosis outbreaks.

MATERIALS AND METHODS

Bacterial strains and media. The L. monocytogenes strainsused in the study and their sources are listed in Table 1. Theseisolates were obtained from the Health Protection Branch ofHealth Canada and the Laboratory Center for Disease ControlCanada. Multilocus enzyme electrophoresis typing of strainswas conducted under the same conditions in the one laboratory(28). Strains were cultured in tryptic soy broth (TSB; Sigma, St.Louis, Mo.), Listeria enrichment broth (LEB; Difco, Detroit,Mich.), or brain heart infusion (BHI) broth (Difco), as notedbelow.The bacteria to be used for mouse inoculations were cul-

tured in BHI broth at 37°C for 20 h. Cells were washed withpotassium phosphate-buffered saline (0.01 M; pH 7.4) andwere resuspended in sterile water to approximate concentra-tions of 10"' CFU/ml.

Cell lysis. The Listeria strains were grown at 37°C in TSB toan optical density at 600 nm of 1.5. A 250-p.l aliquot of theculture was pelleted by centrifugation at 12,000 x g for 5 min,and the medium was discarded. The cell pellet was washed withsaline (0.9% [wt/vol] NaCl) and was centrifuged. The bacterialpellet was then resuspended in 100 of 1 x PCR buffer (20mM Tris-HCl [pH 8.4], 50 mM KCl, 4 mM MgCl2). A 4-p.laliquot of lysozyme (10 mg/ml) was added, and the mixture wasincubated at 37°C for 30 min. A 1-p.l aliquot of proteinase K(20 mg/ml) was added, and the mixture was incubated at 55°C

for 1 h. The solution was then boiled for 10 min and was usedfor RAPD reactions and dot blot hybridization.

Amplification conditions. A kit of 20 different primers (kitA) was purchased from Operon (Alameda, Calif.). The reac-tions were performed in a 1 x PCR buffer containing 20 mMTris-HCl (pH 8.4), 50 mM KCl, 4 mM MgCl,, 100 F.M (each)deoxynucleotide triphosphate, 0.2 p.M primer, I [l of celllysate, and 1 U of Taq DNA polymerase (GIBCO BRL, GrandIsland, N.Y.). RAPD amplification conditions were modifiedslightly from those of Williams et al. (35). RAPD amplifica-tions were carried out by using a Hybaid TRI thermal cycler.The cycling parameters were 45 s at 94°C, 1 min at 36°C, and2 min at 72°C for a total of 45 cycles. RAPD products wereanalyzed by using a 1.0% agarose gel and TBE (89 mMTris-borate, 2 mM EDTA [pH 8.2]) buffer at 100 V for 45 min.

Dot blot and Southern hybridizations. A 10-plI aliquot ofunrestricted lysate was denatured by the addition of 10 RI of0.8 M NaOH at 37°C for 10 min. A 90-plI aliquot of water wasadded to the denatured lysate mixture, which was then neu-tralized by the addition of 100 .I1 of 2 M ammonium acetate.The 200-,ul solution was transferred to a nylon membrane(Micron Separations, Inc., Westborough, Mass.) with theBio-Dot microfiltration unit (Bio-Rad, Rockville Centre,N.Y.). DNA was immobilized on the nylon membrane bybaking the membrane at 65°C for 30 min, and the DNA wasused for the dot blot hybridization.Chromosomal DNA for Southern hybridization was pre-

pared as described by Pitcher et al. (30). Purified chromosomalDNA was then digested with ApoI (New England Biolabs,Beverly, Mass.), and the restricted fragments were separated ina 1.0% agarose gel and TBE buffer. The DNA was denaturedby soaking the gel in 4 volumes of 1.0 M NaCl-0.5 M NaOHtwice for 10 min. The gel was neutralized by soaking it in 4volumes of 0.5 M Tris-HCl (pH 7.5)-1.5 M NaCl twice for 10min. The DNA was transferred to a nylon membrane bycapillary transfer blotting and was then immobilized on themembrane by baking at 65°C for 30 min.The RAPD product to be used as a probe was excised from

the agarose gel and was purified by using the USBioclean MPDNA purification kit (United States Biochemicals, Cleveland,Ohio). The purified DNA was used as a template for a secondPCR, with the same primer used in the initial amplification, toincorporate digoxigenin-11-dUTP. Conditions were identicalto those of the initial PCR, except that the concentration ofdTTP was reduced to 80 p.M and 20 p.M digoxigenin-1 1-dUTP(Boehringer Mannheim, Indianapolis, Ind.) was added. Thedigoxigenin-11-dUTP-labelled RAPD band was excised from a1.0% low-melting-point agarose TBE gel (International Bio-technologies Inc., New Haven, Conn.), and a 100-plI aliquot ofwater was added to the excised DNA. The mixture was boiledfor 10 min, and a 10-pl aliquot was used as a probe for dot blotand Southern hybridizations. The molecular weight markers(bacteriophage X HindIII) were 3' end labelled by using theKlenow fragment of DNA polymerase I and dUTP-1 1-digoxi-genin. The hybridization and nonradioactive detection wereconducted according to the protocol provided with the alkalinephosphatase-based Genius 3 Nucleic Acid Detection Kit(Boehringer Mannheim).Mouse inoculations. Five-week-old mice were injected in-

traperitoneally with a 250-p.l aliquot of a 10-mg/ml solution ofcarrageenan type II (Sigma). After 24 h, three mice wereinoculated intraperitoneally with i0' CFU of L. monocytogenesand two mice were inoculated perorally with 109 CFU of L.monocytogenes, administered with a sterile 20-gauge 1.5-in.(3.8-cm) animal-feeding needle. The three mice that wereinoculated intraperitoneally died within 48 h postinoculation.

VOL. 32, 1994

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

1282 CZAJKA ANDBATC

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

FIG. 1. RAPD patterns for the brie cheese isolate PP-395 and the corresponding clinical isolate PP-447. Lanes I and 20, bacteriophage A DNAdigested with EcoRI and HindIII (selected molecular size markers [in kilobases] are noted on the left); even-numbered lanes represent the patternsproduced by isolate PP-395; odd-numbered lanes represent the patterns produced by the clinical isolate PP-447. The primers OPA-1 throughOPA-5 and OPA-7 through OPA-10 were used to generate the RAPD profiles for each pair of isolates and are shown in sequential order.

The two mice that were inoculated perorally were killed bycarbon dioxide asphyxiation after 72 h.The brain, liver, and spleen from each mouse were asepti-

cally removed by dissection, macerated, and individually cul-tured in 10 ml of TSB at 37°C for 12 h. The cultures were thenstreaked on Listeria agar (Oxford formulation; Unipath Ltd.,Basingstoke, United Kingdom). Listeria colonies were identi-fied by blackening the agar after 18 h of incubation at 37°C.

RESULTSFood and clinical isolates from four outbreaks of listeriosis

were analyzed in the study. At least one isolate was recoveredfrom each of the suspected foods and was hypothesized to bethe causal strain responsible for the corresponding outbreak.The isolates from the outbreak in 1981 involving coleslaw, theone in 1985 involving Mexican-style cheese, and the one in1989 involving alfalfa tablets were serotype 4b, ET 1. Therewas one additional isolate, PP-861, recovered from the alfalfatablets that was serotype 4b, ET 6 (28), which distinguishedthis isolate from the clinical isolate. The isolates from theoutbreak in 1989 involving brie cheese were serotype 1/2b, ET22 (28) (Table 1).A total of 20 different 10-mer primers (kit A) were used for

the RAPD analysis of the food and clinical L. monocytogenesisolates that were believed to be identical within each of thefour outbreaks. Eighteen of the primers produced clear, repro-ducible patterns for each isolate. OPA-6 produced only one ortwo faint bands that were not consistently observed in repeatedanalysis. OPA-1 9 generated patterns that varied in the numberof fragments produced for an individual lysate from onereaction to the next. The number of RAPD bands produced fora given primer ranged from 2 to 12, with molecular sizesranging from 0.1 to 4.0 kb. The RAPD patterns generated withmost primers were consistently produced for isolates of thesame serotype. These results correspond to those of previousstudies, which reported that a selected group of primers yieldsRAPD patterns that are the same for representatives within agiven serotype (7), while other primers yield more discriminat-ing patterns that differ between strains within a serotype (20).For example, primer OPA-1 differentiated the serotype 1/2bisolates from the serotype 4b isolates. The difference in the

OPA-1 RAPD patterns for each serotype was exemplified bythe presence of the 0.8- and 0.85-kb bands unique to theserotype 1/2b isolates (Fig. 1, lanes 2 to 3) and the 0.3-kb bandunique&to the serotype 4b isolates (Fig. 2, lanes 2 to 9). RAPDanalysis of the isolates from the outbreak involving alfalfatablets demonstrated the ability of RAPD analysis to discrim-inate between L. monocytogenes serovars of the same enzymeET as well as between serovars of different ETs. For example,primer OPA-1 (Fig. 2) clearly differentiated the food isolatesPP-295, PP-296, PP-297, PP-298, PP-299, and PP-397 from theclinical isolate PP-448, all of which were serotype 4b and ET 1.Primer OPA-1 also differentiated the alfalfa tablet isolatePP-861, also serotype 4b, but ET 6 (10), from the other alfalfatablet isolates, all of which were serotype 4b, ET 1.

Figure 1 shows that identical RAPD patterns were producedfor the food and clinical isolates involved in the outbreakinvolving brie cheese. Isolates recovered from that outbreakwere serotype 1/2b and ET 22. Primer OPA-1 generatedidentical patterns for the food isolate PP-395 and the clinicalisolate PP-447. A total of five bands ranging in size from 0.8 to1.4 kb were generated for each isolate for OPA-1.The RAPD patterns observed for the isolates from the

outbreaks involving Mexican-style cheese and coleslaw wereidentical for the corresponding pairs of food and clinicalisolates within each outbreak. For example, primer OPA-13generated three distinct bands, approximately 0.9, 1.1, and 1.9kb in size, that were identical for both the Mexican-style cheeseisolate PP-22 and the clinical isolate PP-23 (Fig. 3). The RAPDpatterns for the coleslaw isolate and the related clinical isolatewere identical for all 20 primers. Primer OPA-20, for example,generated clear, identical patterns for the coleslaw isolate PP-2and the clinical isolate PP-37 (Fig. 4). Bands of very lowintensity were not consistently reproducible and were nottaken into consideration when the RAPD patterns were com-pared. Primers OPA-1 through OPA-20 generated identicalpatterns for the Mexican-style cheese and coleslaw isolates,anditherefore could not further segregate the two isolates,further supporting their identities.RAPD analysis of the food and clinical isolates from the

outbreak involving alfalfa tablets revealed that at least twodifferent strains of L. monocytogenes were present in the alfalfa

18 19 20

2.01.30.8

J. CLIN. MICROBIOL.

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

RAPD PATTERNS OF L. MONOCYTOGENES 1283

1 2 3 4 S 6 7 8 9 1011 121314 1516 17 18192i21 ;223 42526

2.00.8

2.00.8

27 28 29 30 31 32 33 34 35 36 37 36 39 40 41 42 43 4546 47 4649 St St 2

FIG. 2. RAPD patterns for L. monocytogenes alfalfa food isolates PP-85, PP-295, PP-296, PP-297, PP-298, PP-299, and PP-397 and thecorresponding clinical isolate PP-448. Lanes 1, 26, 27, and 52, bacteriophage X DNA digested with EcoRI and Hindlll (selected molecular sizemarkers [in kilobases] are noted on the left); lanes 2 to 9, OPA-1; lanes 10 to 17, OPA-2; lanes 18 to 25, OPA-3; lanes 28 to 35, OPA-4; lanes 36to 43, OPA-5; lanes 44 to 51, OPA-7; lanes 2, 10, 18, 28, 36, and 44, PP-448; lanes 3, 11, 19, 29, 37, and 45, PP-397; lanes 4, 12, 20, 30, 38, and46, PP-861; lanes 5, 13, 21, 31, 39, and 47, PP-295; lanes 6, 14, 22, 32, 40, and 48, PP-296; lanes 7, 15, 23, 33, 41, and 49, PP-297; lanes 8, 16, 24,34, 42, and 50, PP-298; lanes 9, 17, 25, 35, 43, and 51, PP-299.

tablets. Isolate PP-397 was serotype 4b and ET 1, which werethe same as those for the clinical isolate PP-448, while isolatePP-861 was serotype 4b, ET 6. In addition to isolates PP-397and PP-861, five additional isolates recovered from the sameprimary enrichment from the alfalfa tablets were analyzed(Table 1). The RAPD patterns for all five additional isolates,PP-295, PP-296, PP-297, PP-298, and PP-299, were identical tothat for isolate PP-397. The patterns for isolate PP-861 werevery similar to those for the other alfalfa tablet isolates, butthey were not identical (Fig. 2). Isolate PP-861 generated thefollowing polymorphisms in the RAPD patterns: OPA-1, 1.3kb; OPA-2, 1.2 kb; OPA-7, 0.7 kb; OPA-9, 0.5 kb; OPA-16, 2.3kb. Thus, the RAPD patterns discriminated isolate PP-861from the other food isolates within the same serotype.

1 2 3 4 66 7 8 9 11

2.01.30.8

Identical RAPD patterns for food isolates PP-397, PP-295,PP-296, PP-297, PP-298, PP-299 and the clinical isolate PP-448were observed for primers OPA-2 through OPA-5 and OPA-7through OPA-20; OPA-6, as mentioned previously, did notgenerate bands for any of the isolates from the outbreakinvolving alfalfa tablets. The RAPD patterns produced byprimer OPA-1 contained a single 1.8-kb polymorphism, desig-nated OPA-1-1.8, that was present in all of the alfalfa tabletisolates except isolate PP-861 (Fig. 2).

Polymorphisms in RAPD patterns can arise from pointmutations within the primer-binding site as well as insertionsor deletions between primer-binding sites. In the outbreakinvolving alfalfa tablets, clinical isolate PP-448 differed fromthe implicated food isolate by the presence of the OPA-1-1.8

Lo 11 12 13 14 15 16 17 18 19 2021 =T."I

FIG. 3. RAPD patterns for the Mexican-style cheese isolate PP-22 and the corresponding clinical isolate PP-23. Lanes 1 and 22, bacteriophageX DNA digested with EcoRI and HindIll (selected molecular size markers [in kilobases] are noted on left); even-numbered lanes represent thepatterns produced by isolate PP-22; odd-numbered lanes represent the patterns produced by isolate PP-23. The primers OPA-I1 through OPA-20were used to generate these RAPD profiles for each pair of isolates and are shown in sequential order.

VOL. 32, 1994

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

1284 CZAJKA ANDBATI

2.01.3

FIG. 4. RAPD patterns for the coleslaw isolate PP-2 and the corresponding clinical isolate PP-37. Lanes 1 and 22, bacteriophage X DNAdigested with EcoRI and HindlIl (selected molecular size markers [in kilobases] are noted on the left); even-numbered lanes represent the patternsproduced by isolate PP-2; odd-numbered lanes represent the patterns produced by isolate PP-37. The primers OPA-1 1 through OPA-20 were usedto generate the RAPD profiles for each pair of isolates and are shown in sequential order.

polymorphism. In order to determine the nature of the OPA-1-1.8 polymorphism, dot blots and Southern hybridizationswere performed on cell lysates and genomic digests, respec-tively. Ten food and four clinical isolates from all four out-breaks were probed in the hybridizations with the L. monocy-togenes food isolate PP-397 generated OPA-1-1.8. Digox-igenin-11-dUTP, incorporated into the fragment by PCR am-plification, was used for colorimetric detection of the hybrids.Dot blot hybridizations revealed that the probe bound only tothe food isolates PP-295, PP-296, PP-297, PP-298, PP-299, andPP-397, while the other eight isolates were negative for probebinding (data not shown). These were the only alfalfa tabletisolates that contained the OPA-1-1.8 RAPD fragment.Genomic DNA digested with ApoI was used for the South-

ern hybridization. OPA-1-1.8 hybridized only to a single frag-ment with a molecular size of approximately 1.8 kb; thisfragment was found in the food isolate PP-397 but not in theclinical isolate PP-448 (Fig. 5). These results indicate that theadditional band is not a result of a small genetic differencebetween the food and clinical isolates, such as a point mutationor a deletion within the primer-binding site, but rather is adeletion of at least 1.8 kb. It is only coincidental that the sizeof the ApoI restriction fragment is identical to the size of theOPA-1-1.8 RAPD product.The food isolate in the outbreak involving alfalfa tablets may

have given rise to the clinical isolate by a spontaneous deletion,which may have included one or both of the OPA-1-bindingsites, during its cycle of infection or subsequent isolation onmicrobiological medium. Experimental conditions were estab-lished in an attempt to simulate the selection conditions duringisolation that may have induced this deletion. Food isolatePP-397 was cultured at 37 and 42°C in either TSB or LEB for12 h and was subcultured for an additional 12 h. This subcul-turing step was repeated 10 times for a total of approximately120 generations. RAPD analysis of 10 isolates recovered after120 generations revealed no loss of the OPA-1-1.8 polymor-phism. Mice were infected with strain PP-397 in order tosimulate infective passage. Three mice were infected intra-peritoneally, and two mice were infected perorally. L. mono-cytogenes was recovered from the livers, spleens, and brains ofthe mice infected intraperitoneally and perorally. Two isolatesrecovered from the livers, spleens, and brains of a mouseinfected intraperitoneally and another mouse infected per-

orally were tested by RAPD analysis. Each of the isolatesrecovered carried the OPA-1-1.8 polymorphism (data notshown).

DISCUSSION

Food and clinical isolates of L. monocytogenes from fourdifferent outbreaks were tested by RAPD analysis to deter-mine whether the original identification of the causative strainrecovered from the food was correct. Reproducible, identicalRAPD patterns were observed for each related food andclinical isolate in the outbreaks involving Mexican-style cheese,coleslaw, and brie cheese. Eighteen of 20 primers generatedclear patterns; primers OPA-6 and OPA-19 gave inconclusive

1 2 3

FIG. 5. Southern hybridization of Apol-digested genomic DNAwith the digoxigenin-labelled 1.8-kb polymorphism produced by alfalfatablet isolate PP-397 with OPA-1. Lane 1, bacteriophage X DNAdigested with HindIll; lane 2, strain PP-397; lane 3, clinical isolatePP-448.

J. CLIN. MICROBIOL.

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

RAPD PAT-FERNS OF L. MONOCYTOGENES 1285

results. Our data substantiate the results of restriction endo-nuclease analysis conducted by Wesley and Ashton (34) sug-gesting that the food isolates from the coleslaw and Mexican-style cheese were the causal sources for the 1981 Canadian andthe 1985 Californian outbreaks, respectively. The RAPD anal-ysis of the isolates implicated in the outbreak involving briecheese also resulted in identical RAPD profiles for the foodand clinical isolates, substantiating the previous confirmationof that causal source (31).RAPD analysis did not differentiate the food isolates from

the outbreaks involving coleslaw and Mexican-style cheese.These data are contradictory to the results obtained by Wesleyand Ashton (34), suggesting that the two isolates were differentstrains. Differences in restriction patterns can result fromsimilar genetic changes that produce RAPD polymorphisms,such as deletions or insertions between restriction sites andpoint mutations within the restriction site. One additionalhypothesis for differences in the restriction patterns is thepossibility of either the acquisition or the loss of a DNAmethyltransferase. DNA methyltransferase genes are oftenincorporated into bacteriophages (16) and plasmids (23). Theloss or gain of a lysogenic phage or plasmid containing amethyltransferase gene may occur under different environmen-tal conditions. This type of change may result during the cycleof infection or isolation from the contaminated food. Twoisolates of the same strain, growing under different conditions,may become divergent in their methylation patterns because ofthe different selective pressures. Changes in the methylationpattern from one isolate to the next may result in differentrestriction patterns. The sources of the differences in restric-tion patterns are unknown, and without further analysis, thegenetic basis for these differences remains unclear. The num-ber or extent of genetic differences which define two isolates asbeing two different strains is not easily resolved (8). However,the detection of relatively subtle differences will increase as thesensitivity of subtyping assays increases. It is also possible thatthe primers used for RAPD analysis did not bind to areasadjacent to or within the polymorphism, and no differencebetween isolates was detected.RAPD analysis of the food and clinical isolates from the

outbreak involving alfalfa tablets resulted in patterns that wereidentical for all primers tested except OPA-1. PolymorphismOPA-1-1.8 was observed only in the food isolates. Bands ofvery low intensity varied between reactions. The inability toconsistently visualize these bands may have been due to thelimited sensitivity of ethidium bromide staining. Another pos-sible explanation for the lack of reproducibility of these verylow intensity bands is that the primer may have preferentiallybound to primary target sites (exact matches) during the initialamplification, thereby creating more of these target sites forlater amplification cycles, resulting in bands of greater inten-sity. This essentially titers out primer that would have bound tothe secondary target sites (one or more mismatches) generat-ing the lower-intensity bands. In addition, the fidelity of primerbinding is dependent on PCR conditions, such as the magne-sium chloride concentration. An increase in the magnesiumchloride concentration to 4 mM resulted in an increase in thenumber and intensity of RAPD bands (24). The increase in thenumber of RAPD bands is a result of the primer binding totargets with one or more mismatches. Concentrations ofmagnesium chloride of less than 4 mM resulted in the disap-pearance of the OPA-1 -1.8 polymorphism produced by OPA-1with the alfalfa tablet isolates, serotype 4b, ET I (data notshown).The nonisotopically labelled primer OPA-1-1.8 bound to

only the food isolates in both the dot blot and Southern

hybridizations under stringent hybridization conditions, sug-gesting that the OPA-1-1.8 polymorphism comprises DNA thatis completely absent from the clinical isolat.- This observationnegates the possibility that the polymorphism was the result ofa small genetic change such as a point mutation within theprimer-binding site. The possibility of a larger genetic alter-ation such as an insertion or deletion between primer-bindingsites is unlikely. Theoretically, even under high-stringencyconditions, a probe comprising a sequence of only 100 consec-utive nucleotides identical to the target sequence should showa positive result for the clinical isolate (5).These hybridization data suggest that the OPA-1-1.8 poly-

morphism is a segment of DNA that is unique to the foodisolates PP-295, PP-296, PP-297, PP-298, PP-299, and PP-397.Experimental tests were conducted in an attempt to induce thisdeletion. L. monocytogenes often contains plasmids (27), whichmay be a possible source for RAPD bands. The loss ofplasmids can occur quite readily, thereby explaining the loss ofa RAPD fragment. Isolates recovered from the alfalfa tabletswere repeatedly subcultured in TSB or LEB in an attempt toinduce the deletion of the OPA-1-1.8 polymorphism. TSB wasused as a growth medium to determine whether the cause ofthis deletion was a simple case of repeated culturing undernonselective conditions. LEB was used as a selective mediumin an attempt to cure the strain of any plasmids resulting fromthe presence of acriflavine, a known plasmid-curing agent (11).Temperatures of 37 and 42°C were also tested as possibleselective agents that might induce plasmid loss. RAPD analysisof the individual isolates after 120 generations resulted inpatterns that were identical to the original patterns for isolatescontaining the OPA-1-1.8 polymorphism. Although this doesnot eliminate plasmid curing as the source of the polymor-phism, it suggests that plasmid loss may not be as frequent oras easily induced by most isolation protocols for L. monocvto-genes as has been suspected.A second hypothesis suggests that the deletion of this

sequence may have been selected for during the infectiveprocess. The transformation of specific types of Pneumococcusspecies during infection demonstrates the induction of inher-itable and specific alterations in cell structure and function.The attenuated, nonencapsulated, avirulent variant was trans-formed in vitro to a fully encapsulated, virulent strain, whichwas indicated by the resulting infectivity (2). Another possibil-ity is the transposition of an insertion element. Insertionelements have been directly linked to the reversible expressionof a virulence antigen in Citrobacterffreundii, which occurs at ahigh frequency (26). A high rate of antigenic variation has beenobserved in Borrelia hermsii, resulting in frequent changes ofserotype within one strain. Antigenic variation in B. hermsii isa direct result of DNA sequence rearrangement (22). This typeof frequent genetic recombination is a possible source of DNAloss. Examples of virulence and antigen switching such as thesesupport the possibility of an infection-induced change withinthe alfalfa tablet isolate containing the OPA-1-1.8 fragment,giving rise to the clinical isolate.

In an attempt to simulate infection conditions, mice wereimmunocompromised and inoculated intraperitoneally andperorally. Mice inoculated intraperitoneally developed symp-toms of listeriosis and died within 48 h, confirming thevirulence of the isolates. The brains, livers, and spleens wereremoved from the mice, and L. monocytogenes was successfullyisolated from all three organs from mice inoculated intraperi-toneally and perorally. However, all isolates carried the OPA-1-1.8 polymorphism. This suggests that the cause of thedeletion may not be due to the selective agents present during

VOL. 32, 1994

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

1286 CZAJKA AND BA7F[

infection. Although the food and clinical strains are closelyrelated, they appeared to be distinct and stable strains.

Previous RAPD typing studies have involved the use of onlyone primer to type Listeria strains (19, 20), and the discrimi-nating ability of primer HLWL74 is limited. Many of theRAPD profiles were similar, differing only in one or two bandsof a total of seven or eight bands. The primer discriminatesamong most isolates of different serotypes; however, twogroups of isolates, each group having representatives of fourdifferent serotypes, had identical RAPD patterns. This indi-cates the need for multiple primers when typing L. monocyto-genes isolates. We have previously demonstrated that the use ofas few as two primers clearly discriminates between isolates ofdifferent serotypes (7). The use of multiple primers generatesa greater number of discriminatory bands, increasing confi-dence in the accuracy of the results.

Preliminary tests in our laboratory of various lysis methodshave revealed that the boiling of cells may not providesufficient lysis. The use of boiled lysates generated variableresults from one amplification to the next (data not shown).We have also seen variable results in RAPD patterns whendifferent numbers of cells were used for the lysate. Variabilityin the efficiency of cell lysis leads to variability in the amount oftemplate available for primer binding. Changes in templateconcentration result in the appearance and disappearance ofRAPD bands (24, 25). The method for RAPD amplificationsof Mazurier et al. (19) did not include cell lysis prior to PCRamplification. Cells were added directly to the reaction tubeand were heated at 94°C for 4 min (19). Differences in theamount of cells lysed from one reaction to the next could beanother possible source of the absence of one or two RAPDbands.RAPD analysis provides a method for the classification of L.

monocytogenes strains into groups more specific than serotypesand possibly even more specific than ET. RAPD analysisperformed with selected primers can identify and differentiateclosely related strains of L. monocytogenes. Previous studieshave demonstrated that clinical isolates of L. monocytogenesserotypes 1/2a, 1/2b, 1/2c, 4a, and 4b can be further subdividedinto 45 ETs (29). RAPD profiles clearly differentiated the twoisolates PP-397 and PP-861, which were of the same serotypebut of a different ET. Five of a total of 20 primers generateddifferent RAPD patterns for each isolate. Twenty-five percentof all primers used were able to discriminate between twostrains of different ETs. Furthermore, 1 of 20 primers discrim-inated between two isolates that were of the same ET.

If the possible number of RAPD patterns were estimated onthe basis of an average of six bands per primer and differencesin RAPD patterns were scored on the basis of only a singleband change, there would be a possible total of 720 differentpatterns for one primer. If changes in more than one band orband sizes were scored, the total number of possible bandswould increase significantly. The large number of potentialpolymorphic bands suggests that RAPD fingerprinting is a verysensitive method of isolate discrimination.The isolate recovered from the alfalfa tablets was previously

assumed to be the causative agent. However, results of RAPDanalysis conflicted with this initial indictment. Although thefood isolate was clearly virulent, the actual food source thatgave rise to this case of listeriosis remains to be identified. It isalso possible that the causative strain was present in, but notrecovered from, the alfalfa tablets. RAPD analysis of the foodand clinical isolates from the outbreaks involving Mexican-style cheese, coleslaw, and brie cheese substantiated the orig-inal confirmation of the causal agent implicated in eachoutbreak. The RAPD technique provides a fast and simple

method for the epidemiological confirmation of food-bornebacterial outbreaks.

ACKNOWLEDGMENTS

This work was supported by the Northeast Dairy Foods ResearchCenter; a grant from the Cornell Center for Advanced Technology inBiotechnology, which is sponsored by the New York State Science andTechnology Foundation, a consortium of industries, and the NationalScience Foundation; and New York State Hatch Funds.We thank Pina Fratamico (USDA-ARS-ERRC) for assistance with

the mouse lethality studies and Pearl Peterkin and William Yan of theHealth Protection Branch of Health Canada for providing the L.monocytogenes strains tested in the study.

REFERENCES1. Audurier, A., A. G. Taylor, B. Carbonelle, and J. McLauchlin.

1984. A phage typing system for Listeria monocytogenes and its usein epidemiological studies. Clin. Invest. Med. 7:229-232.

2. Avery, 0. T., C. M. Macleod, and M. McCarty. 1944. Studies onthe chemical nature of the substance inducing transformation ofpneumococcal types. J. Exp. Med. 79:137-158.

3. Baloga, A. O., and S. K. Harlander. 1991. Comparison of methodsfor discrimination between strains of Listeria monocytogenes fromepidemiological surveys. Appl. Environ. Microbiol. 57:2324-2331.

4. Boerlin, P., J. Rocourt, and J.-C. Piffaretti. 1991. Taxonomy of thegenus Listeria by using multilocus enzyme electrophoresis. Int. J.Syst. Bacteriol. 41:59-64.

5. Bolton, E. T., and B. J. McCarthy. 1962. A general method for theisolation of RNA complementary to DNA. Proc. Natl. Acad. Sci.USA 48:1390.

6. Brousseau, R., A. Saint-Onge, G. Prefontaine, L. Masson, and J.Cabana. 1993. Arbitrary primer polymerase chain reaction, apowerful method to identify Bacillus thuringiensis serovars andstrains. Appl. Environ. Microbiol. 59:114-119.

7. Czajka, J., N. Bsat, M. Piani, W. Russ, K. Sultana, M. Wiedmann,R. Whitaker, and C. A. Batt. 1993. Differentiation of Listeriamonocytogenes and Listeria innocua by 16S rDNA and intraspeciesdiscrimination of Listeria monocytogenes strains by random ampli-fied polymorphic DNA polymorphisms. Appl. Environ. Microbiol.59:304-308.

8. Daniels, D. L. 1990. Constructing encyclopedias of genomes, p.43-51. In K. Drlica and M. Riley (ed.), The bacterial genome.American Society for Microbiology, Washington, D.C.

9. Estela, L. A., J. N. Sofos, and B. B. Flores. 1992. Bacteriophagetyping of Listeria monocytogenes cultures isolated from seafoods. J.Food Prot. 55:13-17.

10. Farber, J. M., A. 0. Carter, P. V. Varughese, F. E. Ashton, andE. P. Ewan. 1990. Listeriosis traced to the consumption of alfalfatablets and soft cheese. N. Engl. J. Med. 322:338.

11. Farber, J. M., and P. I. Peterkin. 1991. Listeria monocytogenes, afood-borne pathogen. Microbiol. Rev. 55:476-511.

12. Fekete, A., J. A. Bantle, S. M. Hailing, and R. W. Stich. 1992.Amplification fragment length polymorphism in Brucella strains byuse of polymerase chain reaction with arbitrary primers. J. Bacte-riol. 174:7778-7783.

13. Fleming, D. W., S. L. Cochi, K. L. MacDonald, J. Brondum, P. S.Hayes, B. D. Plikaytis, M. B. Holmes, A. Audurier, C. V. Broome,and A. L. Reingold. 1985. Pasteurized milk as a vehicle of infectionin an outbreak of listeriosis. N. Engl. J. Med. 312:404-407.

14. Gellin, B. G., and C. V. Broome. 1989. Listeriosis. JAMA 261:1313-1320.

15. Howard, P. J., K. D. Harsono, and J. B. Luchansky. 1992.Differentiation of Listeria monocytogenes, Listeria innocua, Listeriaivanovii, and Listeria seeligeri by pulsed-field gel electrophoresis.Appl. Environ. Microbiol. 58:709-712.

16. Lange, C., M. Noyer-Weidner, T. A. Trautner, M. Weiner, andS. A. Zahler. 1991. M-H21, a multispecific 5C-DNA methyltrans-ferase encoded by Bacillus amyloliquefaciens phage H2. Gene100:213-218.

17. Linnan, M. J., L. Mascola, X. D. Lou, V. Goulet, S. May, C.Salminen, D. W. Hird, M. L. Yonekura, P. Hayes, R. Weaver, A.Audurier, B. D. Plikaytis, S. L. Fannin, A. Kleks, and C. V.

J. CLIN. MICROBSIOL.

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

RAPD PATTERNS OF L. MONOCYTOGENES 1287

Broome. 1988. Epidemic listeriosis associated with Mexican-stylecheese. N. Engl. J. Med. 319:823-828.

18. Loessner, M. J., and M. Busse. 1990. Bacteriophage typing ofListeria species. Appl. Environ. Microbiol. 56:1912-1918.

19. Mazurier, S. I., A. Audurier, N. Marquet-Van der Mee, S. Noter-mans, and K. Wernars. 1992. A comparative study of randomlyamplified polymorphic DNA analysis and conventional phagetyping for epidemiological studies of Listeria monocytogenes iso-lates. Res. Microbiol. 143:507-512.

20. Mazurier, S. I., and K. Wernars. 1992. Typing of Listeria strains byrandom amplification of polymorphic DNA. Res. Microbiol. 143:499-505.

21. McLauchlin, J., A. Audurier, and A. G. Taylor. 1986. The evalu-ation of a phage-typing system for Listeria monocytogenes for usein epidemiological studies. J. Med. Microbiol. 22:357-365.

22. Meier, J. T., M. I. Simon, and A. G. Barbour. 1985. Antigenicvariation is associated with DNA rearrangements in a relapsingfever Borrelia. Cell 41:403-409.

23. Monod, M., C. Denoya, and D. Dubnav. 1986. Sequence andproperties of PIM-13 a macrolide-lincosamide-streptogramin Bresistance plasmid from Bacillus subtilus. J. Bacteriol. 167:138-147.

24. Munthali, M., B. V. Ford-Lloyd, and H. J. Newbury. 1992. Therandom amplification of polymorphic DNA for fingerprintingplants. PCR Methods Appl. 1:274-276.

25. Muralidharan, K., and E. K. Wakeland. 1993. Concentration ofprimer and template qualitatively affects products in random-amplified polymorphic DNA PCR. BioTechniques 14:362-364.

26. Ou, J. T., L. S. Baron, F. A. Rubin, and D. J. Kopecko. 1988.Specific insertion and deletion of insertion sequence 1-like DNAelement causes the reversible expression of the virulence capsularantigen Vi of Citrobacter freundii in Escherichia coli. Proc. Natl.

Acad. Sci. USA 85:4402-4405.27. Perez-Diaz, J. C., M. F. Vicente, and F. Baquero. 1982. Plasmids in

Listeria. Plasmid 8:112-118.28. Peterkin, P. I. 1993. Personal communication.29. Piffaretti, J. C., H. Kressebuch, M. Aeschbacher, J. Bille, E.

Bannerman, J. M. Musser, R. K. Selander, and J. Rocourt. 1989.Genetic characterization of clones of the bacterium Listeria monlo-cytogenes causing epidemic disease. Proc. Natl. Acad. Sci. USA86:3818-3822.

30. Pitcher, D. G., N. A. Saunders, and R. J. Owen. 1989. Rapidextraction of bacterial genomic DNA with guanidium thiocyanate.Lett. Appl. Microbiol. 8:151-156.

31. Schlech, W. F., P. M. Lavigne, R. A. Bortolussi, A. C. Allen, E. V.Haldane, A. J. Wort, A. W. Hightower, S. E. Johnson, S. H. King,E. S. Nicholls, and C. V. Broome. 1983. Epidemic listeriosis-evidence for transmission by food. N. Engl. J. Med. 308:203-206.

32. Schwartz, B., C. A. Ciesielski, C. V. Broome, S. Gaventa, G. R.Brown, B. G. Gellin, A. W. Hightower, and L. Mascola. 1988.Association of sporadic listeriosis with consumption of uncookedhot dogs and undercooked chicken. Lancet ii:779-782.

33. Schwartz, B., D. Hexter, C. V. Broome, A. W. Hightower, R. B.Hirschhorn, J. D. Porter, P. S. Hayes, W. F. Bibb, B. Lorber, andD. G. Faris. 1989. Investigation of an outbreak of listeriosis: newhypotheses for the etiology of epidemic Listeria monocytogen2esinfections. J. Infect. Dis. 159:680-685.

34. Wesley, I., and F. Ashton. 1991. Restriction enzyme analysis ofListeiia monocytogenes strains associated with food-borne epidem-ics. Appl. Environ. Microbiol. 57:969-975.

35. Williams, J. G. K., A. R. Kubelik, K. J. Livak, J. A. Rafalski, andS. V. Tingey. 1990. DNA polymorphisms amplified by arbitraryprimers are useful as genetic markers. Nucleic Acids Res. 18:653 1-6535.

VOL. 32, 1994

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from