Vol. 57, No. 9

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Sept. 1991, p. 2576-2580

0099-2240/91/092576-05$02.00/0

Sensitive and Specific Detection of Listeria monocytogenes in Milk and Ground Beef with the Polymerase Chain Reaction ELIZABETH J. GOLSTEYN THOMAS, ROBIN K. KING, JACK BURCHAK, AND VICTOR P. J. GANNON* Animal Diseases Research Institute, Agriculture Canada, P.O. Box 640, Lethbridge, Alberta, Canada TJJ 3Z4

Received 7 May 1991/Accepted 1 July 1991

A sensitive and specific method for detection of Listeria monocytogenes in milk and ground-beef samples is described. It consists of culturing samples in listeria enrichment broth (LEB) and subculturing them from LEB to listeria plating media, followed by DNA extraction and species-specific detection of the organism by using the polymerase chain reaction (PCR). In developing the L. monocytogenes PCR assay, five oligonucleotide primers complementary to the nucleotide sequence of the listeriolysin 0 gene were synthesized and used in amplification experiments. PCR products of the predicted size, based on nucleotide sequence information, were generated with DNA from all of 72 L. monocytogenes strains with five different primer pairs. DNA from Listeria ivanovii, Listeria innocua, Listeria seeligeri, Listeria welshimeri, Listeria grayi, and Listeria murrayi strains and a panel of 47 bacterial strains representing 17 genera did not generate PCR products with the primer pairs employed. As little as 1 pg of L. monocytogenes DNA could be detected with the assay. To determine the most sensitive culture protocol to use in conjunction with the PCR assay, milk (10 ml) and ground-beef (25 g) samples were inoculated with L. monocytogenes at concentrations ranging from 0 to 105 CFU ml-l or g-1, as appropriate for the sample. PCR assays on DNA extracted from growth on listeria plating media, inoculated with 24-h LEB sample cultures, were most sensitive, allowing detection of as little as 0.1 CFU of L. monocytogenes m-l or gof milk and ground beef, respectively.

Listeria monocytogenes is associated with meningoencephalitis, septicemia, and abortion in humans (15). Pregnant women, newborns, and immunocompromised individuals are particularly susceptible to infection; however, in some instances, apparently healthy individuals develop clinical disease after ingestion of food contaminated with this pathogen (30, 31). There have been at least five serious food-borne outbreaks of listeriosis reported in North America and Europe in recent years (10, 12, 30, 31). The wide distribution of this pathogen in nature and its ability to proliferate at refrigeration temperature make elimination of the organism from foods very difficult. Recent attention has focused on rapid detection systems for L. monocytogenes. While cultural methods provide the "gold standard," these are labor intensive and time consuming. A more rapid method which specifically identifies L. monocytogenes is required. DNA probes complementary to the listeriolysin 0 gene (hlyA) (7, 28, 33) and other L. monocytogenes-specific target genes such as the iap locus (6, 8, 11, 19, 21) and the dth gene (27) have been used successfully in hybridization studies to detect the organism. However, at least 105 to 106 copies of the target sequence are required for colony hybridization assays (23). Recently, a hybridization assay, which recognizes 16S rRNA, has been reported for detection of Listeria spp. (20) and made commercially available. The advantage of this test is its increased sensitivity of detection over other DNA probes, given that the number of target rRNA molecules per bacterial cell is approximately 10,000. The disadvantage of this system is that it is genus rather than species specific. Recently, polymerase chain reaction (PCR) assays based on the amplification of target DNA sequences in the listeriolysin 0 (3, 4, 9) and dth (32) genes have been reported. *

Corresponding author. 2576

However, use of these assays in the detection of L. monocytogenes in foods was not examined or the sensitivity of the assay was low and results were inconsistent. In this study, we describe a rapid and sensitive method for detecting as few as 1 CFU in the original milk or ground-beef samples based on selective enrichment of L. monocytogenes by culture and PCR amplification of the listeriolysin 0 gene. MATERIALS AND METHODS Bacterial strains, culture conditions, and media. Bacterial strains used in this study are listed in Table 1. These include the following: (i) 72 L. monocytogenes strains originally isolated from foods and clinical cases in humans and animals; (ii) 4 Listeria ivanovii, 14 Listeria innocua, 2 Listeria seeligeri, 2 Listeria welshimeri, 1 Listeria grayi, and 1 Listeria murrayi strains; and (iii) a panel of 47 isolates from other bacterial species representing 17 genera. These were kindly provided by J. Farber (Bureau of Microbial Hazards, Ottawa, Ontario, Canada), S. Messier (Health of Animals Laboratory, St. Hyacinthe, Quebec, Canada), S. Read (Health of Animals Laboratory, Guelph, Ontario, Canada), B. McMullin (University of Lethbridge, Lethbridge, Alberta, Canada), H. Ross (University of Calgary, Calgary, Alberta, Canada), and L. Andreachuk (Foothills Hospital, Calgary, Alberta, Canada). For routine culture of bacterial strains, brain heart infusion broth was used, and for selective culture of L. monocytogenes, listeria enrichment broth (LEB) and listeria plating medium (LPM) were used (13). L. monocytogenes A4, originally isolated from meat, was used in testing the sensitivity of the PCR assay and in experiments with milk and ground-beef samples. In the testing of the PCR assay on foods, pasteurized 2% partly skim milk and freshly ground beef were purchased from a local retail outlet. Listeria spp. could not be detected by culture in either food product. These were inoculated

TABLE 1. Bacterial strains used in this study No. of strains

Bacterial strain

Listeria species L. monocytogenes L. ivanovii L. innocua L. seeligeri L. welshimeri L. grayi L. murrayi

Other bacteria Agrobacterium tumefaciens Bacillus cereus Bacillus megaterium Bacillus subtilis Clostridium tetani Enterobacter cloacae Erysipelothrix rhusiopathiae Escherichia coli Klebsiella pneumoniae Micrococcus lysodeiktius Pasteurella haemolytica Proteus vulgaris Pseudomonas aeruginosa Pseudomonas fluorescens Pseudomonas testosteroni Rhodococcus equi Salmonella montevideo Salmonella typhimurium Sarcina aurantiaca Sarcina lutea Serratia marcescens Staphylococcus aureus Staphylococcus epidermidis Streptococcus agalactiae Streptococcus faecalis Streptococcus group A Streptococcus pneumoniae Streptococcus pyogenes

2577

AMPLIFICATION OF LISTERIOLYSIN 0 GENE

VOL. 57, 1991

Tested

Hemolytic

PCR positive

72 4 14 2 2 1 1

72 4 0 2 0 0 0

72 0 0 0 0 0 0

1 4 1 2 1 1 1 3 1 1 1 1 2 1 1 1 1 2 1 1 1 4 2 1 3 1 3 4

0 4 0 2 1 0 0 0 0 0 1 0 2 0 0 0 0 0 0 0 0 4 0 1 3 1 3 4

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

with 0 to 105 CFU of L. monocytogenes ml-' or g-1, as appropriate for the sample. After this, 10 ml of each milk sample was mixed with 90 ml of LEB and 25 g of each ground-beef sample was mixed with 225 ml of LEB by use of a stomacher. A 0.1-ml aliquot of each LEB-sample mixture (LEBO) was plated onto LPM agar either immediately (LEBO+LPM) or after incubation overnight at 37°C (LEB24+LPM). LPM plates were incubated overnight at 370C. DNA extraction. For testing the sensitivity and species specificity of the PCR assay, DNA was extracted from bacteria grown in brain heart infusion broth as described by Ausubel et al. (1). After culture of milk and ground-beef samples (which had been inoculated with L. monocytogenes) in LEB or on LPM agar, DNA was extracted from the bacterial growth by using the following procedures. Bacteria were removed from the LPM plates by using a dry cotton swab and suspended in 1 ml of phosphate-buffered saline (pH 7.4). A 0.1-ml aliquot of lysis buffer (10 mM EDTA, 100 mM Tris [pH 8.0]) containing 7.5 mg ml-1 of lysozyme and 750 U ml-1 of mutanolysin was added to 0.3 ml of LEB cultures or bacterial suspensions from LPM agar. These mixtures were incubated at 37°C for 30 min. A 0.1-ml volume of lysis buffer containing 5 mg ml-'

TABLE 2. Nucleotide sequence of oligonucleotide primers used in this study Location Orienta-Sequence (5'--3') content within Primer

~~~~~G+C tinttn(%)

GTG AA GGT AC GAT AT TGC TC TCG TC a From published hlyA gene for listeriolysin 0 (26).

LL1 LL3 LL4 LL5 LL6

Forward Forward Reverse Forward Reverse

GAC ATT CGC AAC CTG

ATT GCG CAC CTA TAA

CAA AAA ACT TCC GCC

GTT TTT TGA AGG ATT

35 35 47 53 47

gene

(bp)a

340-357 652-669 874-891 372-389 622-639

of proteinase K and 50 mg ml-1 of Sarkosyl (sodium lauryl sarcosinate) was then added to each sample, and these were incubated for an additional 15 min at 37°C. DNA was precipitated from samples by adding 50 ,ul of 3 M sodium acetate and 1 ml of ice-cold absolute ethanol and collected by centrifugation. The pellets were dried and resuspended in distilled water for use in PCR assays. PCR assays. Synthetic oligonucleotide primers used in the study are listed in Table 2. These consist of three forward (LL1, LL3, LL5) and two reverse (LL4, LL6) primers complementary to the published listeriolysin 0 gene sequence (26). Oligonucleotides were either purchased from the Regional DNA Synthesis Laboratory (University of Calgary, Calgary, Alberta, Canada) or synthesized by using a model 391 DNA synthesizer (PCR-Mate; Applied Biosystems). PCR assays on bacterial DNA were performed in 100-,ul volumes containing 2.0 to 3.0 mM MgCl2 (depending on the primer combination), 10 mM Tris hydrochloride (pH 8.3), 50 mM KCI, 1 ng of template DNA (unless otherwise specified), 0.2 mM each dATP, dGTP, dCTP, and dTTP, 1 ,uM each primer, and 2.5 U of Thermus aquaticus (Taq) DNA polymerase.

PCR assays were performed on bacterial DNA with forward and reverse oligonucleotide primer combinations in a DNA thermal cycler (Perkin-Elmer Cetus) by using 30 cycles of 1 min at 94°C, 1 min at 55°C, and 2 min at 72°C, and a final incubation at 72°C for 5 min. A reaction mix without DNA and one with 1 ng of L. monocytogenes total DNA were included as controls in each experiment. In one experiment, from 0 to 10 ,ug of herring sperm DNA (Boehringer-Mannheim) was added to PCR mixtures. PCR products (15 [lI) were analyzed by submarine gel electrophoresis by using 1.2% agarose gels containing ethidium bromide. RESULTS

Specificity of the PCR. Five combinations of forward and reverse primers (LL1 and LL4, LL1 and LL6, LL3 and LL4, LLS and LL4, and LLS and LL6) generated single PCR products with DNA from all of the tested L. monocytogenes strains (72 of 72). These were of the size predicted based on the published nucleotide sequence of the listeriolysin 0 gene for the respective primer combinations, namely, 560, 299, 240, 520, and 267 bp (Fig. 1). All of the PCR products were of approximately equal intensity on agarose gel electrophoresis. The electrophoretic profile obtained with primer combination LL5 and LL4 for Listeria DNA is presented in Fig. 2. As illustrated, a 520-bp product was detected in PCRs with DNA from L. monocytogenes but not from strains from the other Listeria species tested. No PCR

2578

APPL. ENVIRON. MICROBIOL.

GOLSTEYN THOMAS ET AL.

A A

FIG. 1. PCR amplification of purified L. monocytogenes DNA by using different primer combinations. Amplification products for the different primer combinations with 10 ng of L. monocytogenes DNA as template were analyzed by electrophoresis on a 1.2% agarose gel. Lanes: A, 1-kb ladder (Bethesda Research Laboratories, Inc.); B, primers LL1 and LL4; C, primers LL1 and LL6; D, primers LL3 and LL4; E, primers LL5 and LL4; F, primers LL5 and LL6.

products were evident when the five primer combinations were used in PCR assays with DNA from other Listeria species or DNA from a panel of 47 strains from 17 other bacterial genera (Table 1). Sensitivity of the PCR. A PCR gene product was detected with as little as 1 pg of L. monocytogenes total DNA (Fig. 3A). The sensitivity of the assay was increased to 100 fg when the entire PCR-amplified DNA sample was recovered by ethanol precipitation and loaded onto the gel (data not shown). Formation of a PCR product with 1 ng of L. monocytogenes was completely inhibited by 10 ,ug of herring sperm (HS) DNA, and HS DNA concentrations of from 5 to 9 ,ug decreased the intensity of PCR products on agarose gels

(Fig. 3B).

AB

CH DE FG

FIG. 2. Specificity of the PCR assay by using DNA from different Listeria species. Amplification products obtained with 10 ng of template DNA derived from different Listeria species were analyzed by electrophoresis on a 1.2% agarose gel. Lanes: A, 1-kb ladder; B, L. monocytogenes; C, L. ivanovii; D, L. innocua; E, L. seeligeri; F, L. welshimeri; G, L. grayi; H, L. murrayi.

B

FIG. 3. (A) Sensitivity of the PCR assay in detecting DNA from L. monocytogenes A4. Lanes: A, 1-kb ladder; B, 1 ,ug of template DNA; C, 100 ng of template DNA; D, 10 ng of template DNA; E, 1 ng of template DNA; F, 100 pg of template DNA; G, 10 pg of template DNA; H, 1 pg of template DNA; I, 100 fg of template DNA; J, 10 fg of template DNA; K, 1 fg of template DNA. (B) Inhibition of gene product formation in the presence of various amounts of herring sperm DNA. Amplification products obtained with 1 ng of L. monocytogenes A4 DNA in the presence of various amounts of herring sperm (HS) DNA were analyzed by electrophoresis on a 1.2% agarose gel. Lanes: A, 10 p.g of HS DNA; B, 9 jig of HS DNA; C, 8 ,ug of HS DNA; D, 7 jig of HS DNA; E, 6 jig of HS DNA; F, 5 jigofHS DNA; G, 2.5 jigofHS DNA; H, 1 jigofHS DNA; I, 0.1 Rg of HS DNA; J, no HS DNA; K, 10 jig of HS DNA but no template DNA.

Detection of L. monocytogenes in milk and ground beef by PCR. Results of the PCR assays with the primer combination LL5 and LL4 on DNA extracted from LEB and LPM cultures of milk samples inoculated with L. monocytogenes are shown in Fig. 4. For milk samples, PCR products were detected with DNA extracted from LEBO+LPM plates (Fig. 4B), LEB24 (Fig. 4C), and LEB24+LPM plates (Fig. 4D). The minimum number of L. monocytogenes detected in PCR assays with DNA from the respective cultures was 2.5 x 105, 2.5, and 0.25 CFU ml-'. No PCR products were detected with DNA extracted from LEBO milk samples (Fig. 4A). Similar results were obtained for the ground-meat samples inoculated with L. monocytogenes (data not shown). No PCR products were detected with DNA extracted from LEBO and LEBO+LPM plates. However, as few as 1.2 and 0.12 CFU of L. monocytogenes g-1 were detected in PCR assays with DNA extracted from LEB24 and LEB24+LPM plates, respectively.

VOL. 57, 1991

FIG. 4. PCR amplification of DNA extracted from cultures of milk samples inoculated with L. monocytogenes A4. Panels A to D are photographs of 1.2% agarose gels of PCR products obtained by using the following culture methods: A, LEBO; B, LEBO+LPM; C, LEB24; D, LEB24+LPM. Lanes A to I, PCR products obtained from cultures of milk samples inoculated with 2.5 x 106, 2.5 x 105, 2.5 x 104, 2.5 x 103, 2.5 x 102, 25, 2.5, 0.25, and 0 CFU of L. monocytogenes, respectively; lane J, no template DNA; lane K, 1 or 10 ng of L. monocytogenes DNA.

DISCUSSION Listeriolysin 0 has been shown to be necessary for full virulence of L. monocytogenes in mice (5, 14, 18, 29). Since all clinical strains of L. monocytogenes possess this attribute (16, 17), it would appear to be a clinically relevant marker for

identification. The PCR primer combinations used in this study were specific for L. monocytogenes, with no detectable amplification of DNA from 24 strains of six other Listeria species, including two species which produce hemolysins related to listeriolysin 0 (24). Similarly, PCR products were not present after amplification of DNA from 47 other nonListeria strains. This included strains of Bacillus cereus (four strains), Streptococcus pyogenes (four strains), and Streptococcus pneumoniae (three strains), i.e., bacterial species which also produce hemolysins related to listeriolysin 0 (25). The PCR assay overcomes the test sensitivity barrier, which limits the application of nucleic acid probes, by amplification of target DNA sequences present in low copy numbers. By using PCR, we were able to detect as little as 1 pg of L. monocytogenes DNA, equivalent to approximately 100 bacteria (2). This level of sensitivity is approximately 20,000 times greater than that reported by Bessesen et al. (3) for PCR assays with L. monocytogenes listeriolysin 0 gene as a target but is comparable to that reported by Deneer and Boychuk (9). Wernars et al. (32) detected between 1 and 10 CFU by using the dth gene as their target. The reason for the differences in PCR sensitivity in these studies is unclear but may be related to the gene target used. However, it is difficult to compare studies which use DNA versus number of CFUs in determining assay sensitivity, since dead as well as viable cells may be present and each CFU may result from many cells rather than a single cell. Direct detection of L. monocytogenes and other pathogens in food samples by PCR presents a number of technical difficulties as demonstrated in this study and those of Bes-

AMPLIFICATION OF LISTERIOLYSIN 0 GENE

2579

sesen et al. (3) and Wernars et al. (32). The levels of L. monocytogenes in foods such as ground beef may be low relative to the large numbers of other bacteria present. Aerobic plate counts of >108 CFU g-' have been reported (22). The pathogen may also be nonuniformly distributed in the food product. Nucleic acids from other bacteria and the sample itself may inhibit PCR assays. In this study, 10 ,g of HS DNA completely inhibited amplification of target DNA. Other sample components, ranging from factors which affect cation concentrations to heparin, could also inhibit the PCR assay. These factors limit the sample size on which PCR assays can be performed (0.3 ml of LEB or bacterial suspension in this study) and require that some method of nucleic acid extraction be performed. By using direct PCR, Bessesen et al. (3) were only able to detect 105 CFU ml-' of L. monocytogenes in milk and cerebrospinal fluid, whereas Wernars et al. (32) found that direct detection of L. monocytogenes in soft cheeses was not consistent with the detection limits ranging from 103 CFU per 0.5 g of cheese to no PCR product in the presence of 108 CFU per 0.5 g. Use of selective enrichment media overcomes the difficulties inherent in direct PCR by increasing the amount of target DNA and diluting out non-Listeria DNA and other inhibitors of PCR which may be present in samples. It also ensures that only DNA obtained from viable L. monocytogenes is detected. However, this increased PCR sensitivity is achieved at the expense of time. After 24 h of incubation in LEB, as few as 25 CFU of L. monocytogenes were detected in the milk or ground-beef sample, but the sensitivity of the assay was increased 10-fold (i.e., detection of