INTRODUCTION
The presence of Listeria spp. in foods, particularly ready-to-eat products, is of significant public health importance. One member of this genus, L. monocytogenes, has been responsible for numerous outbreaks of foodborne illnesses with symptoms ranging from potentially deadly infections such as gastroenteritis and meningitis to septic abortions in pregnant women [Reference Farber and Peterkin1–Reference Liu3]. The biggest threat to food safety by this organism is due to its hardiness and ability to resist common methods of food preservation [Reference Lado, Yousef, Ryser and Marth4]. With the exception of the rarely occurring L. ivanovii, which is more often associated with animals than humans [Reference Liu3, Reference Bubert5], other species of Listeria are apathogenic including L. seeligeri, L. innocua, L. welshimeri and L. grayi [Reference Liu3, Reference Schmid6]. L. innocua, although harmless to humans, is of practical importance as it is closely related to L. monocytogenes, and is used as an indicator of the presence of the more harmful species [Reference Bubert5, Reference Kamat and Nair7, Reference Liu, Puri and Demirci8].
The need to identify Listeria to the species level is therefore of utmost importance to food manufacturers. Biochemical and serological-based methods have been conventionally used to identify this organism at the genus and species levels. However, these traditional methods are labour intensive and time consuming. With the advent of polymerase chain reaction (PCR)-based diagnostic assays, it is now possible to cost-effectively and quickly confirm identification of different species of Listeria [Reference Bansal9, Reference Fratamico, Kawasaki and Wilson10]. Multiplex PCR utilizing primers that target specific markers such as the iap or prs genes allows for the simultaneous identification and differentiation of potentially pathogenic L. monocytogenes strains from other species [Reference Bubert5, Reference Chen and Knabel11–Reference Zeng13].
PCR-based techniques such as repetitive extragenic palindromic PCR (rep-PCR) can also be used to genotype bacterial isolates and give useful information on genetic relatedness of strains within a species [Reference Rademaker, de Bruijn, Caetano-Anolles and Gresshoff14]. Rep-PCR is based on generating multiple DNA fragments by amplification of regions between repetitive elements present in bacterial genomes. Distance measures among isolates can be obtained based on presence or absence of amplified fragments (loci) using cluster analysis methods [Reference Rademaker, de Bruijn, Caetano-Anolles and Gresshoff14]. Primers based on REP, BOX and, ERIC elements have been extensively used for genotyping bacteria, including Listeria [Reference Gilot15].
The presence of Listeria spp. was previously reported in both raw meat samples and post-processed ready-to-eat products at a processing plant in Trinidad [Reference Gibbons16]. However, there is very little published data on the prevalence of this pathogen in locally produced, ready-to-eat meats sold at retail outlets. A previous study detected L. monocytogenes in one of 70 delicatessen meat samples collected from local supermarkets [Reference Hosein17] but no information was given on the origin or description of the positive sample.
This study investigated the occurrence and genetic relatedness of Listeria spp. in retail products (chicken frankfurters, chicken bologna, bacon) of two locally produced brands (A and B) and the processing plant environment of one of the brands (B). These were the two most popular brands of locally processed meats in Trinidad based on a survey conducted by interviewing managers of supermarkets (data not shown) prior to sample collection and analysis. The plants producing these brands are located on different regions of the island, ~80 km apart. Both are privately owned and are among the largest meat-processing plants in Trinidad. Only the processing environment of plant B was included in the study since approval was not obtained from the management of plant A to conduct investigations at that plant.
METHOD
Isolation of Listeria spp
A total of 480 samples of locally processed bacon, chicken frankfurters and chicken bologna were collected (October 2005 to November 2006) from eight grocery stores in diverse areas in Trinidad and analysed for the presence of Listeria spp. using conventional biochemical and serological methods described by Pagotto et al. [Reference Pagotto18]. The samples collected were equally divided between the three products, two brands (A and B) and the grocery stores. Plant B was visited on two separate occasions during the processing of each of the three products. During each visit, raw ingredients, in-process food samples and finished products were collected and analysed for the presence of Listeria spp. together with swabs of environmental surfaces, equipment, workers' gloves and coats as indicated above. Listeria monocytogenes strain ATCC 7644 was used as a control in all experiments. A single isolate from each positive sample was stored as frozen culture at −80°C in brain heart infusion broth containing 25% glycerol, for further analysis.
DNA extraction and PCR
After confirmation of the presence or absence of haemolysis in blood agar, a single colony of each isolate was selected and DNA was then extracted using the UltraClean™ Microbial DNA Isolation kit (Mobio Laboratories Inc., USA) according to the manufacturer's instructions. Confirmation of the genus Listeria was done by PCR amplification of a ~370-bp region of the prs gene using the Listeria forward (5′-GCTGAAGAGATTGCGAAAGAAG-3′) and Listeria reverse (5′-CAAAGAAACTTGGATTTGCGG-3′) primers [Reference Doumith19]. Each reaction (25 μl) contained ~10 ng DNA, 2·5 μl 10×PC2 buffer (supplied by the manufacturer), 1·5 mm MgCl2, 0·2 mm dNTPs, 1 pmol of each primer and 3·125 U KlenTaq polymerase (DNA Polymerase Technology Inc., USA). PCR was carried out in a Techne Touchgene Gradient Thermocycler (Techne, USA) using the following conditions: initial denaturation at 94°C for 2 min followed by 30 cycles of 95°C for 15 s, 58°C for 30 s and 72°C for 45 s, with a final extension for 5 min at 72°C. Reaction mixtures and conditions for the determination of Listeria spp. by multiplex PCR were similar as above but primers targeting regions of the iap gene were used as described by Bubert et al. [Reference Bubert5]: MonoA (5′-CAAACTGCTAACACAGCTACT-3′), an upstream primer for the detection of L. monocytogenes; Ino2 (5′-ACTAGCACTCCAGTTGTTAAAC-3′), an upstream primer for the detection of L. innocua; Siwi2 (5′-TAACTGAGGTAGCGAGCGAA-3′), an upstream primer for the detection of the L. seeligeri–L. welshimeri–L. ivanovii group (Siwi group); and Lis1B (5′-TTATACGCGACCGAAGCCAAC-3′), a fixed downstream primer for all species. With these multiplex primers, strains of L. monocytogenes were expected to give an amplified fragment of ~660 bp, L. innocua ~870 bp and the Siwi group ~1·2 kb [Reference Bubert5].
For rep-PCR, the BOXA1R primer (5′-CTACGGCAAGGCGACGCTGACG-3′) [Reference Versalovic20] was used. Each PCR reaction (25 μl) contained ~10 ng template DNA, 2·5 μl PC2 reaction buffer (supplied by the manufacturer), 2·5 mm MgCl2, 0·2 mm dNTPs, 5 pmol of BOXA1R primer and 3·125 U KlenTaq DNA polymerase (DNA Polymerase Technology Inc.). Cycling conditions were: an initial denaturation for 7 min at 95°C, 30 cycles of 94°C for 1 min, 51°C for 1 min, 65°C for 8 min, and a final extension at 65°C for 15 min [Reference Versalovic20].
Products of Listeria-specific PCR (5 μl) and rep-PCR (9 μl ) were separated by electrophoresis in 1·5% and 1·0% agarose gels, respectively, and photographed on a UV trans-illuminator [Reference Sambrook and Russel21]. Isolates were scored for the presence or absence of specific bands based on comparison to a 1-kb ladder (Invitrogen Life Technologies, USA).
Non-parametric χ2 tests were applied to the multiplex data using SPSS software, version 15.0 (SPSS Inc., USA). The binary DNA fingerprinting data was subjected to Cluster analysis with average linkage and Euclidean distance (Minitab statistical software package, version 14.0; Minitab Inc., USA) to generate dendrograms showing relationships between isolates.
RESULTS
Ninety of the 480 (18·75%) retail samples, 24 food samples, and environmental surfaces from plant B were positive for the presence of Listeria spp. based on the conventional isolation and identification methods. Of the 114 single isolates obtained from positive samples, 110 (96·5%) isolates were confirmed as belonging to Listeria spp. by amplification of the ~370-bp prs gene fragment. Two isolates from retail samples and two from plant B did not give the expected ~370-bp band.
The results of the multiplex PCR showed a significantly (χ2, P<0·001) higher number of isolates (64/110) having bands characteristic of L. innocua compared to L. monocytogenes (39/110) and members of the Siwi group (7/110). Six of the 39 isolates (15·4%) confirmed as L. monocytogenes were non-haemolytic on blood agar medium. All of these isolates were recovered from bacon products of brand A. Six of 64 L. innocua isolates (9·4%) were also haemolytic on blood agar. Of these, four were obtained from raw chicken bologna mixture from plant B and the remaining two isolates originated from two retail samples from plant A: one bacon and one chicken frankfurter.
Eighty-eight of the 90 isolates from retail samples gave positive PCR results for Listeria spp. and of these, 46 (52·3%) belonged to L. innocua, 39 (44·3%) to L. monocytogenes and three (3·4%) to the Siwi group (Table 1). Overall, Listeria spp. were detected in 18·3% of retail samples with prevalence rates of 9·6%, 8·1% and 0·6% for L. innocua, L. monocytogenes and the Siwi group, respectively. A significantly (χ2, P<0·001) higher number of the isolates were from brand A (68/88) compared to brand B (20/88). There was no statistically significant (χ2, P>0·05) difference in the distribution of species in plant A products; however, plant B products had significantly (χ2, P<0·05) more samples which were positive for L. innocua than the other species.
Table 1. Distribution of Listeria spp. in retail samples of brand A and brand B products
The majority (66/88) of the Listeria isolates from retail samples were from bacon, which had a significantly (χ2, P<0·001) higher prevalence of the organism in both brands (Table 1). Bacon also mostly accounted for the presence of L. monocytogenes from brand A products (χ2, P<0·001) as well as L. innocua from both brands under study (χ2, brand A: P<0·05; brand B: P<0·001). The Siwi group was only found in brand B bacon. No L. monocytogenes isolate was found in retail items from plant B and no isolate of the Siwi group was detected in plant A products. There were no significant (χ2, P>0·05) differences in association of Listeria spp., L. monocytogenes, L. innocua, or Siwi group with grocery stores and locations.
Of the 24 isolates obtained from plant B using conventional methods, 91·7% (22/24) were confirmed as Listeria spp. by amplification of the prs fragment. Most of these isolates (18/22) were L. innocua while the remaining (4/22) belonged to the Siwi group (χ2, P<0·05). Of these isolates, most (19/22) originated from in-process food samples and 13·6% (3/22) were obtained from equipment during bacon and chicken bologna pre-cooking processes (P=0·001) (Table 2). No Listeria spp. was detected in the post-cooking environment and no L. monocytogenes was found in the entire plant environment. The Siwi group was only detected during bacon production: in meat on the injector during the first visit, in liquid cure on the second visit and in pumped meat during both sampling occasions.
Table 2. Species of Listeria found in plant B's manufacturing environment during the processing of bologna and bacon
BOX–PCR fingerprinting
Analysis of rep-PCR banding patterns revealed very diverse populations of the different Listeria spp. obtained from the retailed meats and the plant B environment. Isolates from the largest L. innocua group separated into two broad clusters in addition to two outlier isolates that were more genomically related to the L. monocytogenes reference isolate (Fig. 1). There were seven subgroups with 2–6 isolates having similarity levels of 100%. Most of the isolates (27/35) linked to plant B (retail grocery products, in-process food or equipment) were present in group 1, with the dominant product being bacon followed by bologna. The majority (14/18) of the isolates from in-process food and equipment clustered together in one subgroup, with approximately equal numbers of the isolates in this subgroup coming from samples collected during the processing of bacon and frankfurters. Most (19/29) of the isolates from plant A retail products were in group 2 (Fig. 1). The majority of these isolates were from bacon followed by frankfurters. There was no clear trend of distribution of isolates from retail products based on grocery stores from which the products were collected.
Fig. 1. Dendrogram showing clusters among Listeria innocua isolates. Group 1 isolates are coloured blue in the dendrogram and group 2 isolates green.
The L. monocytogenes isolates clustered into two major groups in addition to one outlier isolate G_1_h9_8 III A-8 and the Listeria control forming separate branches (Fig. 2). All the isolates originated from plant A grocery products, with the majority (35/39) from bacon and the remainder from frankfurters and bologna. Group 1 isolates were from six different grocery stores and included two subgroups with three and seven isolates that had identical DNA fingerprints. Group 2 isolates were from seven different grocery stores and had four groups which contained 2–9 isolates with identical DNA fingerprints. The subgroups containing isolates with identical fingerprints generally had isolates originating from different stores, for example, the seven isolates in the larger subgroup in group 1 came from four different stores and were mainly from bacon products, with the exception of one isolate from chicken frankfurters. Similarly, the largest subgroup in group 2 had nine isolates from four different stores which were mainly from bacon items together with one from frankfurters and one from bologna.
Fig. 2. Dendrogram showing clusters among Listeria monocytogenes isolates. Group 1 isolates are coloured blue in the dendrogram and group 2 isolates green.
Within the Siwi group, only one subgroup contained two isolates with 100% similarity (Fig. 3). Both of these isolates originated from bacon food samples obtained from the plant environment, each from a separate sampling visit. The processing plant isolates generally clustered in a separate group from the grocery isolates, which were more diverse.
Fig. 3. Dendrogram showing clusters among isolates of the L. seeligeri–L. welshimeri–L. ivanovii group.
DISCUSSION
The results of the study show the occurrence of typical as well as atypical Listeria spp. in retail ready-to-eat meats and in-process food samples in Trinidad. Although there was a strong concurrence observed between haemolysis on blood agar and L. monocytogenes detected by molecular methods, certain strains were not haemolytic on blood agar plates. These strains may have been non-haemolytic variants which have been previously reported and attributed to a genetic mutation in the listeriolysin gene [Reference Allerberger22, Reference Strom23]. A few L. innocua isolates also showed haemolysis on blood agar, which Volokhov et al. [Reference Volokhov24] suggested may be as a result of retention of genes from L. monocytogenes, its possible ancestral predecessor.
The study has shown that consumers of locally processed, ready-to-eat meats may be at risk due to contamination by Listeria spp. including L. monocytogenes. The detection of this pathogen led to a voluntary recall of chicken frankfurters, spice ham and turkey ham processed at one plant in Trinidad in 2003 [Reference Gibbons16]. The potential threat to public health due to contamination of ready-to-eat meats thus still exist since either L. monocytogenes or the indicator L. innocua [Reference Liu, Puri and Demirci8] was found in all three product types from the two different brands investigated (Table 1). The 8·1% prevalence rate of L. monocytogenes in retail samples from this study was similar to the findings of a study conducted in Greece, which also found that 8·1% of samples of ready-to-eat meat products were contaminated with L. monocytogenes and that bacon had the highest level of association with the pathogen [Reference Angelidis and Koutsoumanis25]. However, the prevalence rate determined in this study was higher than the 3–5% contamination rates found for various categories of ready-to-eat meats from retail markets in Edmonton, Canada [Reference Bohaychuk26].
Both plants A and B, but particularly the former, need to revise their quality assurance programme. Of the three products investigated, bacon was the most prone to contamination by Listeria spp. This may be attributed to its high fat content which could have protected bacteria throughout manufacture, as well as poor quality assurance programmes implemented in the manufacturing operations [Reference Senhaji and Loncin27–Reference Wong, Wijewickreme and Kitts30]. Brand A products had a higher risk of exposing consumers to Listeria compared to brand B products. The fact that most of the Listeria spp. and all of the L. monocytogenes isolates came from products manufactured by plant A leads to the inference that the quality assurance programme in this plant was less effective than the programme in plant B. However, it must be noted that the study only included a limited number of samples over a limited period of time. Thus, further monitoring may be needed to ascertain whether there may be higher risks associated with products from plant B and other plants in Trinidad as well as imported brands.
Among grocery stores, the Siwi group was only found in brand B bacon products. The strains were found again in raw material and equipment during the manufacture of bacon in plant B, which suggests a linkage in contamination from plant to grocery. The persistence and survival of specific strains of this group in the plant environment, as well as their ability to contaminate different batches of product could also be inferred from the finding that two strains with 100% similarity were obtained in food samples from different sampling visits.
The presence of Listeria spp. in raw materials such as the liquid cure and the raw chicken meat indicates the possibility that contamination of retail products could have originated from the raw material. Hinton et al. [Reference Hinton, Cason and Ingram31] reported survival of bacteria during processing as well as cross-contamination of post-processed broiler carcasses from raw materials. Contamination of sausages and cold cuts has also been attributed to the presence of pathogens in raw material [Reference Nesbakken, Kapperud and Caugant32, Reference Sartz33].
No finished products at plant B were positive for Listeria spp. despite the organism's presence on environmental surfaces, raw materials and retail items from grocery stores. If in fact there were low levels of contamination of finished products in the plant, it is also possible that the organism may not have been detected due to limited sensitivity of the analytical methods. However, bacterial levels could increase to within the limits of detection if unsuitable storage conditions existed in grocery stores since Listeria spp. is known to be able to grow at relatively low temperatures [Reference Azizoglu34]. It must be noted that the existence of Listeria on surfaces is a serious risk for cross-contamination of finished products and is suggestive of the need for a better quality assurance programme. The minimum infective dose of L. monocytogenes in humans has not been established, although a review of the literature by Farber & Peterkin [Reference Farber and Peterkin1] indicated infection of healthy individuals occurred due to consumption of food contaminated with 2·7×106 organisms/g and as little as 102–104 organisms/g for immunocompromised individuals. Thus, even low levels of contamination by L. monocytogenes at the plant may be a potentially serious health risk to susceptible individuals.
On observation of the DNA fingerprints, it was found that isolates which originated from the processing plant and those from retail samples did not generally have identical banding patterns. However, most of the L. innocua isolates segregated based on brand (Fig. 1). Additionally, isolates from the processing samples in plant B clustered together with isolates from retail items from this plant. This together with the fact that the retail samples came from several different grocery stores suggest that there is a link between the plant and contamination of the products. Other studies have also shown contamination of processed meats with L. monocytogenes and spoilage organisms have been due to cross-contamination from the manufacturing environment [Reference Dykes, Cloete and von Holy35, Reference Samelis36]. Further influence of plant conditions on retail samples could be corroborated in this study as no L. monocytogenes was found in plant B and no L. monocytogenes species were found in brand B retail samples.
Some L. innocua strains obtained from brand A retail products were more closely related to isolates from plant B's samples, as seen in three subclusters which had isolates from brand A and brand B samples with 100% similarity (Fig. 1). Strains of Listeria are known to occur widely in different environments and could be commonly associated with raw meat from different areas and sources [Reference Wesley, Ryser and Marth37]. It is possible that both plants A and B may have common suppliers for imported raw materials resulting in the similar strains being present in both products.
The environmental tolerance of different Listeria spp., coupled with the nutritional composition of different meats may have also influenced which species or strains were found in a particular product. Boyer et al. [Reference Boyer38] showed that L. monocytogenes was significantly more resilient to environmental stresses than L. innocua. This, coupled with the tendency of fat from pork to protect bacteria, could possibly explain in part, the relatively high prevalence of L. monocytogenes found in bacon [Reference Lin, Cao and Chen29, Reference Wong, Wijewickreme and Kitts30].
With respect to the cluster analysis of L. monocytogenes, the dendrogram (Fig. 2) indicates the possible existence and persistence of similar strains of the pathogen in plant A's manufacturing environment and the potential for these to contaminate several different products. In two subgroups with 100% similarity, strains which were mostly obtained from bacon were also present in chicken frankfurters and bologna. These particular strains may have survived in niches in the plant and could have been transferred to other products by employees or inadequately cleaned communal equipment. The fact that the L. monocytogenes with identical DNA fingerprints were recovered from different grocery samples (Fig. 2) further supports the link of product contamination to the processing environment.
There is also evidence of association of specific strains of Listeria spp. with plant B's environment. Strains of L. innocua showing 100% similarity were obtained on two separate sampling occasions about 1 week apart from pre-cooked bologna food samples. Considering that each piece of machinery is cleaned on a daily basis, it is possible that this specific strain of L. innocua could have persisted over time in biofilms [Reference Henning and Cutter28] on equipment and the environment. Therefore, through either cross-contamination or contact with contaminated machinery bacterial cells could have been successfully transferred to new batches of food material during bologna manufacture. Alternatively, considering that sampling occasions were taken within a relatively close space of time, the same batch of raw material contaminated with that specific Listeria strain could have been used to manufacture bologna within a 2-week period. Similarly, one L. innocua subgroup with 100% similarity consisted of isolates which originated from two different brand B retail products: one chicken frankfurters and one bologna (Fig. 1). This supports the idea of bacterial Listeria surviving as biofilms [Reference Henning and Cutter28] since the manufacture of both products utilized common plant equipment. These scenarios complement a similar study performed in a large meat-processing plant in Trinidad which attributed the presence of biofilms and lapses in good sanitary practices to the production of items contaminated with Listeria [Reference Gibbons16].
The Siwi group was the least prevalent in both retail and plant samples. This species group has been documented to occur very rarely in nature and this was reflected in its limited presence during this study [Reference Bubert5]. However, the presence of the Siwi group may be of some concern as L. ivanovii, a member of this group, is considered pathogenic even though it rarely occurs in humans [Reference Liu3].
Further research is needed to determine the extent of Listeria contamination of ready-to-eat meats in Trinidad by including other locally produced brands as well as imported brands. However, from this limited study, a clear risk to consumers has been identified together with a possible link between processing plants and quality of retail products. Processors should improve the quality of their products by stringent implementation or reinforcements of programmes such as Good Manufacturing Practices, Good Hygiene Practices and Hazard Analysis and Critical Control Point programmes [Reference Nesbakken, Kapperud and Caugant32] in order to enhance the safety of consumers.
ACKNOWLEDGEMENTS
This study was supported by a financial grant from the School for Graduate Studies and Research, The University of the West Indies, St Augustine, Trinidad.
DECLARATION OF INTEREST
None.
