Occurrence, Diversity of Listeria spp. Isolates from Food and Food-Contact Surfaces and the Presence of Virulence Genes

This study evaluates the hazards posed by foodborne bacteria of the Listeria genus by analyzing the occurrence, diversity and virulence of Listeria spp.in food and food-manufacturing plants. Seventy-five isolates obtained from the routine analysis of 653 samples taken by three diagnostic laboratories in Northern Italy were genotypically differentiated by Repetitive Extragenic Palindrome (rep) PCR, with the GTG5 primer identified by sequencing the 16S rRNA gene and examined by specific PCR tests for the presence of L. monocytogenes virulence determinants occasionally found to occur in other species of the genus. Within this sample, 76% (n = 57) isolates were identified as L. innocua, 16% (n = 12) as L. monocytogenes, 6.6% (n = 5) as L. welshimeri and 1.3% (n = 1) as L. seeligeri. All L. monocytogenes isolates belonged to the serotype 1/2a and were predicted to be virulent for the presence of the inlJ internalin gene. Potentially virulent strains of L. innocua, L. seeligeri and L. welshimeri, carrying the L. monocytogenesinlA gene and/or hly gene, were identified, and most isolates were found to possess the toxin–antitoxin system mazEF for efficient adaptation to heat shock. Results indicated the need to reinforce food-contamination-prevention measures against all Listeria species by defining efficiently their environmental distribution.


Introduction
Bacteria belonging to the low-G+C facultatively anaerobic, Gram-positive Listeria genus are ubiquitous and can contaminate food products of both animal and plant origin [1].
The human pathogen Listeria monocytogenes can cause severe foodborne infections, including septicemia, meningitis and endocarditis, with fatality rates of 20-30% in the elderly, immunocompromised persons and pregnant women, with infection in the latter resulting in abortion or neonatal infections [2]. It grows at refrigeration temperatures, under vacuum or in a modified atmosphere, tolerates low-pH values and forms sanitation-resistant biofilms in food-processing plants. These characteristics make contamination prevention the main defense against this pathogen [3]. Ready-to-eat (RTE) foods are those with higher risk to transmit the bacterium. In these products, L. monocytogenes must not reach levels higher than 100 CFU/g during their shelf life, while it must be

Positive Samples/Isolates
The numbers missing from the series correspond to isolates for which identification as Listeria spp. was not confirmed by genotypic tests. The number of Listeria spp. isolates correspond to the number of positive samples.
All bacterial strains were streaked onto blood agar (Liofilchem, Roseto degli Abruzzi, TE, Italy) plates to obtain isolated colonies. In this way, isolates were also tested for hemolytic activity. Single colonies were sub-cultured in tubes of Tryptic Soy (TS) broth (Biolife Italiana, Milan, Italy). Cultures were incubated in aerobiosis at 37 • C for 18-24 h. For long-term storage, the cultures were kept at −20 • C in the same medium with the addition of 20% glycerol.

DNA Extraction
DNA was extracted from 2 mL of fresh culture using the Genomic DNA Extraction Kit RBC Bioscience (Diatech Labline, Jesi, AN, Italy), according to the manufacturer instructions. The quantity and integrity of the extracted DNA were checked by comparison with known amounts of Lambda DNA (ThermoFisher Scientific, Rodano, MI, Italy) and for the absence of DNA smears on 1.5% w/v agarose gel in 1× TAE buffer (80 mM Tris-acetate, 2 mM EDTA, pH 8.0) at 120 V stained with 1:10,000 diluted GelRed (Biotium, Società Italiana Chimici, Rome, Italy).
All of the PCR tests were carried out with the EmeraldAmp GT PCR Master Mix Takara Clontech (Diatech, Jesi, Italy) under the conditions described in the respective literature ( Table 2). The amplification products were separated on 1.5% (w/v) agarose gels.

Numerical Analysis of Genotypic Profiles
Genotypic profiles obtained by rep-PCR were analyzed by the BioNumerics V5.10 software (Applied-Maths, Belgium), using the Dice coefficient for pairwise comparison, and the Unweighted Pair Group Method, using Arithmetic Averages (UPGMA) clustering.

Sequencing and Sequence Analyses
Before sequencing, the PCR products were purified using the HiYield Gel/PCR Fragment Extraction Kit RBC Bioscience (Diatech) according to the instructions. The sequencing of both amplicon strands was carried out by Eurofins Genomics (Ebersberg, Germany), with the same primers used for amplification. Sequencing was carried out for species-level identification based on 16S rRNA gene identity and to confirm the PCR product identity in targeted assays of inlA, inlJ, hly, llsA and mazF genes by BLASTn. Sequencing of the 16S rRNA gene was carried out for one isolate from each rep-PCR cluster separated at 85% similarity, a cut-off value chosen on the basis of the lowest similarity exhibited by ten duplicate strains in preliminary experiments.

Typing and Identification of Listeria spp. Isolates
Numerical analysis of rep-PCR profiles enabled the distinction of three main clusters exhibiting comparable internal diversity, each comprising exclusively of L. innocua, L. monocytogenes or L. welshimeri isolates, as ascertained by sequencing the 16S rRNA gene, with the exception of L. welshimeri 81 that fell outside the group formed by the other L. welshimeri isolates. It was observed that the rep-PCR profile of this strain lacked most of the weaker bands observed in the other L. welshimeri profiles, exhibiting only a strong band common to those profiles. One isolate that did not belong to any of the three main clusters was identified as L. seeligeri ( Figure 1). In all cases, identification at the species-level was based on 99% sequence identity with entries in the public domain database.
From Figure 1 and Table 1, it can be observed that isolates deriving from the same sample type showed less than 85% profile similarity, indicating that the Listeria spp. subtypes in the same food category differed. Conversely, highly similar genotypes were shown by isolates from different food categories.
Isolate 5 from a sample of raw sausage appeared highly similar to Isolates 1 and 3 from fresh pork ( Figure 1, Table 1).
Based on genotypic identification, Listeria species were present in 34 out of 257 bovine meat samples, 5 out of 61 pork meat samples, 3 out of 45 fresh fish samples, 6 out of 43 raw cured pork product samples, 6 out of 36 frozen pizza samples, 3 out of 29 pizza dough samples, 3 out of 23 frozen pasta samples, 2 out of 15 roasted peppers samples, 9 out of 97 cheese-making plant swabs, and 6 out of 48 cheese-aging board swabs.
L. monocytogenes was detected in four fresh bovine meat samples, four fresh pork meat samples, two raw sausages, and two swabs from cheese-aging boards.
Variability in Listeria spp. occurrence among the three laboratories for different sample categories indicated an uneven distribution of these bacteria among sampling locations (Table 1).

Pathogenic Potential of the Listeria spp. Isolates
All of the L. monocytogenes isolates were attributed to serotype 1/2a, although none belonged to the major epidemic clone ECIII included in this serotype. L. monocytogenes isolates exhibited β-hemolysis and were all positive for the inlJ-gene-targeted PCR assay, applied only to the L. monocytogenes isolates, being, therefore, classifiable as virulent according to Liu et al. 2003 [14].
Listeria isolates not belonging to the L. monocytogenes species were tested for the presence of L. monocytogenes inlA, hly and llsA genes. These were found in seven isolates (9.3% of all isolates). An amplification product with 100% sequence identity with L. monocytogenes inlA was obtained from L. innocua 74, L. seeligeri 22, and L. welshimeri 86. The L. monocytogenes hly gene was detected in L. innocua 23, 45, and 95 and L. welshimeri 86 and 98. Notably, L. welshimeri 86, endowed with both L. monocytogenes inlA and hly genes, could be highly virulent according to Moura et al. 2019 [10]. However, as all the Listeria spp. isolates, except for L. monocytogenes isolates, it was not hemolytic, indicating either the inability to express hly or that the gene product is not functional. The llsA gene was not detected in any Listeria spp. other than the L. monocytogenes isolate.
The PCR test for the presence of the mazEF toxin-antitoxin system gave an amplification product for all three species tested, though one or two discrepancies were present in the forward-primer-annealing site and three discrepancies were present in the reverse-primer-annealing site for all the species other than the L. monocytogenes. The mazF gene appeared to be absent in L. innocua isolates 31, 35 and 80 and to be most likely mutated at the primer-annealing sites for L. monocytogenes isolates 6 and 10, for which amplification was inefficient with the mazF-specific PCR test. However, most isolates were shown to possess this additional stress-adaptation system.

Discussion
This study presents the results of Listeria spp. occurrence based on the analysis of hundreds of food products and food-contact surfaces in manufacturing plants, and, as such, provides an insight into the distribution of these bacteria and to consumer exposure. With differences among the food categories, Listeria spp. appeared to be frequent in all food matrices and production plants, indicating that safety measures to prevent contamination must be improved, especially since innocuous species are considered indicators of L. monocytogenes co-occurrence [18].
A discrepancy between physiologic and molecular identification was observed in this study, since 27 of the 102 presumptive Listeria spp. provided by the laboratories were not confirmed to belong to the Listeria genus. It must be underlined that this constitutes a problem for the detection and enumeration of Listeria spp. in foods, since the EN ISO Standards 11290-1 and -2 often make it difficult to efficiently detect and enumerate Listeria spp. apart from L. monocytogenes in foods [19]. In particular, in this study, some presumptive Listeria spp. were genotypically identified as Bacillus spp. and Enterococcus species.
The application of rep-PCR with the GTG5 primer to different species in this study highlighted the usefulness of this technique in differentiating isolates of L. monocytogenes, L. innocua, and L. welshimeri. Its application in new studies regarding different Listeria species could lead to the definition of profile types to be used as references for a rapid preliminary identification of new isolates at the species-level, simultaneously allowing intra-species clustering. The possible correspondence of the intra-species clusters defined by rep-PCR with intra-species genetic variants should be investigated. The intra-species relationships defined by this genotyping method showed that highly similar genotypes belong to isolates from different food categories, meaning that some Listeria spp. subtypes are not associated with specific food products and are distributed across diverse environments, purportedly for their better ability to persist in processing plants. These subtypes should be thoroughly characterized genotypically and physiologically and preferentially used to test the efficiency of sanitation procedures.
An L. monocytogenes subtype, possibly derived from fresh meat, was identified in raw fermented sausages, an RTE food category. The highly similar isolates found in pork meat and sausages might represent a L. monocytogenes subtype able to survive in ready-to-eat cured pork products and should be investigated with genotyping methods with a higher resolution. A genetically closely related strain was isolated from bovine meat, indicating that the subtype comprising these isolates is distributed in different environments and is tolerant to the sausage-ripening process. This highlighted that the risk of L. monocytogenes growth in raw cured pork products should be assessed and that more efforts must be devoted to preventing the presence of L. monocytogenes in pork meat and processing plants by identifying specific contamination routes.
The monitoring of raw meat for the presence of L. monocytogenes through the application of genotyping methods with high discrimination power could elucidate whether this is the origin of processing-plant contamination, or whether inappropriate production and cleaning procedures favor the persistence of environmental L. monocytogenes subtypes in manufacturing plants. A similar approach should be applied to all RTE food categories in order to reduce contamination risk.
The isolation of L. monocytogenes serotype 1/2a in this study is in agreement with reports of its wide distribution in food and food-processing environments [12,20,21] and the consequently numerous listeriosis outbreaks that have, until recently, been caused by this serotype [22][23][24][25][26].
In this investigation, the genetic analysis of virulence characters highlighted that food can be a source of potentially pathogenic strains of Listeria spp. belonging to species generally considered to be innocuous. It was found that the L. monocytogenes inlA and hly virulence determinants can be harbored not only, as previously reported, by atypical L. innocua strains, but by L. welshimeri and L. seeligeri isolates as well. Therefore, species identification is not sufficient to estimate the risk associated with the presence of Listeria spp. in food, and both contamination prevention and the identification of contamination sources should be extended to all Listeria species. Further studies are needed to elucidate if the virulence genes found in this study in Listeria strains belonging to species considered commonly to be innocuous are functional.
The presence in most isolates of the mazEF toxin-antitoxin system, which indicates efficient heat-shock-adaptation capacity [15], underlines the need to reinforce contamination prevention to reduce the risk that these bacteria pose to consumers.