Presence and Persistence of Listeria monocytogenes in the Danish Ready-to-Eat Food Production Environment

: Listeria monocytogenes is an ubiquitously occurring foodborne bacterial pathogen known to contaminate foods during the production processes. To assess the presence and persistence of L. monocytogenes in Danish ready-to-eat (RTE) food production companies in response to a Listeria awareness campaign, the production environment of selected companies were sampled in 2016 and in 2020. Whole genome sequencing (WGS) was performed to characterize the isolates ( n = 50, plus 35 isolates obtained from the routine surveillance during 2016–2020), including investigation of the presence of virulence, persistence and resistance genes. The number of companies that tested positive by culture was 17/39 (43.6%) in 2016 and 11/34 (32.4%) in 2020, indicating a limited effect of the campaign. WGS analyses of the 85 isolates showed that the most common sequence types (STs) were ST8 and ST121. The single nucleotide polymorphism (SNP) analysis showed that isolates coming from the same company and belonging to the same ST exhibited <10 SNP differences regardless of the sampling year and whether the samples came from the environment or products, indicating the persistence of the speciﬁc STs. Several prevalent STs were found in clinical cases concurrently, including genetically similar isolates. This highlights the issue of persistent L. monocytogenes in the food production environment and the need for improved risk communication and mitigation strategies.


Introduction
Listeria monocytogenes is an opportunistic foodborne pathogenic bacterium that can cause severe clinical disease in humans and animals (listeriosis). Despite the low number of confirmed human cases, listeriosis is considered one of the most serious foodborne diseases in the EU due to its high mortality and hospitalization rate [1]. In Denmark, there are approximately 40-60 laboratory confirmed listeriosis cases with up to five outbreaks per year [2] and its disease burden has been estimated as the third highest among seven foodborne pathogens, after campylobacteriosis and salmonellosis [3].
Listeria monocytogenes is a facultative anaerobic bacterium with the ability to survive and grow outside the animal host cells. It is ubiquitous in the natural environment such as in soil and water and can be transmitted to the food chain through various routes. Unlike other foodborne pathogens, L. monocytogenes can multiply under various "stressful" conditions such as pH ranging from 4.6 to 9.2, relatively low water activity (a w of 0.90) and low temperatures [4][5][6], thus enabling the bacterium to survive under various food processes and storage. Additionally, the pathogen can form biofilms on surfaces and tolerate disinfectants and antimicrobial agents [7]. Sufficient heating can eliminate the pathogen during food processing, but food products may become re-contaminated after the lethal treatment due to L. monocytogenes being present in the food processing environment. Ready-to-eat (RTE) food products that either receive no listericidal treatment or become re-contaminated during packing steps and have a composition that allows for growth of L. monocytogenes within the refrigerated shelf-life, constitute the main source of listeriosis outbreaks [7,8]. diversity, persistence of genetically similar strains (here defined as strains with <20 SNP differences) over the years and presence of genes encoding for antimicrobial/biocide resistance and virulence factors. Finally, putative links between the isolates from the production environment and human cases were surveyed to elucidate the significance of presence and persistence of L. monocytogenes in the production environment in response to the Listeria awareness campaign.

Bacterial Isolates and Other Relevant Information
Food production companies, which produce known L. monocytogenes risk products (RTE fish and meat products) in different regions of Denmark, were selected for the study in 2016 based on information from the DVFA surveillance program. The sampling design aimed for recruitment of 20 fish and 20 meat production companies. The project was repeated in 2020 with the aim of taking samples from the same companies. A replacement company was selected for sampling in 2020 if the company from 2016 was not available.
Ten swab samples from the food production environment (equipment and surfaces) were taken from each of the food companies. Sampling was carried out by experienced samplers (official food inspectors), according to the accredited procedure Guidelines on sampling the food processing area and equipment for the detection of Listeria monocytogenes Version 3-20 August 2012 [22]. Product contact and non-contact surfaces, including from the drains, were represented from all the companies. Swab samples were sent to the DVFA's accredited laboratory (Ringsted, Denmark) and processed according to EN/ISO 11290-1 method for identification of L. monocytogenes by culture. One isolate from each positive sample was selected for WGS and stored frozen (−80 • C) in Tryptone Soy Broth (Oxoid, Fisher Scientific, Kamstrup, Denmark) supplemented with 30% glycerol.
In addition to this study's environmental sampling, the joint DVFA and DTU archive was searched for L. monocytogenes isolates, which came from products and environmental swabs collected from the same companies as part of routine surveillance in the period 2016-2020. Only the isolates with available WGS data were included in the analyses.

Whole Genome Sequencing
DNA extraction followed the standard DVFA protocol. Briefly, L. monocytogenes were revived from frozen stock (−80 • C) by streaking on Tryptic Soy blood agar (Statens Serum Institut, Copenhagen, Denmark) for 24 h at 37 • C. Single colonies were sub-cultured in Tryptic Soy Broth (Oxoid, Fisher Scientific, Roskilde, Denmark) for 18 h at 37 • C. DNA was extracted from this culture using the Easy-DNA kit (Invitrogen, Waltham, MA, USA). DNA concentrations (0.18-0.28 ng/µL) were measured using the Qubit dsDNA BR Assay Kit (Invitrogen). Sequencing was done using MiSeq sequencing machines (Illumina Inc., San Diego, CA, USA) and the Nextera XT Library preparation protocol for paired-end reads of 250 bp. The WGS data were then processed using an internal quality-control (QC) pipeline at the Center for Genomic Epidemiology, Technical University of Denmark. This pipeline performs trimming of low-quality and adaptor sequences and de novo assembly using SPAdes.

Characterization of Isolates in Terms of Virulence, Resistance and Plasmids
The assembled sequences were screened for the presence of genes encoding antimicrobial and other resistance genes (stress, heat and biocide) using AMRFinderPlus v.3.10.14 with a default setting of minimum coverage of 90% and minimum identity of 90% [27]. In addition, eight genes encoding for biocide resistance (Tn6188 (which includes qacH and tetR genes), qacA, qacC, qacH, bcrABC, emrC, emrE and mdrL) were manually searched using ABRicate with minimum 70% coverage and 90% identity [28].

SNP Analyses
Single nucleotide polymorphisms (SNP) calling and construction of SNP matrices were performed using the CSIPhylogeny v.1.4 [32], where the paired-end trimmed reads were mapped to the L. monocytogenes EGD-e chromosome, complete genome (NC_003210.1) as a reference. The SNPs were selected with the following criteria: (1) a minimum of 10% relative depth at SNP positions, (2) a minimum distance of 10 bps within each SNP, (3) SNP quality of >30 and (4) mapping quality of >25. Phylogenetic trees based on SNP matrices were visualized by iTOL v.6 [33].

Human Listeriosis Cases during the Study Period
Listeria monocytogenes is a notifiable human disease in Denmark. Regional, national and international food and water borne outbreaks are monitored and investigated by the Central Outbreak Group (DCUG) in Demark. The group is made up of members from the Statens Serum Institute (SSI), the Danish Veterinary Food Administration (DVFA) and National Food Institute (DTU) and meets once a week to ensure a coordinated effort to manage outbreaks. Listeria monocytogenes isolates from patients are handled by SSI for typing, while sampling and sequencing of food and production environment are conducted jointly by DVFA and DTU, as described in Lassen et al. [10]. Outbreaks are reported to EFSA and in the Annual Report on Zoonoses. Possible links between the isolates from this study and human cases were searched by screening these public sources [2,[34][35][36][37].

Summary of the Bacterial Isolates
A total of 777 environmental samples were obtained from 53 companies, of which 32 and 20 samples were positive by culture for L. monocytogenes in 2016 and 2020, respectively. An overview of the samples and companies are shown in Table 1, with more details provided in Supplementary Table S1 Table S1). Of the 52 L. monocytogenes isolates, the quality of WGS data from two isolates were found to be suboptimal. The remaining 50 L. monocytogenes isolates (30 from 2016 and 20 from 2020), which came from 24 companies, were used for further WGS analyses ( Table 2 and Table S1). Table 2. Identification of additional Listeria monocytogenes isolates from the production environment and products from the 53 selected food production companies. Isolates were obtained during routine surveillance performed by the Danish Food and Veterinary Administration during 2016-2020. Abbreviations: Env = environment, Prod = product. No additional strains were identified for 2016 and from the production environment data in 2020.
Additional database search for L. monocytogenes, which had been isolated during the DVFA's routine surveillance of the participating companies in 2016-2020, resulted in 35 additional isolates (21 product and 14 environmental, Table 2), none of which came from the 2016 surveillance program. In total, 85 isolates with WGS data from 27 companies were used for further WGS analyses.

Distribution of STs and SNP Analyses
The metadata of the 85 isolates and selected output from the Listeria PasteurMLST database are shown in Supplementary Table S3. Sixteen STs of lineages I and II were identified; ST1, ST6, ST87, ST296 and ST416 belong to lineage I and the rest to lineage II (Table 3). An overview of the isolates and their origin, i.e., company, sampling year and source (environment vs. product) and STs are summarised in Figure 1. The STs, which were isolated over two or more years from the same company were: ST7, ST8, ST121, ST399 and ST451. Ten out of 27 companies harboured isolates of different STs (Table 3). Table 3. Overview of the STs of 85 Listeria monocytogenes isolates obtained from the food processing environment and products between 2016 and 2020.   The phylogenetic tree based on SNP analyses showed clear separation according to lineages and STs ( Figure 2). The isolates belonging to the same ST and coming from the same company showed only small SNP differences (<10 SNP) regardless of sampling year and source (environment vs. product), e.g., ≤6 SNP differences among the nine ST451 isolates. Isolates which belonged to the same ST but came from different companies showed greater SNP differences, generally >50. In the case of ST121 (Figure 3a), a closer look revealed that company no. 2 harboured L. monocytogenes isolates belonging to this ST, which exhibited between 0 and 9 SNP differences regardless of their year of isolation (2017 and 2020). ST8 (Figure 3b) did not show as clear clustering among the isolates from within a company, i.e., isolates from company no. 4 exhibited between 1 to 55 SNP differences, while the SNP difference between company no. 4 and no. 12 was only 34. The phylogenetic tree based on SNP analyses showed clear separation according to lineages and STs ( Figure 2). The isolates belonging to the same ST and coming from the same company showed only small SNP differences (<10 SNP) regardless of sampling year and source (environment vs. product), e.g., ≤6 SNP differences among the nine ST451 isolates. Isolates which belonged to the same ST but came from different companies showed greater SNP differences, generally >50. In the case of ST121 (Figure 3a), a closer look revealed that company no. 2 harboured L. monocytogenes isolates belonging to this ST, which exhibited between 0 and 9 SNP differences regardless of their year of isolation (2017 and 2020). ST8 (Figure 3b) did not show as clear clustering among the isolates from within a company, i.e., isolates from company no. 4 exhibited between 1 to 55 SNP differences, while the SNP difference between company no. 4 and no. 12 was only 34.    (16)(17)(18)(19)(20) and the last two digits indicate the company ID; no highlight and grey highlight indicate environmental and product samples, respectively; STs are showed in the right column with colour; and SNP differences between selected isolates are shown as numbers on the right.

Characterisation of the Selected Virulence, Resistant Genes and Plasmids
The presence of selected virulence and resistance genes as well as plasmids corresponded well with the STs (Figure 4). Internalin A was found in all isolates except for one isolate, which is likely to have contained the gene but only showed 62.6% coverage due to the location at the start of a contig (1503 bp). Three of the ST9 isolates had PMSC mutation type 29 (allele inlA_47), while the last isolate contained PMSC type 11 (allele inlA_44) according to the current typing overview of reported PMSCs in inlA [38][39][40]. All ST121 isolates contained PMSC mutation type 6 (allele inlA_49) [41]. ST6 isolates were found to  (16)(17)(18)(19)(20) and the last two digits indicate the company ID; no highlight and grey highlight indicate environmental and product samples, respectively; STs are showed in the right column with colour; and SNP differences between selected isolates are shown as numbers on the right.

Characterisation of the Selected Virulence, Resistant Genes and Plasmids
The presence of selected virulence and resistance genes as well as plasmids corresponded well with the STs (Figure 4). Internalin A was found in all isolates except for one isolate, which is likely to have contained the gene but only showed 62.6% coverage due to the location at the start of a contig (1503 bp). Three of the ST9 isolates had PMSC mutation type 29 (allele inlA_47), while the last isolate contained PMSC type 11 (allele inlA_44) according to the current typing overview of reported PMSCs in inlA [38][39][40]. All ST121 isolates contained PMSC mutation type 6 (allele inlA_49) [41]. ST6 isolates were found to have a deletion of 9 bp from positions 2212-2220 in the gene, which may or may not produce a functional inlA. ammonium compounds (QACs) (qacC, bcrABC, emrC, emrE), were found.
All isolates contained resistant genes against fosfomycin (fosX) and lincosamide (lmo0919). ATP-dependent protease (ClpL)-encoding clpL gene, which increases heat resistance in L. monocytogenes, was found in ST 121 and ST 9 isolates. The clpL gene was on the same contig as the replicon genes of pLM5578 (except for one of ST121 isolate) and pLM33 of ST9 isolates, suggesting that the clpL gene is likely to be encoded on the pLM5578 and pLM33 plasmids. Likewise, many of the cadC genes were found to be on contigs that were associated with known plasmid genes. However, cadC genes were also identified in isolates belonging to ST14 and ST155, while no plasmid replicons were found in these isolates.  LIPI-1 was present in all isolates, while LIPI-3 was found in five isolates of ST1, ST6 and ST416, all of which belonged to lineage I. LIPI-4 was found in the remaining lineage I strains (ST87 and ST296).
Out of the eight genes for biocide resistance investigated, qacA, mdrL were found in all isolates. Tn6188, which includes qacH, was found in 21 isolates (all isolates from ST121 and 2 of ST9). No isolates were found to contain qacH resistance genes without the concurrent presence of Tn6188. None of the other genes, which confer tolerance to quaternary ammonium compounds (QACs) (qacC, bcrABC, emrC, emrE), were found.
All isolates contained resistant genes against fosfomycin (fosX) and lincosamide (lmo0919). ATP-dependent protease (ClpL)-encoding clpL gene, which increases heat resistance in L. monocytogenes, was found in ST 121 and ST 9 isolates. The clpL gene was on the same contig as the replicon genes of pLM5578 (except for one of ST121 isolate) and pLM33 of ST9 isolates, suggesting that the clpL gene is likely to be encoded on the pLM5578 and pLM33 plasmids. Likewise, many of the cadC genes were found to be on contigs that were associated with known plasmid genes. However, cadC genes were also identified in isolates belonging to ST14 and ST155, while no plasmid replicons were found in these isolates.

Possible Links to Danish Human Listeriosis Cases
According to published reports and communication with DCUG, 10 different STs had been implicated in human cases in Denmark during the study period ( Table 4). Five of these STs were also found in the studied RTE food companies. A closer look at the WGS from human ST451 isolates from 2014 and 2020 showed that there were <10 SNP differences between isolates from these clinical cases and food production isolates from the present study. Moreover, isolates taken from the environment in 2016 were genetically nearly identical (≤6 SNPs) to those collected from the products in 2020. Taking that information together, it seems plausible that this specific L. monocytogenes ST451 strain has persisted over time in the production environment and may have contributed to human cases, with a 6-year interval.

Discussion
In the present study, L. monocytogenes were found from the production environment of 17/39 (43.5%) companies in 2016 and 11/34 (32.4%) in 2020, indicating that the prevalence was slightly lower in 2020 but that L. monocytogenes was still found in one third of the participating RTE food processing facilities. It should be noted that the difference in prevalence between the two years cannot be compared directly since only 20 companies participated in both years and the number of samples per company were variable. Additionally, sampling in 2020 was more focused in the area after heat treatment rather than the area where the raw materials were handled. This may have resulted in the lower number of positives in 2020.
L. monocytogenes isolates belonging to 16 different STs were found in the environmental and food product samples from the RTE food companies (Figure 1). In Norway, STs belonging to clonal complex (CC) 7, CC8, CC9, CC121 were the most prevalent [16] and, likewise, CC9 and CC121 were the most frequently found strain types in Switzerland [42]. In our study, ST8 and ST121 were the most prevalent isolates and ST7 were found in lesser extent, but no isolates belonging to CC9 (e.g., ST9 and ST477) were found. It has been hypothesized that ST9 strains are highly adapted to meat production environments [17,42]. Despite our limited sample size, ST9 was in low prevalence in the Danish meat production environment. While each company was sampled several times, it is possible that all STs were not captured since WGS were performed on only one of the colonies taken randomly from the culture plates from each positive sample. It should be noted that selective media may also select for specific STs during culture [43].
Results from the present study showed that the same STs can be isolated over long periods from the same companies; the isolates belonging to the same ST and obtained from the same companies were genetically similar (<10 SNP) regardless of the sampling years or whether the samples were taken from the products or the environment (Figure 3). Wang et al. [44] have shown that the probability of two isolates with <20 SNP differences coming from the same facility is high (70%) and suggested 20 SNP to be used as a threshold for calling isolates identical in L. monocytogenes outbreak and traceback investigations. On the contrary, Fagerlund et al. [17] found small SNP differences (7-11 SNPs) between two isolates from two different facilities, making it difficult to deduce the origin of an outbreak unless the two facilities were linked through an unrealized common source of contamination. Supplementation with wgMLST analyses would provide further information, as wgMLST includes accessory genes, while SNP analyses exclude variable elements such as plasmids and prophages [17,45]. However, this was beyond the scope of the present study. The isolates belonging to the same ST and coming from the same companies in our study clustered together, while larger SNP differences were found among the ST isolates taken from different companies (Figures 2 and 3a). ST8 was an exception and will be discussed further below. Taken together, the low number of SNP differences found among isolates from individual RTE food companies in the present study is a strong indicator that L. monocytogenes isolates have persisted in the production environment over time, and transfer between the environment and food products occurred during the production.
While isolates of the same STs from different companies showed generally larger SNP differences (>50 SNP), highly similar (34 SNPs) ST8 isolates were obtained from companies no. 4 and 12 ( Figure 3b). Interestingly, ST8 isolates from company no. 4 showed large genetic variation (up to 55 SNP), making it difficult to evaluate if the ST8 isolates persisted and mutated over time, or there were repeated introductions of ST8 L. monocytogenes into the RTE food company. A recent Norwegian study also reported finding genetically similar isolates belonging to STs such as ST451 and ST37 in different natural environments, geographical locations and clinical cases [46], indicating that our knowledge on the transmission of the bacterium through the natural and urban environment is limited.
Currently, all STs of L. monocytogenes are considered as pathogenic and control measures are taken as soon as L. monocytogenes are found. However, there seem to be variations among different strains in terms of pathogenicity and persistence and more genes that are suspected to play an important role for virulence have been identified in recent years. A recent report from FAO/WHO [20] has proposed sub-type specific risk assessments for virulence ranking of L. monocytogenes strains. The ranking is based on the information on lineage and presence/absence of LIPI-1/3/4 and inlA PMSC, with proposed grouping as follows (highest to lowest risk): (1) Lineage I strains with LIPI-1, 3, or 4 with a full length inlA; (2) strain of any lineage with complete and functional LIPI-1 and with full length inlA; and (3) any L. monocytogenes strain with a truncated inlA. The proposed ranking is still in its infancy and not validated. However, if it is optimised, validated and implemented, this ranking could potentially lead to optimisation of resource allocations in risk management and less food waste. Considering our dataset as an example, the truncated inlA were present in only 23 out of 85 isolates (ST9 and ST121), making categorizing them as the lowest risk ranking group 3, while the rest are classified as either risk ranking 1 (4 STs) or 2 (9 STs). Highly similar ST451 (lineage II) isolates, which belong to the suggested risk ranking group 2, were putatively associated with concurrent human cases in our example. Although lineage I is considered as more virulent and risk ranking is higher compared to other lineages, lineage II strains do still cause sporadic clinical cases, as shown in Table 4. More studies on virulence genes, validation of the virulence gene expressions, as well as the genes' associations within the sub-types are needed to evaluate the usefulness of the proposed ranking system.
All the isolates, which were present in the RTE food companies over two or more years, belonged to lineage II in the present study. In addition, tolerance to quaternary ammonium compounds (QACs) conferred by qacH gene were found in the isolates with apparently reduced virulence or truncated inlA, which are considered lower risk according to the proposed risk ranking. This seems to support the common notion that persistent strains are generally equal to or less virulent than non-persistent strains [21,47]. Some of the isolates, which were present over two or more years e.g., ST451, did not seemingly have any additional biocide and resistant genes than qacA and mdrL genes. However, admittedly our gene search was not extensive and other genetic mechanisms may be at play. There is still limited knowledge related to the mechanisms of persistence and virulence among different strains and such knowledge will be useful in terms of risk assessment and management of L. monocytogenes in the food production environment.
Results from the present study are in agreement with previous studies [17,48] that L. monocytogenes STs with few SNP differences are able to colonise RTE food companies, thereby demonstrating the need for a multi-facetted approach to manage the risk of L. monocytogenes in RTE products. The DVFA's Listeria awareness campaign launched in 2015 was a risk communication strategy aimed at reducing the prevalence and outbreaks caused by the bacterium through education of food producers, food inspectors and consumers, including susceptible segments of the population. Despite this campaign, our study shows that the overall prevalence of L. monocytogenes in high risk RTE facilities was largely unchanged between 2016 and 2020. Moreover, cases of listeriosis in Denmark are not going down, with recorded cases having increased from 39 in 2016 to 86 in 2022 (median 58, equivalent of 1 case per 100.000 inhabitants) [49]. Difficulties in reaching specific target groups and finding the best way to communicate the risk of listeriosis to different consumers is well documented [50]. In this review, it was also noted that while the internet is the most common way to communicate food safety information, it may not be the best medium, depending on the target group. The DVFA's campaign relied in part on the internet. It appears that new ways to reach the industry and consumers are needed to communicate about topics such as the importance of consumption prior to the best-before dates for RTE products; industrial and domestic hygiene; suitable HACCP plans; and application of L. monocytogenes growth controlling hurdles in the form of listericidal treatments (e.g., heat treatment), or (re)formulation of products (e.g., addition of organic acids; see Dalgaard & Mejlholm [51]).

Conclusions
The study presented some evidence for persistence of L. monocytogenes in the RTE food production environment. Transfer of L. monocytogenes between the environment and products seems to occur during food production, as isolates with the same ST from the same company were genetically almost identical (<10 SNP) regardless of sampling year and whether the isolates came from products or production environment. Clinical isolates genetically similar to the persistent strains isolated from the food production environment were identified. Prevalence of L. monocytogenes remained comparable between 2016 and 2020 samplings, which, taken together with the increasing trend in listeriosis cases in Denmark, may indicate that the current risk communication strategy is not working, despite the DVFA's intensive Listeria awareness campaign and availability of comprehensive information resources on the agency's website. Risk assessment and management of L. monocytogenes require a multi-actor approach. To optimise eradication, control and risk identification, further research is needed in terms of risk communication strategies to reach stakeholders, as well as mitigation strategies, which are based on an understanding of the mechanism of persistence and virulence.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/hygiene3010004/s1, Table S1: Summary of participating companies and the number of samples for the study; Table S2: List of databases and reference genes used for the study; Table S3: Metatdata of the isolates used for the study with the selected output from the Listeria PasteurMLST database.