Prevalence and Population Diversity of Listeria monocytogenes Isolated from Dairy Cattle Farms in the Cantabria Region of Spain

Simple Summary The origin and prevalence of Listeria monocytogenes was studied in dairy cattle farms in order to examine its diversity and determine its possible persistence in manure. The utilization of manure for agricultural purposes is common in many countries. While properly treated and managed manure is an effective and safe fertilizer, foodborne illness outbreaks can occur, as many of the most prominent foodborne pathogens are carried by healthy livestock. It is, therefore, necessary to study the origin and persistence of zoonotic agents in general and of L. monocytogenes in particular, in order to avoid recirculation in farms and reduce risk for human populations. Abstract Listeria monocytogenes is an opportunistic pathogen that is widely distributed in the environment. Here we show the prevalence and transmission of L. monocytogenes in dairy farms in the Cantabria region, on the northern coast of Spain. A total of 424 samples was collected from 14 dairy farms (5 organic and 9 conventional) and 211 L. monocytogenes isolates were recovered following conventional microbiological methods. There were no statistically significant differences in antimicrobial resistance ratios between organic and conventional farms. A clonal relationship among the isolates was assessed by pulsed field gel electrophoresis (PFGE) analysis and 64 different pulsotypes were obtained. Most isolates (89%, n = 187) were classified as PCR serogroup IVb by using a multiplex PCR assay. In this case, 45 isolates of PCR serogroup IVb were whole genome-sequenced to perform a further analysis at genomic level. In silico MLST analysis showed the presence of 12 sequence types (ST), of which ST1, ST54 and ST666 were the most common. Our data indicate that the environment of cattle farms retains a high incidence of L. monocytogenes, including subtypes involved in human listeriosis reports and outbreaks. This pathogen is shed in the feces and could easily colonize dairy products, as a result of fecal contamination. Effective herd and manure management are needed in order to prevent possible outbreaks.


Introduction
Listeria monocytogenes is a Gram-positive, non-spore-forming, motile, facultative anaerobic, rod-shaped bacterium that is an opportunistic pathogen. It causes listeriosis, a severe foodborne infection, leading to blood, brain or fetal infections in humans. Ruminants discussed, bearing in mind that information on the molecular level of the foodborne bacteria can help in anticipating outbreaks and designing effective control strategies in the event of possible disease emergence. The PCR serogroups and pulsotypes of all isolates, as well as the WGS analysis and the antimicrobial susceptibility profiles of a selection of 45 isolates of PCR serogroup IVb were performed. According to the results of this study, the L. monocytogenes analyzed encode virulence factors associated with human disease, capable of surviving at least 35 days in livestock waste and 70 days on pasture. The information presented here helps to increase our understanding of the on-farm life cycle of L. monocytogenes.

Study Design
The sampling was performed during 2017-2019 and included 14 dairy cattle farms (5 organic and 9 conventional) from 8 different sites in the Cantabria region (north coast of Spain). Samples were taken from forage, water, raw tank milk, the tank milk filters, fresh feces, stored manure and soil. In total, 424 samples were collected and tested for L. monocytogenes.

Sample Collection, Processing and L. monocytogenes Isolation
The samples were collected using sterile material, transported to the laboratory on ice within 2 h from the collection and subjected to detection of L. monocytogenes immediately upon arrival using 10 g for solid samples; 3 L filtered through a 0.5 nm sterile filter for water; 100 mL centrifuged at 8500 rpm/10 min for milk. Rectal swabs and filters from the milk tank were placed directly in the selective media. Samples were treated according to the ISO11290-1:2017 method (https://www.iso.org/standard/60313.html, accessed on 1 January 2018) with modifications. Briefly, for an initial selective enrichment, each sample was inoculated in 100 mL half Fraser broth (HFB) (Oxoid, Basingstoke, UK) for 24 h at 30 • C. A second enrichment followed, by inoculation of 100 µL of the suspension of the first enrichment in 10 mL of Full-strength Fraser broth (FFB) (Oxoid, Basingstoke, UK) for 24 h at 37 • C. Next, 100 µL were spread onto agar Listeria according to Ottaviani and Agosti (ALOA) (bioMérieux, Marcy-l'Étoile, France) and modified Oxford agar (MOX) (Oxoid, Basingstoke, UK) and incubated for 24 h at 37 • C. Presumptive Listeria spp. colonies from each ALOA and MOX plates were re-streaked to tryptic soy agar supplemented with 0.6% yeast extract (TSA-YE) plates (bioMérieux, Marcy-l'Étoile, France), incubated for 24 h at 37 • C and subjected to microscopic examination, Gram staining, glucose fermentation, oxidase, catalase and API tests (bioMérieux, Marcy-l'Étoile, France), according to the manufacturer's instructions.

Decline of L. monocytogenes in Manure
The experiment was performed as previously described [28] and included a tank with manure not inoculated with L. monocytogenes (used as negative control), a tank with manure inoculated with L. monocytogenes strain CIP103575 (https://www.pasteur.fr/en/ cip-distribution, accessed on 1 January 2019; positive control) and samples inoculated with L. monocytogenes strain MS6507 isolated from cattle feces (this study), all repeated six times (n = 6). The L. monocytogenes strains used were grown at 37 • C in tryptic soy broth supplemented with 0.6% yeast extract (TSB-YE) (bioMérieux, Marcy-l'Étoile, France). Before manure inoculation, L. monocytogenes strains were centrifuged in order to remove growth media, washed and resuspended in 0.1% peptone buffer (Oxoid, Basingstoke, UK) to a density of 10 9 CFU/mL. Inoculants were providing 10 6 CFU/g of manure wet weight, a quantity that represents a very high pathogen load and represents the worstcase scenario in terms of pathogen load. The presence of L. monocytogenes was monitored by survival curves generated by plate counting in CHROMagar TM Listeria plates (https: //www.chromagar.com/en/product/chromagar-listeria, accessed on 1 January 2019).

Decline of L. monocytogenes in Pasture Crops
The experiment was performed as previously described [28] using compost infected with L. monocytogenes MS6507. The infected compost was applied to crops and the experiment was performed in a germination chamber under controlled conditions of temperature and light [28]. The presence of L. monocytogenes on the surface of plants was monitored on CHROMagar TM Listeria plates for the generation of the survival curve and the calculation of CFU/g of plant, using 3 g of pasture crops as described in [28].

Pulse-Field Gel Electrophoresis (PFGE) Typing
The clonal groups were identified by pulsed-field agarose gel electrophoresis (PFGE). The PFGE typing results were analyzed following an optimized PulseNet standardized protocol [31], as previously described [32]. L. monocytogenes strain H2446 was used as reference [33]. DNA extraction was carried out on agarose plugs slices in a conventional manner. DNA digestion was performed with two different restriction enzymes, AscI and ApaI (New England BioLabs, Ibswich, MA, USA), and the generated restriction fragments were separated by electrophoresis in a clamped homogeneous electric field (CHEF-DRII) system (BIO-RAD Laboratories, Hercules, CA, USA), with the following running parameters: gradient of 6 V/cm, angle 120 • , temperature of 14 • C, initial switch time 4 s, final switch time 40 s and run time 24 h. Using the Bionumerics software (Version 4.5, Applied Maths, Kortrijk, Belgium), a database of the ApaI and AscI patterns obtained using PFGE was constructed with all the isolates of L. monocytogenes found in the different surveys conducted. These ApaI and AscI patterns were used to assign a new type of combined PFGE according to the standardized optimized PulseNet protocol for L. monocytogenes [31] and the PulseNet Bionumerics manual (http://www.pulsenetinternational.org/protocols/bionumerics/, accessed on 1 January 2019). The Dice correlation coefficient was applied to identify similarities between the PFGE types with a tolerance of 1.5% and an optimization of 0.5%, generating a single dendrogram using the Unweighted-Pair Group Matching Algorithm (UPGMA).

Antimicrobial Susceptibility Test
The antimicrobial susceptibility of the 45 whole-genome sequenced isolates was determined by broth microdilution according to the Clinical and Laboratory Standards Institute (CLSI) [34]. In this procedure, a panel of 8 antimicrobial agents was used: ampicillin (AMP), ciprofloxacin (CIP), erythromycin (ERY), gentamicin (GEN), tetracycline (TET), vancomycin (VA), meropenem (MEM) and cefoxitin (FOX). All antibiotics were purchased from Sigma-Aldrich, Darmstadt, Germany. Staphylococcus aureus ATCC29213, Enterococcus faecalis ATCC29212 and L. monocytogenes ATCCBAA-67 were used as quality control strains. MIC breakpoints were those established by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) [35]. As in the case of FOX, CIP, GEN, TET and VA, there are no interpretive criteria available for L. monocytogenes, we applied the breakpoint defined by EUCAST for Staphylococcus spp. Isolates were considered multidrug-resistant (MDR) when they showed non-susceptibility to at least one agent in three or more antimicrobial categories [36].

DNA Extraction and Sequencing
The total DNA from the L. monocytogenes isolates was purified with the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) and sequenced on a MiSeq device using reagents kit v3 for 2 × 300 paired-end libraries (Illumina, San Diego, CA, USA) as previously described [37].

Nucleotide Accession Numbers
The raw fastq files, annotated assembly and L. monocytogenes isolates information are all available under the BioProject ID PRJNA855628.

Statistical Analyses
All statistical analyses were carried out using GraphPad Prism version 9.4.0 (Graph-Pad Software, San Diego, CA, USA). To investigate the association between positive L. monocytogenes samples and source frequencies, the χ 2 and Fisher s exact test was performed. In linear regression analyses, the t-test was used to determine whether the slope of the regression line differs significantly from zero. In all cases, differences were considered statistically significant at p < 0.05.

Environmental Sampling, Incidence and L. monocytogenes Isolation
A total of 424 samples were collected and processed from 14 different dairy cattle farms (5 organic and 9 conventional) in the Cantabria region (Northern Spain, Atlantic Coast). Samples were obtained from forage, silage, concentrate, water, raw tank milk and tank milk filters, fecal samples, manure and soil from within and around the stable. Fecal samples were collected from both unhealthy (diarrhea/infection/miscarriage) and apparently healthy dairy cows (without symptoms). The majority of samples (91%, n = 387) were positive for Listeria spp., of which 54% (n = 211) were positive for L. monocytogenes (Table 1). Strains were isolated by conventional microbiological plating methods and presumptive L. monocytogenes isolates were tested by multiplex PCR-based serogrouping, as described in Materials and Methods. Each of the 14 farms yielded at least one positive L. monocytogenes sample. Environmental samples obtained from sites around the farms related to feces (soil, slurry spreader and slurry drain) and fecal samples (stored manure and fresh feces) showed a higher percentage of L. monocytogenes and Listeria spp. in general in comparison with the feed samples and the raw dairy samples (χ 2 and Fisher s exact test p < 0.0001, Table 1). L. monocytogenes in particular, and Listeria spp. in general, were found in both organic and conventional farms with no statistical difference between them (χ 2 and Fisher s exact test p = 0.0960, Table 2). The full list of the L. monocytogenes isolates is shown in the Supplementary Materials (Table S1).

Transmission of L. monocytogenes
To assess the relationship of L. monocytogenes prevalence between sample categories, farm-level prevalence scatter plots were generated, and regression analyses were performed as described before [52]. Scatter plots were constructed for all possible relationships between samples, but only the environmental samples versus stored manure regression analysis had a slope significantly different from zero (p < 0.05) ( Figure 1). Since the prevalence of L. monocytogenes in stored manure is higher than in environmental samples ( Table 1) and assuming that transmission is expected to happen from high-contaminated to less-contaminated sites, these data indicate that L. monocytogenes is possibly transmitted through the manure to the farm ecosystem.
farm-level prevalence scatter plots were generated, and regression analyses were performed as described before [52]. Scatter plots were constructed for all possible relationships between samples, but only the environmental samples versus stored manure regression analysis had a slope significantly different from zero (p < 0.05) ( Figure 1). Since the prevalence of L. monocytogenes in stored manure is higher than in environmental samples ( Table 1) and assuming that transmission is expected to happen from high-contaminated to less-contaminated sites, these data indicate that L. monocytogenes is possibly transmitted through the manure to the farm ecosystem.

Decline in L. monocytogenes in Livestock Waste
One of the L. monocytogenes isolates (MS6507, isolated from rectal swab sample from cow) was selected to assess its ability to survive in manure. L. monocytogenes CIP103575 was used as a control strain. Liquid manure samples were inoculated, and survival curves were created. L. monocytogenes was detected until 28 days without enrichment. After that point, enrichment was performed before plating to confirm the presence/absence of L. monocytogenes, which was detected for one more week, until 35 days had passed. Negative control tanks always yielded negative results for L. monocytogenes. The results are shown in Figure 2A, expressed as the average ± standard deviation (n = 6).

Decline in L. monocytogenes in Livestock Waste
One of the L. monocytogenes isolates (MS6507, isolated from rectal swab sample from cow) was selected to assess its ability to survive in manure. L. monocytogenes CIP103575 was used as a control strain. Liquid manure samples were inoculated, and survival curves were created. L. monocytogenes was detected until 28 days without enrichment. After that point, enrichment was performed before plating to confirm the presence/absence of L. monocytogenes, which was detected for one more week, until 35 days had passed. Negative control tanks always yielded negative results for L. monocytogenes. The results are shown in Figure 2A, expressed as the average ± standard deviation (n = 6).

Decline in L. monocytogenes in Pasture Crops
The stored manure was inoculated with L. monocytogenes and applied to pasture crops in a controlled laboratory setting. L. monocytogenes was detected until 15 days without enrichment. After that point, enrichment was performed before plating to confirm the presence/absence of L. monocytogenes. The results are shown in Figure 2B, expressed as the average ± standard deviation (n = 4). Maximum survival time on crops was surprisingly high, reaching 70 days.

PCR Serogroups
The 89% of the isolates (n = 187 out of 211) were identified as PCR serogroup IVb, 5% (n = 11 out of 211) as IIa and 6% (n = 13 out of 211) as serotype IIb. All three PCR serogroups were found in both organic and conventional farms with no statistical difference (χ2 and Fisher´s exact test p = 0.2479). Farms had at least two different PCR serogroups, except one organic farm in which only IVb was found.

Pulsed-Field Gel Electrophoresis (PFGE)
A total of 211 L. monocytogenes isolates were characterized using PFGE in order to elucidate the genetic relationship among them. The PFGE profiles were analyzed and compared using BioNumerics. The PFGE analysis using ApaI and AscI yielded 62 and 40 restriction profiles, respectively (representative PFGE profiles are shown in Figure 3A).

Decline in L. monocytogenes in Pasture Crops
The stored manure was inoculated with L. monocytogenes and applied to pasture crops in a controlled laboratory setting. L. monocytogenes was detected until 15 days without enrichment. After that point, enrichment was performed before plating to confirm the presence/absence of L. monocytogenes. The results are shown in Figure 2B, expressed as the average ± standard deviation (n = 4). Maximum survival time on crops was surprisingly high, reaching 70 days.

PCR Serogroups
The 89% of the isolates (n = 187 out of 211) were identified as PCR serogroup IVb, 5% (n = 11 out of 211) as IIa and 6% (n = 13 out of 211) as serotype IIb. All three PCR serogroups were found in both organic and conventional farms with no statistical difference (χ2 and Fisher s exact test p = 0.2479). Farms had at least two different PCR serogroups, except one organic farm in which only IVb was found.

Pulsed-Field Gel Electrophoresis (PFGE)
A total of 211 L. monocytogenes isolates were characterized using PFGE in order to elucidate the genetic relationship among them. The PFGE profiles were analyzed and compared using BioNumerics. The PFGE analysis using ApaI and AscI yielded 62 and 40 restriction profiles, respectively (representative PFGE profiles are shown in Figure 3A). The combination of the two restriction enzymes resulted in 64 restriction PFGE types or pulsotypes with 53% of the pulsotypes (34 out of 64) isolated from unique sites and not repeated among different farms ( Figure 3B). In addition, the total of the farms showed a high diversity of pulsotypes, as more than one pulsotype was found within the same farm. On the other hand, isolates with the same pulsotype were isolated from geographically distant farms (Supplementary Materials Table S1). The molecular basis of the clonal relatedness and the virulence genes repertoire of 45 L. monocytogenes isolates was examined by whole genome sequencing (WGS). For the selection of the samples to be sequenced, the following criteria were employed: isolates should be of PCR serogroup IV, should represent different pulsotypes and should have  The molecular basis of the clonal relatedness and the virulence genes repertoire of 45 L. monocytogenes isolates was examined by whole genome sequencing (WGS). For the selection of the samples to be sequenced, the following criteria were employed: isolates should be of PCR serogroup IV, should represent different pulsotypes and should have been isolated from farms with information available about antibiotics use. Summary and statistics for the genome sequencing of the L. monocytogenes isolates are detailed in the Supplementary Materials (Table S2). The pangenome (the set of all genes that are present in the analyzed dataset) consisted of 4240 genes, with a core-genome (the pool of conserved genes, which are represented in all genomes included in the analysis) of 2545 (60%) genes; softcore-genome (the set of genes present in at least 95% of genomes analyzed) of 83 (3%); shell-genome (the pool of genes moderately common in the pangenome; more than 15% but less than 95%) of 434 (10%) and cloud-genome (the genes present in less than 15% of the genomes analyzed) of 1178 (27%). Plasmids were not detected in any of the isolates. Nevertheless, genes involved on conjugation and/or mobilization of plasmids were detected in 25 of the 45 isolates, using a custom blast database in BLASTx [50] and MacSyFinder [51]. The results are given in Supplementary Materials (Tables S3 and S4).
In silico MLST determination was performed, and showed that the 45 sequenced L. monocytogenes isolates belonged to one of the following 12 different sequence types (STs): ST1, ST2, ST4, ST6, ST54, ST59, ST217, ST388, ST389, ST489, ST666 and one novel sequence type ST2921 with the following housekeeping allele combination: abcZ(1), bglA(1), cat (12) (Table 3). In the approximately-maximum-likelihood phylogenetic trees from the alignments of the accessory and the core genes ( Figure 4A,B, respectively), the isolates are grouped according to the CCs that they belong. served genes, which are represented in all genomes included in the analysis) of 2545 (60%) genes; softcore-genome (the set of genes present in at least 95% of genomes analyzed) of 83 (3%); shell-genome (the pool of genes moderately common in the pangenome; more than 15% but less than 95%) of 434 (10%) and cloud-genome (the genes present in less than 15% of the genomes analyzed) of 1178 (27%). Plasmids were not detected in any of the isolates. Nevertheless, genes involved on conjugation and/or mobilization of plasmids were detected in 25 of the 45 isolates, using a custom blast database in BLASTx [50] and MacSyFinder [51]. The results are given in supplementary material (Tables S3  and S4).

In Silico Analysis of Antimicrobial Resistance Genes and In Vitro Antimicrobial Susceptibility Testing
Only four antimicrobial resistance genes, present in all of the 45 L. monocytogenes isolates, were detected: fosX (involved in fosfomycin resistance and detected with ResFinder and CARD database), lin (involved in lincomycin resistance and detected with CARD and ARG-ANNOT databases), norB (conferring quinolone resistance and detected with CARD database) and mprF (which protects against cationic peptides and detected with CARD database).

In Silico Analysis of Virulence Genes
Virulence genes were detected in silico using TORMES ® version 1.0 by screening the L. monocytogenes isolates genomes against the Virulence Factors Data Base (VFDB) [49] and using the L. monocytogenes strain EGD-e as reference. For Listeria pathogenicity island 4 (LIPI-4) the L. monocytogenes strain CLIP80459 was used as reference and was detected using a custom blast database and using BLASTx version 2.12.0 [50]. The majority of virulence genes were shared across the 45 sequenced L. monocytogenes genomes ( Figure 5). All the sequenced isolates harbored inlA, inlB, inlC, inlJ and inlK, encoding proteins of the "internalin" family of L. monocytogenes, which interact with distinct host receptors to promote infection of human cells [53]. An additional gene encoding a member of the internalin family, inlF, was also present in most isolates (39 out of 45). All sequenced isolates contained the pathogenicity island LIPI-1, a 9 kb DNA fragment composed of six genes (prfA, mpl, plcA, plcB, actA and hly), the products of which are required for the intracellular life cycle of L. monocytogenes [54]. Most isolates (38 out of 45) harbored a 6 kb DNA fragment corresponding to the pathogenicity cluster LIPI-3 composed of eight genes (llsA, llsB, llsD, llsG, llsH, llsP, llsX and llsY) which encodes listeriolysin O, a pore-forming toxin involved in virulence [55]. On the other hand, the pathogenicity island LIP-4, a system strongly associated with the central nervous system and placental infections, was only found in 9 isolates. Other genes involved in virulence, present in all sequenced isolates are prsA2 gene, coding a critical post-translocation secretion chaperone [56] and loci encoding the Clp stress tolerance mediators, ClpC, ClpE and ClpP, involved in intraphagosomal survival [57][58][59]. There was no statistically significant difference in antimicrobial resistance ratios between organic and conventional farms (χ 2 and Fisher´s exact test p = 0.2337). The results are given in supplementary material (Table S3). Of the 6 MDR isolates, only one [MS6499(CC1)] was isolated from an organic farm; the other five [MS6484(CC1), MS6488(CC388), MS6490(CC666), MS6485(CC388) and MS6501(CC1)] were isolated from conventional farms.

In Silico Analysis of Virulence Genes
Virulence genes were detected in silico using TORMES ® version 1.0 by screening the L. monocytogenes isolates genomes against the Virulence Factors Data Base (VFDB) [49] and using the L. monocytogenes strain EGD-e as reference. For Listeria pathogenicity island 4 (LIPI-4) the L. monocytogenes strain CLIP80459 was used as reference and was detected using a custom blast database and using BLASTx version 2.12.0 [50]. The majority of virulence genes were shared across the 45 sequenced L. monocytogenes genomes ( Figure  5). All the sequenced isolates harbored inlA, inlB, inlC, inlJ and inlK, encoding proteins of the "internalin" family of L. monocytogenes, which interact with distinct host receptors to promote infection of human cells [53]. An additional gene encoding a member of the internalin family, inlF, was also present in most isolates (39 out of 45). All sequenced isolates contained the pathogenicity island LIPI-1, a 9 kb DNA fragment composed of six genes (prfA, mpl, plcA, plcB, actA and hly), the products of which are required for the intracellular life cycle of L. monocytogenes [54]. Most isolates (38 out of 45) harbored a 6 kb DNA fragment corresponding to the pathogenicity cluster LIPI-3 composed of eight genes (llsA, llsB, llsD, llsG, llsH, llsP, llsX and llsY) which encodes listeriolysin O, a pore-forming toxin involved in virulence [55]. On the other hand, the pathogenicity island LIP-4, a system strongly associated with the central nervous system and placental infections, was only found in 9 isolates. Other genes involved in virulence, present in all sequenced isolates are prsA2 gene, coding a critical post-translocation secretion chaperone [56] and loci encoding the Clp stress tolerance mediators, ClpC, ClpE and ClpP, involved in intraphagosomal survival [57][58][59]. Only two isolates [MS6670 (CC59) and MS6677 (CC489)] harbored ami gene, which codes for an autolysin required for adhesion to eukaryotic cells [60]. However, the rest of the isolates (43 out of 45) contained truncated versions of ami gene leading to a truncated protein with 772 or 607 amino acids instead of 934 as in the canonical Ami sequence ( Figure 6A). The same two isolates [MS6670 (CC59) and MS6677 (CC489)] that harbored ami gene, contained aut gene also, which codes for another autolysin, required for entry into eukaryotic cells [61]. The shorter aut variant LMOF2365_RS00075 [62] was found in all the rest of the sequenced isolates. Four isolates belonging to CC6 (MS6492, MS6666, MS6668 and MS6669) showed a deletion of six amino acids in the shorter Aut variant ( Figure 6B). Only two isolates [MS6670 (CC59) and MS6677 (CC489)] harbored ami gene, which codes for an autolysin required for adhesion to eukaryotic cells [60]. However, the rest of the isolates (43 out of 45) contained truncated versions of ami gene leading to a truncated protein with 772 or 607 amino acids instead of 934 as in the canonical Ami sequence ( Figure 6A). The same two isolates [MS6670 (CC59) and MS6677 (CC489)] that harbored ami gene, contained aut gene also, which codes for another autolysin, required for entry into eukaryotic cells [61]. The shorter aut variant LMOF2365_RS00075 [62] was found in all the rest of the sequenced isolates. Four isolates belonging to CC6 (MS6492, MS6666, MS6668 and MS6669) showed a deletion of six amino acids in the shorter Aut variant ( Figure 6B).

Discussion
Listeria monocytogenes is a facultative foodborne pathogen and several listeriosis outbreaks have been linked to its presence in the farm environment, either because of fresh vegetables fertilized with infected manure or because of the consumption of contaminated fresh dairy products [21,63]. The results of the present study demonstrate that samples from sites related to feces (slurry spreader, stable floor, slurry drain, manure and fresh feces) had a higher prevalence of L. monocytogenes than the ones found in feed and raw dairy samples (Table 1). These data support the hypothesis that on-farm transmis-

Discussion
Listeria monocytogenes is a facultative foodborne pathogen and several listeriosis outbreaks have been linked to its presence in the farm environment, either because of fresh vegetables fertilized with infected manure or because of the consumption of contaminated fresh dairy products [21,63]. The results of the present study demonstrate that samples from sites related to feces (slurry spreader, stable floor, slurry drain, manure and fresh feces) had a higher prevalence of L. monocytogenes than the ones found in feed and raw dairy samples (Table 1). These data support the hypothesis that on-farm transmission appears to be due to ingestion of feed contaminated with L. monocytogenes and, afterward, fecal shedding of L. monocytogenes in the environment by the bovine hosts, both with and without clinical disease [8]. In a way, livestock contributes to the increase and spread of L. monocytogenes into the farm environment, a hypothesis that is consistent with previous studies [52,64].
Several studies conducted under controlled laboratory conditions confirmed that L. monocytogenes can grow on plants [65][66][67]. The main drawback of most of those studies is the absence of other microorganisms. As a result, these experiments are best-case scenarios demonstrating that L. monocytogenes can colonize the surface of most plants. Whether or not L. monocytogenes colonizes plants internally is still a matter of debate, and conflicting reports are available [67][68][69]. Nevertheless, data available so far favor the hypothesis that L. monocytogenes can utilize nutrients from the plants to multiply and survive on plant surfaces. This hypothesis could explain the extended survival time of L. monocytogenes in crops, as shown in Figure 2B. Even though the experiment was performed in controlled laboratory conditions and the presence of rhizosphere and phyllosphere microbiome was limited, the maximum survival time of L. monocytogenes on crops was surprisingly high (70 days). The results of the decline of L. monocytogenes in pasture crops in combination with its decline in manure (Figure 2A) demonstrate that the life cycle of this pathogen between dairy animals and their surroundings is continuous. In order to reduce the contamination risk, effective herd management aiming the reduction of intestinal carriage, and manure treatment and management, when used as fertilizer in grassland and cropland, should be applied [8].
In this study, three PCR serogroups were isolated: IIa (lineage II) and IIb and IVb (lineage I). Comparatively few studies have isolates from natural environments and/or animals characterized to PCR serogroup and PFGE subtype. In general, according to previous studies, L. monocytogenes lineage I and lineage II isolates seem to be similarly prevalent [18,52,70], though contamination patterns of lineage I and II may differ [23]. Nevertheless, in this study, 95% of the isolates belonged to lineage I. Usually, linage I isolates display higher tolerance to low pH [71], which can be advantageous for the survival of L. monocytogenes in the gastric acidic environment and may account for their association with the fecal samples. Taking into account the prevalence of lineage I isolates in this study, it can be an additional indicator that the presence of L. monocytogenes in the farm environment is due to fecal contamination. Similar results were observed in poultry carcasses [23].
L. monocytogenes isolates were characterized using PFGE in order to elucidate their genetic resemblance. A total of 211 isolates were subtyped by PFGE, using ApaI and AscI endonucleases resulting in 64 PFGE types or pulsotypes ( Figure 3). The pulsotypes were separated by their PCR serogroup, results that are in accordance with previous studies [72,73] that confirmed relationships between serotype and PFGE patterns and reveal that the L. monocytogenes population of dairy farms in the Cantabria region is genetically diverse. From these 64 pulsotypes, 45 isolates were selected for whole genome sequencing, employing the following criteria: isolates should be of PCR serogroup IVb, represent different pulsotypes and be isolates from farms with information available about the antibiotics use. The reason why the PCR serogroup IVb was selected instead of IIa and IIb, which were also isolated from the dairy farms, is because the PCR serogroup IVb strains are commonly implicated in outbreaks and clinical human listeriosis cases. Previous studies showed that serotype 4b displays the highest pathogenicity compared to other serotypes in zebrafish embryos [74] and highest virulence than other serotypes according to reports that used mouse and Galleria mellonella as a model for identifying virulence determinants [75,76]. These features make the PCR serogroup IVb strains a very interesting candidate for further analysis at the genomic level.
A diversity of STs was found, as the 45 sequenced isolates were grouped in 12 different STs. Except for a new ST [ST2921(CC54)], isolated from an organic dairy farm, the rest of the L. monocytogenes isolates sequenced in this study included STs already reported. The most prevalent ST (Table 3) was ST1 (29% of all the sequenced isolates), followed by ST6 (15% of all the sequenced isolates), both of them related to outbreaks and clinical cases [77,78]. The prevalence of ST1/CC1 in ruminant-associated L. monocytogenes isolates has also been reported in another recent study conducted in Spain [79]. Interestingly, clones ST1/CC1, ST2/CC2, ST4/CC4, and ST6/CC6 constitute almost half of the sequenced L. monocytogenes isolates of this study (18 out of 45). The predominance of those clones in dairy products and ruminant feces has been already demonstrated and as they invade more efficiently the gut, they probably could lead to a higher fecal shedding and release of L. moonocytogenes in the environment [80]. Of particular concern is the isolation of ST388 (CC388), as it has been identified as the cause of a severe outbreak that took place in 2019 in Spain (data from the Spanish Ministry of Health). These findings highlight the importance of surveillance programs in farm animals.
Antibiotic resistance is one of the major concerns of public health since the percentage of infections due to resistant bacteria is increasing [81]. From the 45 isolates tested for their minimum inhibitory concentration (MIC), only six were characterized as multidrugresistant. All 45 isolates were resistant to cefoxitin, an expected result, since Listeria spp. are naturally resistant to cephalosporins [2,82]. Overall, our results confirm the susceptibility of L. monocytogenes. However, some isolates were found resistant to antimicrobials used to treat human cases of listeriosis, such as ampicillin and tetracycline, and even though the antimicrobial resistance among L. monocytogenes isolates is still low, there are reports of increasing resistance. This is an important issue since it is known that L. monocytogenes can acquire and/or transfer resistance genes through horizontal transfer processes in the intestinal microbiota [83]. In a recent study, 61% of L. monocytogenes isolates from a meat facility contained conjugative and/or mobilizable plasmids [84]. In contrast, no plasmids were detected in the isolates described here. Nevertheless, genes responsible for plasmid conjugation and/or mobilization were detected by MacSyFinder, with protein profiles mostly belonging to the MOBP1 Hidden Markov Models (HMMs) profile (Table S4). The MOBP1 HMMs profile targets relaxases from the MOB P conjugative transfer system [85], a diverse superfamily divided in seven subfamilies (MOB P1 -MOB P7 ) [86]. The conjugative transfer systems are classified in six MOB families: MOB F , MOB H , MOB Q , MOB C , MOB P and MOB V , with MOB P being the most diverse as it can be detected in plasmids from various incompatibility groups [86]. Since proteins from the MOB P conjugative transfer system is present in many conjugative and mobilizable plasmids [87] we can only hypothesize that conjugative and/or mobilizable plasmids might be present in the isolates studied, but it was not possible to detect them with the tools used.
The presence of core genome virulence factors and pathogenicity islands was evaluated for the 45 sequenced isolates in this study ( Figure 5). Four Listeria pathogenicity islands (LIPI) have been verified thus far in the literature, involved in the invasion, survival and colonization of Listeria in host tissues. Three of them have been described for L. monocytogenes (LIPI-1, LIPI-3 and LIPI-4), whereas LIPI-2 has only been detected in L. ivanovii [88]. LIPI-1 contains six genes: hly, prfA, plcA, plcB, mpl and actaA [89,90]. It encodes virulence factors that enable bacteria to escape from the vacuole (endosome or phagosome), proliferate in the cytosol and spread to the adjacent cells. LIPI-3 is composed of eight genes: llsA, llsB, llsD, llsG, llsH, llsP, and llsX. It encodes listeriolysin S (LLS), a haemolysin acting as a bacteriocin able to alter the host intestinal microbiota [91]. LIPI-4 encodes a cellobiose-family phosphotransfer system (PTS) and is involved in neural and placental infection [76]. Each of the 45 isolates sequenced in this study had at least one LIPI, with LIPI-1 being the one present in all strains. LIPI-3 was absent in ST2/CC2, as reported before [74], and also in CC59 and CC388. Isolates from CC4 and CC217 had all three LIPI, indicating the hypervirulence nature of those isolates. Originally the presence of LIPI-4 was reported only in strains from CC4 [76] and later other strains such as ST217/CC217 and ST388/CC388 were reported for carrying LIPI-4 [92,93], confirming the results of the present study.
In addition to the three pathogenicity islands described above, L. monocytogenes expresses a variety of virulence factors that are essential for its survival and persistence in the gastrointestinal tract and sequenced isolates contained various genes associated with virulence. For example, all isolates harbored the virulence genes encoding for internalins A, B, C, and J (inlA, inlB, inlC and inlJ), some cover surface proteins important in bacterial adhesion and invasion [94]. The inlK gene was also present in all of the sequenced isolates, encoding for a protein, InlK, which is involved in the autophagic recognition escape of L. monocytogenes [95]. Another gene encoding for a member of the internalin family of surface proteins, inlF, was found in 39 out of 45 isolates; its translational product, InlF, was reported of playing a role in L. monocytogenes colonization of the brain in vivo [96]. The mechanism of pathogenesis in L. monocytogenes involves a diversity of virulence proteins and the mechanisms leading to virulence are not completely clear yet. However, the presence of inlA, inlB, inlC, inlF, inlJ and inlK genes suggests the virulence potential of the L. monocytogenes isolates described in this study. Surprisingly, only two of the sequenced isolates harbored the ami gene, coding the Ami protein, involved in attachment to host cells and bacterial colonization of hepatocytes [60,97]. After closer inspection, all of the remaining 43 sequenced isolates not containing the ami gene, contained a truncated ami gene, leading to either a truncated 772-or a 607-amino acid Ami variants instead of the 934-amino acid wild type Ami protein ( Figure 6A). Since the initial search of virulence genes was performed by screening the genome against Virulence Factors DataBase (VFDB, [49]) by using ABRIcate and accepting only hits with coverage and/or identity above 90%, the truncated version of ami was not originally detected but revealed after using a custom blast database and using BLASTx [50]. This deletion has been already reported by other authors [74]. The same pattern was observed for the aut, a gene that encodes a surface protein with autolytic activity required for invasion into eucaryotic cells and as a consequence for virulence in vitro [61]. After the first analysis, aut gene was found absent in all sequenced isolates, except in two isolates that also harbored the full-length ami gene ( Figure 5). After closer inspection, it was revealed that the rest of the sequenced isolates harbored the shorter aut variant LMOF2365_RS00075 already reported by other authors [62]. A new variation of the aut allele was found in four isolates from CC6 (MS6492, MS6666, MS6668 and MS6669) showing a deletion of six amino acids ( Figure 6B). Whether or not the truncated ami gene and aut allele found in the isolates described in this study lead to non-virulent or less virulent strains in vivo, remains to be tested and was beyond the objectives of the present study. In any case, our data demonstrate that dairy farms constitute a reservoir for possibly hypervirulent L. monocytogenes, which are shed in the feces and could easily colonize dairy products, as a result of fecal contamination.

Conclusions
An analysis of L. monocytogenes isolated from environmental sites of livestock and livestock farms has shown that several L. monocytogenes, responsible for human infection, circulate in the biosphere and agricultural systems and might contribute to the spread of these pathogens throughout the food chain, thereby posing a major health challenge. A significant portion of isolates recovered are from the same CCs as those frequently isolated from human clinical cases and outbreaks on a global scale and many of those strains encode virulence factors associated with serious illness. As the SARS-CoV2 outbreak has demonstrated, the spread of infectious diseases in humans from animal reservoirs represents a major public health risk and it is expected that zoonotic diseases will occur more often due to climate change. This is the main pillar of the "One Health" approach, which accepts that human health is tightly connected with animal health and the environment. Considering the ability of L. monocytogenes to survive and adapt in different ecosystems and hosts, which makes it an important archetype of the "One Health" axis, effective herd and manure management and respect of the animal welfare are needed to prevent foodborne illnesses. Farmers should pay attention especially to the transmission from animal to animal through fecal-oral routes, usually via manure contamination of the pasture or silage with the microorganism. Feeding the animals with good quality silage, avoiding any other rotten vegetation and isolation of sick animals following good hygiene and sanitation on the farm is also important. Thereafter, food safety programs throughout the food production chain (from farm to fork) are needed to prevent foodborne illnesses.