L. monocytogenes CIP 110868 (serogroup IVb) was isolated from pig manure. The strain was grown at 30°C in a nutrient broth supplemented with 0.3% glucose (NBG). Three independent inocula were prepared by inoculating 100 ml of NBG to an O.D_{600 nm} of 0.04 and incubating overnight at 30°C to an O.D_{600 nm} of 0.4. The cultures were then centrifuged at 5,000 g for 5 min and the pellets were suspended in 0.8% NaCl solution. After a second washing step, the cells were suspended in saline to reach a final concentration of 5 × 10⁹ CFU/ml.

Soil and lagoon effluent were collected in sterile flasks from a pig manure treatment processing unit located in Brittany (France). The treatment consists of pre-storage and centrifugation of the manure to remove phosphorus. The liquid phase, which is biologically treated to reduce the level of nitrogen, is then left to settle. The liquid fraction in the settling tank is sent to an open-air lagoon where it is stored for 9–12 months. The sample of silty topsoil (0–20 cm) was collected near the lagoon. The two matrices were stored at 4°C until analysis.

Soil extract was prepared according to Piveteau et al. (2011). Briefly, 500 g of soil were mixed in 750 mL of water and autoclaved at 130°C for 1 h. The soil suspension was centrifuged at 10,000 g for 20 min and the supernatant was filtered on Whatman® 3 MM paper. The soil extract obtained after filtration was used after autoclaving at 120°C for 20 min.

The lagoon effluent extract was prepared by centrifugation of the effluent at 10,000 g for 10 min. The supernatant was filtered through Whatman® 3 MM paper and Whatman® qualitative filter paper Grade 2 (8 μm). The extract was then filter-sterilized on cellulose ester membrane pore size 3, 1.2, 0.8, 0.45, and 0.22 μm and transferred to sterile flasks. Sterility was confirmed by measuring culturable bacteria on plate count agar (PCA) incubated at 30 and 37°C for 48 h. Culturable bacteria were not found in either the soil extract or in the effluent.

Microcosms (30 ml) of both the soil extract and the lagoon effluent were prepared in triplicate.

Soil extract and lagoon effluent microcosms were inoculated with L. monocytogenes CIP 110868 to reach a final concentration of 5 × 10⁸ CFU/ml. Immediately after inoculation, the microcosms were incubated under static conditions at 21°C for 20 min (T1) and 24 h (T2). The temperature of 21°C was selected to mimic the average summer temperature in piggery lagoons in Brittany. In each treatment, three independent microcosms were inoculated with three independent inocula.

Total RNA was extracted from the culture broth (T0) and from the inoculum after 20 min (T1) and 24 h (T2) of incubation. RNAs were immediately stabilized by treating 2 ml of the bacterial cultures with RNA protect bacterial reagent (Qiagen, France) according to the manufacturer's instructions. Three independent replicates were prepared and treated separately. Four RNA extractions were performed of each replicate.

RNA was isolated according to the protocol described by Piveteau et al. (2011). Briefly, bacteria were harvested by centrifugation at 5,500 g for 5 min. Pellets were suspended in 700 ml of RLT buffer supplemented with 1% β-mercaptoethanol and 0.2 g of RNase-free glass beads (100 μm). Cells were mechanically disrupted in a Fast Prep®-24 instrument (MP Bio, France) with four cycles (6 m/s, 30 s). Total RNA was isolated using the Qiagen RNeasy kit (Qiagen, France) according to the manufacturer's instructions. After treatment with DNase I (DNAse on column, 10 U, Qiagen), RNA was eluted in 40 μl of RNase-free water. The four eluates of the same replicate were then pooled and treated with RQ1 RNase free DNase (Promega, France) in RNasin (40 U, Promega). RNAs were purified and concentrated using the RNeasy MinElute cleanup kit (Qiagen, France) according to the manufacturer's instructions. The quality and purity of the RNAs were determined spectrophotometrically and by capillary electrophoresis using the Bioanalyser 2100 (Agilent, France).

All RNAs were stored at −80°C until analysis.

RNA stabilization, 16S, 23S, and 5S RNA depletion, library preparation (“strand specific” method), clustering and sequencing (illumina HiSeq2000) were performed by Sistemas Genómicos (Valence, Spain) according to the manufacturer's instructions. Sequenced reads were quality-checked using FastQC. Reads were mapped against the most recent version of the L. monocytogenes EGD-e genome provided by NCBI database, and quantified. The statistical analysis performed by Sistemas Genómicos is described below. The differential expression study between groups of samples was achieved using Python and R statistic packages. The gene differential expression was studied with the algorithm proposed by DESeq2 (Anders and Huber, 2010; Anders et al., 2013) using a negative binomial distribution as dispersion model. Using the functional annotation from Uniprot database (http://www.uniprot.org/), Sistemas Genómicos realized a hyper-geometric test using different statistical packages from statistical platform R to performed the functional enrichment study. T-statistics and p-values (FDR adjusted p-value of 0.05) were calculated to identify differentially expressed genes. The fold changes at T1 and T2 were calculated by comparing results with expression levels at time 0.

Genes with at least a three-fold change were considered significantly different in the interpretation applied in this study. Finally, functional studies and differential analyses were performed between samples. Genes differentially transcribed were grouped according to their functional category (Glaser et al., 2001).

The transfer of L. monocytogenes CIP 110868 into the lagoon effluent caused substantial changes of the transcriptome. Within 20 min, 1,076 genes were differentially transcribed. After 24 h, the number of up or down regulated genes remained stable, with 1,034 genes identified with a different transcript level, indicating that transcriptome reshaping took place within minutes following contact with the effluent (Table S1). A high proportion of operons were differentially transcribed (Table 2). Seventy-four complete operons including 218 genes were differentially expressed after 20 min of incubation, which represents 20% of the set of genes with different transcript levels. After 24 h, the proportion increased to 26% with 76 operons including 268 genes. A large set of transcriptional regulators was differentially transcribed at T1 and T2 (functional category 3.5.2, Table 3, Table S1). Among the 209 regulators described in the genome of L. monocytogenes EGD-e (Glaser et al., 2001), 76 (36%) were differentially transcribed after 20 min and 83 (40%) after 24 h. Interestingly, most had a lower transcript level. These results confirm the pivotal role of the extensive regulatory repertoire in the ability of L. monocytogenes to survive in various environments and to adapt its physiology (Gandhi and Chikindas, 2007). Therefore, it is not surprising that a high proportion of transcriptional regulators and response regulators was detected in the set of differentially transcribed genes. Table 4 lists the differentially transcribed regulators for which biological functions have been described. Similarly, high proportions of the regulons (from 31.4 to 59.4%) of sigma factors and other transcriptional regulators were differentially transcribed after 20 min and 24 h (Table 5). This suggests these regulons may be important for adaptation to the lagoon environment.

A large repertoire of differentially transcribed genes illustrated the metabolic adaptation of L. monocytogenes CIP 110868 to its environment within the first minutes of incubation (Table 3). Transporters required for the use of available resources were highly represented, suggesting that a readjustment of the transcriptome occurred rapidly in order to access resources available in the lagoon environment. After 20 min of incubation in the lagoon, a total of 125 genes associated with transport were differentially transcribed. This represents 38% of the genes identified in the genome of L. monocytogenes EGD-e as being devoted to carbohydrate uptake (Glaser et al., 2001; Fichant et al., 2006). These genes included 42 phosphoenolpyruvate-dependent phosphotransferase systems (PTS systems) and 54 ATP-binding cassette transporters (ABC transporters). Among them, respectively 29 and 30 genes had higher transcript levels. These genes encoded proteins involved in glycerol (lmo1539), galactitol (lmo2667), fructose (lmo0039, lmo0400 and lmo0358), and maltose-maltodextrin (lmo2123-24-25) uptake. Moreover, arabinose and lactose transcriptional regulators (lmo2851 and lmo1727, respectively) and a xylose repressor (lmo0178) were identified in the set of genes with higher transcript levels. In addition, three known activators of PTS systems (lmo2138, lmo2099, and lmo2668) were also found to be upregulated. Genes encoding transporters found in the set of genes with lower transcript levels were mainly related to the transport of metal cations but none to the transport of carbohydrates.

The same pattern was observed after 24 h of incubation, except that the percentage of genes encoding transporters decreased from 38 to 31%.

Throughout the experiment, genes involved in cell surface rearrangement were differentially transcribed. The functional categories “cell wall” and “cell surface proteins” were highly represented. Most of the genes were downregulated. For example, seven genes encoding teichoic acid synthesis had lower transcript levels. Similarly, the transcript level of genes encoding cell surface proteins also decreased. This included peptidoglycan bound proteins and internalins. As reviewed by Bierne and Cossart (2007), the adjustment of the cell envelope is of major importance for coping with new environmental conditions. Cell surface proteins fulfill many functions as diverse as bacterial growth, peptidoglycan metabolism, sensing of and protection from environmental stresses, adhesion and invasion of host cells and signaling. For example, teichoic acids, which are directly linked to the structure of the outer cell wall, enable reinforcement of the cell wall of L. monocytogenes when environmental conditions are unfavorable. The products of both genes of the tagGH operon are involved in translocation in teichoic acids. The operon, controlled by a σA-dependent promoter, was first described in Bacillus subtilis for which it is essential for cell growth (Lazarevic and Karamata, 1995). In L. monocytogenes, the operon plays a role in protecting cells against the biocide benzethonium chloride (Casey et al., 2014). Additional genes involved in the biosynthesis and the glycosylation of teichoic acids, such as ggaB, lmo1077, lmo1085, gtcA, tagB, and tagO, were also found in the set of genes with lower transcript levels. It is important to note that these genes were differentially transcribed with a fold change in the region of −1,000.

The variations in transcript levels observed in these functional categories point to the need for a readjustment of the cell surface to adjust to the new environment.

lmo0441, pbpA, pbpB, lmo2229, and lmo1438 genes were positively transcribed. These penicillin-binding protein encoding genes have been shown to be involved in cell wall functioning (Rismondo et al., 2015), cell division during growth (Krawczyk-Balska et al., 2012) and resistance to β-lactam antibiotics (Van de Velde et al., 2009; Collins et al., 2012). Moreover, the penicillinase repressor lmo0252 had a lower transcript level, suggesting deactivation of repression and the production of penicillinase. Similarly, the transcriptional repressor (lmo2690) from the TetR family, which targets positive regulators of multi-drug efflux pumps, had a lower transcript level, indicating that the repression of the pumps may be lifted in the effluent. We also found the tetracycline resistance protein coding gene lmo0839 in the set of genes with higher transcript levels. The response regulator lmo1022, which resembles the LiaSR two-component system of B. subtilis known to be induced when antibiotics interfere with the cytoplasmic membrane (Fritsch et al., 2011), was found to be activated. Finally, lmo1967 was also activated. lmo1967, first described for toxic ion resistance, is required for optimum resistance to nisin and other antimicrobial molecules (oxacillin, cefuroxime, cefotaxime, methicillin, and bacitracin; Collins et al., 2010). These data suggest that mechanisms of protection against antibiotics had been triggered. Conversely, other genes (lmo0872, lmo1409, and lmo1617), also implicated in antibiotic resistance and in the functioning of multi-drug efflux transporter, were downregulated.

Antibiotics (mainly oxytetracycline and chlortetracycline) are frequently administered to pigs and animals may excrete from 25 to 90% of the administered antibiotics as parent compounds. As a consequence, high concentrations of antibiotics are detected in manures (Qian et al., 2016). For example, oxytetracycline and chlortetracycline have been detected at concentrations in the order of mg per kg of dry weight in manure samples (Widyasari-Mehta et al., 2016). As a result, it is possible that lagoon effluents (liquid fraction of biologically treated manure) are also contaminated by antibiotics and that L. monocytogenes had to face these molecules in the lagoon effluent, as our results suggest.

Metal ions are commonly detected in piggery effluents (Xiong et al., 2010; Hölzel et al., 2012). For example in the lagoon effluent used in this study, copper and zinc were detected at concentrations of 16 and 55 mg/L respectively (data not shown). The upregulation of lmo1478, a transcriptional regulator of the MerR family, within the first minutes of incubation in the effluent, suggests that L. monocytogenes responded to exposure to metal ions. In contrast, genes encoding efflux systems for cadmium/arsenic and for zinc/arsenic transport were downregulated.

The introduction of L. monocytogenes in harsh environments triggers the regulation of genes involved in stress responses (Piveteau et al., 2011; Casey et al., 2014).

A battery of genes involved in stress response was detected among genes with higher transcript levels throughout the course of the experiment. Briefly, the expression of the cold shock genes cspD and cspL increased. Upregulation of the opuC operon (opuCA, opuCB, opuCD) suggests a response to osmotic stress. This system is involved in the uptake of the osmoprotectants glycine-betaine, carnitine, and choline. These molecules have previously been shown to play a role in the osmoregulation and cold adaptation of L. monocytogenes (Fraser et al., 2000). The GltC activator also had a higher transcript level, suggesting that oxidative stress occurred in the lagoon environment. The LexA SOS system repressor and the repressor of class III stress genes Lmo0229 were in the set of genes with lower transcript levels. Finally, it is important to note that σB, which plays a central role in the stress response of L. monocytogenes and the RsbU phosphatase involved in the regulation of σB, had higher transcript levels. Among the 216 genes of the σB regulon (Hain et al., 2008), 79 were differentially transcribed after 20 min of incubation and 76 after 24 h.

All these modifications suggest that incubation in the lagoon triggered responses to environmental stresses. Conversely, regulators involved in adaptation to heat shock (Lmo2107; van der Veen et al., 2007) and acid shock (Lmo1618; Phan-Thanh and Mahouin, 1999; Rea et al., 2004) were in the set of genes with lower transcript levels. Some of these variations may be explained by the conditions under which inocula were prepared. Indeed, the pH of the broth and temperature used as reference differed from the conditions of the lagoon.

The number of genes dedicated to the metabolism of amino acids and related molecules increased with time, from 55 genes differentially transcribed after 20 min to 63 genes after 24 h. These genes are particularly involved in the synthesis of leucine and isoleucine (ilvDBC, leuABCD, ilvA), histidine (hisEIA, hisJ, hisC), threonine (thrC), glutamate and glutamine (serC, glnA), alanine (daI), glycine (lmo1349-50, glyA), arginine (argFDBJC, argGH), lysine (dapF), and tryptophan (trpABCD; Figure S1). In addition, several riboswitches (Tbox, lysine, glycine, TPP, purine, pyrimidine, glmS riboswitch), which regulate amino acid biosynthesis, were also found in the set of upregulated genes (Table S2).

After 24 h of incubation in the lagoon environment, the genes involved in motility and chemotaxis were highly activated. In this functional category, 28 genes had higher transcript levels after 24 h but not after 20 min. Among the 28 genes, we identified the well-described flaA encoding flagellin (Grundling et al., 2004), motAB encoding flagellar motor rotation proteins (Michel et al., 1998) and fli, flg, and flh genes. Interestingly, these genes belong to two operons, all of whose genes were also activated (42 genes). Among them cheYA encodes the two component system responsible for the regulation of chemotaxis in L. monocytogenes (Flanary et al., 1999; Williams et al., 2005). Motility may be a strategy used by cells to cope with their environment. Several of these genes (for example flaA and motAB) are thermoregulated, with higher expression at 25°C (saprophyte life in the environment) and lower expression at 37°C (intracellular life; Michel et al., 1998; Grundling et al., 2004).

As for some genes involved in motility, temperature is an environmental cue that affects PrfA activity and expression of its regulon (Freitag et al., 2009; de las Heras et al., 2011). During the saprophytic lifestyle of the bacteria (at temperatures below 30°C), the secondary structure of the mRNA of prfA forms a hairpin structure which masks the ribosome binding region. In this conformation, the thermosensor represses the translation of prfA mRNAs. prfA was downregulated throughout the course of our experiment. Among the 70 genes referenced in the PrfA regulon (Milohanic et al., 2003; Chaturongakul et al., 2011), 41 had different transcript levels after 20 min and 29 after 24 h. As a result of introducing L. monocytogenes in an environmental matrix, genes specifically involved in the adaptation of the pathogen to the intracellular environment were downregulated. Internalin-coding genes inlB, inlC, inlE, inlG, and inlH were in the set of genes with lower transcript levels. Similarly, lmo0754, lmo2067, and lmo0446, which encode bile acid hydrolases in response to acid stress, lmo0445, a candidate regulator of virulence genes (Raengpradub et al., 2008) and lmo2200, an ortholog regulator of ohrR in B. subtilis, which is essential for intracellular growth (Chatterjee et al., 2006) were downregulated.

Genes of the ComG (comGA and comGB) and ComE (comEB and comEC) operons were found to have a higher transcript level in the lagoon environment both at 20 min and 24 h. They are similar to B. subtilis comG and comE operons, which encode proteins enabling he development of competence and the acquisition of extracellular DNA (Dubnau, 1997). The MecA regulatory protein, which represses competence genes (Borezee et al., 2000), was in the set of genes with lower transcript levels. While the Com master activator is disrupted by a Listeria-specific prophage in some strains including EGD-e, other strains harbor the full version of comK. Although competence has never been demonstrated in L. monocytogenes, Rabinovich et al. (2012) showed that excision of the prophage can allow functional transcription of comK and that the role of Com systems in Listeria might have diverged and could be useful for phagosome escape and virulence. Alternatively, competence may be possible under some environmental conditions yet to be identified such as lagoons. Some isolates might be competent even though signals inducing competence have not been identified (Buchrieser et al., 2003).

Whatever the incubation time, a wide range of ncRNAs was detected in the sets of differentially transcribed targets (Table S2) most of which presented a lower transcript level. Like in other bacterial models, ncRNAs play a central role in the adaptation of L. monocytogenes to its environment. So far, a role for the non-coding part of the genome has been reported in the regulation of virulence (Mandin et al., 2007; Loh et al., 2009; Toledo-Arana et al., 2009; Mraheil et al., 2011; Nielsen et al., 2011; Mellin and Cossart, 2012), in the transport and metabolism of sugars (Mandin et al., 2007; Nielsen et al., 2011), in the transport of metal ions (Wurtzel et al., 2012), in the synthesis of vitamins (Mellin and Cossart, 2012), in the formation of biofilm (Peng et al., 2016) and in the regulation of motility (Mellin and Cossart, 2012). In our study, we found that Rli28, Rli31, Rli38, Rli49, and Rli50, which are involved in invasion of the host's blood system, in the regulation of virulence factors (Toledo-Arana et al., 2009), in intracellular multiplication in mouse macrophage P338D1 and complete virulence in Galleria mellonella models (Mraheil et al., 2011), had lower transcript levels after incubation in the lagoon environment. Similarly, the virulence regulators LhrC1, LhrC2, LhrC3, and LhrC4 riboswitches were deactivated. Conversely, Rli60 was found in the set of ncRNAs with higher transcript levels. This ncRNA has recently been reported to play a role in the response (growth and biofilm formation) of L. monocytogenes EGD-e to environmental stresses (temperature, pH, osmotic pressure, alcohol; Peng et al., 2016).

To extract more information on the response of L. monocytogenes to the lagoon environment, we next focused on genes for which at least 10-fold increase was observed. A total of 568 genes met this criterion (fold change >10 = 183 genes and fold change < −10 = 385 genes). Among the genes with a 10-fold higher transcript level, functional categories 1.2 (transport/binding proteins and lipoproteins), 2.1.1 (specific pathways), 3.5.2 (regulation), and 5.2 (similar to unknown proteins from other organisms) were the most widely represented. Interestingly, the genes lmo0049 (agrD) and lmo0050 (agrC), encoding respectively the propeptide AgrD (precursor of the auto-inducing peptide) and the histidine kinase AgrC, were in this set of genes. AgrD and AgrC are a part of the Agr communication system (Rieu et al., 2007). This indicates that communication between bacteria might have been induced in the lagoon environment. As the Agr communication system is involved in the overall adaptation of the cell (Rieu et al., 2007; Riedel et al., 2009; Vivant et al., 2014) by integrating environmental cues, it was not surprising to find a marked increase in the transcript levels of the agrD and agrC genes. The functional categories highly represented in the set of genes with fold changes lower than −10 differed. We identified categories 3.5.2 (regulation), 4.3 (phage-related functions), 5.2 (similar to unknown proteins in other organisms), and 6 (no similarity). Genes with phage-related functions are often described in studies involving differential analysis (Piveteau et al., 2011; Casey et al., 2014). As described in the study by Piveteau et al. (2011), transcript levels of phage-related genes decreased within the first minutes of incubation (42 of the 48 phage-related genes at 20 min) and throughout the course of the experiment (36 genes at 24 h). However, the function of these modifications is still not known. In our study, the majority of the genes in the “phage-related functions” category were related to the Listeria phage A118.

Smaller variations in the transcriptome were observed after incubation of L. monocytogenes CIP 110868 in the soil than in the lagoon (Table S3). We detected 210 genes with different transcripts level after 20 min of incubation and 284 genes after 24 h of incubation in the soil extract (Table 2, Figures 2C,D). Another difference between the sets of differentially transcribed genes from lagoon and soil is that, after incubation of L. monocytogenes CIP 110868 in soil extract, fold changes were low (Table 1). Indeed, of the 210 differentially transcribed genes at 20 min, only three had fold changes >10 (one transporter and two unknown proteins). On average, fold changes ranged from −3.5 to 4.4. A similar pattern was observed after 24 h of incubation, when 35 of the 284 genes were highly differentially transcribed (fold change >10 or < −10). These genes were mainly dedicated to transport, mobility and unknown functions.

As described in the lagoon environment, complete operons were differentially transcribed during incubation in soil extract (Table 2). Transcription and response regulators were differentially transcribed but their proportion was smaller than in the lagoon environment. The percentage of regulators differentially transcribed was 5.7 and 4.6% of the genes differentially transcribed at 20 min and 24 h in the soil, whereas these proportions were 7.1 and 8% in the lagoon. The proportion of genes under the control of regulators and sigma factors was also smaller. Interestingly, the proportion of genes of the σC, σL, and CodY regulons in the set of differentially transcribed genes increased after 24 h (Table 5).

Surprisingly, genes with higher transcript levels differed completely at T1 and T2 (Figure 2C). In contrast, 37 genes with lower transcript levels were common at the two incubation times. They encoded unknown proteins and the internalin InlC.

Like in the lagoon environment, activation of a large panel of genes involved in motility was observed after 24 h of incubation (Table 6), including the regulator CheY. This confirms that motility is a strategy used by L. monocytogenes during its saprophytic lifestyle. For example, motility offers the opportunity to access a wide range of nutrients.

Incubation of L. monocytogenes for 20 min in the soil extract led to the deactivation of transporters dedicated to fructose, sorbitol, and ribulose and the activation of ABC transporters of metal ions. Moreover, transporters of spermidine and putrescine were also present in the set of genes with higher transcript levels. These two polyamines are involved in numerous biological processes and are essential for biofilm formation (Burrell et al., 2010; Wortham et al., 2010; Sakamoto et al., 2012). As confirmation that L. monocytogenes was introduced in a natural environment, a PTS system dedicated to the transport of lichenan, a polysaccharide isolated from lichen, was also activated.

The proportion of genes from category transport and binding protein was higher at 24 h (Table 6). Our results illustrate the requirement to acquire substrates available in the soil extract. This included plant residues such as cellobiose (lmo2708, lmo2762-63 and lmo2765), mannitol (lmo2651), maltose and maltodextrin (lmo2122-26), rhamnose, and rhamnulose (lmo2847-49), residues of arthropods and fungi such as chitin (lmo0105 and lmo1883) and residues of algae and plant exudates such as galacticol (lmo2665-67). The upregulation of transporters to acquire substrates from various sources in soil has already been described (Piveteau et al., 2011).

Lesser variations of transcripts of genes related to stress response were detected from soil than from the lagoon environment, as only a few genes dedicated to stress response were differentially transcribed. They are mainly involved in the adaptation of L. monocytogenes to changes in temperature. Broth and incubation temperatures differed by about 10°C in our experiment, which explains the response to thermal stimuli. We identified as downregulated the genes encoding the cold shock proteins CspB and CspL and the heat shock protein GroES. The transcriptional regulator NmlR involved in oxidative stress and in tolerance to acid (Supa-amornkul et al., 2016) was found to be activated while the general stress protein encoded by lmo1601 was downregulated. L. monocytogenes is ubiquitous in the environment. The pathogen has been detected in soils (Locatelli et al., 2013a) and its ability to survive in this environment for long periods has been repeatedly demonstrated (Dowe et al., 1997; Moshtaghi et al., 2009; McLaughlin et al., 2011; Locatelli et al., 2013b).

The virulence master regulator PrfA, the internalin InlC, and a bile acid hydrolase (lmo2067) were downregulated 20 min after L. monocytogenes was introduced in the soil extract. This is consistent with the saprophytic lifestyle of the bacteria. These observations are consistent with the observations made by Piveteau et al. (2011). In their study, transcript levels of virulence determinants including PrfA and internalins decreased 18 h after L. monocytogenes was transferred in soil extract microcosms.

As confirmation that virulence factors were downregulated within the first minutes of incubation, ncRNAs involved in the regulation of virulence (Rli31, Rli50, and Rli47) were in the set of ncRNAs with lower transcript levels after 20 min of incubation (Table S4). Rli31 and Rli50 are required for full virulence in murine, larva, and macrophage infection models (Mraheil et al., 2011), and Rli47 (sreB) is a trans-acting ncRNA which controls the expression of the regulator PrfA by binding the untranslated region of its mRNA. Seven ncRNAs were found at both incubation times (Table S4) and all had lower transcript levels. Their functions are not yet described in the literature, except for RliI, which is involved in the regulation of the metabolism and the transport of carbohydrates (Mandin et al., 2007). Interestingly, after 24 h, no ncRNAs with higher transcript levels were detected, while a larger proportion had lower transcript levels. Among them, we mainly identified the four LhrC (1, 2, 3, and 4) riboswitches and Tbox and SAM riboswitches.

As observed in the lagoon effluent, several genes from functional category phage-related functions were in the set of genes with lower transcript levels within the first minutes of incubation in the soil extract (Table 6). These targets matched with genes coding proteins from bacteriophage A118.

Regardless of the incubation time (20 min and 24 h), a large fraction of the genes differentially expressed in the soil extract (42–78%) were shared with the lagoon effluent (Figure S2). The adaptation of L. monocytogenes CIP 110868 to the natural environment led to similar readjustments of the transcriptome (Figure 3). Whether L. monocytogenes was inoculated in the lagoon effluent or in the soil extract, genes encoding proteins involved in motility and chemotaxis and in the transport of carbohydrates were the most widely represented in the set of genes with higher transcript levels. This is commonly reported in the literature, except that nutrients acquired by bacteria differ depending on the environment. Within genes with lower transcript levels, in both conditions, we mainly found genes in the “phage-related functions” category. It is worth mentioning that a large proportion of the genes differentially transcribed after the introduction of L. monocytogenes CIP 110868 in the natural matrices, encoded proteins with unknown functions. This represents from 25% of the set of genes with higher transcript levels to 60% of the set of genes with lower transcript levels. Finally, the proportion of genes from the PrfA, CodY, and CtsR regulons increased between 20 min and 24 h both in the lagoon and in the soil.

Comparison of differentially transcribed genes in soil and lagoon led to the identification of specific variations. Three functional categories (cell wall, competence, and metabolism of amino acids) were highly represented in the lagoon environment but not in the soil. Differences in the composition of the two matrices could explain such results. For example, the higher concentrations of detrimental molecules such as antibiotics and metal ions in lagoon effluents might trigger upregulation of genes leading to the adjustment of the cell wall for protection against toxic compounds.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.