Investigation of the roles of AgrA and σ^B regulators in Listeria monocytogenes adaptation to roots and soil

ABSTRACT

Little is known about the regulatory mechanisms that ensure the survival of the food-borne bacterial pathogen Listeria monocytogenes in the telluric environment and on roots. Earlier studies have suggested a regulatory overlap between the Agr cell–cell communication system and the general stress response regulator σ^B. Here, we investigated the contribution of these two systems to root colonisation and survival in sterilised and biotic soil. The ability to colonise the roots of the grass Festuca arundinacea was significantly compromised in the double mutant (∆agrA∆sigB). In sterile soil at 25°C, a significant defect was observed in the double mutant, suggesting some synergy between these systems. However, growth was observed and similar population dynamics were shown in the parental strain, ΔagrA and ΔsigB mutants. In biotic soil at 25°C, viability of the parental strain declined steadily over a two-week period highlighting the challenging nature of live soil environments. Inactivation of the two systems further decreased survival. The synergistic effect of Agr and σ^B was stronger in biotic soil. Transcriptional analysis confirmed the expected effects of the mutations on known Agr- and σ^B-dependent genes. Data highlight the important role that these global regulatory systems play in the natural ecology of this pathogen.

Soil survival, roots colonization, AgrA, σ^B, Listeria monocytogenes, transcription regulators

INTRODUCTION

Listeria monocytogenes is a robust food-borne pathogen capable of surviving in a wide range of niches outside the host (O'Byrne and Karatzas 2008; NicAogain and O'Byrne 2016; Dorey et al. 2019). Soil is a natural reservoir of L. monocytogenes and transfer from contaminated sources such as sewage, animal manure and decaying plant vegetation to soil was observed (Welshimer 1960; Welshimer and Donker-Voet 1971; Fenlon, Wilson and Donachie 1996; Nightingale et al. 2004; Locatelli et al. 2013; Strawn et al. 2013; Vivant et al. 2013b). However, populations of L. monocytogenes in bulk soil are generally low (MacGowan et al. 1994; Dowe et al. 1997), with an estimated prevalence of less than 10⁴ CFU per gram of dry soil throughout the French territory (Locatelli et al. 2013). L. monocytogenes can persist in soil over a period of time (Ivanek et al. 2009; McLaughlin et al. 2011; Locatelli et al. 2013). Soil chemical properties and natural microbiota as well as geographical and meteorological conditions may affect survival and persistence of L. monocytogenes (Ivanek et al. 2009; McLaughlin et al. 2011; Strawn et al. 2013). The presence of L. monocytogenes in crop production has often been reported (Brandl 2006), since it can persist on the surface of plants throughout the cultivation period (Jablasone, Warriner and Griffiths 2005). Nonetheless, in comparison to in vivo conditions, little is known about the regulatory mechanisms that underlie survival of this pathogen over its saprophytic lifestyle (Dorey et al. 2019).

The Agr cell–cell communication system of L. monocytogenes is required for optimal adaptation and survival in soil (Piveteau et al. 2011; Vivant et al. 2014). This system was reported to activate chitin hydrolysis through repression of the small RNA (sRNA) LhrA known to repress chiA translation (Paspaliari et al. 2014). ChiA belongs to the chitinolytic system of L. monocytogenes. Chitin degradation liberates nutrient sources (Gooday 1990; Gooday 1990; Beier and Bertilsson 2013). Likewise, σ^B activates production of chitinases, as its absence results in lower chiA and chiB expression and a decrease in L. monocytogenes chitinolytic activity (Larsen, Leisner and Ingmer 2010). Differential transcriptome analyses during adaptation to soil showed large transcriptome alterations when AgrA was not functional. Over 500 protein-coding genes involved in cell envelope and cellular processes, and an extensive repertoire of sRNAs were differentially transcribed (Vivant et al. 2015). Comparatively, lower transcriptomic differences were recorded in sterilized soil, suggesting that activation of the AgrA system is required for optimal re-shaping of the transcriptional landscape when L. monocytogenes encounters complex biotic conditions (Vivant et al. 2015).

The σ^B factor controls the general stress response, coordinating transcription of approximately 10% of the genome of L. monocytogenes (Chaturongakul et al. 2011). Studies concerning the role of σ^B during L. monocytogenes life in outdoors environments are scarce. It is expected to contribute to resistance to stresses that L. monocytogenes may encounter in telluric environments (Dorey et al. 2019). Indeed, several genes from the σ^B regulon are differentially transcribed during incubation in soil extracts (Piveteau et al. 2011). In synthetic soil, inactivation of σ^B resulted in growth defect of L. monocytogenes and the transfer of the pathogen from contaminated soil to crops was reduced (Gorski, Duhe and Flaherty 2011).

Direct link between Agr and σ^B systems was reported in Staphylococcus aureus, where σ^B decreases transcription of RNAIII, the regulatory RNA effector of the Agr system (Bischoff, Entenza and Giachino 2001). Although no direct association between these two systems has been reported in L. monocytogenes, differential transcription of several genes of the σ^B regulon when AgrA is not functional suggests crosstalk between the two systems (Garmyn et al. 2012). The biological function of the σ^B-dependent sRNA Rli47 is yet undetermined (Mujahid et al. 2012), however, its involvement in adaptation to outdoors has been suggested, perhaps involving crosstalk between σ^B and AgrA regulons, since transcription of Rli47 was up regulated in a ΔagrA L. monocytogenes mutant during soil survival (Vivant et al. 2015). In this study, we investigated the contribution of the general stress response (σ^B) and cell-cell communication system (AgrA) to root colonisation and soil fitness of L. monocytogenes. The response to the extrinsic factors ‘soil microbiota’ and ‘water content’ were also assessed.

MATERIALS AND METHODS

Bacterial strains, plasmids, primers and culture media

Listeria monocytogenes EGD-e (Glaser et al. 2001), ΔagrA (Rieu et al. 2007) and ∆sigB (Marinho et al. 2019) strains were used in this study (Table 1). A two-step integration/excision procedure with the pMAD vector (Arnaud, Chastanet and Debarbouille 2004) was used to construct ΔagrAΔsigB in-frame deletion. For this, sigB was deleted in the ∆agrA mutant (DG125A). All primers and plasmids used for this construction are listed in Table 1. Cloning was performed in One Shot TOP10 Chemically Competent Escherichia coli (Invitrogen). Electro-competent cells of L. monocytogenes were prepared and transformed according to Monk, Gahan and Hill (2008). Gene deletions were checked by sequencing PCR products using gene-flanking primers. Whole-genome sequencing was performed (MicrobesNG) and genomes were analysed by Breseq (Deatherage and Barrick 2014). Rifampin-resistant strains were used to facilitate recovery from unsterilized soil, since some soil microorganisms have the ability to grow on Listeria selective media. Isogenic rifampicin resistant (RIF^R) strains EGD-e L9 and DG125A6 were used in this study (Vivant et al. 2014). Transduction of ϕLMUP35–EGD-e L9 preparations (Hodgson 2000) was implemented to transfer rifampicin resistance (RIF^R) in all other deletion mutants; transformants were selected in TSA + RIF 200 μg.ml⁻¹ (Sigma-Aldrich). Rifampicin resistant strains were cultured on trypton soy broth (TSB) (CONDA) and Polymyxin-Acriflavin-Lithium-Chloride-Ceftazidime-Aesculin-Mannitol (PALCAM) agar medium (AES Chemnex) supplemented with 200 μg.ml⁻¹ rifampicin (Sigma-Aldrich) according to Lemunier et al. (2005). All strains used in this study are listed Table 1.

Table 1.

Primers, plasmids and strains used in this study.

STRAINS
Name	Source or reference	Genotype
EGD-e	Glaser et al. 2001	Listeria monocytogenes EGD-e serotype 1/2a.
L9	Lemunier et al. 2005	EGD-e, RIFR.
DG125A	Rieu et al. 2007	EGD-e ∆agrA.
DG125A6	Rieu et al. 2007	DG125A, RIFR.
CM003	This work	EGD-e ∆agrA∆sigB.
CM004	This work	CM3, RIFR.
COB262	C. Gahan	EGD-e ∆sigB gene.
CM007	Marinho et al. 2019	EGD-e ∆sigB gene.
CM008	This work	CM7, RIFR.
TOP 10	Invitrogen	OneShot TOP 10 Electrocomp E. coli.
PLASMIDS
Name	Source or reference	Further information
pMAD	Arnaud, Chastanet and Debarbouille 2004	Amp and ery resistance, β-galactosidase gene, oripE194ts, oripBR322.
pCM002	This work	pMAD_ΔsigB, insert of 701 bp spanning the upstream and downstream flanking regions of sigB with a Δ561 bp (64_624del).
PRIMERS
Cloning
Name	Sequence (5′ → 3′)	Further information
agrA_F	GATAATGCGGTTGAAGCGGC	Forward primer for upstream flanking region of agrA.
agrA_R	TACCCGGTCATGCAGCAAAT	Reverse primer for downstream flanking region of agrA.
sigB_F	GCGCCGAATCAAAGAGTTAG	Forward primer for sigB.
sigB_R	CACCAACAACATCAAGTAACG	Reverse primer for sigB.
sigB1_F	CGTATCGGTGGGCTAGGTT	Forward primer to confirm plasmid integration.
sigB2_R	TCCACTTCCTCATTCTGCAA	Reverse primer to confirm plasmid integration.
sigB3_F	GGGGAGGATCCATGCATTTAAAGGGGAAG	Forward primer for upstream flanking region of sigB. BamHI restriction enzyme site.
sigB4_R	CCCCGAATTCTTTTATGGCGTCAACAGTCG	Reverse primer for downstream flanking region of sigB. EcoRI restriction enzyme site.
pmad_flk_F	CCATCAGACGGTTCGATCTT	Forward primer upstream insert on pMAD.
pmad_flk_R	TTTGCGCATTCACAGTTCTC	Reverse primer downstream insert on pMAD.
RT-qPCR
rli47_RTqPCR_F	TTACGGGCGCTCATTTATTC	Forward primer for the gene rli47.
rli47_RTqPCR_R	AACATTCGGAAAACGGTGAG	Reverse primer for the gene rli47.
opuCA_RTqPCR_F	CATCGATAAAGGAGAATTTG	Forward primer for the gene apuCA.
opuCA_RTqPCR_R	CATAACCAATTGAGCGTCTTAG	Reverse primer for the gene opuCA.
agrA_RTqPCR_F	AAGAACGGATTTCCAATGATCA	Forward primer for the gene agrA.
agrA_RTqPCR_R	CGCTGTCTCAAAAAACAAGATATC	Reverse primer for the gene agrA.
sigB_RTqPCR_F	CTATATTTGGATTGCCGCTTAC	Forward primer for the gene sigB.
sigB_RTqPCR_R	CAAACGTTGCATCATATCTTC	Reverse primer for the gene sigB.
drm2_RTqPCR_F	ACACACCTTTCCGCGTATTC	Forward primer for the gene drm.
drm2_RTqPCR_R	GCTCGCTGGTTTATTTCCAA	Reverse primer for the gene drm.

Root colonization kinetics

Seeds of tall fescue grass (Festuca arundinacea), a model often used to study interactions of plants and microorganisms (Monk et al. 2009; White et al. 2015; Borowick et al. 2019) were disinfected by washing with 3% sodium hypochlorite at 56°C for 1 h, rinsed three times with sterile water and dried at RT. The efficacy of disinfection was assessed by incubating seeds on TSA plates at 25°C and the percentage of germination of dry disinfected seeds was checked in white agar plates (Thermo Scientific™ Oxoid™ Agar Bacteriological) incubated at 30°C for three to four days. Disinfected seeds were germinated for three to four days at 30°C in white agar plates. Seedlings were aseptically transferred to Hoagland mineral solid medium (8 g/L Vitro agar (CondaLab), 0.5 mM KH₂PO₄, 5 mM KNO₃, 1 mM MgSO₄7H₂O, 0.0025 mM Ca(NO₃)₂ 4 H₂O, 0.05 mM H3BO3, 0.1 µM CoCl₂ 7 H₂O, 0.05 µM CuSO₄ 5 H₂O, 0.015 mM ZnSO₄ 7 H₂O, 0.0025 mM Kl, 0.05 mM MnSO₄ H₂O, 0.003 MoNa and 0.05 mM Fe EDTA) with five seedlings per square plate (20 × 120 × 17 mm, Greiner Bio-One). All Hoagland plates were incubated in a plant growth climatic chamber under cycles of 16 h light at 23°C and 8 h dark at 18°C for one week. Bacterial inoculums cultured in TSB medium at 25°C were harvested at exponential phase (OD₆₀₀ = 0.4), washed twice by centrifuging (7000 ×g, 10 min, RT) and suspending the cell pellet in deionised water. The concentration of these inoculums was estimated by OD₆₀₀ reading. For each plant, 200 µL of inoculum adjusted to 5 × 10⁶ CFU.mL⁻¹ was spread uniformly on the roots and allowed to dry horizontally for one hour. Inoculated Hoagland plates were incubated vertically in the plant growth climatic chamber (16 h light at 23°C, and 8 h dark at 18°C) and the population of root-associated L. monocytogenes was enumerated at suitable intervals over one week (on days 0, 1, 2, 4 and 7). At each sampling time, one plate was used per condition. The root system of the five plants was aseptically dissected, weighted and transferred into a vial with 10 mL of TS buffer. 0.7 g of sterile glass beads (≥106 µM acid washed glass beads, Sigma-Aldrich) was added and the vials were vortex for 2 min to remove bacteria from the roots. L. monocytogenes populations were enumerated by serial plating on TSA plates incubated at 25°C.

Preparation of Soil microcosms and L. monocytogenes kinetics

Soil was collected at an experimental farming unit in France (Epoisses, 47° 30′ 22.1832″ N, 4° 10′26.4648″). It is a clay loamy soil with a pH of 7.15. Five soil cores (0–20 cm) were sampled from three locations 20 m apart and pooled in a composite sample. Soils were sieved to 0.4 mm. Aliquots were sterilised by γ-radiation (45 kGy; Ionisos, Dagneux). Uninoculated samples of ionised soil were plated out on TSA agar and incubated at 25°C for 48 h to validate soil sterility. Sterilised and non-sterile soil microcosms were prepared. For each soil sample, triplicate microcosms were prepared by adding 50 g of soil to each of three sterile 180 mL capped plastic tubes (Dutscher, France).

Bacterial cultures were grown statically in TSB at 25°C until exponential phase (OD₆₀₀ = 0.4) or early stationary phase (OD₆₀₀ = 1.0) in the dark. Inoculums were washed twice by centrifuging (7000 × g, 10 min, RT) and suspending cell pellets in deionised water. The concentration of these inoculums was estimated by OD₆₀₀ reading. Each soil microcosm was inoculated to achieve a final concentration of 2 × 10⁶ CFU.g⁻¹ of soil. Soil's water-holding capacity (WHC) was determined (Blazka and Fisher 2014), and the volume of the inoculum was adjusted in order to reach a final soil moisture content of 40% or 60% of the soil's WHC. After inoculation, soils were stirred with sterile spatula, and incubated in the dark at 25°C for two weeks. The dynamics of L. monocytogenes populations in soil were followed immediately after inoculation and periodically over 14 days. One gram (equivalent dry soil) was sampled and added to 9 ml TS-buffer (1% Tryptone, 8.5% NaCl) with 0.7 g of sterile glass beads (≥106 µM acid washed glass beads, Sigma-Aldrich) and vortexed for 2 min. The mixture was then serially diluted and numerated on selective PALCAM agar supplemented with 100 μg.mL⁻¹ cycloheximide and 100 μg.mL⁻¹ rifampicin.

RNA extraction from soil

Cells were lysed by adding 2 g of inoculated soil in sterile lysis tubes containing 4 g of 106 µM beads, 5 g 1.4 mm ceramic beads, 8 × 4 mm glass beads (Sigma-Aldrich). A total of 500 µL of RNAprotect (Qiagen) and 3.5 mL lysis buffer (2 g casein, 20 mL 1 M sodium phosphate (pH 8), 2 mL 5 M NaCl, 10 mL 0.5 M EDTA (pH 8)) was added and tubes were homogenised in the FastPrep instrument (MP biomedicals) 4 × 30 s setting of 6 m/s. Samples were incubated for 5 min on ice in between runs. After 15 min of incubation at 70°C, tubes were centrifuged at 12 000 × g for 5 min at 4°C. The aqueous phase was collected and 850 µl phenol-chloroform-isoamyl alcohol (125:24:1, pH = 4.5) was added to each tube and homogenized for RNA extraction. After centrifugation (14 000 × g, 5 min, 4°C), the aqueous phase was transferred to a new tube. This step was repeated. 800 µL chloroform was added to the clear supernatant and homogenized. Tubes were centrifuged at 14 000 × g for 5 min at 4°C and the aqueous supernatant transferred to a new tube. RNA was precipitated overnight at −80°C by adding 1 mL of precipitation solution (30% polyethylene glycol 6000, 1.6 M NaCl). Samples were centrifuged for 30 min at 16 000 × g at 4°C and the supernatant aspirated. A total of 800 µL of ice cold 75% ethanol was added to the pellet and mixed well. Samples were centrifuged full speed for 10 min at 4°C. Ethanol was aspirated and pellets were left to dry at RT for 30 min. Dry pellets were then dissolved and pooled in 200 µL nuclease free water. Samples were further purified using Zymobiomics RNA mini kit (Zymo Research). The quantity and quality of extracted RNA was measured by NanoDrop (Thermo Fisher Scientific) and agarose gel electrophoresis.

cDNA synthesis and Reverse Transcriptase-Quantitative Polymerase Chain Reaction
(RT-qPCR)

Fifty micrograms of total RNA were DNase-treated with TURBO DNA-free kit according to the manufacturer instructions (Invitrogen). First-strand cDNA was synthesized using SuperScript IV First-Strand Synthesis System (Invitrogen). The cDNA quantity was determined by Qubit fluorometer (Invitrogen) following manufacturer's recommendations. Three sets of cDNA were analysed, originating from three sets of RNA extracted from triplicate cultures. RT-qPCR was performed using QuantiTect SYBR Green PCR Kit (Qiagen) and specific primer sets designed for rli47, ilvA, opuCA and drm (Table 1). Primers efficiency was calculated prior to sample runs. Samples were run on a LightCycler 480 System (Roche) with an initial step at 95°C for 15 min, 45 cycles of 15 s at 95°C, 15 s at 53°C and 30 s at 72°C, a melting curve was drawn for 5 s at 95°C, 1 min at 55°C followed by increases of 0.11°C.s⁻¹ until 95°C, and cooling for 30 s at 40°C. Cycle quantification values were calculated by the software LightCycler 480 Software version 1.5.1 (Roche) and the Pfaffl relative expression formula (Pfaffl 2001; Pfaffl et al. 2002). Relative expression ratio was used to analyse RT-qPCR results with drm (phosphopentomutase) as reference gene (Rieu et al. 2007). Experiments were carried out in three biological replicates, each in technical triplicates.

Statistical analysis

All experiments were performed with technical and biological triplicates. The significance of the differences in patterns of survival in sterilized and biotic microcosms was assessed by two-way ANOVA with Sidak's multiple comparisons test. The same test was used to estimate whether or not the differences in gene transcript levels were statistically different between mutants and the parental strain. Statistical differences between inoculums, growth phases and soil water content were analysed by paired Student t-tests.

RESULTS AND DISCUSSION

Deletion of agrA and sigB resulted in an additive reduction on L. monocytogenes ability to colonise roots

Plant roots can contribute to the spread of bacteria in soil (Kuiper et al. 2004). To determine if alterations of either the Agr communication system, the general stress response regulator σ^B or both affect L. monocytogenes ability to colonise roots, an in vitro study was conducted in roots of F. arundinacea after inoculation of the parental strain and deletion mutants. The population of the parental strain increased within the first day of incubation and stabilised at 10⁶ CFU.root⁻¹ until the end of the experiment (Fig. 1). These results are consistent with the ability of L. monocytogenes to colonise other plant models such as Arabidopsis thaliana (Milillo et al. 2008), Hordeum vulgare (Kutter, Hartmann and Schmid 2006) and several vegetable types (Gorski, Palumbo and Nguyen 2004; Jablasone, Warriner and Griffiths 2005; Gorski, Flaherty and Duhe 2008; Kljujev et al. 2018). Contaminated plant material can be a route of entry of L. monocytogenes in the food chain, from where it can ultimately reach the consumer (NicAogain and O'Byrne 2016). Although it is expected that large doses of L. monocytogenes (∼10⁶ CFU) are required to cause infection, listeriosis outbreaks caused by consumption of contaminated fresh products have been reported (McLaughlin et al. 2011). In this study, the population on roots is consistent with plants acting as possible vectors of L. monocytogenes.

Figure 1.

Listeria monocytogenes colonization of roots of Festuca arundinacea. Logarithmic values of CFU per root on the mutant strains are presented. Each coloured line represents the behaviour profile of each strain, parental (blue), ∆agrA (red), ∆sigB (purple) and ∆agrA∆sigB (green) mutants. The data represents three biological replicates (n = 15 plants) with three technical independent repetitions. Error bars indicate standard deviation. Asterisks represent P-values (* = P-value < 0.05, ** = P-value < 0.001, *** = P-value < 0.0001) calculated using a two-way ANOVA with Sidak's multiple comparisons test.

Deletion of agrA or sigB did not significantly affect the initial growth of L. monocytogenes on roots, however statistically significant reduction of the  ΔsigB mutant compared to parental strain was observed at day 7 (Fig. 1), suggesting a role for σ^B once the population has settled on the root. Impaired attachment and colonisation of radish tissue when σ^B was inactivated was reported previously (Gorski, Duhe and Flaherty 2011). Results with the double mutant were different. After initial growth during the first day of incubation, the population of ∆agrA∆sigB mutant declined until the end of the experiment, and dropped back to its initial level (Fig. 1). This additive effect of agrA and sigB deletions suggested synergy between the two systems during colonisation of the roots. Sequencing indicated additional amino acid changing point mutations in coding regions of the deletion mutants relatively to the parental strain: one point mutation in the ∆agrA mutant (DG125A), lmo1950 (G369→T) and 11 point mutations in the double mutant (CM003), rpsR (G8→A), lmo0178 (C964→T), prs (C763→A), ldh (G548→T), sigH (G507→T), lmo0748 (T230→C), lmo1066 (C407→A), lmo1444 (C31→T), lmo2003 (G130→A), lmo2360 (C898→A) and lmo2684 (C962→T). These SNPs were randomly distributed across the genomes and in genes encoding proteins of diverse functions. Selecting clones of the double mutant happened to be difficult and several attempts were necessary. One can hypothesize that construction of the mutant with the extra deletion counter-selected additional mutations counterbalancing fitness problems when both AgrA and σ^B are inactivated.

Listeria monocytogenes populations increased in sterilised soil, but growth was affected by the double mutation

Sterilised soil microcosms were inoculated with exponential phase cells of L. monocytogenes parental and mutant strains. The population of the parental strain increased over the first day of incubation and the population remained stable until the end of the experiment (Fig. 2). This is in agreement with previous reports on the ability of L. monocytogenes to persist in sterilised soil as a result of inactivation of microbial communities (Dowe et al. 1997; Moshtaghi, Garg and Mandokhot 2009; McLaughlin et al. 2011; Piveteau et al. 2011; Locatelli et al. 2013; Vivant, Garmyn and Piveteau 2013a; Vivant et al. 2014). However, the water content affected the final level of the population and higher maximal populations were recorded at 60% WHC (Fig. 2).

Figure 2.

Growth kinetics of L. monocytogenes strains in sterilized soil microcosms. Logarithmic values of CFU per gram of soil are presented. Each coloured line represents the behaviour profile of each strain, parental (blue), ∆agrA (red), ∆sigB (purple) and ∆agrA∆sigB (green) mutants. The data represents three biological replicates with three technical independent repetitions. Error bars indicate standard deviation. Asterisks represent P-values (* = P-value < 0.05, ** = P-value < 0.001, *** = P-value < 0.0001) calculated using a Two-way ANOVA with Sidak's multiple comparisons test.

Deletion of either agrA or sigB did not affect growth in sterilised soil at 40% WHC but differences between the ∆agrA∆sigB and the parental populations were highly significant from day 1 until the end of incubation (P-value < 0.0001) (Fig. 2). At the higher water content, this phenotype was limited and again, the single mutants and the parental strain had similar growth kinetics. These results suggested that while inactivation of either the Agr system or the general stress response did not affect the dynamics of L. monocytogenes population, the simultaneous loss of the two systems resulted in impaired growth, especially when the water content was low.

In biotic soil, functional AgrA and σ^B are required for maximum survival of L. monocytogenes and further growth defect was recorded after simultaneous inactivation of the two systems.

Different cell growth phases can affect the survival of L. monocytogenes to external factors (Utratna et al. 2014). Because the Agr system is a communication system, cell density has a strong impact on its expression and activity. For this reason, an exponential phase inoculum was chosen to allow comparison with the authors’ previous data (Vivant et al.2013a, 2015). Listeria monocytogenes goes through numerous changes during its life cycle (Wen et al. 2009), some of which might affect the survival of this pathogen to stresses it might encounter in the telluric environment. However, selecting this growth phase also encompasses limitations as exponential phase inoculum is likely to only poorly simulate situations in the natural environment under complex biotic conditions. After determining the ability of exponential cells to colonize roots and to grow in sterilised soil, we aimed to assess if the growth phase would affect survival in the telluric environment. Indeed, the impact of the L. monocytogenes growth phase was one of the parameters tested for soil kinetics experiments.

The presence of indigenous microflora resulted in different population dynamics in soil (Fig. 3). First, no growth was observed in any strain, under any of the experimental conditions tested. The population of the parental strain was stable during the first two days of incubation, and declined dramatically thereafter (four log). The deletion of sigB affected survival of both exponentially grown and stationary phase cells under every condition tested (Fig. 3), even if induction of σ^B activity is known to occur during transition to stationary phase (Utratna et al. 2012). This highlighted a beneficial contribution of the σ^B system to L. monocytogenes soil survival. Similar results were observed in the ∆agrA mutant population, as inactivation of the Agr system resulted in decreased L. monocytogenes populations in all biotic soil conditions tested (Fig. 3)

Figure 3.

Dynamics of (A), exponential and (B), stationary growth phase L. monocytogenes populations in biotic soil microcosms at 40% and 60% of total water holding capacity. Logarithmic values of CFU per gram of soil are presented. Each coloured line represents the behaviour profile of each strain, parental (blue), ∆agrA (red), ∆sigB (purple) and ∆agrA∆sigB (green) mutants. The data represents three biological replicates with three technical independent repetitions. Error bars indicate standard deviation. Asterisks represent P-values (* = P-value < 0.05, ** = P-value < 0.001, *** = P-value < 0.0001) calculated using a Two-way ANOVA with Sidak's multiple comparisons test.

A dramatic decline of the population of the double mutant was observed (Fig. 3A). Interestingly, despite the significant decline in survival of the ∆agrA and ∆sigB mutant populations relative to the parental, this sharper reduction in the double mutant population suggested that some synergy might occur between these two systems during soil survival. Differences in population dynamics according to the water content were not significant in biotic soil (Table 2). Likewise, cells growth phase did not impact L. monocytogenes soil survival in any of the water contents tested (Table 2), except that stationary phase cells of the parental population showed a more resilient phenotype in soil at 40% T_WHC than exponentially grown cells over the first seven days of this experiment (Fig. 3A and B). Soil moisture might affect microbial communities by controlling available oxygen, substrates and water in the pore space. The conclusions of several studies on the effect of the water content on soil microbial communities are not unanimous. While some suggested that water content is a major determinant of the soil's microbial community composition and biomass (Borowik and Wyszkowska 2016; Li et al. 2017), others did not find significant effects (Bossio and Scow 1995; Gordon, Haygarth and Bardgett 2008; Wu et al. 2010).

Table 2.

Statistical analysis of differences in strain survival resulted by growth phase and soil's water content. Differences were calculated using paired t-tests (t = t-value; df = degrees of freedom; * = P-value < 0.05). T_WHC, Soil's total water holding capacity.

	Inoculum growth phase (Exponential vs Stationary phase)						Soil's water content (TWHC = 40% vs TWHC = 60%)
	TWHC = 40%			TWHC = 60%			Exponential phase			Stationary phase
Paired t test	t	df	P-value	t	df	P-value	t	df	P-value	t	df	P-value
Parental	2.804	5	0.03780 *	0.2097	5	0.8422	2.115	5	0.08000	0.2918	5	0.7822
∆agrA	1.574	5	0.1764	0.3635	5	0.7311	1.935	5	0.1108	2.036	5	0.09400
∆agrA∆sigB	0.5313	5	0.6179	0.9068	5	0.4061	0.5599	5	0.5997	2.350	5	0.06560
∆sigB	1.356	5	0.2331	0.7014	5	0.5143	0.1793	5	0.8648	0.03131	5	0.9762

Previous studies showed that impairment of AgrA reduced adaptation of L. monocytogenes to the soil environment (Vivant et al. 2014; Vivant et al. 2015). The AgrA regulon includes genes responsible for the transport and metabolism of amino acids, motility and chemotaxis and genes that code regulators (Garmyn et al. 2012). In the biotic soil environment, over 500 protein-coding genes involved in cell envelope, cellular processes and resistance to antimicrobial peptides, as well as an extensive repertoire of sRNAs were differentially transcribed (Vivant et al. 2015). Together with our results, these studies provide evidence that activation of the Agr system may be critical when L. monocytogenes has to face unfavourable conditions in the telluric environment. Moreover, σ^B activity benefits L. monocytogenes resistance to stresses such as osmotic, acid and oxidative stresses (NicAogain and O'Byrne 2016; Dorey et al. 2019). Data suggest a contribution of the general stress response to face unfavourable conditions that L. monocytogenes may encounter in soil. The present data with agricultural soil confirm previous results from commercial potting soil experiments (Gorski, Duhe and Flaherty 2011).

It is striking that these unfavourable conditions are mainly microbiome driven as σ^B inactivation was detrimental only in live soil. The biotic environment is the parameter with the stronger impact on the ability of L. monocytogenes to survive in soil (Dowe et al. 1997; Moshtaghi, Garg and Mandokhot 2009; McLaughlin et al. 2011; Locatelli et al. 2013; Vivant et al. 2013b; Vivant et al. 2014). Interactions between L. monocytogenes and the background soil microbiota may include competition for resources and/or antibiosis and  σ^B could be important to face such stress conditions, although to our knowledge this has yet to be investigated. Interestingly,  σ^B contributes to survival in the gut (revised by Gahan and Hill 2014), another environment characterised by a rich and complex microbiota. One could postulate that functional  σ^B is required to cope with these microbiota-induced stresses. Abundance, diversity and structure of microbial communities are key parameters of these competitions (Vivant, Garmyn and Piveteau 2013a). Soil protozoa and nematodes could facilitate survival of L. monocytogenes by adding to dispersion of the pathogen (Vivant, Garmyn and Piveteau 2013a). Nonetheless, antimicrobial (Undabarrena et al. 2016; Helfrich et al. 2018) and bacteriocin-like compounds (Sharma, Gupta and Gautam 2014) produced by autochthonous bacteria, as well as the presence of predatory microorganisms (Olanya and Lakshman 2015) are inhibitory to L. monocytogenes, preventing it from thriving in most outdoor environments. Taken together, our results suggest that there might be functional redundancy in the regulation of these two key mechanisms and loss of one could be compensated to some extent by the other when the population adjusts its physiology to environmental conditions. It suggests both AgrA and σ^B cooperatively modulate L. monocytogenes transcriptome to fine-tune adaptation for optimal growth and survival. However, when both regulators are impaired, L. monocytogenes may fail to adapt appropriately.

While in sterile soil σ^B inhibits agrA transcript levels, agrA transcription is restored in the biotic microcosms

Transcripts of rli47, sigB, agrA and opuCA were quantified in L. monocytogenes parental, ∆agrA, ∆sigB and ∆agrA∆sigB mutant strains during incubation in sterilized and biotic soil, and compared to the parental strain (Fig. 4). As expected, no sigB or agrA transcripts were detected in the ∆sigB and ∆agrA mutants, respectively, nor both in the ∆agrA∆sigB mutant. Significantly lower relative transcript levels of rli47 and opuCA were found in the ∆sigB and ∆agrA∆sigB mutants, thus confirming the σ^B-dependent transcription of these genes (Toledo-Arana et al. 2009).

Figure 4.

Impact of sigB and/or agrA deletion on rli47, sigB, agrA and opuCA transcript levels in (A), sterilised and (B), biotic soil. Quantification of rli47, sigB, agrA and opuCA was accessed in the ∆agrA, ∆agrA∆sigB and ∆sigB populations relative to the parental strain by RT-qPCR and normalized to drm. Each coloured bar represents the normalised gene expression of each mutant strain, ∆agrA (red), ∆sigB (purple) and ∆agrA∆sigB (green) relative to the parental strain. Logarithmic fold change values on the mutant strains are relative to the parental strain. The data represents three biological replicates with three technical independent repetitions. Error bars indicate standard deviation. Asterisks represent P-values (* = P-value < 0.05, ** = P-value < 0.001, *** = P-value < 0.0001) calculated using a two-way ANOVA with Sidak's multiple comparisons test.

The sRNA Rli47 was shown to be responsible for specifically repressing isoleucine biosynthesis as a way to restrict growth under harsh conditions, potentially contributing to the survival of L. monocytogenes (Marinho et al. 2019). Unlike what was reported in a previous study (Vivant et al. 2015), the differences in the relative transcript levels of rli47 in the ∆agrA mutant during adaptation to sterilised and biotic soil were not significant. This may be a result of the utilization of different soil types between the two studies. Further investigations are required in order to assess the exact role of the sRNA Rli47 in L. monocytogenes soil adaptation. Significantly lower relative transcript levels of opuCA were found in the ∆agrA background during incubation in soil. This was previously reported in complex medium by Garmyn et al. (2012), where the relative transcripts of the opuCABCD operon was lower when AgrA was not functional.

During adaptation to sterilized soil, relative transcript levels of agrA (log₂FC = 5.0 ± 0.30, P-value < 0.0001) and sigB (log₂FC = 0.82 ± 0.33, P-value ≥ 0.05) were increased in the ∆sigB and ∆agrA populations, respectively (Fig. 4A). However, during adaptation to biotic soil, these differences were not significant (P-value ≥ 0.05) (Fig. 4B). These results further suggest direct and/or indirect repression of agrA by σ^B. In the presence of the soil microbiota, environmental cues may trigger overexpression of agrA in the line of the tight regulation of its activity according to the conditions of the habitat. Although non-significant, the levels of sigB transcripts were higher in the ∆agrA mutant in sterilised soil and no differences were recorded under biotic soil conditions. Altogether, these results suggest that interconnections between cell-cell communication and general stress response might be taking place. Altering either of these systems did affect fitness of L. monocytogenes and inactivation of the two regulators resulted in an even stronger defect preventing the cells to cope with the biotic environment. Overall, our results point to a model where σ^B activity results in a repression of Agr transcription. This was observed in sterilised soil only. In biotic soil, sensing of environmental cues might trigger activation of the Agr system, resulting in increased transcription of the agr operon, despite σ^B activity. Taken together, these results suggest that AgrA and σ^B global regulatory systems play important roles in the natural ecology of L. monocytogenes. This will need to be further confirmed with more isolates of L. monocytogenes tested in a wider range of edaphic conditions.

FUNDING

This work was supported by the European Union's Horizon 2020 Research and Innovation Program under the Marie Skłodowska-CurieGrant agreement No. 641984.

Conflicts of interest. 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.