Modulation of Campylobacter jejuni Motility, Adhesion to Polystyrene Surfaces, and Invasion of INT407 Cells by Quorum-Sensing Inhibition

Campylobacter jejuni is a major foodborne pathogen, and the LuxS-mediated quorum-sensing (QS) system influences its motility, biofilm formation, invasion, host colonization, and virulence. QS therefore represents a target for the control of C. jejuni. The aim of this study was to investigate the correlation of QS inhibition with changes in C. jejuni motility, adhesion to polystyrene surfaces, and adhesion to and invasion of INT407 cells. This was achieved by studying (i) the luxS-deficient mutant and (ii) treatment of C. jejuni with 20 natural extracts as six essential oils, 11 ethanolic extracts, and three pure compounds. Compared to the wild-type, the ΔluxS mutant showed decreased motility, adhesion to polystyrene surfaces, and invasion of INT407 cells. The anti-QS effects of the treatments (n = 15/20) were assayed using Vibrio harveyi BB170 bioluminescence. Moderate positive correlation was shown between C. jejuni QS reduction and reduced motility (τ = 0.492, p = 0.024), adhesion to polystyrene surfaces (τ = 0.419, p = 0.008), and invasion (r = 0.394, p = 0.068). The best overall effect was achieved with a Sedum rosea (roseroot) extract, with 96% QS reduction, a 1.41 log (96%) decrease in adhesion to polystyrene surfaces, and an 82% decrease in invasion. We show that natural extracts can reduce motility, adhesion to polystyrene surfaces, and invasion of INT407 cells by C. jejuni through modulation of the LuxS (QS) system.


Introduction
Campylobacter jejuni is the leading cause of the most commonly reported bacterial gastroenteritis, campylobacteriosis, worldwide. It is considered an important food-safety hazard, and the major route for its spread is via poultry. Campylobacteriosis manifests as acute watery diarrhea, cramps, and fever, and it is associated with the development of the severe neurological condition known as Guillain-Barre syndrome [1][2][3]. The number of confirmed campylobacteriosis cases in European countries is more than 250,000 per year, and due to the overuse of antibiotics in veterinary and human medicine, more than 50% of C. jejuni isolates from poultry are now resistant to antibiotics [4]. The risk of resistant C. jejuni spreading through the food chain is therefore high, and thus the acquisition of resistant strains during infection would be more likely. Although campylobacteriosis rarely ends in death, when an infection is caused by a resistant strain, the likelihood of adverse effects increases by >5-fold [2,4,5]. Altogether, these numbers are alarming, and they represent a major burden for human health and national economies. The estimated total annual cost of campylobacteriosis and its consequences has

Essential Oils and Pure Compounds
All of the essential oils were commercially obtained from companies in Slovenia, as: juniper  Table S1).
The plant materials were dried and ground before the ethanolic extractions with 150 mL 96% denatured ethanol (Roth, Germany) added to 20 g dried material. The suspensions were heated until boiling, stirred for 30 min, and filtered (Rotilabo pleated paper filters; Roth, Germany) to remove undissolved particles. The solvent was evaporated off using a rotary evaporator (Rotavapor; Büchi, Flawil, Switzerland) at 40 • C and 175 bar. The post-distillation material was prepared with the same protocol but used after the distillation process. Finally, extracts were dried fully using a nitrogen flow and stored at −20 • C. The roseroot (E8) was provided by Bucar F. (Department of Pharmacognosy, Institute of Pharmaceutical Sciences, University of Graz, Graz, Austria) and was prepared as described by Alperth et al. [29].

Determination of Minimal Inhibitory Concentrations
The minimal inhibitory concentrations (MICs) against C. jejuni of all of the treatments were initially determined using the broth microdilution method, as described previously [30]. For this purpose, all stock solutions were prepared in dimethylsulphoxide (DMSO) at the stock concentration of 40 mg/mL and then further diluted in MHB. The final concentration of DMSO used in the assays did not exceed 2.5% in MHB with appropriate DMSO controls also included. Treatments with 0.25× MICs were chosen for the AI-2 QS inhibition bioassays (Supplementary Table S1).

Autoinducer-2 Bioassay
To determine the influence of these treatments on C. jejuni QS, AI-2 bioassays were performed. The changes in C. jejuni AI-2 activities or QS were inferred from the changes in the measured luminescence of the V. harveyi BB170 reporter strain in relative luminescent units (RLU).
Campylobacter jejuni 11168 and C. jejuni 11168∆luxS [19] cultures in MHB were adjusted to OD 600 0.1 (5 × 10 7 CFU/mL). Here, the treatment stock solutions were prepared as 100× the concentrations of 0.25× MIC for each treatment in 100% DMSO. Then, 50 µL of each treatment stock was added to 5 mL C. jejuni cultures (at OD 600 0.1) for the final concentration of 0.25× MIC. Cultures treated with 1% DMSO were used as controls. The cultures were incubated under microaerobic conditions at 42 • C for 24 h. To obtain the cell-free supernatants (CFSs) for the following bioassay, samples of 2 mL were taken and filter sterilized using 0.2 µm syringe filters (Sartorius, Göttingen, Germany).
The AI-2 bioassays were performed as previously described [27,28] with some modifications. These QS inhibition bioassays were carried out using the biosensor reporter strain V. harveyi BB170 [27], which was grown for 16 h at 30 • C and 150 rpm mixing to approximately 10 7 CFU/mL and used at the final concentration of 5 × 10 3 CFU/mL in AB medium. The C. jejuni CFSs were added to the suspensions of the reporter strain to a final concentration of 10% (v/v) (i.e., 20 µL CFS added to 180 µL reporter strain suspension). Sterile medium was used as the blank [10% (v/v) MHB, 90% (v/v) AB medium] and the negative control was 10% (v/v) MHB and 90% (v/v) Vibrio harveyi suspension. Kinetic measurements were carried out for the bioluminescence signals of the V. harveyi BB170 that were produced as a result of the QS signal that originated from the presence of the CFSs. The relative luminescence signals were measured at 15 min intervals over 20 h at 30 • C in white microtiter plates (Nunc, Thermo Fisher Scientific, Roskilde, Denmark incubated in a microplate reader (Varioskan Lux; Thermo Scientific, Waltham, MA, USA). The relative luminescence signals were interpreted as the QS signal in the C. jejuni CFSs (i.e., higher signals indicated higher concentrations of QS signaling molecules in the CFSs produced by C. jejuni).
When added to MHB, V. harveyi produces a background luminescence signal that increases with the concentration of the culture. To find the most stable point of signal production, the CFSs from C. jejuni 11168∆luxS, a mutant that cannot produce the quorum-sensing signal AI-2, and fresh MHB were used as the negative controls. The time point when the V. harveyi signal production after addition of MHB and C. jejuni 11168∆luxS was still the same but differed from C. jejuni 11168 CFSs (i.e., after 14.5 h incubation) was used for the calculation of the QS signals attributed to C. jejuni (Supplementary Figure S1).
The experiments were performed as three independent biological replicates and three technical replicates. The data presented are means ± standard deviation from three biological replicates.

Motility Assay
Campylobacter jejuni 11168 and C. jejuni 11168∆luxS were resuspended in MH broth, and the OD 600 was adjusted to 0.1. The treatments were added to these C. jejuni liquid cultures to the final concentrations of 0.25× MIC in 1% DMSO. Control cultures were treated with 1% DMSO. The prepared cultures were incubated at 42 • C under microaerobic conditions for 16 h. After these incubations, 1 µL of each culture was dropped onto 0.4% MH agar and incubated at 42 • C under microaerophilic conditions for 72 h. The diameter of each colony that formed was measured in mm. The experiments were carried out as three technical replicates, where one plate represented one technical replicate, and in three biological replicates on separate occasions. Data are means ± standard deviation of the biological replicates. The differences in the colony diameters (mm) were determined, and the effectiveness of the treatments on C. jejuni motility inferred from the changes in colony diameters (∆mm).

Adhesion to Polystyrene Surfaces
The treatments were carried out in liquid cultures of C. jejuni 11168 and C. jejuni 11168∆luxS (at OD 600 0.1) at the final concentrations of 0.25× MIC and 1% DMSO in MHB. Control cultures were treated with 1% DMSO. Then, 100 µL of the prepared cultures were added to a 96-well polystyrene microtiter plate (untreated). The plate was incubated at 42 • C under microaerophilic conditions for 24 h. After the incubation, the wells were washed with phosphate-buffered saline (PBS) three times and treated in an ultrasonic bath (Iskra Pio, Šentjernej, Slovenia) for 10 min. Then, 10-fold serial dilutions were plated onto MH agar. The numbers of adhered C. jejuni were determined as CFU/well. The experiments were carried out in three technical replicates, where one well was one technical replicate, and in three biological replicates on separate occasions. Data are means ± standard deviation of the biological replicates. The effectiveness of the treatments on C. jejuni adhesion was inferred from the differences between the adhered C. jejuni (CFU/well) of the treated cultures compared to the control.

Invasion and Adhesion of INT407 Cells
The adherence to and invasion of human intestinal epithelial cells by C. jejuni 11168 and C. jejuni 11168∆luxS after the treatments were determined using the INT407 cell line (ATCC), according to Negretti and Konkel [31], with some modifications. The effectiveness of the treatments for C. jejuni QS inhibition was inferred from the change in adhered/invaded C. jejuni (CFU/mL) in the INT407 cell cultures compared to the untreated control.
For the bacterial culture preparation, C. jejuni 11168 and C. jejuni 11168∆luxS were suspended in MH broth, and the OD 600 was adjusted to 0.1. Extracts were added to liquid cultures at the final concentration of 0.25× MIC in 1% DMSO. Control cultures were treated with 1% DMSO. The cultures were then incubated at 42 • C under microaerobic conditions for 16 h. After the incubations, the cultures were centrifuged (6000× g, 20 • C) and washed twice with PBS. The final cultures were resuspended in DMEM1% (DMEM with 1% fetal bovine serum, 1% HEPES, 1% glutamax), and the OD 600 adjusted to 0.3. These suspensions were then diluted 1:10 in DMEM1% to reach an OD 600 of 0.03 or 3 × 10 7 CFU/mL and were then used to infect the INT407 cells.
The plates with the INT407 cells were washed twice with DMEM1%, inoculated with 1 mL of the prepared C. jejuni suspensions, and incubated 3 h at 37 • C under 5% CO 2 . For determination of the total C. jejuni, the plates were washed three times with PBS, and the INT407 cells were lysed with 200 µL 0.1% Triton X-100 (Sigma, MD, USA) for 5 min at 37 • C. Following the incubation, 800 µL PBS was added to each well, and to determine the CFU/mL of C. jejuni, 10-fold serial dilutions were performed with plating onto MHA plates and incubation at 42 • C under microaerobic conditions. For determination of the internalized C. jejuni, the plates were washed twice with DMEM1%, and 1 mL DMEM1% supplemented with 100 mg/L gentamycin (Gipco, Life Technologies, Grand Island, NY, USA) was added. The plates were incubated for an additional 3 h at 37 • C in 5% CO 2 . The plates were then washed, the INT407 cells lysed, and the numbers of internalized C. jejuni determined as described above. The number of adhered cells was determined by subtracting the number of internalized cells from the total C. jejuni concentration. The concentrations of the adhered and the internalized C. jejuni are reported as log 10 CFU/mL.
The experiments were carried out in three technical replicates, where one replicate was one well in the plate, and in three biological replicates on separate occasions. Data are means ± standard deviation of the biological replicates.

Statistical Analysis
For all the data, before statistical tests were performed, normality was tested with Shapiro-Wilk and Kolmogorov-Smirnov tests. All the CFU/mL and the RLU values were log transformed before analysis. To determine the statistical significance between treated and control cultures, Student's t-tests or Mann-Whitney U-tests were performed, depending on the normality distribution of the data. For correlation analysis, only mean values and not replicates were used for each treatment. Correlations between quorum sensing (luminescence), invasion and adhesion (log 10 CFU/mL), motility (diameter), and adhesion to polystyrene surfaces (log 10 CFU/well) were tested using Pearson's or Kendall's Tau correlation tests, depending on the normality distribution of the data.
All the analyses were performed using the SPSS software, version 22 (IBM Corp., Armonk, NY, USA).

Comparison of Wild-Type C. jejuni 11168 with the luxS-Deficient Mutant
To evaluate the influence of QS inhibition caused by the loss of the luxS gene that is responsible for AI-2 signal production in C. jejuni, wild-type C. jejuni 11168 and the luxS-deficient mutant were tested for motility on soft agar, adhesion to polystyrene surfaces, and adhesion to and invasion of INT407 cells.
Compared to the wild-type C. jejuni 11168, the C. jejuni 11168 ∆luxS mutant strain showed a small but significant decrease in motility of 4.7% (3.67 mm; p = 0.018) and significantly lower adherence to polystyrene surfaces of 98.8% (1.82-log units; p < 0.001), and although adhesion on INT407 cells was not significantly altered, the invasion of the INT407 cells by the C. jejuni 11168∆luxS mutant was significantly lower than for the wild-type by 1.10 log units (92.1%; p = 0.003) (Figure 1).
Microorganisms 2020, 8, x FOR PEER REVIEW 6 of 14 motility (diameter), and adhesion to polystyrene surfaces (log10 CFU/well) were tested using Pearson's or Kendall's Tau correlation tests, depending on the normality distribution of the data. All the analyses were performed using the SPSS software, version 22 (IBM Corp., Armonk, NY, USA).

Comparison of Wild-Type C. jejuni 11168 with the luxS-Deficient Mutant
To evaluate the influence of QS inhibition caused by the loss of the luxS gene that is responsible for AI-2 signal production in C. jejuni, wild-type C. jejuni 11168 and the luxS-deficient mutant were tested for motility on soft agar, adhesion to polystyrene surfaces, and adhesion to and invasion of INT407 cells.
Compared to the wild-type C. jejuni 11168, the C. jejuni 11168 ΔluxS mutant strain showed a small but significant decrease in motility of 4.7% (3.67 mm; p = 0.018) and significantly lower adherence to polystyrene surfaces of 98.8% (1.82-log units; p < 0.001), and although adhesion on INT407 cells was not significantly altered, the invasion of the INT407 cells by the C. jejuni 11168ΔluxS mutant was significantly lower than for the wild-type by 1.10 log units (92.1%; p = 0.003) (Figure 1).

Inhibition of Quorum Sensing of C. jejuni with Essential Oils, Ethanolic Extracts, and Pure Compounds
Inhibition of C. jejuni QS was examined using the AI-2 bioluminescence assays, where the C. jejuni CFSs, which contained the QS signals, were added to the biosensory strain V. harveyi BB170. Reduction of C. jejuni QS by these essential oils, ethanolic extracts, and pure compounds was seen as a decrease in the luminescence signal where the CFSs of treated cultures were added in comparison to the control CFS.

Inhibition of Quorum Sensing of C. jejuni with Essential Oils, Ethanolic Extracts, and Pure Compounds
Inhibition of C. jejuni QS was examined using the AI-2 bioluminescence assays, where the C. jejuni CFSs, which contained the QS signals, were added to the biosensory strain V. harveyi BB170. Reduction of C. jejuni QS by these essential oils, ethanolic extracts, and pure compounds was seen as a decrease in the luminescence signal where the CFSs of treated cultures were added in comparison to the control CFS.
Most of these treatments (n = 15/20) significantly lowered the C. jejuni QS signals compared to the control (Figure 2). The luminescence signal decreases ranged from 27% to 96% (Table 1) with the strongest effect achieved by the ethanolic extract E8 (roseroot). One of the essential oils (EO2, oregano), two of the ethanolic extracts (E1, oregano, predistilled; E11, rosemary), and two of the pure compounds (P1, carvacrol; P2, rosmarinic acid) did not have any significant impacts on C. jejuni QS.

Modulation of C. jejuni Motility
The C. jejuni 11168 and the C. jejuni 11168∆luxS motilities were examined in cultures on soft agar that was treated with these essential oils, ethanolic extracts, and pure compounds. The motility of the C. jejuni wild-type was significantly reduced by three of the essential oils (EO1, juniper; EO3, cloves; EO5, lavender), three of the ethanolic extracts (E4, oregano, flowers; E6, nettle, post-distillation; E8, roseroot), and two of the pure compounds (P2, rosmarinic acid; P3, γ-terpinene). The E8 treatment (S. rosea, roseroot) had the strongest effect, as it decreased the C. jejuni 11168 colonies by 27 mm (35%) ( Figure 3A). Although statistical significance was not reached due to the high variability between the experiments here, the motility appeared to be decreased by >15% also by essential oils EO2 (oregano) and EO4 (rosemary) and by ethanolic extract E11 (rosemary) ( Figure 3A, Table 1, Table S2).

Modulation of C. jejuni Motility
The C. jejuni 11168 and the C. jejuni 11168ΔluxS motilities were examined in cultures on soft agar that was treated with these essential oils, ethanolic extracts, and pure compounds. The motility of the C. jejuni wild-type was significantly reduced by three of the essential oils (EO1, juniper; EO3, cloves; EO5, lavender), three of the ethanolic extracts (E4, oregano, flowers; E6, nettle, post-distillation; E8, roseroot), and two of the pure compounds (P2, rosmarinic acid; P3, γ-terpinene). The E8 treatment (S. rosea, roseroot) had the strongest effect, as it decreased the C. jejuni 11168 colonies by 27 mm (35%) ( Figure 3A). Although statistical significance was not reached due to the high variability between the experiments here, the motility appeared to be decreased by >15% also by essential oils EO2 (oregano) and EO4 (rosemary) and by ethanolic extract E11 (rosemary) ( Figure 3A, Table 1, Table S2.).
Unlike the wild-type, the motility of the ΔluxS mutant strain was not significantly changed by the treatments compared to the control; overall, none of these treatments showed any reduction of the motility of C. jejuni 11168ΔluxS by >10% (Supplementary Figure S2A).  Unlike the wild-type, the motility of the ∆luxS mutant strain was not significantly changed by the treatments compared to the control; overall, none of these treatments showed any reduction of the motility of C. jejuni 11168∆luxS by >10% (Supplementary Figure S2A).

Modulation of C. jejuni Adhesion to Polystyrene Surfaces
The number of cultivable C. jejuni 11168 and C. jejuni 11168∆luxS that were adhered to a polystyrene surface after a 24 h incubation with or without these treatments with essential oils, ethanolic extracts, and pure compounds (at 0.25× MIC) were determined as a measure of the C. jejuni biofilm formation on polystyrene surfaces. The adhesion of the wild-type was significantly reduced by nine of these treatments (n = 9/20): four essential oils (EO1, juniper; EO3, cloves; EO4, rosemary; EO5, lavender); four ethanolic extracts (E2, oregano, post-distilled; E3, oregano, leaves; E4, oregano, flowers; E8, roseroot); and one of the pure compounds (P3, γ-terpinene). The reductions caused by the treatments ranged from 0.13 to 1.89 mean log units (26% to 99%) ( Figure 3B, Table 1, Table S2), where the greatest reduction was achieved by essential oil EO1, from juniper. The treatments showed no significant effect on C. jejuni 11168∆luxS adhesion to polystyrene surfaces ( Figure S2B).

Modulation of Adhesion to and Invasion of INT407 Cells by C. jejuni 11168
The effects of these treatments of C. jejuni 11168 and 11168∆luxS on their adhesion to and invasion of INT407 cells was investigated. C. jejuni cultures were incubated together with these treatments (at 0.25× MIC) for 16 h, then washed and added to the INT407 epithelial cells. In this way, only treated C. jejuni were added without transfer of essential oils, ethanolic extracts, or pure compounds to the INT407 cells. Thus, the changes in adhesion or invasion can be considered to result from phenotypic changes of the C. jejuni cultures due to these treatments.
The adhesion of the treated C. jejuni 11168 cultures to the INT407 cells did not significantly differ from the control for the majority of these treatments (n = 17/20). Indeed, only treatments E4 (oregano, flowers), E7 (winter savory), and P3 (γ-terpinene) showed significantly lower C. jejuni adhesion by 0.67, 0.58, and 0.62 mean log units, respectively ( Figure 3C, Table S2). The C. jejuni 11168∆luxS adhesion to INT407 cells was not influenced by any of these treatments (Supplementary Figure S2C).
The invasion of the INT407 cells by C. jejuni 11168 was, on the other hand, significantly modified by most of these treatments (n = 15/20; Figure 3D). The reduced invasion ranged from 0.57 to 1.02 mean log units, with P1 (carvacrol) showing the greatest effect (91%). The five of these 20 treatments that had no significant effects on invasion were four ethanolic extracts (E4, oregano flowers; E5, nettle pre-distilled; E6, nettle post-distilled; and E10, yarrow pre-distilled) and one pure compound (P2, rosmarinic acid) ( Figure 3D, Table 1). The invasion of C. jejuni 11168∆luxS was not affected by these treatments (Supplementary Figure S2D).
It should be noted that none of the treatments achieved the same effects on invasion reduction as did the knock-out mutation of the luxS gene.

Correlation Analysis
The correlations between the QS reductions of the C. jejuni (luminescence signal, logRFU) with the motility (diameter, mm), the adhesion to and invasion of INT407 cells (logCFU/mL), and the adhesion to polystyrene surfaces (logCFU/mL) after these treatments with essential oils, ethanolic extracts, and pure compounds were examined. Only means for luminescence, colony diameters, and C. jejuni logCFU/well were used here. In this way, we determined the correlation of treated cultures without the noise from the error between the experiments. This analysis showed significant, although moderate, positive correlations of luminescence (i.e., QS) with C. jejuni 11168 motility (τ = 0.492, p = 0.024) and adhesion to polystyrene surfaces (τ = 0.419, p = 0.008). These suggest that the decrease in luminescence, and thus the C. jejuni QS, results in decreased C. jejuni motility on soft agar and adhesion to a polystyrene surface.
These data also showed moderate positive correlation of luminescence and invasion of the INT407 cells (r = 0.394), although this remained a trend as statistical significance was not reached (p = 0.068). However, this might suggest an influence of the reduced C. jejuni QS on their invasion of these INT407 epithelial cells.
In the present study, we investigated the modulation of the C. jejuni QS system by (i) knockout mutation of the luxS gene, which is responsible for AI-2 signal production, and (ii) treatments with 20 natural extracts (essential oils, ethanolic extracts, pure compounds). We studied the changes that these modulations might induce in terms of the C. jejuni motility on soft agar, adhesion to polystyrene surfaces, and adhesion to and invasion of human INT407 epithelial cells.
The luxS-deficient mutant showed statistically significant reduction in motility on soft agar (4.7%), adhesion to polystyrene surfaces (1.47-log reduction), and invasion of INT407 cells (1.10-log reduction) compared to the wild-type C. jejuni. Decreased motility of the luxS-deficient mutant compared to the wild-type was also reported by Elvers and Park [13], Jeon et al. [16], Holmes et al. [17], Quiñones et al. [18], and Plummer et al. [19]. Furthermore, Jeon et al. [16] and Holmes et al. [17] showed that the expression of the flaA gene, which is crucial for flagella-mediated motility, is significantly reduced in the luxS-deficient mutant. Elvers and Park [13] also showed somewhat weaker invasion of Caco-2 cells by the luxS-deficient mutant compared to the wild-type (37.5% lower) and no effect on adhesion. Quiñones et al. [18], on the other hand, showed significant reduction in adhesion to Leghorn male hepatoma (chicken hepatoma) cells by the luxS mutant compared to the wild-type.
The response of C. jejuni 11168 and the knockout mutant C. jejuni 11168∆luxS to 20 natural formulations, namely six essential oils, 11 ethanolic extracts, and three pure compounds that are often found in plant materials, were compared using the concentration of 0.25× MIC to ensure they did not influence C. jejuni growth.
The effects of these treatments on the C. jejuni LuxS system were measured with the bioluminescence assay, whereby a reduction in the C. jejuni QS can be inferred from a reduction in the luminescence measured. Significant C. jejuni QS reduction was achieved by the majority of these treatments (n = 15/20; Figure 2). Indeed, a number of these treatments have been previously investigated as QS inhibitors in various organisms, although not in C. jejuni. Juniper [32], lavender [33], cloves [34], thyme [35], and oregano [36] have been reported to show anti-QS activities, which we confirmed here in C. jejuni. Nettle [37] and yarrow [38] had no effects on E. coli QS and Chromobacterium violaceum, respectively, although in the present study they reduced the AI-2 levels (and thus QS) in C. jejuni. In C. violaceum, carvacrol has also been reported to have anti-QS activity [39], although this was not shown here for C. jejuni QS.
Castillo et al. [25] reported that citrus extracts reduced C. jejuni AI-2 activity in this bioluminescence assay, which correlated with its motility and biofilm formation. Furthermore, these citrus extracts reduced C. jejuni invasion of and adhesion to HeLa cells and the expression of the virulence genes cadF and ciaB [26]. On the other hand, Bezek et al. [27] showed good C. jejuni QS reduction with Euodia ruticarpa but no connection to C. jejuni biofilm formation. This indicates high sensitivity of the C. jejuni LuxS system, and thus QS signaling, to outside stressors, such as treatments with natural extracts. However, it does not conclusively confirm the effects of C. jejuni QS inhibition on the other phenotypic changes in C. jejuni.
We showed here that the motility of C. jejuni 11168 was significantly reduced by only eight of these 20 treatments, while the motility of C. jejuni 11168∆luxS was not affected by any of them. Interestingly, the greatest effect on motility was achieved by the S. rosea (roseroot) ethanolic extract (E8), which also had the largest effect on C. jejuni QS. Indeed, a moderate but significant positive correlation of the luminescence, and thus QS, and C. jejuni 11168 motility was found after these treatments (τ = 0.492, p = 0.024). This suggests that treatments with these essential oils, ethanolic extracts, and pure compounds can modulate C. jejuni motility through modulation of the LuxS/QS system. C. jejuni motility is a promising target for reduction of C. jejuni host colonization, as a loss of motility greatly reduces its colonization [40,41].
Flagella-mediated motility and QS also have important roles in C. jejuni biofilm formation, because non-motile mutants and luxS-deficient mutants have shown weaker biofilm formation compared to the wild-type [42,43]. As a measure of C. jejuni biofilm formation on polystyrene surfaces, we determined here the number of cultivable C. jejuni that adhered to the walls of a polystyrene plate. The adhesion of C. jejuni 11168 to the polystyrene surface was significantly reduced by nine of these treatments with essential oils, ethanolic extracts, and pure compounds. The juniper essential oil (EO1) had the strongest effect here, as it reduced C. jejuni adhesion by 1.89-log units (approximately 99%). The same strong anti-adhesion effect of juniper essential oil (99% reduction) was also demonstrated by Klančnik et al. [44] using qPCR quantification of C. jejuni adhesion.
Significant moderate positive correlation was found here between luminescence (i.e., QS) and adhesion of C. jejuni to polystyrene surfaces after these treatments. This thus indicates that reduced luminescence as AI-2 activity for C. jejuni corresponds to reduced adhesion.
The adhesion of the luxS mutant was lower compared to the wild-type, which again suggests involvement of the LuxS system of C. jejuni in its adhesion to polystyrene surfaces. Consequently, the modulation of the LuxS system results in reduced adhesion of C. jejuni.
The influence in the present study of these treatments with essential oils, ethanolic extracts, and pure compounds on C. jejuni invasion of and adhesion to INT407 epithelial cells was tested. Only three of these treatments of C. jejuni 11168 reduced their adhesion to the INT407 cells (E4, oregano, flowers; E7, winter savory; P3, γ-terpinene), and none reduced the adhesion of C. jejuni 11168∆luxS. Although the adhesion of C. jejuni to the INT407 epithelial cells was generally not affected, the majority of these treatments (n = 15/20) reduced the C. jejuni 11168 invasion of the INT407 cells. None of the treatments achieved the same level of effect as the knock-out mutation of the luxS gene. This was not surprising, as none of the treatments reduced the C. jejuni QS signal by 100%. The effects of these treatments on C. jejuni 11168 and the absence of effects on C. jejuni 11168∆luxS show that modulation of the LuxS system by outside stressors can result in changed C. jejuni invasiveness, which in this case was caused by most of these treatments with essential oils, ethanolic extracts, and pure compounds. Furthermore, a moderate, although not statistically significant, positive correlation was shown between C. jejuni AI-2 activity and C. jejuni invasiveness of the INT407 epithelial cells (r = 0.394, p = 0.068). Lower cell invasiveness of the luxS mutant on Caco-2 cells and no effect on adhesion were previously reported by Elvers and Park [13]. This shows the sensitivity of C. jejuni to the loss or inhibition of the QS system in terms of epithelial cell invasion reduction but not adhesion.
The treatment that showed the greatest C. jejuni QS reduction was the S. rosea (roseroot) ethanolic extract (E8), and this also had the greatest effect on C. jejuni motility and was among the best in terms of reduced invasion of INT407 cells and adhesion to polystyrene surfaces by C. jejuni. The roseroot extract is composed of different cyanogenic glycosides, phenylethanoids, procyanidin and catechin derivatives, phenylpropanoids, monoterpene alcohols, and flavonoids [29], and isolated compounds of this extract should be considered in the search for new anti-QS compounds against C. jejuni. Also, rosmarinic acid (P2) was among the weakest in terms of effects on C. jejuni QS, motility, and adhesion to polystyrene surfaces. The moderate positive correlation of the C. jejuni QS signal activity with C. jejuni invasion of epithelial cells, motility, and adhesion to polystyrene surfaces with these treatments and the absence of any effects with the ∆luxS mutant indicate a role for the LuxS system in the regulation of these phenotypes. This now defines the LuxS system as an interesting target for the control of C. jejuni and also that such plant-derived substances are potential QS inhibitors and thus potential anti-Campylobacter agents.

Conclusions
The beneficial effects of used treatments against the wild-type C. jejuni and the lack of effect against the luxS-deficient mutant strongly suggest the influence of these treatments on the LuxS system. The positive, although moderate, correlation of QS signal reduction with reduced invasion of epithelial cells, motility, and adhesion to polystyrene surfaces links these phenotypic changes. However, as not all the changes in QS seen here resulted in the changed phenotypes, other mechanisms of action of such plant extracts remain to be considered before their use for reduction of C. jejuni in vivo.
We confirmed here that modulation of the C. jejuni LuxS system with essential oils, ethanolic extracts, and pure compounds can result in reduced C. jejuni invasion of INT407 epithelial cells and motility and adhesion to polystyrene surfaces, and that these phenotypic changes correlate with the measured AI-2 activity. To confirm the anti-QS effects of these treatments, further studies are warranted in C. jejuni. Better quantitative measurement systems and the complementation of the C. jejuni luxS mutant as well as the addition of a pure AI-2 control would further confirm that these natural plant-derived essential oils, ethanolic extracts, and pure compounds act as true QS modulators.