Clusters of Lactobacillus Strains from Vegetal Origins Are Associated with Beneficial Functions: Experimental Data and Statistical Interpretations

Nine strains of Lactiplantibacillus plantarum and one strain of Lacticaseibacillus paracasei that were recently isolated from prickly pears, fresh figs and blackberries, which are traditionally and largely consumed fruits in Kabylia (north of Algeria), were studied here for their antagonism and antioxidant properties as well as for production of exopolysaccharides. With respect to their inhibitory properties, these strains were tested against three food representative pathogens including Escherichia coli ATCC 8739, Staphylococcus aureus 2S6 and Listeria monocytogenes 162. The antagonism of these pathogens was attributable to lactic acid production, present in the cell free supernatant, at concentrations ranging from 9 to 16.74 g/L. The anti-adhesive properties observed on polystyrene or eukaryotic Caco-2 cells were exerted in a strain dependent-manner. Indeed, the scores obtained ranged from 27% to 75% for S. aureus 2S6, 54% to 95% for L. monocytogenes 162, and 50% to 97% for E. coli ATCC 8739. The co-aggregation of these Lactobacillus strains with the aforementioned target bacteria appeared to be exerted in a strain-dependent manner, with noticeably the upmost rate for Lb. paracasei FB1 on S. aureus 2S6. Interestingly, these novel Lactobacillus strains were able to produce a large amount (315.55 to 483.22 mg/L) of exopolysaccharides, and showed a significant scavenging activity on the 2,2-di-phényl-2-picrylhydrazyle (DPPH) synthetic free radical with rates of 51% to 56%. Of note, the highest antioxidant activity was observed for Lb. paracasei FB1 using the culture supernatants, intact cells or the intracellular extract. The statistical analysis of these data using the principal component analysis (ACP) enabled us to establish three distinct clusters with potential applications as bioprotective and/or probiotic agents, following further evaluation.


Introduction
Foodborne pathogens represent a major health risk for consumers [1], causing a variety of foodborne diseases such as abdominal pain, diarrhea, fever, low blood pressure, vomiting and other gastrointestinal (GIT) symptoms [2,3]. Foodborne pathogens are considered as one of the most critical public health concerns spreading worldwide [4][5][6]. This unwanted spread of foodborne pathogens can negatively affect the whole economy. Indeed, according to the American Control Diseases Center (CDC) and FoodNet reports, bacterial pathogens such as Listeria monocytogenes, Escherichia coli, cultures were deposited on MRS plates, and incubated for 18 h at 37 • C. Afterwards, 10 mL of BHI agar (8 g agar/L) previously inoculated with the target strain, at 10 6 CFU/mL, were added and incubated again for 18 h at 37 • C. After incubation, the plates were inspected, any inhibition zone around the wells or the spots was recorded, and diameters were measured [33].

Quantification of the Lactic Acid
The quantification of lactic acid produced by Lactobacillus strains was performed by high performance liquid chromatography (HPLC) by spectra system P1000XR (Thermo Fisher Scientific, Waltham, MA, USA) using a Fast Fruit Juice Column (50 mm × 7.8 mm, Phenomenex, Torrance, CA, USA). The mobile phase used was H 3 PO 4 (0.05%, w/w), with a flow rate of 0.8 mL/min and a temperature of 55 • C. Supernatants from Lactobacillus cultures were collected after 8, 18 and 24 h of incubation at 37 • C in MRS broth, then centrifuged (8000× g, 10 min, 4 • C), and filtered through a filter of 0.2 µm/pore size. The volume of injected sample was 25 µL, a calibration curve of pure lactic acid (Sigma-Aldrich) with concentrations of 1, 5, 10, 15 and 20 g/L was carried out to establish the correlation between data obtained from chromatographic peaks area and produced lactic acid concentration. The peak corresponding to lactic acid, which is eluted at 5.32 min [34], was identified using the Azur software.

Biofilm Formation Assessment of Lactobacillus Strains on Polystyrene Tissue Culture Plates (TCP)
To assess the biofilm formation of the studied strains to polystyrene plate, a semi quantitative method was used, as previously described [35]. Briefly, 100 µL of each culture of Lactobacillus (10 8 CFU/mL), grown in MRS broth, were added to the wells of sterile 96-well microplates already filled with 100 µL of tryptic soy broth (TSB) (Difco, Detroit, MI, USA). The microplates were left for 15 min under gentle stirring before being incubated at 37 • C. After 24 h, the cultures were aspirated and the non-adherent cells were removed by two washes of the wells with phosphate buffered saline (PBS, 10 mM, pH 7.2). Subsequently, 200 µL of 96% ethanol (Sigma-Aldrich, St Louis, MO, USA) were added to each well in order to fix the adherent cells. After 15 min of fixation, the wells were drained, dried and then stained with 0.1% (w/v) crystal violet (Biochem Chemopharma, QC, Canada) for 30 min. The stained cells were washed twice with 200 µL of PBS before extracting the dye with 200 µL of 96% ethanol. The number of cells was quantified using a microplate reader (ELX800, BioTek, Winooski, VT, USA) by measuring the absorbance (A) at 630 nm. According to the recommendations of Stepanović et al. [36], these strains were classified into four categories. Considering Ac as the absorbance of the control (sterile TSB), the following interpretations were applied; A ≤ Ac: non-adherent (non-biofilm producer), 2Ac ≥ A > Ac: weakly adherent (weak biofilm producer), 4Ac ≥ A > 2Ac: moderately adherent (moderate biofilm producer), and strongly adherent (strong biofilm producer): A > 4Ac.

Exopolysaccharide Production
The exopolysaccharide (EPS) production of Lactobacillus strains was evaluated according to the method described by van Geel-Schutten et al. [37]. Briefly, Lactobacillus strains were grown for 72 h at 37 • C in 30 mL of MRS broth supplemented with 2% (w/v) glucose. Bacterial cells were removed by centrifugation (6000× g for 20 min, 20 • C) and two volumes of 95% (v/v) cold ethanol (Sigma-Aldrich) were added to one volume of untreated CFS and maintained at 4 • C for 24 h to precipitate EPS. Then, the obtained precipitates were recovered by centrifugation (2000× g, 15 min, 4 • C), washed with distilled water and dried at 60 • C until constant weight was reached. The dried weight was then measured to determine the amount of EPS produced by the Lactobacillus strains [29].

Inhibition of Biofilm Formation by Lactobacillus CFS
Firstly, 50 µL of non-neutralized CFSs of Lactobacillus strains grown in MRS broth, and 50 µL of E. coli ATCC 8739, S. aureus 2S6 or L. monocytogenes 162 at 10 6 CFU/mL, prepared as above-indicated were mixed and added to the wells of sterile 96-well microplates containing 100 µL of sterile TSB medium (Difco, Detroit, MI, USA). The microplates were left under gentle stirring for 15 min before their incubation for 24 h at 37 • C [35]. Of note, the tests were performed in triplicate and sterile MRS broth was used as a negative control. According to Shokri et al. [38], to evaluate the pathogen biofilm removal by CFSs of lactobacilli, the aforementioned target strains were grown at 37 • C for 24 h in the microplates wells, allowing them to form biofilms. Afterwards, 100 µL of non-neutralized CFSs from Lactobacillus strain were added to the wells and then incubated again for 4 h, at 37 • C. Each test was done in triplicate and sterile MRS broth (100 µL) was used as a negative control. Biofilm reduction was measured using the same steps as for the biofilm quantification previously reported [35].

Ultra-Structure Alterations as Visualized by Transmission Electron Microscopy (TEM)
Cultures of E. coli ATCC 8739, S. aureus 2S6 and L. monocytogenes 162 incubated at 37 • C for 18 h were centrifuged (8000× g, 10 min, 4 • C), and the resulting cells were suspended in non-neutralized CFS of Lb. plantarum 2F8 cultures, which presented antibacterial and anti-adhesive activities as well as upmost lactic acid production. Sterile MRS broth (pH 6.5) was used as a control. All suspensions were incubated for an additional 18 h at 37 • C. Samples were collected from each suspension and the cells were recovered by centrifugation (8000× g, 10 min, 4 • C) as a small pellet. For transmission electron microscopy (TEM), the pelleted cells were fixed with 2.5% (v/v) glutaraldehyde solution and 0.1 M (v/v) of cacodylate buffer (pH 7.4) and prepared on a Formvar film of 300 square mesh, nickel grid (EMS FF300-Ni). The TEM images were recorded on a JEOL JEM 2100FX TEM instrument (Jeol, Tokyo, Japan) equipped with a GATAN CCD Orius 200D camera (Gatan, Pleasanton, CA, USA) at an acceleration voltage of 200 KV.

Co-Aggregation Test
The co-aggregation experiences were carried out as previously described by Kos et al. [39]. Briefly, the pathogenic strains E. coli ATCC 8739, S. aureus 2S6 and L. monocytogenes 162 as well as Lactobacillus strains were cultured at 37 • C for 18 h in BHI and MRS broth, respectively. After centrifugation (8000× g, 10 min, 4 • C), the pelleted cells were washed twice with sterile PBS (10 mM, pH 7.2), and the cells were re-suspended in PBS to a final concentration of about 10 8 CFU/mL. Samples of 2 mL of Lactobacillus and pathogen suspensions were mixed by vortexing for 30 s in glass test tubes. Tubes containing 4 mL each pathogen suspension or Lactobacillus strain alone were considered as controls. Absorbance (A) was measured immediately and after 2 h of incubation at 37 • C. The following formula was used to calculate the co-aggregation percentage: where A represents the absorbance, x and y represent each strain in the control tubes, and (x + y) represents their mixture.

Inhibition of Pathogenic Strains Adhesion to Caco-2 Cells by Lactobacillus Strains
The human colorectal adenocarcinoma Caco-2 cells were used for the adhesion inhibition assays [40]. The cells were grown at 37 • C in presence of 5% CO 2 in Dulbeco's modified Eagle medium (DMEM) containing 4.5 g/L of glucose and supplemented with L-glutamine (2 mM), penicillin (100 U/mL), streptomycin (100 µg/mL), 10% of heat-inactivated fetal bovine serum (FBS) and 1% (v/v) non-essential amino acids. All these reagents were provided by PAN-Biotech GmbH (Aidenbach, Germany). The adhesion inhibition assays were carried out as described by Bendali et al. [41] including some adjustments. The 24-well tissue culture plates were used to prepare monolayers of Caco-2 cells. The wells were inoculated by 4.10 4 Caco-2 cells per well, and the plates were incubated for 7 days. Two different protocols were used in order to discriminate competition/exclusion of E. coli ATCC 8739, S. aureus 2S6 and L. monocytogenes 162 with/by lactobacilli.
For exclusion tests, Lactobacillus strains at 10 8 CFU/mL, washed with 1 mL of PBS and resuspended in DMEM without serum or antibiotics, were added to Caco-2 cell monolayers and incubated for 90 min at 37 • C (5% CO 2 ). Afterwards, non-adherent Lactobacillus strains were removed by washing twice with PBS, and E. coli ATCC 8739, S. aureus 2S6 or L. monocytogenes 162 at 10 7 CFU/mL, prepared in the same conditions as Lactobacillus strains, were added and incubated for an additional 2 h at 37 • C. Here, bacterial charges were added to Caco-2 cells monolayer taking into account the multiplicity of infection (MOI) ratio, in agreement with a previously report [42]. Of note, the MOI ratio of lactobacilli was 1:100 (Caco2: lactobacilli), while that for pathogens was 1:10 (Caco2: pathogens).
For competition tests, Lactobacillus (10 8 CFU/mL) and pathogen (10 7 CFU/mL) strains, both prepared as previously described, were mixed and added to the Caco-2 monolayers and incubated for 2 h at 37 • C. Then, the Caco-2 monolayers were washed twice with 500 µL of PBS and incubated with 200 µL of Trypsin/EDTA (Gibco) for 15 min to remove Caco-2 cells with adherent bacteria.
After exclusion and competition tests, the enumeration of adherent E. coli ATCC 8739, S. aureus 2S6 and L. monocytogenes 162 cells was performed on specific media for each strain; eosine methylene blue (EMB), Chapman Stone agar and Palcam agar, respectively. Pathogen adhesion rates were calculated in reference to the control (well containing pathogens without Lactobacillus strains) which represents 100% adhesion.

CFS, Intact Cells and Intracellular Cell-Free Extract Scavenging Activity on Free Radical DPPH
Lactobacillus strains cultures were inoculated into MRS broth and incubated at 30 • C for 18 h and the overnight culture was centrifuged (8000× g at 4 • C for 10 min). The cell-free supernatants were subjected to 2,2-di-phényl-2-picrylhydrazyle (DPPH) free radical scavenging assay, as described by Sharma et al. [43]. A volume of 500 µL of cell-free supernatants was added to 3 mL of freshly prepared DPPH (5 mg/100 mL of methanol), mixed by vortexing and incubated for 30 min in the dark. After 30 min, absorbance was measured at 515 nm with non-inoculated MRS broth as blank. Ascorbic acid was used as the synthetic and natural standard. The percentage of radical scavenging activity was calculated according to the equation: Cells of lactobacilli were harvested by centrifugation (8000× g at 4 • C for 10 min) after 18 h of incubation at 30 • C. For the preparation of intact cells, cells were washed three times with PBS (10 mM, pH 7.2) and resuspended at 10 9 log CFU/mL in this same buffer. To prepare intracellular cell-free extracts, cell pellets were quickly washed twice with deionized water and resuspended in the same solution before transferring to NucleoSpin ® Bead Tubes Type B (Macherey-Nagel, Duren, Germany). Tubes were homogenized using FastPrep-24 5G (MP Biomedicals, Santa Ana, CA, USA) for 3 cycles of 30 s, with cooling on ice bath for 5 min between each cycle. Cell debris was removed by centrifugation at 11,000× g at 4 • C for 10 min, allowing the recovery of a supernatant containing intracellular cell-free extract [44,45]. As previously described by Lin and Chang, with slight modification, 800 µL of a milliliter of intact cells or intracellular cell-free extract were mixed with 1 mL of freshly prepared DPPH solution (0.004%, w/v in methanol) and the scavenging ability was calculated as stated above using PBS as a blank.

Statistical Analysis
Differences between samples were calculated using one way ANOVA and the post hoc Tukey test (p < 0.05) XL-STAT (version 2009, Addinsoft, Paris, France), and data were expressed as a mean ± standard error calculated from at least three independent experiments. The resulting data were analyzed by principal component analysis (PCA) using FactoMineR software. The principal component analysis was performed using version R 3.5.2 (www.r-project.org, R foundation for statistical computing).

Lactobacillus Strains Displayed Antibacterial Properties through Different Mechanisms
The ten Lactobacillus strains inhibited noteworthy results regarding the growth of the three pathogenic strains used here, according to the results from the spot-on-lawn method (Table 1). Nevertheless, this antagonism was attributed to the non-neutralized CFS based on the well diffusion method. Furthermore, the diameters of the inhibition zones obtained for all Lactobacillus strains were similar, regardless of the considered pathogen, except for Lb. paracasei FB1 strain, which exhibited lower inhibition zone diameters (Table 1).
Results are expressed as mean ± standard deviation n = 3, means with different lowercase letters were significantly different (p < 0.05) based on Tukey's test.

Lactic Acid Quantification
The absence of inhibition when the neutralized CFS was used indicated that antibacterial activity was due to the pH lowering, most likely due to the production of lactic acid. Related to that, we noticed that Lb. plantarum F3 and 2F8 strains produced the greatest amounts of lactic acid, reaching 16.74 g/L after 24 h of growth. However, the lowest production was registered for Lb. paracasei FB1 and Lb. plantarum FB3 with 10.73 and 9.64 g/L, respectively (Table 2). Results are expressed as mean ± standard deviation n = 3, means with different lowercase letters were significantly different (p < 0.05) based on Tukey's test.

Adhesive Capacity of Lactobacillus and Pathogenic Strains on Polystyrene Microplates
The results illustrated in Figure 1 highlight the capabilities of Lactobacillus strains to adhere and form biofilms under the tested conditions. Consequently, the absorbencies recorded for Lactobacillus Foods 2020, 9, 985 7 of 19 strains ranged from 0.26 to 1.78 ( Figure 1). Nine out 10 strains were strongly adherent according to the classification proposed by Stepanović et al. [36], whereas Lb. plantarum FB3 appeared to be moderately adherent. Remarkably, the most adherent strains were Lb. plantarum FB13 followed by Lb. plantarum NCA4. Regarding the pathogenic strains, the absorbencies recorded for E. coli ATCC 8739, L. monocytogenes 162 and S. aureus 2S6 were 0.366, 0.422 and 1.368, respectively. Therefore, S. aureus 2S6 was thereof considered as strongly adherent, while E. coli ATCC 8739 and L. monocytogenes 162 were moderately adherent.

Adhesive Capacity of Lactobacillus and Pathogenic Strains on Polystyrene Microplates
The results illustrated in Figure 1 highlight the capabilities of Lactobacillus strains to adhere and form biofilms under the tested conditions. Consequently, the absorbencies recorded for Lactobacillus strains ranged from 0.26 to 1.78 ( Figure 1). Nine out 10 strains were strongly adherent according to the classification proposed by Stepanović et al. [36], whereas Lb. plantarum FB3 appeared to be moderately adherent. Remarkably, the most adherent strains were Lb. plantarum FB13 followed by Lb. plantarum NCA4. Regarding the pathogenic strains, the absorbencies recorded for E. coli ATCC 8739, L. monocytogenes 162 and S. aureus 2S6 were 0.366, 0.422 and 1.368, respectively. Therefore, S. aureus 2S6 was thereof considered as strongly adherent, while E. coli ATCC 8739 and L. monocytogenes 162 were moderately adherent.

The CFS from Lactobacillus Strains Were Able to Prevent the Pathogens Biofilm Formation
The non-neutralized CFS from all tested Lactobacillus strains were able to prevent the adhesion and subsequently the biofilm formation of E. coli ATCC 8739, S. aureus 2S6 and L. monocytogenes 162 based on the data obtained with the semi quantitative TCP method ( Figure 2A). The adhesion inhibition rates ranged between 77.23% ± 1.02% and 89.40% ± 0.41%. The significantly most important percentage of inhibition was registered for Lb. plantarum FB3 (89.40% ± 0.41%) against S. aureus 2S6 biofilm formation. In addition, these Lactobacillus strains were able to remove the formed biofilms after 4 h of contact with the non-neutralized CFS ( Figure 2B). Therefore, the biofilm removal levels ranged from 52.60% ± 0.42% to 78.02% ± 1.91%, and the significantly highest rate of removal was shown for Lb. paracasei FB1 against L. monocytogenes 162 biofilm.
percentage of inhibition was registered for Lb. plantarum FB3 (89.40% ± 0.41%) against S. aureus 2S6 biofilm formation. In addition, these Lactobacillus strains were able to remove the formed biofilms after 4 h of contact with the non-neutralized CFS ( Figure 2B). Therefore, the biofilm removal levels ranged from 52.60% ± 0.42% to 78.02% ± 1.91%, and the significantly highest rate of removal was shown for Lb. paracasei FB1 against L. monocytogenes 162 biofilm. The rates are the means of three independent experiments. The error bars represent the standard deviations. Columns without common letter are significantly different (P < 0.05) using Tukey's test.

Exopolysaccharide Production
Lactobacilli strains were found to be able to produce 315.55 to 483.22 mg/L of EPS (Table 2). Furthermore, production of EPS in significantly (P < 0.05) higher amounts was displayed by Lb. plantarum NCA4 (483.22 mg/L), followed by Lb. plantarum F2 (454.12 mg/L) and F3 (453.32 mg/L). On the other hand, Lb. paracasei FB1 was found to be the lowest EPS producer, with only 315.55 mg/L.

Exopolysaccharide Production
Lactobacilli strains were found to be able to produce 315.55 to 483.22 mg/L of EPS (Table 2). Furthermore, production of EPS in significantly (p < 0.05) higher amounts was displayed by Lb. plantarum NCA4 (483.22 mg/L), followed by Lb. plantarum F2 (454.12 mg/L) and F3 (453.32 mg/L). On the other hand, Lb. paracasei FB1 was found to be the lowest EPS producer, with only 315.55 mg/L.

Cellular Ultra-Structural Alterations of Pathogens Treated with CFS Examined by Transmission Electron Microscopy (TEM)
TEM was used to observe the ultra-structural modifications, if any, of pathogens cells treated with CFS from Lactobacillus strains. The data gathered showed the disruption of the cell wall structure and the condensation of ribosomes. Therefore, bacterial death resulted from the wall's mechanical strength and osmotic lysis. The cytoplasm content of treated E. coli cells was agglutinated and appeared as globules of variable size. Nevertheless, the membrane was easily observed as being devoid of any break or clear cut ( Figure 3A(2)). Notably, the treated S. aureus cells did not show any particular intracellular damage. The cell wall was altered with a "frayed" appearance, indicating a possible peptidoglycan alteration ( Figure 3B(2)). Regarding L. monocytogenes, cells exhibited an altered cell wall, separation of membrane layers and leakage of intracellular contents ( Figure 3C(2)).
Foods 2020, 9, x FOR PEER REVIEW 9 of 20 TEM was used to observe the ultra-structural modifications, if any, of pathogens cells treated with CFS from Lactobacillus strains. The data gathered showed the disruption of the cell wall structure and the condensation of ribosomes. Therefore, bacterial death resulted from the wall's mechanical strength and osmotic lysis. The cytoplasm content of treated E. coli cells was agglutinated and appeared as globules of variable size. Nevertheless, the membrane was easily observed as being devoid of any break or clear cut ( Figure 3A(2)). Notably, the treated S. aureus cells did not show any particular intracellular damage. The cell wall was altered with a "frayed" appearance, indicating a possible peptidoglycan alteration ( Figure 3B(2)). Regarding L. monocytogenes, cells exhibited an altered cell wall, separation of membrane layers and leakage of intracellular contents ( Figure 3C(2)).

The Co-Aggregation Ability of Lactobacilli with Pathogens
The co-aggregation of the ten-studied Lactobacillus strains with the three pathogens (E. coli ATCC 8739, S. aureus 2S6 and L. monocytogenes 162) is shown in Figure 4. Notably, the co-aggregation registered for Lb. paracasei FB1 after 2 h of incubation was significantly (p < 0.05) stronger with S. aureus (28.34% ± 2.56%), followed by those obtained for E. coli (23.39% ± 2.12%), and L. monocytogenes (20.10% ± 2.04%). Interestingly, Lb. plantarum NCA4 showed co-aggregation with L. monocytogenes but not with E. coli ATCC 8739 and S. aureus 2S6. Similarly, Lb. plantarum FB13 did not show co-aggregation with S. aureus 2S6. Columns without a common letter are significantly different (P < 0.05) using Tukey's test.

Anti-adhesive Activity of Lactobacillus Strains towards Pathogens on Caco-2 Cells
The adhesion rates to human epithelial Caco-2 cells of L. monocytogenes 162, S. aureus 2S6 and E. coli ATCC 8739 were 8.40%, 7.25% and 7%, respectively ( Figure 5). However, when the adhesion experiment was performed, after incubating either pathogens with Lactobacillus strains, the adhesion levels decreased significantly (P < 0.05), as depicted in Figure 6. Regarding the competition method, when Lactobacillus strains were incubated simultaneously with pathogens, the adhesion rates decreased significantly (P ˂ 0.05), ranging from 6.84% to 49.09% for E. coli ATCC 8739, 9.05% to 45.74% for L. monocytogenes 162, and 30.46% to 72.92% for S. aureus 2S6 ( Figure 6A). Based on the exclusion method, the adhesion appeared to further decrease. Indeed, when Lactobacillus strains were added to Caco-2 cells for 90 min before pathogen incorporation, we observed that rates were between 2.61% and 43.55% for E. coli ATCC 8739, 5.13% and 40.89% for L. monocytogenes 162 and 24.32% and 65% for S. aureus 2S6, as illustrated in Figure 6B. The lowest adhesion rate was reported following addition of Lb. plantarum F2 strain, with 2.61%, 5.13%, and 24.32% for E. coli ATCC 8739, L. monocytogenes 162 and S. aureus 2S6, respectively. The presence of Lb. paracasei FB1 exhibited a weak adhesion score, with only 3.97% in the presence of E. coli ATCC 8739, and this score reached 8.63% in the presence of L. monocytogenes 162 and 36.05% for S. aureus 2S6. Columns without a common letter are significantly different (p < 0.05) using Tukey's test.

Anti-Adhesive Activity of Lactobacillus Strains towards Pathogens on Caco-2 Cells
The adhesion rates to human epithelial Caco-2 cells of L. monocytogenes 162, S. aureus 2S6 and E. coli ATCC 8739 were 8.40%, 7.25% and 7%, respectively ( Figure 5). However, when the adhesion experiment was performed, after incubating either pathogens with Lactobacillus strains, the adhesion levels decreased significantly (p < 0.05), as depicted in Figure 6. Regarding the competition method, when Lactobacillus strains were incubated simultaneously with pathogens, the adhesion rates decreased significantly (p < 0.05), ranging from 6.84% to 49.09% for E. coli ATCC 8739, 9.05% to 45.74% for L. monocytogenes 162, and 30.46% to 72.92% for S. aureus 2S6 ( Figure 6A). Based on the exclusion method, the adhesion appeared to further decrease. Indeed, when Lactobacillus strains were added to Caco-2 cells for 90 min before pathogen incorporation, we observed that rates were between 2.61% and 43.55% for E. coli ATCC 8739, 5.13% and 40.89% for L. monocytogenes 162 and 24.32% and 65% for S. aureus 2S6, as illustrated in Figure 6B. The lowest adhesion rate was reported following addition of Lb. plantarum F2 strain, with 2.61%, 5.13%, and 24.32% for E. coli ATCC 8739, L. monocytogenes 162 and S. aureus 2S6, respectively. The presence of Lb. paracasei FB1 exhibited a weak adhesion score, with only 3.97% in the presence of E. coli ATCC 8739, and this score reached 8.63% in the presence of L. monocytogenes 162 and 36.05% for S. aureus 2S6.      The rates are the means of three independent experiments. The columns without a common letter are different significantly (p < 0.05) using Tukey's test.

Principal Component Analysis and Summary Hierarchical Classification of the Ten Studied Lactobacillus Strains
The results of the PCA performed on the ten lactobacilli strains using 24 variables show that 81.14% of the total variation was distributed in four dimensions (Table S1), and the two most representative dimensions were Dim 1 (36.48%) and Dim 2 (25.66%) ( Figure 8A). Thus, it was found that the five variables that most contributed to the construction of this PCA, and consequently to the projection of individuals (lactobacilli) in Figure 8B, were anti-E. coli activity (spot method), an antiadhesive effect against L. monocytogeneson Caco-2 cells, DPPH scavenging activities of CFS, acidification (pH) and EPS quantification ( Figure 8A). Hierarchical principal component classification was performed based on the generated data in this study. Consequently, the lactobacilli were grouped into three main clusters ( Figure 8C). The first cluster is composed of six strains, which are Lb. plantarum F2, 2F8, M10, M12, F3 and FB13. The second is represented by three strains that are Lb. plantarum NCA3, NCA4 and FB3. The variables (Table S2) allowing the classification of these nine strains into two groups are the acidification (pH) and the antibacterial activity against S. aureus using the well diffusion method. Indeed, the strains in the first group are significantly the most acidifying (P < 0.05) with the greatest antibacterial activity (P < 0.05 against E. coli, S. aureus and L. monocytogenes (well-diffusion method) and against E. coli (spot method). Meanwhile, the second group of these

Principal Component Analysis and Summary Hierarchical Classification of the Ten Studied Lactobacillus Strains
The results of the PCA performed on the ten lactobacilli strains using 24 variables show that 81.14% of the total variation was distributed in four dimensions (Table S1), and the two most representative dimensions were Dim 1 (36.48%) and Dim 2 (25.66%) ( Figure 8A). Thus, it was found that the five variables that most contributed to the construction of this PCA, and consequently to the projection of individuals (lactobacilli) in Figure 8B, were anti-E. coli activity (spot method), an antiadhesive effect against L. monocytogenes on Caco-2 cells, DPPH scavenging activities of CFS, acidification (pH) and EPS quantification ( Figure 8A). Hierarchical principal component classification was performed based on the generated data in this study. Consequently, the lactobacilli were grouped into three main clusters ( Figure 8C). The first cluster is composed of six strains, which are Lb. plantarum F2, 2F8, M10, M12, F3 and FB13. The second is represented by three strains that are Lb. plantarum NCA3, NCA4 and FB3. The variables (Table S2) allowing the classification of these nine strains into two groups are the acidification (pH) and the antibacterial activity against S. aureus using the well diffusion method. Indeed, the strains in the first group are significantly the most acidifying (p < 0.05) with the greatest antibacterial activity (p < 0.05 against E. coli, S. aureus and L. monocytogenes (well-diffusion method) and against E. coli (spot method). Meanwhile, the second group of these strains is less acidifying and exhibited a weaker antagonistic activity compared to Lb. plantarum strains contained in the first group. However, the third cluster is represented by one strain, namely Lb. paracasei FB1, which is significantly (p < 0.05) less acidifying, and with lower EPS production and antibacterial activities against E. coli and S. aureus (using spot method) compared to the other two clusters. This third cluster was significantly (p < 0.05) related to (i) high DPPH scavenging activities with CFS and intracellular extract, (ii) co-aggregation with E. coli, S. aureus and L. monocytogenes, and (iii) removing S. aureus and L. monocytogenes biofilms. Remarkably, the strains in the first and the third cluster had the most significant (p < 0.05) antiadhesive activities against E. coli and L. monocytogenes on Caco-2 cells.
Foods 2020, 9, x FOR PEER REVIEW 13 of 20 strains is less acidifying and exhibited a weaker antagonistic activity compared to Lb. plantarum strains contained in the first group. However, the third cluster is represented by one strain, namely Lb. paracasei FB1, which is significantly (P < 0.05) less acidifying, and with lower EPS production and antibacterial activities against E. coli and S. aureus (using spot method) compared to the other two clusters. This third cluster was significantly (P ˂ 0.05) related to (i) high DPPH scavenging activities with CFS and intracellular extract, (ii) co-aggregation with E. coli, S. aureus and L. monocytogenes, and (iii) removing S. aureus and L. monocytogenes biofilms. Remarkably, the strains in the first and the third cluster had the most significant (P ˂ 0.05) antiadhesive activities against E. coli and L. monocytogenes on Caco-2 cells.

Discussion
The antagonism and antioxidant effects of Lactobacillus strains are key elements in the selection of new strains predicted to be used as bio-preservative agents or qualified as probiotic candidates. Increasing studies have tended to use principal component analysis to select the promising strains [46,47], allowing researchers to group strains with similar characteristics within the same cluster.

Discussion
The antagonism and antioxidant effects of Lactobacillus strains are key elements in the selection of new strains predicted to be used as bio-preservative agents or qualified as probiotic candidates. Increasing studies have tended to use principal component analysis to select the promising strains [46,47], allowing researchers to group strains with similar characteristics within the same cluster.
The inhibitory properties allocated to Lactobacillus strains are due to the production of different antimicrobial substances, such as organic acids that are usually used as food preservatives. Of note, acetic, citric and lactic acids were steadily reported and considered as the most studied organic acids. All of these organic acids are capable of controlling pathogens and stopping their proliferation. They can hamper their capability to form biofilms, or they can act on their quorum signaling pathways [48]. Here, ten novel Lactobacillus strains isolated from traditional Algerian fruits were assessed for their inhibitory properties against Gram-positive and Gram-negative pathogenic bacteria, which are representative of main foodborne pathogens. Consequently, the inhibitory properties of these Lactobacillus strains were shown to be associated mainly with the production of lactic acid, and exerted in a strain-dependent manner. Of note, these Lactobacillus strains of vegetal sources could produce up to 15 g/L of lactic acid, except for Lb. plantarum FB3 and Lb. paracasei FB1 strains, which produced less than 10.75 g/L. These data fit with those usually reported in the literature. Production of organic acids has several advantages, and was shown, in different studies, to prevent contamination of food during the fermentation process [49][50][51]. Regarding their mode of action, organic acids decrease the intracellular pH of the targeted microorganisms, causing the inhibition of cell growth. More precisely, the addition of organic acids causes proton accumulation in the cytoplasm, which exceeds the buffering levels of the cells and their ability to pump out protons through the H + -ATPase transport system [13].
Moreover, all Lactobacillus strains considered in this study were able to produce large quantities of EPS, compared to other typical lactobacilli isolated from diverse sources [52,53]. The EPS structure can act as a capsule bound to the cell surface, allowing protection against toxic agents and stress conditions found during desiccation, or osmotic stress and acidity conditions [54,55].
To gain insight on these strains, we established here the role of non-neutralized CFSs obtained from Lb. plantarum and Lb. paracasei cultures in preventing the initial attachment of pathogenic bacteria such as E. coli ATCC 8739, S. aureus 2S6 and L. monocytogenes 162 to a polystyrene surface. The addition of the non-neutralized CFS has a lasting effect on the biofilm formation by these foodborne pathogens. In direct line, Hossain et al. [56] reported that Lb. plantarum, Lb. curvatus and Lb. sakei isolated from kimchi impeded the biofilm formation by a clinical L. monocytogenes strain. Similarly, Mahdi et al. [43] showed the potential of a Lb. plantarum strain to prevent the biofilm formation by E. coli strains [52]. On the other hand, Cotar et al. [57] reported the role of organic acids, mainly that of lactic acid, produced by Lb. paracasei subsp. paracasei strain, in the inhibition of the biofilm formation by Pseudomonas aeruginosa. All these examples delineate collectively the attributes of Lactobacillus strains in controlling the biofilm formation by pathogenic bacteria, through the production of organic acids. Nevertheless, these studies failed in showing the anti-adhesive properties of the Lactobacillus strains used.
Furthermore, we examined here the morphology and the cell structure of pathogenic bacteria following their treatments with the non-neutralized CFS using TEM. We then noticed clear changes in the global morphology, even in the cell ultra-structure of the aforementioned target bacteria, when they were treated with the CFS Lb. plantarum 2F8 strain that was characterized as one of the upmost strains for lactic acid production. Moreover, S. aureus 2S6 and E. coli ATCC 8739 were morphologically altered, whereas L. monocytogenes 162 exhibited a disruption of its cell membrane integrity. Similar alterations were previously observed for human pathogenic bacteria including staphylococci, enterotoxigenic enterobacteria, Candida albicans, Acinetobacter baumannii and L. monocytogenes treated with different Lactobacillus species like Limosilactobacillus fermentum, Lactobacillus jensenii, Lacticaseibacillus rhamnosus and Lacticaseibacillus paracasei species [58][59][60].
The Lactobacillus strains used here prevented the adhesion of E. coli ATCC 8739, S. aureus 2S6 and L. monocytogenes 162 on eukaryotic Caco-2 cells, when used at an appropriate MOI ratio. In direct line with this, Yu et al. [61] reported the preventive effect of lactobacilli against enteric pathogenic bacteria using the same ratio of 10:1 (lactobacilli/pathogens). The anti-adhesive property of these Lactobacillus strains was dependent on many factors, including the conditions of incubation of Lactobacilli and pathogens during the competition and exclusion assays. Of note, the invasion of eukaryotic Caco-2 cells by L. monocytogenes, which is an intracellular pathogen, can be as well affected by these adherent Lactobacillus strains. The data gathered from the exclusion test are promising since they pointed out a clear protective effect against these pathogens, mainly by impeding their adhesion on the Caco-2 monolayer model, which strengthens our former report [34]. Of note, the abilities to hamper pathogens adhesion to the host tissues is as a key hurdle, and can therefore prevent their invasion process or limit some of their adverse effects, such as the destabilization of intercellular junctions [62].
In agreement with probiotic selection, Lactobacillus strains are known to prevent the adhesion of pathogenic bacteria to biotic surfaces by employing different mechanisms, other than the production of bioactive metabolites [63]. Among these mechanisms, the co-aggregation considered as a key strategy [39,64] is of major interest. Here, the co-aggregation rate varied widely from one Lactobacillus strain to another. Thus Lb. plantarum NCA4 and FB13 strains lack this function, while Lb. paracasei FB1 displayed a co-aggregation rate of over 20% with all target bacteria tested. In good agreement with this, Lactobacillus strains were shown to control through the co-aggregation process the microenvironment around the pathogenic cells [64]. Beganović et al. [65], Ferrando et al. [66] reported that surface proteins of Lactobacillus strains were involved in both auto-aggregation and co-aggregation with pathogens. Thus, co-aggregation, as above-indicated, is a strategic function which can be involved in competitive and exclusion mechanisms, leading thereof to a reduction in pathogenic load during infections [67].
Among the interesting probiotics aspects, some Lactobacillus strains have been reported to have antioxidant properties by decreasing the level of oxidants such as reactive oxygen species (ROS) [68]. The DPPH assay is generally used in vitro to determine the scavenging activity, and is one of the most sensitive, common and reliable methods [69]. In this study, the lactobacilli antioxidant capacity was focused on intact cells, cell-free supernatants and intracellular extracts. The important DPPH scavenging rates using cell-free supernatants are consistent with those reported by Sharma et al. [43]. Otherwise, the DPPH scavenging activity of intact cells was greater than the rates recorded for the intracellular extracts, as reported in previous studies [70,71]. The intracellular extract can have a significant antioxidant postbiotic effect [72,73]. Many studies have revealed that antioxidant activity of Lactobacillus strains might be linked to their production of cell-surface compounds, e.g., lipoteichoic acid and EPS, and to antioxidant enzymes, such as superoxide dismutase, NADH-oxidase and NADH-peroxide, and heterologous non-haem catalase [74,75].

Conclusions
Here, we characterized the antibacterial, anti-adhesive and antioxidant properties of ten Lactobacillus strains from vegetal sources, and confirmed their potential to produce lactic acid and EPS in a strain-dependent manner. In addition, we revealed, using TEM, accurate ultra-structure damages caused by lactic acid produced by these strains. These Lactobacillus strains exhibited capabilities to adhere on human Caco-2 cells, as well as on inanimate surfaces; this mechanism stands as a hurdle to stop the proliferation of pathogens such as E. coli, S. aureus and L. monocytogenes. Collectively, the results from this study open up new possibilities for application of these new strains as potential probiotics or bioprotective cultures. For the probiotic application, they can be used, after further assessments, to protect human or animal GIT microbiota from malevolent pathogens and oxidative stress. For their application as bioprotective cultures, they can indeed be employed as a hurdle mean, mainly in processed foods from vegetal origin to replace chemicals and ascertain their safety. For both applications, further in vivo and in situ evidence is needed, and this will be our next focus.
Supplementary Materials: The following are available online at http://www.mdpi.com/2304-8158/9/8/985/s1. Table S1: Correlation, contribution and representation of the variables with the dimensions of the summary PCA based on factor loadings. Table S2