In Vitro Study of Probiotic, Antioxidant and Anti-Inflammatory Activities among Indigenous Saccharomyces cerevisiae Strains

Nowadays, the interest toward products containing probiotics is growing due to their potential health benefits to the host and the research is focusing on search of new probiotic microorganisms. The present work was focused on the characterization of indigenous Saccharomyces cerevisiae strains, isolated from different food matrixes, with the goal to select strains with probiotic or health-beneficial potential. A preliminary screening performed on fifty S. cerevisiae indigenous strains, in comparison to a commercial probiotic strain, allowed to individuate the most suitable ones for potential probiotic aptitude. Fourteen selected strains were tested for survival ability in the gastrointestinal tract and finally, the strains characterized for the most important probiotic features were analyzed for health-beneficial traits, such as the content of glucan, antioxidant and potential anti-inflammatory activities. Three strains, 4LBI-3, LL-1, TA4-10, showing better attributes compared to the commercial probiotic S. cerevisiae var. boulardii strain, were characterized by interesting health-beneficial traits, such as high content of glucan, high antioxidant and potential anti-inflammatory activities. Our results suggest that some of the tested S. cerevisiae strains have potential as probiotics and candidate for different applications, such as dietary supplements, and starter for the production of functional foods or as probiotic to be used therapeutically.


Introduction
Probiotics are defined by the World Health Organization (WHO) and the Food and Agriculture Organization of the United Nations (FAO) as "live microorganisms which, when administered in adequate amounts, confer a health benefit on the host" [1]. The term has been also applied to the supplements that are added to animal feed in order to improve the health of the animal for the beneficial effects on their gut microbiota.
In the last decades, the use and the request of the commercial probiotics are growing in consequence of the worldwide interest to their positive impact on the human health. In fact, many clinical data confirm therapeutic effects of the probiotics in the treatment of gastrointestinal diseases (irritable bowel syndrome, gastrointestinal disorders, elimination of Helicobacter pylori, diarrhoeas) [2] and prophylaxis of respiratory and urogenital infections [3]. Probiotics play a significant role towards allergic reactions as atopic dermatitis, and in the regulation of common metabolic diseases (obesity, insulin resistance syndrome, type 2 diabetes) [4,5]. Scientific studies also report their anti-cancer activity and the enhance of the body's immune response (immunomodulation) by the activation of specific genes of isolated from different food matrixes, with the goal to select strains with probiotic or health-beneficial potential. At this purpose, a preliminary screening was performed on fifty S. cerevisiae indigenous strains, isolated from different food matrixes, in order to evaluate their probiotic aptitude as tolerance to gastrointestinal conditions (pH, temperature), antibiotic resistance and hydrophobicity. Fourteen selected strains were tested for their behaviour after simulated transit into gut and only three strains (4LBI-3, TA4-10, LL-1) in comparison to the probiotic commercial strain Sb, were analyzed to evaluate functional activities, such as the content of glucans, antioxidant and anti-inflammatory activities.

Origin of Yeast Strains
In this work, fifty indigenous S. cerevisiae strains (non-commercial strains), belonging to UNIBAS Yeast Collection (UBYC), University of Basilicata (Potenza, Italy) and previously isolated from food environments, were studied (Table 1). Furthermore, two commercial strains, EC1118 (Lallemand Inc., Toulose, France), a winemaking starter, and a probiotic S. cerevisiae var. boulardii strain (coded as Sb) and isolated from the commercial product Codex (Zambon, Italy), were used. The strains were maintained at −80 • C in a glycerol solution (Sigma, Milan, Italy); before each experiment, yeast strains were grown in YPD (yeast extract 1%, bacteriological peptone 2%, glucose 2%, agar 2%, Oxoid, Hampshire, UK) at 26 ± 2 • C for 24 h. Table 1. Origin of the fifty indigenous Saccharomyces cerevisiae strains tested in this study.

Growth at Human Body Temperature
This test was performed in YPD-agar medium; each strain was inoculated as spot with 5 µL of cell suspension with a concentration of around 10 6 cells/mL, measured at 600 nm by spectrophotometer (SPECTROstar Nano, BMG Labtech, Ortenberg, Germany). Plates were aerobically incubated at 37 ± 2 • C and 26 ± 2 • C (as positive control) for 48 h. Cell growth level was evaluated by comparing the size and thickness of each strain spot. All the experiments were performed in duplicate. This experiment was performed by inoculating sterile tubes containing 10 mL of YPD broth, adjusted to different pH (2.5, 3.0, 3.5, 4.0, and 7.2) with approximately 4.8 × 10 6 cells/mL of each strain and incubated at 37 ± 2 • C. Positive controls were obtained by inoculating the different yeasts in non-modified pH YPD-broth (pH = 6.55). Yeast growth was evaluated after 24 of incubation as absorbance at 600 nm. The analyses were performed in duplicate and data were elaborated as Growth Index (GI) following Bevilacqua et al. [51] GI = (Abss/Absc) × 100% where Abss is the absorbance of the samples at different pH, whereas Absc is the absorbance of the positive control. GI values were classified as follows: • GI < 50%: Yeast Inhibition • 50% < GI < 75%: Partial Inhibition • GI > 75%: Growth Similar to the Control

Hydrophobicity of the Cells Surface
Hydrophobicity was determined as the ability of the yeast strains to adhere to hydrocarbons and it was expressed as Microbial Adhesion to Solvents (MATS), following the protocol reported by Chelliah et al. [52], with some modifications. Yeast cultures in stationary phase (inoculated in YPD broth for 48 h) were concentrated by centrifugation at 5000 rpm for 10 min, washed twice with PBS (Potassium phosphate buffer 0.1 M, pH 7.4, Merck, Darmstadt, Germany) and suspended again in 3 mL of the same buffered solution. The OD at 600 nm of the suspension was adjusted with PBS (pH 7.4) to a value of 1.0 (A0). 0.2 mL of toluene (Merck, Darmstadt, Germany) was added to 1 mL of this suspension (1 × 10 7 cell/mL) and mixed for 2 min. After 1 h of incubation at 30 ± 2 • C, the aqueous phase was removed from the upper solvent layer containing the cells and the OD was determined again at 600 nm (A1). The MATS was calculated as:

Antibiotic Resistance
Antibiotic-resistance was carried out with the Kirbye Bauer method according to the NCCLS protocol [53]. Five different antibiotics were used: streptomycin (ST), chloramphenicol (CH), gentamicin (GE), erythromycin (ER), and oxytetracycline (OX) (Sigma, Milan, Italy). Tests were performed on YPD agar added with five different concentrations of each antibiotic: 40, 80, 100, 130 and 150 µg/mL, in comparison to the control without antibiotic addition. Cell suspensions were adjusted to 0.2 of Optical Density (OD) (10 6 cells/mL), and 8 µL of each strain were dropped on the antibiotic-YPD plates, which were incubated at 37 ± 2 • C for 48 h. Cell survival and growth level on the different concentrations of the five antibiotics were qualitatively evaluated, by comparing the size and thickness of each strain spot with the corresponding control. All the experiments were performed in duplicate.

Screening for Potentially Probiotic Traits: Simulation of the Transit into Gut
The experiment was performed according the method of Picot and Lacroix [54], with some modifications. Yeasts were grown in YPD broth for 48 h, then cells were washed twice and suspended into sterile distilled water. For this analysis, in order to simulate the sequential digestion, two different solutions were prepared: Gastric Juice: 0.26 g/L of pepsin (porcine gastric mucosal, Sigma, Milan, Italy) were dissolved in HCl 0.1 N and the final pH of the solution was adjusted to 1.9 with 1 N NaOH; Intestinal fluid: 1.95 g/L of pancreatin (porcine pancreas, Sigma, Milan, Italy) were dissolved in a phosphate buffer solution (NaH 2 PO 4 /Na 2 HPO 4 0.02 M, pH 7.5), Foods 2022, 11, 1342 5 of 18 and added with 3 g/L of bile extract (bile extract porcine, Sigma, Milan, Italy), dissolved in sterile distilled water; and the final pH was adjusted to 7.5 with 1N NaOH.
Each strain, at a concentration of about 2.0 × 10 7 cells/mL, was inoculated in 50 mL of sterile distilled water for 15 min at 4 • C. Then, the pH was adjusted to 1.9 with HCl 1 N and the volume was diluted to 60 mL with distilled water; subsequently, 20 mL of the pepsin solution were added and the samples were incubated at 37 ± 2 • C for 3 h (simulation of the stomach transit). Every 30 min, an aliquot of sample (2 mL) was picked up and kept in ice before the inoculation on the plates. After 3 h, the reaction was stopped by increasing the pH to 7.5 with 1 N NaOH, and added with 5 mL of a concentrated sodium phosphate buffer and 2 mL of the bile salt. The pH was adjusted to 7.5 with 1 N NaOH and distilled water was added to obtain a final volume of 90 mL; after that, 10 mL of the pancreatic solution were added and incubated at 37 ± 2 • C for 24 h to simulate the intestinal transit. Every 30 min, a sample of 2 mL was picked up and kept in ice before the spreading on the plates. All the experiments were performed in duplicate. Data were elaborated as Viability Index (VI): where Nt is yeast cell concentration for each time and N0 the initial inoculum.

Evaluation of Functional Activities on Four Selected Strains
Three indigenous strains, selected on the basis of previous screening (TA4-10, LL-1, 4LBI-3), in comparison to the probiotic strain Sb, were tested for evaluation of further biological activities, such as the content of glucans, antioxidant activity and cytotoxicity assay.

Total Antioxidant Activity
The total antioxidant activity, evaluated on the basis of the scavenging activity of the 1,1-diphenyl-2-picrylhydrazyl (DPPH, Sigma, Milan, Italy) free radical, was determined by the protocol described by Chen et al. [14] and Capece et al. [49], with some variations. Briefly, the cells pellets, recovered by centrifugation (1200 rpm, 5 min) of 1 mL of yeast culture, grown on YPD broth for 24 h, was added with 1 mL of methanolic DPPH radical solution (0.2 mM in methanol, Sigma, Milan, Italy). The mix was vortexes and then incubated for 30 min at room temperature in darkness. The reaction tubes were centrifuged (1200 rpm, 5 min) and the absorbance at 517 nm was measured (in triplicate) on 300 µL of the supernatant. The blank used was pure methanol. The scavenging ability was defined as follows: Scavenging rate (%) = [1 − (A sample/A blank)] × 100%

Determination of Glucans
Total glucans and β-glucans contents were analyzed by using a commercially available kit (Megazyme, Wicklow, Ireland), an enzymatic method based on acidic hydrolysis of the β-glucans, followed by enzymatic digestion to glucose [55]. For determination of total glucans, the yeast cells were aerobically grown in YPD broth at 30 • C for 24 h in order to reach the exponential phase of cultures. The yeast cultures were stopped at 2 × 10 7 cells/mL and harvested by centrifugation at 4700 rpm for 10 min. Yeast pellets (about 100 mg) were suspended in 2.0 mL of 12 M HCl, stirred at 30 • C for 1 h to hydrolyse total glucans, following the manufacture's instruction. In details, each tube was added with 10 mL of water and placed in a boiling water bath for 2 h, after that the tubes were cooled at room temperature and added with 200 mM sodium acetate buffer (pH = 5). An aliquot of the sample was centrifuged at 1500 rpm for 10 min and 0.1 mL of extracts was added with 0.1 mL of a mixture of exo-1,3-β-glucanase (20 U/mL) and β-glucosidase (4 U/mL) in 200 mM sodium acetate buffer. The mixture was incubated at 40 • C for 60 min, after that each tube was added with 3.0 mL of glucose oxidase/peroxidase (GOPOD) reagent and incubated at 40 • C for 20 min. The absorbance of all samples was measured at 510 nm against the blank. The second step of enzymatic kits analysis was addressed to the determination of Alfa-Glucans (α-Glucans). In this step, 100 mg of yeast pellet were mixed with 2 mL of 2 M KOH; after stirring in ice-water, the samples were added with 8 mL of 1.2 M sodium acetate buffer (pH 3.8), 0.2 mL of amyloglucosidase (1630 U/mL) plus invertase (500 U/mL) and the mixture was incubated in a water bath at 40 • C for 30 min. After this step, the samples were centrifuged at 1500 rpm for 10 min and 0.1 mL of supernatants was added with 0.1 mL of sodium acetate buffer (200 mM, pH 5.0) plus 3.0 mL of GOPOD reagent and incubated at 40 • C for 20 min. In the last step, the absorbance of all samples was measured at 510 nm against the reagent blank. The β-Glucan content was calculated as difference between Total Glucans and α-Glucans by using the Megazyme Mega-CalcTM software.2.4.3. (Megazyme, Wicklow, Ireland) Cytotoxicity Assay and Anti-inflammatory Activity.

Cytotoxicity Assay
The effects of TA4-10, LL-1, Sb, 4LBI-3 on U937/PMA cell proliferation were determined using a Millipore Scepter™ handheld automated cell counter (Merck Millipore, Darmstadt, Germany) 72 h after incubation, according to the manufacturer's instructions [58]. Briefly, cells were cultured into 96-well plates (5 × 10 4 cells/well) in RPMI (CTRL) or in RPMI where the four strains were grown for 24 h. After 72 h, U937/PMA cells were collected in 1.5 mL microfuge tubes and counted.

ROS and NO Detection
To evaluate reactive oxygen species (ROS) and nitric oxide (NO) levels, U937/PMA cells were triggered by 200 ng/mL of LPS in the presence or absence of RPMI where the four strains were grown for 24 h. After 24 h, ROS and NO concentrations were measured by using 6-carboxy-2 ,7 -dichlorodihydrofluorescein diacetate (DCF-DA, Thermo Fisher Scientific, San Jose, CA, USA) and -amino-5-methylamino-2 ,7 -difluorofluorescein diacetate (DAF-FM diacetate, Thermo Fisher Scientific, San Jose, CA, USA), respectively, as previously described [59]. Briefly, cells were collected in tubes and counted. After centrifugation, the cell pellet was resuspended in PBS to have 10 5 cells/100 µL. Then, 10 µM DCF-DA or 2.24 µM DAF-FM diacetate were added. After 30 min of incubation in the dark at 37 • C, a 100 µL portion of sample was transferred to triplicate groups of wells on a black microtiter plate and the fluorescence was revealed using a GloMax plate reader (Promega, Madison, WI, USA).

Data Analysis
Data regarding the strain tolerance to different pH and the hydrophobicity of the cell surface were analyzed by stacked chart. Statistical significance of differences between the strains for the simulation of the transit into gut, was determined using one-way analysis of variance (ANOVA) and the homogeneity of variance was verified applying the Levene's test. Differences were considered as significant with a p value ≤ 0.05. Results regarding cytotoxicity assay and ROS and NO detection anti-inflammatory activity were shown as means ± SD of, at least, three independent experiments and were analyzed by one-way analysis of variance (ANOVA). Differences were considered as significant (p < 0.05), very significant (p < 0.01), and highly significant (p < 0.001).

Preliminary Characterization
In the preliminary phase, fifty indigenous S. cerevisiae strains, and two commercial strains, EC1118 and Sb (as positive control strain) were characterized by different tests, such as growth at human body temperature, tolerance to different pH levels, hydrophobicity of the cell surface and antibiotic resistance. - Growth at human body temperature: The temperature is an important physical parameter to evaluate the probiotic aptitude of the microorganism. In this study, the experiments were carried out at 37 • C, the typical temperature of the human body, and all the strains were able to grow at 37 • C (data not shown). These results agree with previous findings, reporting the ability of yeast strains to grow at human body temperature [22], although yeasts usually show a trend toward surviving to lower temperatures than lactic acid bacteria [61], and confirm that indigenous yeast strains might be an interesting source for the selection of new potential probiotic strains. -Strain tolerance to different pH: As regards the tolerance to pH, different pH values (2.5, 3.0, 3.5, 4.0, and 7.2) were investigated in order to evaluate the ability of tested yeasts to survive at the pH of the gastro-intestinal environment. In fact, the probiotics must survive harsh conditions during their passage through the intestinal tract [11]. All the strains exhibited a growth similar to the control (GI > 75%) to pH 3.5, 4.2, and 7.2, whereas the lowest pH values (2.5 and 3.0) influenced the yeast growth. In particular, at pH 2.5, 15 strains were inhibited (GI < 50%), 31 strains (59.61%) showed a partial inhibition (50% < GI < 75%), whereas only 6 strains exhibited a good tolerance to low pH (GI > 75%). At pH 3.0, 13 strains exhibited a partial inhibition (50% < GI < 75%) and 39 strains (75%) grew similar to the control (GI > 75%) ( Figure 1). The highest tolerance to low pH, with a GI > 90%, were found in five strains, namely FD-3, M5-15 (both isolated from sourdough), TA4-10, CA10-4sc2 (both isolated from grape must) and 9-15G, isolated from honey. Furthermore, the majority of indigenous strains analyzed in this step was more tolerant to low pH level than Sb, the commercial probiotic strain. Resistance to acidic conditions was not surprising since the strains studied in this work were isolated from low pH environments, such as grape must, or substrate in which they coexist with lactic acid bacteria, such as sourdough [62]. -Hydrophobicity of the Cell Surface: Hydrophobicity was defined as an interaction between microbial and host cells, mediated by cell-surface proteins and lipoteichoic acids [63]. As a consequence, this is one of the main criteria for the selection of strains with potential probiotic activity, because the microbial strain included in the probiotic category must have the ability to adhere to the intestinal mucosa for colonization and modulation of the immune system against pathogens [64]. The hydrophobic ability of the yeasts was reported in Figure 1. The results highlighted that only two strains (BP2-33, and SA10-19) showed 0% MATS (no hydrophobic activity), while the majority of the strains (35) exhibited a MATS in a range between 50% to 75%, which were much higher than the level found in S. cerevisiae var. boulardii strain used as control, confirming other results reporting better hydrophobicity level in indigenous strains than in control strains [65,66]. -Hydrophobicity of the Cell Surface: Hydrophobicity was defined as an interaction between microbial and host cells, mediated by cell-surface proteins and lipoteichoic acids [63]. As a consequence, this is one of the main criteria for the selection of strains with potential probiotic activity, because the microbial strain included in the probiotic category must have the ability to adhere to the intestinal mucosa for colonization and modulation of the immune system against pathogens [64]. The hydrophobic ability of the yeasts was reported in Figure 1. The results highlighted that only two strains (BP2-33, and SA10-19) showed 0% MATS (no hydrophobic activity), while the majority of the strains (35) exhibited a MATS in a range between 50% to 75%, which were much higher than the level found in S. cerevisiae var. boulardii strain used as control, confirming other results reporting better hydrophobicity level in indigenous strains than in control strains [65,66].
-Antibiotic Resistance: The fifty-two strains were tested for the resistance to five antibiotics, at five different doses (40, 80, 100, 130, and 150 µg/mL) and all the strains survived in the presence of the highest dose of the tested antibiotics (150 µg/mL); in fact, no inhibition zone was observed around the spotted strains (data not shown). The high resistance of the S. cerevisiae strains to these antibiotics provide them advantage over bacteria in patients undergoing antibiotic treatment for therapeutic use [67]. Furthermore, this characteristic is of great interest also for potential application of these strains as probiotic sources in animal feed in consequence of potential presence of antibiotics residues in animal feed.
On the basis of the results obtained by preliminary characterization, 14 strains, exhibiting the best combination of the tested parameters, were chosen for further analysis; the characteristics of the selected strains are summarized in Table 2. Four strains (4LBI-3, LL-1, P4 and TA4-10) showed the highest percentage of MATS (>75%).
-Antibiotic Resistance: The fifty-two strains were tested for the resistance to five antibiotics, at five different doses (40,80,100,130, and 150 µg/mL) and all the strains survived in the presence of the highest dose of the tested antibiotics (150 µg/mL); in fact, no inhibition zone was observed around the spotted strains (data not shown). The high resistance of the S. cerevisiae strains to these antibiotics provide them advantage over bacteria in patients undergoing antibiotic treatment for therapeutic use [67]. Furthermore, this characteristic is of great interest also for potential application of these strains as probiotic sources in animal feed in consequence of potential presence of antibiotics residues in animal feed.
On the basis of the results obtained by preliminary characterization, 14 strains, exhibiting the best combination of the tested parameters, were chosen for further analysis; the characteristics of the selected strains are summarized in Table 2.

Screening for Potentially Probiotic Traits: Simulation of the Transit into Gut
In the second phase of the study, the 14 selected strains were tested for their potential probiotic activity. In this step, the 14 selected strains were tested for their survival level to the simulated transit into gut, by evaluating the tolerance level to gastric and pancreatic juices, which are important criteria in the selection of probiotic microorganisms, allowing them to survive during the transit across the human digestive tract. The resistance to gastric juice was evaluated by calculating the Viability Index (VI, %) in Gastric Juice (GJ), simulating of the stomach conditions, after 0.5 h and 2 h of incubation ( Figure 2). After 30 min, the SC5-25 strain exhibited the lowest VI (87.35% ± 0.92), whereas the highest VI (100%) was found in four strains (CA10-4sc2, EC1118, M4 and TA4-10). After 2 h, the exposure time to stomach acidic conditions recommended by different authors [22,68], the surviving level of the strains remained almost unchanged for the majority of the strains, whereas a slight increase of VI was observed for some of them, such as M5-15, Sb, SC5-25.

Screening for Potentially Probiotic Traits: Simulation of the Transit into Gut
In the second phase of the study, the 14 selected strains were tested for their potentia probiotic activity. In this step, the 14 selected strains were tested for their survival level to the simulated transit into gut, by evaluating the tolerance level to gastric and pancreati juices, which are important criteria in the selection of probiotic microorganisms, allowing them to survive during the transit across the human digestive tract.
The resistance to gastric juice was evaluated by calculating the Viability Index (VI %) in Gastric Juice (GJ), simulating of the stomach conditions, after 0.5 h and 2 h of incu bation ( Figure 2). After 30 min, the SC5-25 strain exhibited the lowest VI (87.35% ± 0.92) whereas the highest VI (100%) was found in four strains (CA10-4sc2, EC1118, M4 and TA4-10). After 2 h, the exposure time to stomach acidic conditions recommended by dif ferent authors [22,68], the surviving level of the strains remained almost unchanged fo the majority of the strains, whereas a slight increase of VI was observed for some of them such as M5-15, Sb, SC5-25.  As regards the simulation of the intestinal transit, high variability among the analyzed yeasts was found for strain survival level to the simulated intestinal fluid (Figure 3). After 3 h, all the strains showed a VI > 50%, except the control strain Sb (45.36% ± 3.25) and CD2-6sc2 (47.89% ± 3.25). In particular, five strains exhibited a VI > 80% [EC1118 (84.80% ± 7.15), M5-15 (86.56% ± 7.55), P4 (89.05% ± 7.36) and TA4-10 (84.90% ± 7.63)] and LL-1 strain showed the highest VI (93.77% ± 8.06). After 24 h, a certain viability loss was observed for all the strains, except for EC1118 (84.80 to 98.30% ± 8.88), LL-1 (93.77 to 99.98% ± 8.95), M3-59 (from 77.61% ± 6.54 to 82.35% ± 7.14) and M4 (from 75.84% ± 6.39 to 85.76% ± 7.39). Two strains exhibited a viability loss of about half between 3 h and 24 h incubation, which were BA-215 (69.59% ± 5.43 to 38.84% ± 2.54) and P4 (89.05% ± 7.36 to 30.00% ± 2.94). These results showed the existence of high differences among the analyzed strains, confirming results previously reported by other authors that probiotic capacities are strain dependent [65,69,70]. The indigenous strains, characterized by higher resistance to digestive enzymes, such as pepsin, pancreatin as well as bile salts, than the commercial probiotic strain, demonstrating to be able to survive GI transit, can represent new microorganisms possibly affecting the microbial balance in the human body. concentration for each time and the initial inoculum and are reported as mean value ± SD of tw independent experiments.
As regards the simulation of the intestinal transit, high variability among the ana lyzed yeasts was found for strain survival level to the simulated intestinal fluid (Figure 3 After 3 h, all the strains showed a VI > 50%, except the control strain Sb (45.36% ± 3.25 and CD2-6sc2 (47.89% ± 3.25). In particular, five strains exhibited a VI > 80% [EC111 (84.80% ± 7.15), M5-15 (86.56% ± 7.55), P4 (89.05% ± 7.36) and TA4-10 (84.90% ± 7.63)] and LL-1 strain showed the highest VI (93.77% ± 8.06). After 24 h, a certain viability loss wa observed for all the strains, except for EC1118 (84.80 to 98.30% ± 8.88), LL-1 (93.77 t 99.98% ± 8.95), M3-59 (from 77.61% ± 6.54 to 82.35% ± 7.14) and M4 (from 75.84% ± 6.39 to 85.76% ± 7.39). Two strains exhibited a viability loss of about half between 3 h and 24 h incubation, which were BA-215 (69.59% ± 5.43 to 38.84% ± 2.54) and P4 (89.05% ± 7.36 t 30.00% ± 2.94). These results showed the existence of high differences among the analyzed strains, confirming results previously reported by other authors that probiotic capacitie are strain dependent [65,69,70]. The indigenous strains, characterized by higher resistanc to digestive enzymes, such as pepsin, pancreatin as well as bile salts, than the commercia probiotic strain, demonstrating to be able to survive GI transit, can represent new micro organisms possibly affecting the microbial balance in the human body. All the data regarding the screening of the 14 S. cerevisiae strains for potentially pro biotic traits were statistically analyzed and used for the matrix plot reported in Figure 4. Thre indigenous strains (4LBI-3, LL-1, TA4-10) were chosen on the basis of high tolerance t bile salts (VI > 75% after 3 h) and high survival level (VI > 95%) after the simulation of th transit into gut, as already reported (Figure 2 and Table 2). Commercial probiotic Sb wa also used in the further characterization as reference strain, although this strain showed the lowest tolerance to bile salts, mainly after 24 h of incubation. All the data regarding the screening of the 14 S. cerevisiae strains for potentially probiotic traits were statistically analyzed and used for the matrix plot reported in Figure 4. Three indigenous strains (4LBI-3, LL-1, TA4-10) were chosen on the basis of high tolerance to bile salts (VI > 75% after 3 h) and high survival level (VI > 95%) after the simulation of the transit into gut, as already reported (Figure 2 and Table 2). Commercial probiotic Sb was also used in the further characterization as reference strain, although this strain showed the lowest tolerance to bile salts, mainly after 24 h of incubation.

Antioxidant Activity and Glucans Content
In the final step of this study, four strains were screened for functional characteristics other than probiotic traits, such as the antioxidant capacity. Several studies have reported the natural antioxidant capacity of yeast cells [14,62,71]. In our study, the in vitro antioxidant activities of the four chosen strain cells, measured as the scavenging effects on DPPH radicals, is reported in Figure 5. All of them showed antioxidant capacity, but at different level; the highest scavenging activity was exhibited by 4LBI-3 cells (49.08% ± 3.05), followed by TA4-10 cells; both the strains showed significantly higher antioxidant activity than the control strain. The strain LL-1 showed the lowest level (36.12% ± 2.47). However, all the four yeast strains showed antioxidant activity, at level comparable with the results obtained in other studies addressed to evaluate the health-beneficial potential of yeast strains isolated from food [22,70]. Although different mechanisms may be involved in antioxidant activity of yeast cells, it is believed that this functional activity is correlated to the content of (1/3)-β-D-glucan and other β-glucans found in the cell wall [72,73], other than cellular compounds, such as some antioxidant enzymes like superoxide dismutase, glutathione peroxidase and catalase [14]. In fact, the β-glucans were described as molecules with different biological activities that can include anti-inflammatory, antioxidative (in terms of free radical scavenging) and immunomodulation. Until now, few studies reporting the effect of source (or strain) of yeast as important factors affecting the β-glucan content are available [74,75]. The amounts of β-glucans extracted from the four selected strains is highly variable among them ( Figure 5), demonstrating that the β-glucan content of a yeast is strain/species dependent [76,77]. Two indigenous strains showed a content significantly higher or similar (TA4-10 and 4LBI-3, respectively) than the content found in the commercial probiotic strain Sb, while LL-1 strain exhibited the lowest amounts of βglucans (7.83 ± 0.68% w/w). Our results confirm the existence of a certain correlation

Antioxidant Activity and Glucans Content
In the final step of this study, four strains were screened for functional characteristics other than probiotic traits, such as the antioxidant capacity. Several studies have reported the natural antioxidant capacity of yeast cells [14,62,71]. In our study, the in vitro antioxidant activities of the four chosen strain cells, measured as the scavenging effects on DPPH radicals, is reported in Figure 5. All of them showed antioxidant capacity, but at different level; the highest scavenging activity was exhibited by 4LBI-3 cells (49.08% ± 3.05), followed by TA4-10 cells; both the strains showed significantly higher antioxidant activity than the control strain. The strain LL-1 showed the lowest level (36.12% ± 2.47). However, all the four yeast strains showed antioxidant activity, at level comparable with the results obtained in other studies addressed to evaluate the health-beneficial potential of yeast strains isolated from food [22,70]. Although different mechanisms may be involved in antioxidant activity of yeast cells, it is believed that this functional activity is correlated to the content of (1/3)-β-D-glucan and other β-glucans found in the cell wall [72,73], other than cellular compounds, such as some antioxidant enzymes like superoxide dismutase, glutathione peroxidase and catalase [14]. In fact, the β-glucans were described as molecules with different biological activities that can include anti-inflammatory, antioxidative (in terms of free radical scavenging) and immunomodulation. Until now, few studies reporting the effect of source (or strain) of yeast as important factors affecting the β-glucan content are available [74,75]. The amounts of β-glucans extracted from the four selected strains is highly variable among them ( Figure 5), demonstrating that the β-glucan content of a yeast is strain/species dependent [76,77]. Two indigenous strains showed a content significantly higher or similar (TA4-10 and 4LBI-3, respectively) than the content found in the commercial probiotic strain Sb, while LL-1 strain exhibited the lowest amounts of β-glucans (7.83 ± 0.68% w/w). Our results confirm the existence of a certain correlation between β-glucans content and antioxidant activity, although other component might influence the antioxidant activity of yeast cells as the strain exhibiting the highest scavenging ability on DPHH (4LBI-3) was not the same characterized by the highest content of β-glucans (TA4-10), as previously reported by other authors [14]. between β-glucans content and antioxidant activity, although other component might influence the antioxidant activity of yeast cells as the strain exhibiting the highest scavenging ability on DPHH (4LBI-3) was not the same characterized by the highest content of βglucans (TA4-10), as previously reported by other authors [14].

Cytotoxicity Assay and Anti-Inflammatory Activity
First, we evaluated the effect of TA4-10, LL1, 4LBI-3 selected yeast strains and Sb control strain on U937/PMA cell proliferation by cell counting. Therefore, U937/PMA macrophage cells were cultured into 96-well plates in RPMI alone (CTRL) or in RPMI where the four selected strains were previously grown for 24 h. The determination of U937/PMA macrophage cell line after 72 h revealed that no strain inhibited U937/PMA cell proliferation ( Figure 6). These results suggest that the three indigenous yeasts, similarly to the commercial probiotic strain, could be considered as non-toxic strains in consequence of their very low cytotoxicity.

Cytotoxicity Assay and Anti-Inflammatory Activity
First, we evaluated the effect of TA4-10, LL1, 4LBI-3 selected yeast strains and Sb control strain on U937/PMA cell proliferation by cell counting. Therefore, U937/PMA macrophage cells were cultured into 96-well plates in RPMI alone (CTRL) or in RPMI where the four selected strains were previously grown for 24 h. The determination of U937/PMA macrophage cell line after 72 h revealed that no strain inhibited U937/PMA cell proliferation ( Figure 6). These results suggest that the three indigenous yeasts, similarly to the commercial probiotic strain, could be considered as non-toxic strains in consequence of their very low cytotoxicity. between β-glucans content and antioxidant activity, although other component might influence the antioxidant activity of yeast cells as the strain exhibiting the highest scavenging ability on DPHH (4LBI-3) was not the same characterized by the highest content of βglucans (TA4-10), as previously reported by other authors [14].

Cytotoxicity Assay and Anti-Inflammatory Activity
First, we evaluated the effect of TA4-10, LL1, 4LBI-3 selected yeast strains and Sb control strain on U937/PMA cell proliferation by cell counting. Therefore, U937/PMA macrophage cells were cultured into 96-well plates in RPMI alone (CTRL) or in RPMI where the four selected strains were previously grown for 24 h. The determination of U937/PMA macrophage cell line after 72 h revealed that no strain inhibited U937/PMA cell proliferation ( Figure 6). These results suggest that the three indigenous yeasts, similarly to the commercial probiotic strain, could be considered as non-toxic strains in consequence of their very low cytotoxicity.

ROS and NO Detection
The next step was addressed to evaluate the potential anti-inflammatory activity of TA4-10, LL-1, 4LBI-3 and Sb strains. To this end, a gram-negative bacterial lipopolysaccharide (LPS), a major component of the outer membrane was used as inducer in U937/PMA macrophage cell line. The levels of two mediators closely linked to the inflammatory response, i.e., reactive oxygen species (ROS), specifically superoxide anion radical, and nitric oxide (NO), were assessed after LPS treatment in the presence or absence of RPMI, where the four strains were previously grown for 24 h. As shown in Figure 7A, the strains that significantly reduced ROS levels are, in order of intensity, LL-1, 4-LBI-3 and TA4-10. Conversely, the reference strain Sb does not seem able to counteract superoxide anion production in LPS-induced U937/PMA macrophage cell line ( Figure 7A). Subsequently, the effect of yeast strains on NO production was tested in U937/PMA macrophage cell line triggered by LPS. Nitric oxide, a key biosignaling molecule produced in both peripheral tissues and the central nervous system, is synthesized from L-arginine to stoichiometric quantities of NO and L-citrulline using molecular oxygen and NADPH as cofactors by a family of enzymes known as nitric oxide synthases (NOSs). In LPS-induced U937/PMA macrophage cell line, TA4-10 and 4LBI-3 significantly reduced NO levels compared to LPS-activated macrophages by 25 and 14%, respectively ( Figure 7B). Also in this trial, the strain less active in inhibiting NO production was the control strain Sb.

ROS and NO Detection
The next step was addressed to evaluate the potential anti-inflammatory activity of TA4-10, LL-1, 4LBI-3 and Sb strains. To this end, a gram-negative bacterial lipopolysaccharide (LPS), a major component of the outer membrane was used as inducer in U937/PMA macrophage cell line. The levels of two mediators closely linked to the inflammatory response, i.e., reactive oxygen species (ROS), specifically superoxide anion radical, and nitric oxide (NO), were assessed after LPS treatment in the presence or absence of RPMI, where the four strains were previously grown for 24 h. As shown in Figure 7A, the strains that significantly reduced ROS levels are, in order of intensity, LL-1, 4-LBI-3 and TA4-10. Conversely, the reference strain Sb does not seem able to counteract superoxide anion production in LPS-induced U937/PMA macrophage cell line ( Figure 7A). Subsequently, the effect of yeast strains on NO production was tested in U937/PMA macrophage cell line triggered by LPS. Nitric oxide, a key biosignaling molecule produced in both peripheral tissues and the central nervous system, is synthesized from L-arginine to stoichiometric quantities of NO and L-citrulline using molecular oxygen and NADPH as cofactors by a family of enzymes known as nitric oxide synthases (NOSs). In LPS-induced U937/PMA macrophage cell line, TA4-10 and 4LBI-3 significantly reduced NO levels compared to LPS-activated macrophages by 25 and 14%, respectively ( Figure 7B). Also in this trial, the strain less active in inhibiting NO production was the control strain Sb. These results confirm the higher potentiality of indigenous strains to be used as potential probiotic than the commercial strain Sb. In fact, intact cells of indigenous strains These results confirm the higher potentiality of indigenous strains to be used as potential probiotic than the commercial strain Sb. In fact, intact cells of indigenous strains were more able against LPS-induced inflammatory systems than intact cell of Sb. It's widely reported the activation of the immune systems by β-1,3/1,6 glucans [78,79]. In our case, we can speculate a correlation between glucan content and anti-inflammatory activity only for the activity toward NO production as the strains possessing the highest ability to reduce the NO levels compared to LPS-activated macrophages were TA4-10 and 4LBI-3, which were the two strains characterized by the highest β-glucans content ( Figure 5).
As regards the effect towards ROS, the highest efficacy was obtained for cells of LL-1 strain, characterized by the lowest β-glucans content, although the antioxidant activity of LL-1 cells was expressed at level comparable with other strains. These data confirm that, in addition to β-glucans, yeast cells contain a complex array of compounds, some of them also derived from cell walls, such as mannanoligosaccharides and glyco-proteins, other than peptides, amino acids, nucleotides, lipids and organic acids, with various capacities for driving pro-inflammatory and anti-inflammatory responses [80]. Some authors [81] found strong anti-inflammatory activity in stimulated human blood cells in β-glucan-depleted products produced by enzyme hydrolysis from brewer's and baker's yeasts.

Conclusions
The indigenous Saccharomyces cerevisiae strains tested in this study, mainly isolated from spontaneous fermented foods, exhibited a high biodiversity level for probiotic traits, confirming that indigenous yeasts represent an interesting source for the selection of new probiotic strains. Three strains, 4LBI-3, LL-1, TA4-10, showed promising probiotic activity; in particular, 4LBI-3 strain, isolated from spontaneous fermentation of Aglianico del Vulture grape must, the typical cultivar of Basilicata region (Southern Italy), might be considered the strain showing the best combination of probiotic and heathy traits. However, all the three S. cerevisiae strains, previously tested for parameters of technological interest, have the potential to be used as starter cultures for the production of functional fermented foods. As regards the therapeutic application, further research should be done to ensure safety and efficiency of these new potential probiotic yeasts.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.