Probiotic characterization and in vitro functional properties of lactic acid bacteria isolated in Thailand

Fourteen lactic acid bacteria from fermented foods and feces of healthy animals in Thailand were characterized for their potential as probiotics. All isolates could survive in simulated gastrointestinal uid (pH 2) and bile salt solution (pH 8) more than 70% and 63%, when compare with initial cell concentration, respectively. Adhesion test showed more than 70% adhesive property an in vitro experiment. The susceptibility assay showed that all isolates were susceptible to amoxicillin, ampicillin, erythromycin, chloramphenicol, clindamycin, imipenem, kanamycin, noroxacin, penicillin, tetracycline and vancomycin. Based on phenotypic and genetic characteristics, they belonged to the genera Lactiplantibacillus, Levilactobacillus, Capanilactobacillus, Pediococcus, Enterococcus, Limosilactobacillus and Lacticaseibacillus. The isolates exhibited antimicrobial ability against pathogenic bacteria; Gram positive strains (Staphylococcus aureus TISTR 1466 and Listeria monocytogenes TISTR 2196) and Gram negative (Escherichia coli TISTR 780, Salmonella enteritidis TISTR 2202 and Salmonella typhimurium TISTR 292). Limosilactobacillus reuteri MF67.1 and Companilactobacillus farciminis R7-1 showed bile salt hydrolase activity. Cell-free culture supernatants of all 14 isolates were screened for immunomodulating effects on Tumor Necrosis Factor Alpha (TNF-α) production. Twelve isolates were able to decrease TNF-α production at different levels, especially Lactiplantibacillus paraplantarum R26-3 and Lacticaseibacillus zeae M2/5 could high inhibit TNF-α production, showing 34 and 29% reduction, respectively. These results suggested that all 14 strains met the general criteria of probiotics and four strains, including Lacticaseibacillus zeae M2/5, Lactiplantibacillus paraplantarum R26-3, Limosilactobacillus reuteri MF67.1 and Companilactobacillus farciminis R7-1, represent interesting candidates for further studies as anti-inammatory (M2/5, R26-3) or cholesterol reducing agents (MF67.1, R7-1) in vivo animal models. MF67.1 and Companilactobacillus farciminis R7-1 showed positive bile salt hydrolase activity (associated with cholesterol-lowering) while Lactiplantibacillus pentosus R26-3 and Lacticaseibacillus zeae M2/5 displayed the highest TNF- α inhibition in macrophages (anti-inammatory) showing 34 and 29% reduction from the control, respectively. The results indicated that the 14 isolates have probiotic properties and are suitable applications in functional foods and health supplements. However, vivo studies are ecacy they or dietary supplements in further study.


Introduction
Nowadays, diet, stress, and modern medical practices (antibiotics and radiotherapy) play important roles in threatening human health (Dunne et al. 2001). In particular, foods containing bene cial microorganisms such as Lactobacillus and Bi dobacterium exhibit a pivotal role in enriching health wellbeing and suppressing disease (Meyer and Stasse-Wolthuis, 2009). Nevertheless, some bacteria such as Escherichia coli, Bacteroides fragilis, Fusobacterium nucleatum, and Clostridium perfringens are potential pathogens as disease sources (Skar et al. 1986). Colonic diseases and colon cancer are reported to be directly in uenced by the diversity of gut microbiota, while consumption of probiotics can suppress obesity, diabetes, and heart disease by balancing the intestinal microbial composition (O'Keefe 2008; Abdelazez et al. 2017). Thus, foods and new creative diets have focused on the prevention of chronic diseases and disorders (Salvetti and O'Toole 2017). Probiotics are living microorganisms that have bene cial effects on the host when administered in an adequate amount (usually at 10 6 -10 7 CFU/g of product) by improving digestion, immune modulation, and intestinal function. They are widely used in food, feed, dairy, and the fermentation industry (FAO/WHO, 2006). Several studies revealed that probiotics are associated with anti-in ammatory, cholesterol-lowering, anti-allergic, enzyme inhibition, anti-hypertensive, lactose intolerance and mood changes (Lorea Baroja et al. 2007;Kumar et al. 2012;Lee et al. 2014;Panwar et al. 2014;Staudacher 2015;Steenbergen et al. 2015). Strains of microorganisms reported as probiotics include lactic acid-producing bacteria (lactobacilli, streptococci, enterococci, lactococci, bi dobacteria), Bacillus, Saccharomyces and Aspergillus (FAO/WHO 2002). These bacterial strains must be easy to consume and inoffensive for ingestion, able to colonize the gut epithelium, adhere to the mucosal membrane, remain stable during storage and survive in the upper GI tract under high acid and bile salt concentration (Verna and Lucak 2010). Several probiotics have received temporary approval from the European Union. Thus, the aim of this study was to identify and characterize the potential of candidate probiotics from natural resources in Thailand such as edible sources, fermented food, soil and fecal samples from healthy animal. The potential probiotic strains were further their functional properties studies and deposited in the TISTR collection to advantage knowledge concerning the sustainable utilization of microorganisms in the supplementary food and animal feed industries.

Selection and identi cation of lactic acid bacteria (LAB)
Samples from Thai-fermented foods and feces from healthy animal (Table 1) were collected and kept in 4°C until the isolation. Ten grams of each sample was suspended in 90 mL of 0.1% peptone water and mixed in a stomacher. One microliter of the suspension was ten-fold serially diluted (10 2 -10 3 ) and 0.1 mL of each diluted sample was spread on TSA and MRS agar supplemented with 0.3% CaCO 3 and incubated under the anaerobic condition (anaerobic jar; Anaerocult® System, Merck, Germany) at 37°C for 24-48 h.
The isolates were initially identi ed based on colony morphology, Gram's reaction, catalase activity, and hemolysis (Igarashi et al. 1999). For further study, non-hemolytic strains were selected and kept at -80°C (in 15% v/v glycerol). Genomic DNA was extracted following Wilson (2001). The 16S rRNA gene sequences were aligned with selected sequences obtained from the EzBioCloud server database (Yoon et al. 2017) and NCBI BLAST program (http://blast.ncbi.nlm.nih.gov/Blast.cgi) by using the CLUSTAL X version 1.81 in BioEdit software (Thompson et al. 1997). Phylogenetic trees were constructed based on the neighbour-joining (Saitou and Nei 1987), maximum-likelihood (Kimura 1980), and maximum-parsimony (Nei and Kumar 2000) methods using MEGA 7 software (Kumar et al. 2016). The con dence values of nodes were evaluated using the bootstrap resampling method with 1,000 replications (Felsenstein 1985).

Acid and bile salt tolerance
Simulated gastrointestinal uid (SGI) was prepared according to the modi ed method of Hyronimus et al. (2000). Brie y, 0.1% of pepsin was dissolved in MRS broth with 0.05% L-cysteine, adjusted to pH 2 with 1 M HCl and 1 M NaOH and sterile-ltered through a membrane (0.2 µm, Life Sciences, Ann Arbor, MI, USA).
The solution was either used immediately or kept in the fridge until required (not longer than 24 h). Bile salt tolerance was determined according to the method of Gilliland et al. (1984). Bile salts at 0.3% were dissolved with MRS broth with 0.05% L-cysteine (pH 8) and sterilized on liquid cycle for 15 minutes. One milliliter overnight cultures in MRS broth were inoculated in 9.0 mL of SGI (pH 2) and 0.3% bile salt solution (pH 8). The sample mixtures were assessed immediately after mixing to determine the viability of candidate probiotics using the pour plate method and then incubated at 37°C for 180 minutes. The viability of bacteria remaining was investigated according to the above methods. Survival rate was calculated as log values of colony-forming units per milliliter (CFU/mL) and calculated according to the following formula: Percentage of survival cell (%) = (Log N 1 / Log N 0 ) × 100 where N 1 is the average of viable cell (CFU/mL) after incubation for 180 minutes and N 0 is the average of viable cell (CFU/mL) at initial incubation time (0 minutes).

Antibiotic susceptibility
Antibiotic susceptibility of the selected isolates was evaluated against large spectra of clinically important antibiotics by the disc diffusion method (Clinical and Laboratory Standards Institute or CLSI, 2019). Eleven antibiotics as amoxicillin (10 µg), ampicillin (10 µg), erythromycin (15 µg), chloramphenicol (30 µg), clindamycin (2 µg), imipenem (10 µg), kanamycin (30 µg), nor oxacin (10 µg), penicillin (10 µg), tetracycline (30 µg) and vancomycin (30 µg) were applied. Cell concentration of each bacterial culture was adjusted equal to McFarland No.1 and seeded onto MRS agar using a sterile cotton swab and allowed to stand at room temperature for 15 minutes. The antibiotic discs were plated onto agar under aseptic conditions. E. coli TISTR 780, and S. aureus TISTR 1466 were used as a positive control. The agar plates were incubated at optimal conditions. Results were measured and compared with the breakpoint value designated by CLSI (2019).

Antimicrobial activity
Bacterial inhibition was determined by agar diffusion assay (Ennahar et al. 2000

Adhesion assay
The ability of each isolated bacteria was tested for adherence to human epithelial cells according to Jacobsen et al. (1999). Brie y, adenocarcinoma cell line (Caco-2) (ATCC, HTB-37) was cultured in DMEM and seeded into a 24-well cell culture plate at density 2×10 5 cells/well and incubated at 37°C in 5% CO 2 with a humidi ed incubator for 14 days. Before assay, complete DMEM was replaced with antibiotic-and serum-free DMEM for 16 h. Caco-2 cells were then washed twice with sterile PBS (pH 7.2). An aliquot of 2 mL of DMEM (without serum and antibiotics) was added to each well and incubated at 37°C for 30 minutes. Cultured cells of each isolated bacteria were harvested by centrifugation (4,000 ×g, 10 minutes, 4°C) and washed twice with sterile PBS. Cell density was adjusted with DMEM (without serum and antibiotics) to 1×10 9 CFU/mL. Then, 1 mL of each isolate suspension was added to each well and incubated at 37°C in a 5% CO 2 atmosphere for 1 h. Finally, Caco-2 cells were washed three times with sterile PBS to remove non-adherence cells according to Roselli et al. (2006). Cells from monolayers were detached by 1% Triton-X-100. The bacterial cell suspension was serially diluted with sterile saline solution and spread on MRS agar. After incubation for 24-48 h at 37°C, adhesion ability was determined using the following formula.
Percentage of Caco-2 cell adhesion = (N 1 /N 0 ) 100 where N 1 = number of bacterial colonies after incubation and N 0 = number of initial bacterial colonies added as a control.
THP-1 cell culture and TNF-α measurement THP-1 monocytic cells (ATCC, TIB 202) were cultured in RPMI-1640 medium (Gibco-Invitrogen, USA) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (FBS; Gibco-Invitrogen, USA) and incubated at 37°C in 5% CO 2 with a humidi ed incubator for 3-7 days. Each of isolated LAB cultures was centrifuged at 3,000 ×g for 15 minutes. Cell-free supernatant was collected and dried under vacuum condition (35°C) and then re-suspended with 500 µL RPMI-1640 and sterile-ltered through a membrane (0.2 µm, Sigma, USA), the sample called as conditioned media. For bioassay, cell viability was assessed by the trypan blue stain exclusion assay. THP-1 monocytic cells were seeded at density 2.5×10 5 cell/mL into a 96-well cell culture plate and incubated at 37°C in 5% CO 2 for 10 minutes before adding 10 µL of conditioned media and 5 µL of 100 ng/ml puri ed lipopolysaccharide (LPS) from E. coli serotype O127:B8 (Sigma, USA). Then, the plates were incubated at 37°C in 5% CO 2 for 4 h. Finally, supernatants were collected by centrifugation at 1000 ×g for 9 min in 4°C for TNF-α measurement. Assays were done three times, in duplicate. TNF-α production was measured using sandwich enzyme-linked immunosorbent assay (ELISA) technique, according to the manufacturer's instructions (R&D Systems, USA). Recombinant human TNF-α was used as standard. Absorbance was measured at 450 nm using a BioTek® Synergy™ HT (Multi-Detection MicroplateReader, USA).

Cell viability assay
Trypan blue exclusion assays were carried out as previously described (Liu et al. 1999

Statistically analysis
Data were statistically analyzed using the Statistical Package for the Social Sciences (SPSS, Statistics version 24.0.0.0) with one-way analysis of variance (ANOVA), while grouping was assessed by Duncan's multiple range tests at a p-value of 0.05 (Duncan 1955). The data were expressed as mean values of triplicates ± standard deviation.

Selection and Identi cation of lactic acid bacteria (LAB)
A total of 14 lactic acid bacteria were isolated and selected from Thai-fermented foods and healthy animal feces. All isolates were preliminarily characterized based on their morphology, Gram staining, and catalase reaction. The results showed there were 2, 8, 3 and 1 isolates in coccus, rod, short rod and long rod shape, respectively.  Table 1 and Fig. 1. These 14 isolates were tested for further probiotic characterization.

Acid and bile torelence Acid and bile salt tolerance
To consider probiotic properties, all isolates were tested for their survivability in human simulated gastrointestinal uid (SGI) (pH 2) and 0.3% bile salt solution (pH 8). After incubation of each of the 14 isolates at SGI pH 2 and 0.3% bile salt solution pH 8 for 3 h, percentage of viable bacteria numbers were evaluated. Results indicated that all isolates showed more than 60% survival with tolerance to acid pH 2 and 0.3% bile salt solution pH 8 (Fig. 2). Survival rates of 14 isolates were higher than 70% in SGI (pH 2), while isolates No. M22-5 and AKM29.1 were sensitive to bile salt solution more than the others at 66 and 68% survival cell, respectively. Survival rates of 3 isolates, including M1/2.1, M2/5 and MF67.1 showed higher values than 90% viable cell after incubation in SGI (pH 2) and 0.3% bile salt solution (pH 8) for 3 h.
The results suggested that these 14 isolates could successfully pass the human stomach and reach to the intestine with at least 60% of the amount of initial cells.

Antibiotic susceptibility
Isolated lactic acid bacteria should not contain antibiotic resistance gene that could be transferred to human pathogens. Antibiotic susceptibility was determined by the disc diffusion method. Antibiotic susceptibility of 14 isolated LAB were tested with 11 antibiotics. Testing of antibiotic susceptibility of the isolated LAB strains revealed that they are resistance to kanamycin (K; 30 µg), nor oxacin (NOR; 10 µg) and vancomycin (VA; 30 µg) and are susceptible to amoxicillin (AML; 10 µg), ampicillin (AMP; 10 µg), erythromycin (E; 15 µg), penicillin (P; 10 µg), chloramphenicol (C; 30 µg), clindamycin (DA; 2 µg), tetracycline (TE; 30 µg), and imipenem (IPM; 10 µg). Only Pediococcus pentosaceus AKB2.8 and Enterococcus durans M29-6 are susceptible to vancomycin (VA; 30 µg). Most of LAB species have antibiotic resistance on aminoglycosides group (kanamycin, streptomycin and gentamycin), quinolones group (nor oxacin, cipro oxacin and nalidixic acid) and vancomycin. This speci c pattern of resistance is considered as intrinsic factor which can not transfer this property to other microorganisms. (Ammor et al. 2008;Gueimonde et al. 2013) The important property of probiotic screening, the isolated LAB should not resistant on Ampicillin, Tetracycline, Penicillin G, Chloramphenicol, Amoxicillin and Erythromycin. Because of the resistant property to these antibiotics is considered as acquired resistant which can transfer this gene or this property to intestinal pathogens (Abriouel et al. 2017). All 14 isolated LAB were sensitive to these antibiotics, therefore the results indicated that all isolates were safe (Table 2).    The ability of probiotic strains to hydrolyze bile salts was also considered for health functional characteristic of probiotics selection. Bile salt hydrolase (BSH) activity helps bacteria to grow and colonize in the intestine by deconjugating bile salts (Begley et al. 2006). The isolate which has BSH activity will show precipitate around other colonies on MRS agar containing 0.5% TDCA. The BSH activity test showed that only 2 isolates (MF67.1 and R7-1) were BSH positive (Table 4). These strains were identi ed as closely to Lactobacillus reuteri and Lactobacillus farciminis at 99.66 and 99.56% similarity, respectively. According to the probiotics, LAB are Generally Recognized as Safe (Rolfe, 2000) and most commonly used because of their signi cant role to the disease prevention or reduction. These LAB isolates could exhibit BSH activity that may be useful to reduce serum cholesterol levels in the patient with hypercholesterolemia and also prevent hypercholesterolemia in normal people (Chae et al. 2013). However this probiotic properties of the isolates should be done for further study in vivo. For TNF-α inhibition assay; -, refers to reduced TNF-α production; +, refers to stimulated TNF-α production compared with the control.
The superscript showed signi cantly different results by multiple comparison using Duncan's method (p-value less than 0.05). The maximum value is represented by the letters 'a' and descending order.

Adhesion assay
Fourtheen LAB showed adhesion rates between 73-100% (Table 4). Isolates AKM29.1, MF58.1, M22/5, and R7-1 showed the highest adhesion rate at 100% but MF62.2 showed the lowest adhesion rates at 73%. From the result suggested that the 14 strains had the ability for adhesion, establishment, and colonization within the gastrointestinal tract (GIT) and this increased their potential for survival. However, further co-culture with pathogens (for inhibition of pathogen adhesion testing) and studies in animal models are required.

TNF-inhibition
TNF-α inhibitory activity of 14 isolated LAB were investigated in THP-1 cells. TNF-α production was measured using the ELISA method. The results revealed that Lacticaseibacillus zeae M2/5 and Lactiplantibacillus pentosus R26-3 showed the highest values at 29 and 34% reduction from the control, respectively. Limosilactobacillus fermentum M1/2.1 and Levilactobacillus brevis MF62.2 was slightly stimulated TNF-α production by increasing 8 and 17% from the control, respectively (Table 4). Nevertheless, other isolates reduced TNF-α production at 6-26% from the control. The viability of THP-1 cell after exposure to the conditioned medium of all 14 strains was more than 80 percent.

Discussion
LAB are considered Generally Recognized as Safe (GRAS) and widely used in food, feed, dairy, and fermentation industries (FAO/WHO 2001). Fourteen isolates were selected from Thai-fermented foods and feces of healthy animal, which are also increasingly considered as reservoir uncharacterized probiotic strains (Ranadheera et al. 2010;Guantario et al. 2018). Morphological and 16S rRNA gene sequencing analysis revealed that all isolates were LAB, Gram positive, non-spore forming, and catalase negative. There were 13 isolates of Lactobacillus and 1 isolates of Enterococcus. Acidity and bile salt tolerance tests suggested that all LAB were more tolerant of SGI pH 2.0 than 0.3% bile salt solution, except for Enterococcus durans M29-6, Levilactobacillus brevis M62.2, and Lactiplantibacillus pentosus R26-3. The high GIT tolerance capabilities of the 14 strains were an important criterion in the selection of potential probiotics. Antimicrobial activity, antibiotics resistance, and potential adhesive properties of candidate probiotics were investigated for a reliable in vitro system. Results indicated that all 14 strains had the ability against both Gram positive and Gram negative pathogens and were susceptible to amoxicillin (AML), ampicillin (AMP), penicillin (P), chloramphenicol (C), clindamycin (DA), tetracycline (TE), imipenem (IPM) and erythromycin (E). Similarly, Sharma et al. (2017) found that LAB is usually sensitive to chloramphenicol, ampicillin, imipenem, meropenem, and erythromycin. Vancomycin was the rst glycopeptide antibiotic used clinically, and this study showed that 12 isolates were resistant to vancomycin, similar to Tulini et al. (2013) and Zhang et al. (2016), while Enterococcus durans M29-6, Pediococcus pentosaceus AKB2.8 and Lactobacillus farciminis R7-1 were sensitive. Some strains of Enterococcus may possess virulence, especially strains that can display a high level of resistance to vancomycin (Shlaes et al. 1989;FAO/WHO 2001). If this resistance is present, transfer to other microorganisms may occur and this could enhance the pathogenesis of such recipients (Noble et al. 1992;Leclercq and Courvalin 1997). Thus, these results suggested that Enterococcus durans M29-6 and other strains can be considered safe. The Caco-2 cell line has been extensively used as a reliable in vitro system to study the adhesion capacity of candidate probiotics (Lievin-Le Moal et al. 2002). Results indicated that the 14 strains displayed a high level of adhesion at more than 70%. However, co-culture with pathogen strains and in vivo experiments are required for antagonistic assessment activity against pathogens. Furthermore, some bene cial bioactivities such as cholesterol reduction, pro-in ammatory cytokine suppression were evaluated. TNF-α suppression is important in alleviating in ammation in a murine model of IBD (Pena et al. 2005), while bile salt hydrolase (BSH) activity is one of the mechanisms involved in cholesterol reduction by bacteria (Taranto et al. 1997). Results indicated that Lactiplantibacillus pentosus R26-3 gave the highest TNF-α inhibition in macrophages. Although lactic acid bacteria are 'Generally Recognized as Safe' (GRAS) Thus, this studies revealed that the 14 LAB had potential probiotic properties such as tolerance to SGI (pH 2.0) and 0.3% bile salt solution. These results investigated that they have been produced with antimicrobial capability against, antibiotics susceptibility, and high adherence to cell lines, with increasing interest for strain-speci c activity on anti-allergic, enzyme inhibition, anti-hypertensive, lactose intolerance, antioxidant, anti-cancer activities. In addition, they can also reduce cholesterol, lipid, sugar or other metabolic substances reducing.

Conclusion
A total of 14 strains of bacteria were obtained from dairy, Thai-fermented foods, edible sources and feces animals. The absence of hemolysis was tested for resistance to biological barriers (acid and bile salts). Fourteen isolates with more than 70% survival rate in SGI and more than 60% survival in 0.3% bile salt solution were selected for further probiotics requirements. All isolates were lactic acid bacteria and had the ability against both Gram positive (L. monocytogenes TISTR 2196) and negative (S. enteritidis TISTR 2202 and S. typhimurium TISTR 292) pathogens. The antimicrobial assay showed that all isolates were susceptible to amoxicillin (AML), ampicillin (AMP), penicillin (P), chloramphenicol (C), clindamycin (DA), tetracycline (TE), and imipenem (IPM) that are commonly used in human and veterinary medicine.
Furthermore, the 14 candidate probiotics showed good adhesive properties in vitro experiments (more than 70% adhesion). However, functional probiotic properties were speci c to strain. The strains; Limosilactobacillus reuteri MF67.1 and Companilactobacillus farciminis R7-1 showed positive bile salt hydrolase activity (associated with cholesterol-lowering) while Lactiplantibacillus pentosus R26-3 and Lacticaseibacillus zeae M2/5 displayed the highest TNF-α inhibition in macrophages (anti-in ammatory) showing 34 and 29% reduction from the control, respectively. The results indicated that the 14 isolates have probiotic properties and are suitable applications in functional foods and health supplements.
However, in vivo studies are required to determine their e cacy before they can be incorporated in foods or dietary supplements in further study.  Viability and survival percentage of Lactobacillus strains incubated in simulated gastric juice (pH 2) and simulated intestinal juice (pH 8)