Screening of GABA-Producing Lactic Acid Bacteria from Thai Fermented Foods and Probiotic Potential of Levilactobacillus brevis F064A for GABA-Fermented Mulberry Juice Production

Gamma-aminobutyric acid (GABA), the inhibitory neurotransmitter, can be naturally synthesized by a group of lactic acid bacteria (LAB) which is commonly found in rich carbohydrate materials such as fruits and fermented foods. Thirty-six isolates of GABA-producing LAB were obtained from Thai fermented foods. Among these, Levilactobacillus brevis F064A isolated from Thai fermented sausage displayed high GABA content, 2.85 ± 0.10 mg/mL and could tolerate acidic pH and bile salts indicating a promising probiotic. Mulberry (Morus sp.) is widely grown in Thailand. Many mulberry fruits are left to deteriorate during the high season. To increase its value, mulberry juice was prepared and added to monosodium glutamate (MSG), 2% (w/v) prior to inoculation with 5% (v/v) of L. brevis F064A and incubated at 37 °C for 48 h to obtain the GABA-fermented mulberry juice (GABA-FMJ). The GABA-FMJ obtained had 3.31 ± 0.06 mg/mL of GABA content, 5.58 ± 0.52 mg gallic acid equivalent/mL of antioxidant activity, 234.68 ± 15.53 mg cyanidin-3-glucoside/mL of anthocyanin, an ability to inhibit growth of Bacillus cereus TISTR 687, Salmonella Typhi DMST 22842 and Shigella dysenteriae DMST 1511, and 10.54 ± 0.5 log10 colony-forming units (CFU)/mL of viable L. brevis F064A cell count. This GABA-FMJ was considered as a potential naturally functional food for human of all ages.


Introduction
Nowadays people are interested in high-quality functional healthy foods which contain valuable bioactive compounds and/or beneficial microorganisms that can promote health. Lactic acid bacteria (LAB), a prominent probiotic, are generally used in food, agroand pharmaceutical industries due to their "Generally Recognized as Safe" (GRAS) property [1]. Besides lactic acid, various organic acids and bioactive compounds are produced by LAB i.e., bacteriocins and γ-aminobutyric acid (GABA).
GABA functions as an inhibitory neurotransmitter in the human brain [2]. Normally, exogenous GABA is believed to be unable to cross the blood brain barrier (BBB), however, some recent research has shown evidence that GABA may pass the BBB or the exogenous GABA may induce the production of endogenous GABA [3]. GABA produced by gut microbiota can affect the central nervous system (CNS) via the enteric nervous system (ENS) and modulate the gut-brain axis [4]. Moreover, GABA has activities as antidepressant [5], anti-diabetic [6,7], anti-hypertension [8], neuroprotecting agent [9], cardiovascular regulation agent [10], lung adenocarcinoma suppressor [11] and rat plasma growth hormone enhancer [12]. The biosynthesis of GABA in LAB is responsible by an enzyme glutamate decarboxylase (GAD) using L-glutamate as a substrate. GAD system is a member of an amino acid-dependent acid resistance (AR) system which is used to maintain the intracellular pH homeostasis through GABA biosynthesis [13].
A fermented beverage (FB) is non-dairy fermented product which is produced by yeast and/or lactic acid bacteria fermentation from different sources such as whey, grain, vegetables or fruit juice. FB can improve nutrition components and become a healthpromoting product [14]. In Thailand, mulberry (Morus spp.) is widely planted but the mulberry fruit market value is quite low. Due to its richness in beneficial bioactive compounds and medicinal functions such as the properties of prevention and treatment of a sore throat, obesity, diabetes mellitus and hypertension [15], investigations to extract the useful compounds as well as to develop healthy food products will add more value to the mulberry fruits and benefit not only gardeners but also customers of all ages. Hence, this study aimed to isolate and screen for effective GABA-producing LAB, to elucidate the human probiotic potential of the selected LAB strain, to investigate mulberry juice properties, to create a product of GABA-fermented mulberry juice (GABA-FMJ) and to analyze the antioxidant and antibacterial properties of the product obtained. The data acquired will provide information about an alternative healthy beverage which could be consumed for prevention and protection of some symptoms and/or diseases both infectious and non-infectious.

Isolation and Screening of γ-Aminobutyric Acid (GABA)-Producing Lactic Acid Bacteria (LAB)
De Man, Rogosa and Sharpe (Merck TM , Darmstadt, Germany) broth and agar containing 2% (w/v) of monosodium glutamate (Ajinomoto TM , Bangkok, Thailand), MRS-MSG, were used as selective media throughout the study. Various kinds of Thai fermented foods including fermented pork (Nham), fermented beef, fermented fish, sausages and fermented vegetables were used for isolation. One gram of each Thai fermented food sample was suspended in 9 mL of sterile normal saline solution (NSS; 0.85% (w/v) NaCl) and mixed thoroughly by a vortex mixer prior to a spread plate technique using MRS-MSG agar. Meanwhile, one gram of each part of mulberry tree (leaves, fruits, branches and stems) was enriched in MRS-MSG broth and aerobically incubated at 37 • C for 24 h before the spread plate technique on MRS-MSG agar. All culture plates were aerobically incubated at 37 • C for 24 h. All morphologically different colonies were re-streaked on MRS-MSG agar to obtain pure cultures. All Gram-positive bacteria with negative catalase test were kept at −20 • C until used [16].

Primary Screening of GABA-Producing LAB Using Thin-Layer Chromatography
Each LAB isolate was grown in MRS-MSG broth and aerobically incubated at 37 • C for 48 h. Each culture broth was centrifuged (Hettich MIKRO 220R, Tuttlingen, Germany) at 9744× g for 10 min at 4 • C to obtain a supernatant which was further screened for GABA by thin-layer chromatography (TLC) method [17,18]. Briefly, one µL of each sample was spotted on an aluminum sheet silica gel 60 F 254 (Merck TM , Darmstadt, Germany). Analytical standard of γ-Aminobutyric acid (Sigma-Aldrich TM , St. Louis, MO, USA) was used as a positive control. A solvent mixture of n-butanol:acetic acid:distilled water (5:3:2) was used [18]. Subsequently, a TLC plate was sprayed with 0.5% (w/v) of ninhydrin solution (Merck TM , Darmstadt, Germany) and then heated at 60 • C for 30 min. A positive band of GABA appeared in red. The TLC positive samples were further confirmed using a method of high-performance liquid chromatography (HPLC).

GABA Quantitative Analysis by Reversed Phase High-Performance Liquid Chromatography (RP-HPLC)
Quantitative analysis of GABA was performed by a reversed-phase high performance liquid chromatography (RP-HPLC) method as described by Cho et al. [9] with slight modi-Microorganisms 2021, 9, 33 3 of 17 fications. Culture broth of each LAB isolate aerobically grown in MRS-MSG broth at 37 • C for 48 h was centrifuged at 9744× g for 10 min at 4 • C. Each culture supernatant was collected and prepared for HPLC analysis as described by Kim and Kim [19] and Ratanaburee et al. [20] with slight modifications. Briefly, 1 mL of supernatant was freeze-dried by a lyophilizer (Labconco, Kansas city, MO, USA) dissolved with 1 mL of a solution mixture of EtOH:DI water:triethylamine (Thermo Fisher Scientific TM , Loughborough, UK) (2:2:1) to which was added 80 µL solution of EtOH:DI water:triethylamine:phenylisothiocyanate (PITC) (Sigma-Aldrich TM , St. Louis, MO, USA) (7:1:1:1). The sample solution was left at ambient temperature for 20 min and then filtered through 0.45 µm nylon filter (CNW Technologies, Shanghai, China). The HPLC system (1200 series, Agilent Technologies, Inc., Santa Clara, CA, USA) equipped with an ultraviolet (UV) detector was used. GABA content was analyzed by Intersil ® ODS-3 column (4.6 × 150 mm, 5 µm) (GL Sciences, Shinjuku-ku, Japan). Mobile phase solution system was comprised of (A); 1.4 mM of sodium acetate (RCI Labscan, Bangkok, Thailand), 0.1% of trimethylamine and 6% of acetonitrile (RCI Labscan), pH 6.1 and (B); 60% acetonitrile. Both mobile phases were filtered through 0.45 µm nylon membrane filter (Filtrex Technologies, Bengaluru, India). The column was eluted for 70 min with a linear gradient of 0-100% at flow rate 1.0 mL/min with mobile phase B [10]. All sample peaks were detected at 254 nm. The authentic GABA standard (Sigma-Aldrich TM ) was used to set a standard curve and the GABA content of each sample was calculated by the equation: y = 14.782x − 0.0408 (R 2 = 0.9998).

Genomic DNA Extraction and Amplification of the Selected GABA-Producing LAB
The genomic DNA of the selected bacterial isolate was extracted using the phenolchloroform extraction method as described by Giraffa et al. [21] with slight modifications. Briefly, the selected bacterial isolate was aerobically grown in MRS-MSG broth at 37 • C for 24 h. To obtain the bacterial cell pellets, 1 mL of culture broth was centrifuged at 9744× g for 10 min at 4 • C and the supernatant was discarded. Subsequently, mixture of phenol:chloroform:isoamyl alcohol (25:24:1) was added to a sample tube and mixed well prior to centrifugation at 21,924× g for 5 min and the supernatant was transferred to a new tube. DNA was precipitated with 3 M of CH 3 COONa (pH 5.2). Isopropanol was added to a sample tube followed by cold absolute ethanol, mixed, incubated at −20 • C for 15 min and then centrifuged at 21,924× g for 10 min. Supernatant was removed and the DNA sediment was washed with 70% ethanol and centrifuged at 21,924× g for 10 min. DNA was dried for 30 min, dissolved with distilled water and kept at −20 • C.

Probiotic
Properties and Characteristics of the Selected GABA-Producing LAB 2.5.1. Acid and Bile Salts Tolerance L. brevis F064A was aerobically grown in MRS broth at 37 • C for 24 h. Tolerance to an acidic condition was examined by adjusting MRS broth to pH 2.0 and pH 3.0 with hydrochloric acid [25]. Tolerance to bile salts was examined in MRS broth supplemented with 0.3% (w/v) of bile salts (Himedia ® Laboratories, Mumbai, India). Initial suspension of L. brevis F064A was prepared up to 10 6 -10 8 colony-forming units (CFU)/mL (OD 625 = 0.08-0.1) before inoculation [26]. Determination of viable cell count was undertaken every 1 h for 6 h on MRS agar as described by Shekh et al. [25] with slight modifications.

Auto-Aggregation Assay
An auto-aggregation assay was performed following the protocol of Rungsirivanich et al. [27]. L. brevis F064A was aerobically grown in MRS broth at 37 • C for 24 h and cell pellets were collected after centrifugation at 3508× g 4 • C for 5 min. Cells were resuspened in 1× phosphate-buffered saline (PBS) and adjusted to have turbidity of 0.1 at 600 nm (OD i ) before standing at ambient temperature. The absorbance of the upper suspension was determined (OD t ) between 1 and 48 h. The result of auto-aggregation percentage was calculated with the following equation: Auto aggregation (%) = (1 − (OD t /OD i )) × 100

Cell Surface Hydrophobicity Assay
The hydrocarbon adherence ability of the selected strain was conducted following the protocol of Rungsirivanich et al. [27]. L. brevis F064A was aerobically grown in MRS broth at 37 • C for 24 h, cell harvested by centrifugation, washed twice with 1× PBS and resuspended in 1× PBS to obtain the turbidity of OD 600 = 0.1 (A initial ). The resuspended cells (5 mL) and xylene (0.5 mL) were added to the test tube and mixed with a vortex mixer. The mixture was left to stand for 30 min at ambient temperature. The bottom aqueous phase was carefully collected and measured for final absorbance (A final ). Cellular surface hydrophobicity percentage was calculated with the following equation: Cellular surface hydrophobicity (%) = ((A initial − A final )/A initial ) × 100.

In Vitro Bacterial Adhesion Assay
Vero cells were propagated in Dulbecco's Modified Eagle Medium (DMEM, Gibco TM , Paisley, UK) containing 10% (v/v) fetal bovine serum (Capricorn Scientific GmbH, Ebsdorfergrund, Germany) and 100 units/mL of penicillin/streptomycin (Caisson Laboratories, Inc., Smithfield, UT, USA) and incubated in CO 2 incubator at 37 • C for 24 h. This assay was slightly modified from Jacobsen et al. [28]. The harvested Vero cells were plated in 6-well plate with microscopic cover glass approximately 10 3 cells/well and incubated in CO 2 incubator at 37 • C for 24 h. Overnight cultures of bacterial cell pellets of L. brevis F064A, E. coli O157:H7 DMST 12743 and Shigella dysenteriae DMST 1511 (as controls) were collected after centrifugation and approximately 10 8 cells/mL were resuspended in DMEM without antibiotics. Media of each well was removed before inoculation of bacterial suspension (1 mL) and new DMEM (1 mL). Plates were incubated in a CO 2 incubator for 1 h. Subsequently, each well was washed 3 times with 1× PBS and fixed with methanol for 5 min. After methanol removal, cells were stained with 0.38% Giemsa stain for 15 min. Each experiment was triplicated. Adhesion capacity of bacteria was counted amount of bacterial cells adherent on 100 Vero cells per well under a compound microscope. Percentage of bacterial adhesion was determined.

Mulberry Juice (MJ) Preparation and GABA-Fermented Mulberry Juice (GABA-FMJ) Production
Mulberry fruit (Morus alba Linn.) was purchased from farmers in Chiang Mai province, Thailand and was frozen at −20 • C before use. Mulberry fruits was blended and crushed to obtain mulberry juice (MJ). MJ was divided into 2 groups as MJ (control) and MJ supplemented with 2% (w/v) of MSG (MJ-MSG). Prior to fermentation, MJ and MJ-MSG were pasteurized at 73 • C for 15 sec before being transferred to sterile Erlenmeyer flasks.
L. brevis F064A was grown in MRS broth at 37 • C for 24 h, centrifuged and prepared to 1 × 10 8 CFU/mL in normal saline solution before inoculation (5% v/v) [20]. All cultures were aerobically incubated at 37 • C for 48 h. Samples were collected for analyses.

Reducing Sugar Content, Brix and pH Analyses of MJ
Reducing sugar was determined by the dinitrosalicylic acid (DNS) method using glucose (0-1000 µg/mL) as a standard as described by Miller [30] with slight modifications.
Total soluble solid ( • Brix) was measured by a portable refractometer (Trans Instruments, Singapore). The pH was measured by a pH meter (OHAUS ® , Parsippany, NJ, USA).

Bacterial Growth and pH Determination of GABA-FMJ
During fermentation, viable cells of L. brevis F064A were evaluated at 0, 24 and 48 h by the 10-fold serial dilution spread plate technique on MRS agar plates. Change of pH during fermentation was analyzed by a pH meter (OHAUS ® ).

GABA Determination of MJ and GABA-FMJ
GABA content of GABA-fermented mulberry juice (GABA-FMJ) was evaluated by the method mentioned in Section 2.3.

DPPH Scavenging Assay
DPPH (2,2-diphenyl-1-picryl-hydrazyl-hydrate) (Sigma-Aldrich TM ), 0.1 mM, was prepared in methanol. Gallic acid (Sigma-Aldrich TM ) solution was dissolved in methanol and used as standard in a range of 0.001-0.01 mg/mL. The standard and samples were reacted with 0.1 mM of DPPH solution. Reaction tubes were incubated in the dark for 20 min and then their absorbance was measured by a spectrophotometer at 517 nm [15]. The half maximal inhibitory concentration of (IC 50 ) of standard and sample were calculated. Afterwards, scavenging activity was calculated following the equation: DPPH scavenging activity = IC 50 standard IC 50 sample (1)

Total Anthocyanin Determination
The pH differential assay was performed for total anthocyanin measurement as described by Hosseinian et al. [31] with slight modifications. Two buffers were prepared, 0.03 M of KCl (pH 1.0) and 0.4 M of sodium acetate trihydrate (pH 4.5). Twenty µL of each sample were added into 180 µL of KCl buffer and 20 µL of sample were added into 180 µL of sodium acetate buffer in different well of 96-well plate. Each sample was analyzed in triplicate. Afterwards, the mixture sample was incubated at ambient temperature for 15 min before absorbance measurement at 510 nm and 700 nm. The concentration of anthocyanin was calculated and reported as Cyanidin-3-glucoside equivalents following the formula: One-hundred µL of each sample were dropped into a hole prepared by a cork borer (diameter of 9 mm) and then aerobically incubated at 37 • C for 24 h. Gentamicin (0.1 mg/mL) was used as a positive control. After incubation, the inhibition zone (mm) was observed [32].

Lipid Peroxidation Inhibitory Activity Assay
A thiobarbituric acid reactive substance (TBARs) assay described by Tarladgis et al. 1960 [33], modified by Djeri and Williams [34] and Semeniuc et al. [35] was used to measure the lipid peroxidation inhibitory activity of L. brevis F064A in unfermented MJ and GABA-FMJ. Malondialdehyde (MDA) solution (0-100 µM) was presented as standard. For standard, 200 µL of each concentration of MDA solution were added with 200 µL of 0.67% (w/v) thiobarbituric acid (TBA) and boiled for 10 min. For sample, 200 µL of sample were mixed well with 400 µL of 10% (w/v) trichloroacetic acid (TCA) and centrifuged at 9744 × g 4 • C for 15 min. Afterwards, 400 µL of supernatant was collected and mixed with 0.67% TBA and boiled for 10 min. After being cooled down at room temperature, mixture solution absorbance was measured at 532 nm. The lipid peroxidation inhibitory activity of sample was calculated following the equation of the standard curve: y = 0.0222x + 0.0473 (R 2 = 0.9987).

Bacterial Isolation, Screening and 16S rRNA Gene Identification of GABA-Producing LAB
We isolated 127 isolates of LAB from some Thai fermented foods and parts of the mulberry tree including fermented pork (63 isolates), fermented beef (8 isolates), fermented fish (14 isolates), fermented vegetables (23 isolates), fermented sausages (14 isolates), mulberry leaves (3 isolates) and mulberry fruits (2 isolates). All of them showed Grampositive and negative catalase tests. After screening by the TLC method, 36 isolates showed the red spot which matched the GABA standard ( Figure 1) and they were selected for the confirmation step by the HPLC method.
Gram-positive and negative catalase tests. After screening by the TLC showed the red spot which matched the GABA standard ( Figure 1) an for the confirmation step by the HPLC method. For quantitative analysis of the selected GABA-producing LAB p GABA contents of all 36 isolates were produced in the range of 0.08 S1). Among 36 isolates, 4 isolates designed as F019A, F032A, F064A an content higher than 1.00 mg/mL, however, the isolated F064A signifi duced higher GABA content at 2.85 ± 0.10 mg/mL than the other isola of all isolates are shown in Figure 2. For quantitative analysis of the selected GABA-producing LAB performed by HPLC, GABA contents of all 36 isolates were produced in the range of 0.08-2.85 mg/mL (Table S1). Among 36 isolates, 4 isolates designed as F019A, F032A, F064A and F087A had GABA content higher than 1.00 mg/mL, however, the isolated F064A significantly (p ≤ 0.05) produced higher GABA content at 2.85 ± 0.10 mg/mL than the other isolates. GABA contents of all isolates are shown in Figure 2.
The isolate F064A which was isolated from Thai fermented sausage (Sai Krok Isan) was identified by 16S rRNA gene sequencing analysis with a length of 1427 bp by the BLAST search program. The 16S rRNA gene sequence of the isolate F064A showed 100% sequence similarity and close to the strain Levilactobacillus brevis KI271266 ( Figure 3). Therefore, the strain F064A was Levilactobacillus brevis F064A (MT846002).
The 16S rRNA gene sequence of the isolate F032A isolated from fermented pork purchased from Lampang province and the isolate F019A isolated from small fermented pork obtained from Chiang Mai province displayed 99.72% sequence similarity to Levilactobacillus brevis KI271266 (Figure 3). Therefore, the strains F032A and F019A were identified as Levilactobacillus brevis F032A (MT846001) and Levilactobacillus brevis F019A (MT846000), respectively. The isolate F087A isolated from fermented meat purchased from Chiang Rai province had 99.79% sequence similarity to Lactiplantibacillus pentosus ATCC 8041T (D79211) and was named Lactiplantibacillus pentosus F087A (MT846003).
GABA contents of all 36 isolates were produced in the range of 0.08-2.85 mg/mL (Table  S1). Among 36 isolates, 4 isolates designed as F019A, F032A, F064A and F087A had GABA content higher than 1.00 mg/mL, however, the isolated F064A significantly (p ≤ 0.05) produced higher GABA content at 2.85 ± 0.10 mg/mL than the other isolates. GABA contents of all isolates are shown in Figure 2.  Figure  2A are considered statistically significant difference at p < 0.05.
The isolate F064A which was isolated from Thai fermented sausage (Sai Krok Isan) was identified by 16S rRNA gene sequencing analysis with a length of 1427 bp by the BLAST search program. The 16S rRNA gene sequence of the isolate F064A showed 100% sequence similarity and close to the strain Levilactobacillus brevis KI271266 (Figure 3). Therefore, the strain F064A was Levilactobacillus brevis F064A (MT846002).
The 16S rRNA gene sequence of the isolate F032A isolated from fermented pork purchased from Lampang province and the isolate F019A isolated from small fermented pork

Acid and Bile Salts Tolerance
At pH 2.0, L. brevis F064A was able to survive after 3 h of incubation with viable cells decreased from 6.78 ± 0.04 (0 h) to 3.64 ± 0.15 log CFU/mL (3 h). They could not survive at pH 2.0 after 4 h of incubation. At pH 3.0, the bacteria survived more than 6 h of incubation and the viable cell count was rather stable from 7.19 ± 0.14 (0 h) to 7.0 6± 0.09 log CFU/mL (6 h). L. brevis F064A survived in 0.3% (w/v) of bile salts more than 6 h of incubation which showed a quantity of viable cells from 6.76 ± 0.08 (0 h) to 6.90 ± 0.08 log CFU/mL (6 h) ( Figure 4).

Acid and Bile Salts Tolerance
At pH 2.0, L. brevis F064A was able to survive after 3 h of incubation with viable cells decreased from 6.78 ± 0.04 (0 h) to 3.64 ± 0.15 log CFU/mL (3 h). They could not survive at pH 2.0 after 4 h of incubation. At pH 3.0, the bacteria survived more than 6 h of incubation and the viable cell count was rather stable from 7.19 ± 0.14 (0 h) to 7.0 6± 0.09 log CFU/mL (6 h). L. brevis F064A survived in 0.3% (w/v) of bile salts more than 6 h of incubation which showed a quantity of viable cells from 6.76 ± 0.08 (0 h) to 6.90 ± 0.08 log CFU/mL (6 h) (Figure 4).

Bacterial Adhesion and Blood Haemolysis
L. brevis F064A presented the highest adhesion ability on Vero cells as 171.8% and 100% when Shigella dysenteriae DMST 1511 and E. coli O157:H7 DMST 12743 were used as controls, respectively ( Figure 6).
L. brevis F064A strain showed no haemolytic activity and was considered as γ-haemolysis type.

Bacterial Adhesion and Blood Haemolysis
L. brevis F064A presented the highest adhesion ability on Vero cells as 171.8% and 100% when Shigella dysenteriae DMST 1511 and E. coli O157:H7 DMST 12743 were used as controls, respectively ( Figure 6).
L. brevis F064A strain showed no haemolytic activity and was considered as γ-haemolysis type.

Characteristics of MJ and GABA-FMJ
3.3.1. Reducing Sugar, Total Soluble Solid, GABA content and pH of MJ Initial parameters of MJ were indicated. Reducing sugar was 46.50 ± 2.35 mg/mL. Total soluble solid (sugar content of sugar solution) was 8.2 °Brix. GABA content was less than 0.10 ± 0.02 mg/mL ( Figure 7A), and initial pH value without any additives was 3.64 ± 0.42.

Bacterial Growth, pH and GABA Content of GABA-FMJ
After inoculation, viable cell count was exponentially increased from 6.68 ± 0.09 (0 h) to 10.54 ± 0.45 log10 CFU/mL at 48 h of incubation and decreased to 8.91 ± 0.31 log10 CFU/mL after 72 h of incubation. Moreover, pH was changed from 3.88 ± 0.01 (0 h) to 4.12 ± 0.02 after 48 h of incubation.

Bacterial Growth, pH and GABA Content of GABA-FMJ
After inoculation, viable cell count was exponentially increased from 6.68 ± 0.09 (0 h) to 10.54 ± 0.45 log 10 CFU/mL at 48 h of incubation and decreased to 8.91 ± 0.31 log 10 CFU/mL after 72 h of incubation. Moreover, pH was changed from 3.88 ± 0.01 (0 h) to 4.12 ± 0.02 after 48 h of incubation.

Discussion
Most sources of GABA-producing LAB isolation are fermented foods or other glutamic acid-consisting sources. In this study, the highest capacity of GABA-producing LAB, L. brevis F064A, was isolated from Thai fermented sausage (Sai Krok Isan). Previous reports have shown many GABA-producing LAB species from different sources such as Lactiplantibacillus plantarum L10-11 was isolated from plaa-som (Thai fermented fish) [36], Lactiplantibacillus pentosus SS6 was isolated from fermented mulberry fruit [37], Companilactobacillus futsaii CS3 was isolated from Thai fermented shrimp or Kung-Som [38], Latilactobacillus sakei A156 was isolated from Korean fermented seafood or Jeot-gal [39], three strains of L. brevis and one of Lactococcus lactis were isolated from Spanish cheese [40], five strains including Lactiplantibacillus plantarum, L. brevis, Leuconostoc mesenteroides, Leuconostoc lactis and Weissella viridescens were isolated from kimchi [19], Companilactobacillus farciminis D323 was isolated from a Myanmar traditional fermented fishery product [17], Lacticaseibacillus paracasei NFRI 7415 was isolated from Japanese traditional fermented fish or Funa-sushi [41]. In addition to fermented foods, GABA-producing LAB as L. plantarum Taj-Apis362 was isolated from honeybees [42]. Some reports also reveal that L. brevis DSM 32386 can produce 0.262 mg/mL of GABA [43] while L. brevis GABA100 produces 27.6 mg/mL of GABA [44]. The GABA-producing capacity of each LAB strain is dependent on glutamate decarboxylase enzyme and many factors.
Nowadays, most probiotic bacteria are classified to the genera Lactobacillus, Bifidobacterium, and include some species of the genera Lactococccus, Enterococcus and Saccharomyces [45]. In this study, L. brevis F064A could tolerate pH 2.0, pH 3.0 and 0.3% bile salts; according to many previous studies such as L. paracasei K5 can tolerate pH 2.0-4.0 at 2 h incubation [46] and C. futsaii can tolerate simulated stomach condition at pH 2.5 for 2 h [47]. Normally, pH value in the human stomach ranges between 1.0-2.0, but can be up to 3.0 or higher with the presence of food. The average concentration of bile salts in humans is 0.3% and food digestion process takes approximately 2-4 h [25,26]. Tolerance to bile salts of Lactobacillus strains is relied on its capability to hydrolyze bile salts by specific enzymes such as bile salt hydrolase (BSH) [26]. L. brevis F064A presented auto aggregation ability less than 50% after 48 h of incubation. High auto-aggregation ability which is strainspecific can provide probiotic bacteria to persist in the intestinal tract and associate with adherence to mucus and epithelial cells. However, some studies have reported that a high hydrophobicity level leads to higher capacity to adhere on Caco-2 cells [48,49]. On the other hand, some research has suggested that lower percentages of auto aggregation can prevent probiotic bacteria from biofilm formation of enteric pathogenic bacteria [48,49]. L. brevis F064A showed a lower hydrophobicity level. The hydrophobicity property indicates the adhesion capacity on epithelial cells of probiotic bacteria but it does not confirm that a high level of hydrophobicity can show a strong adhesion to mucosa or epithelial cells [50,51]. Additionally, the hydrophobicity ability had a high variation even in the same strain of bacteria such as 7 strains of L. fermentum present a variation of hydrophobicity in a range from 0.30-68.81% [51]. Blood haemolysis is one of safety properties for LAB. In this study, L. brevis F064A was unable to damage red blood cells. Normally, blood haemolysis and DNase capacity are presented as virulence factors for pathogenic microorganisms [25].
Antibiotic susceptibility is a criterion of safety attribute evaluation for probiotic bacteria. L. brevis F064A was susceptible to all testing antibiotics but resistant to ticarcillin/clavulanic acid which is a type of penicillin antibiotic. According to Gad et al. [52], lactic acid bacteria is highly susceptible to many beta-lactam antibiotics except penicillin. Ticarcillin is penicillin antibiotic combined with clavulanic acid (beta-lactamase inhibitor) which is generally used against Gram-negative bacilli and enterococci [53].
From the previous study, mulberry juice was individually fermented with 3 strains of Lactobacillus including L. plantarum, L. acidophilus and L. paracasei which improve their antioxidant attributes but there is no study about GABA [54]. Furthermore, mulberry juice has been fermented with a single culture of Saccharomyces cerevisiae SC125 which produced 1.45 g/L of GABAor L. plantarum BC114 which produced 1.03 g/L of GABA while a coculture of starters could increase GABA content to 2.42 g/L [55]. In this study, mulberry juice with 2% of MSG was fermented by L. brevis F064A. The results indicated that GABA-FMJ could increase bacterial growth during fermentation, GABA content, and antioxidant and antibacterial activities while total anthocyanin was decreased (non-significantly) and also showed lipid peroxidation inhibitory. The fermentation of mulberry with LAB are expected to increase many functional compounds [56]. GABA may be increased due to acidic condition of mulberry juice with glutamate which is a substrate to produce GABA via the GAD system to maintain the cytosolic pH of cells. A study of fermented sweet lemon juice by L. plantarum LS5 presents that the cell count is increased from 7.01 ± 0.01 to 8.63 ± 0.38 log CFU/mL at 37 • C for 48 h, the antibacterial activity is improved, the antioxidant activity is increased but the phenolic compound is decreased [57]. Increasing antioxidant activity was agreed with Mousavi et al. [58] who studied fermentation of pomegranate juice fermented with L. plantarum and Lactobacillus acidophilus. Sung et al. [59] shows that antioxidant activity of yogurt supplemented with freeze-dried mulberry increased with no significant level during storage. In this study, increasing antioxidant activity (DPPH assay) and decreasing anthocyanin (Cy-3-gluc) were related to Mousavi et al. [58]. Increasing antioxidant activity is due to the hydrolytic enzyme produced by LAB which can hydrolyze complex phytochemicals to simple structure [54]. Anthocyanin stability during fermentation depends on many factors as type of anthocyanin, light, temperature, pH and enzymes including oxygen. Moreover, hydroxyl and methoxyl groups can decrease anthocyanin stability including that metabolized by a LAB starter [58,60]. The GABA-FMJ product can inhibit growth of some enteric pathogenic bacteria both Grampositive (B. cereus) and Gram-negative (Salmonella Typhi and Shigella dysenteriae) bacteria due to a mixture of some organic acids, low pH, diacetyl compound and some unknown or unidentified compounds from extracellular substances by LAB [61]. Growth inhibition of Gram-positive bacteria was less sensitive than Gram-negative bacteria. However, Kantachote et al. [61] reports that some fermented beverage plants (FBPs) can inhibit growth of B. cereus and S. aureus. In addition, lactic acid can also damage the lipopolysaccharides of the outer membrane of Gram negative bacteria [62]. Malondialdehyde (MDA) is a product of lipid damage [35]. This study suggests that L. brevis F064A can be promoted to increase the value of fermented mulberry juice and will lead to functional food product development.

Conclusions
L. brevis F064A, GABA-producing lactic acid bacteria, isolated from Thai fermented food was considered as a potential human probiotic by its tolerance to acidic condition and bile salt and its ability to adhere to Vero and mucosal cells. Additionally, L. brevis F064A could apply to the fermentation product to improve their bioactive compounds. GABA-FMJ fermented by L. brevis F064A presented an ability to enhance the growth of probiotics, increase GABA content, improve antibacterial and antioxidant activities, and had lipid peroxidation inhibitory activity. GABA-FMJ should be further developed to become a potential functional food product containing high GABA and bioactive compounds for good health of people of all ages.
Author Contributions: Conceptualization, J.K. and N.T.; methodology, J.K., T.C.S. and N.T.; formal analysis and investigation, J.K. and N.T.; data curation, J.K., P.R. and N.T.; writing-original draft preparation, J.K. writing-review and editing, J.K. and N.T.; funding acquisition, N.T. All authors have read and agreed to the published version of the manuscript.