Co-Cultivation of Potential Probiotic Strains Isolated from Water Kefir for Fermented Green Tea Beverage

Abstract

This study aimed to isolate and characterize microorganisms from water kefir beverage for their functional properties. Five lactic acid bacteria (LAB) strains were isolated: three Leuconostoc citreum strains (LB4, LB6, LB13) and two Lactococcus lactis strains (LB5, LB25), identified via 16S rRNA sequencing, along with three Saccharomyces cerevisiae yeast strains (Y7, Y9, Y10), confirmed by 18S rDNA sequencing. Due to the high genetic and phenotypic similarity within each species, one representative strain from each (LB4, LB5, Y9) was selected for further analysis. These strains showed potential probiotic properties, including tolerance to acid and bile, high auto-aggregation, and hydrophobicity. The LAB strains were sensitive to gentamicin, and their supernatants inhibited the growth of tested pathogenic bacteria. The cumulative probiotic potential (CPP) scores were 93.33% for Lc. citreum LB4 and L. lactis LB5, and 100% for S. cerevisiae Y9. Furthermore, the fermentation potential of these strains was evaluated in a green tea beverage using three co-culture formulations. Among the formulations tested, the BF1 beverage, fermented by F1 (40% LB4, 40% LB5, and 20% Y9), demonstrated optimal physicochemical, microbiological, and sensory properties. Notably, while the individual strains did not show anti-inflammatory activity, the BF1 beverage formulation exhibited this effect, suggesting a synergistic interaction during fermentation.

1. Introduction

Water kefir is one of those functional beverages produced by fermenting fruit juices or sugar with water kefir grains, which are symbiotic communities of lactic acid bacteria (LAB), acetic acid bacteria, and yeasts, embedded in a matrix of proteins, lipids, and polysaccharides, called “kefiran”. This matrix not only ensures microbial stability but also contributes to the beverage’s numerous health benefits [1].

The microbial composition of water kefir grains varies based on geographical origin and production methods [2]. In Argentinean water kefir grains, for example, Liquorilactobacillus dominates with 55%, followed by Leuconostoc (18%) and Oenococcus (15%) [3]. In contrast, Malaysian samples have a distinctive microbial profile, including species such as Lentilactobacillus hilgardii, Liquorilactobacillus satsumensis, Lacticaseibacillus zeae, Acetobacter iovaniensis, Acetobacter tropicalis, and Oenococcus oeni [4]. Microbial diversity is also influenced by the fermentation substrate. Zannini et al. [5] demonstrated that Acetobacter and Gluconobacter species are more abundant when water kefir grains are cultured in fruit substrates such as apple.

This diversity has a direct effect on the properties of water kefir. Yeasts, for example, hydrolyze sucrose into glucose and fructose, which bacteria then convert into lactic acid and acetic acid. This metabolic activity not only changes the acidity and taste of the beverage, but also contributes to its preservation. In addition, yeasts produce ethanol and carbon dioxide, which contribute to its foaming [6].

Water kefir is known for its antitumor, antifungal, anti-inflammatory, antioxidant, and antibacterial properties [7]. These effects are attributed to the beneficial microorganisms in the beverage, which enhance its health-promoting qualities. Consequently, water kefir is gaining recognition in the expanding functional beverage market, offering both nutritional and wellness benefits. Several studies have demonstrated the positive properties of strains isolated from kefir and water kefir grains. Lacticaseibacillus paracasei (KEF-w19 strain) has shown high antioxidant activity and the ability to inhibit pathogens [8]. Similarly, Lactobacillus kefiranofaciens has been found to provide various health benefits, including immunomodulation, antimicrobial properties, and positive effects on gut health [9]. Furthermore, Limosilactobacillus fermentum and Lactiplantibacillus plantarum have been shown to produce significantly high amounts of glutamic acid, a compound known for enhancing flavor and supporting various physiological functions [10].

Water kefir production remains largely artisanal, as kefir grains are rarely utilized in industrial fermentation. Scaling up water kefir production faces several major challenges, including difficulties in increasing kefir grain mass, their delicate preservation, and the complex optimization of fermentation parameters involving multiple microbial strains. To overcome these obstacles, the use of a precisely formulated starter culture—easy to produce and capable of ensuring a standardized product—represents a promising solution.

This study aims to develop a starter culture that replicates the properties of a fermented beverage similar to that produced with traditional water kefir grains, while meeting the requirements of industrial production. Isolated and identified microbial strains offer a more practical alternative, ensuring consistency and scalability. In this work, the 16S rRNA, 18S rRNA, and ITS regions were sequenced to identify LAB and yeast strains isolated from a green tea water kefir beverage. Their probiotic potential was evaluated, including survival under gastrointestinal conditions, antimicrobial activity, antibiotic resistance, and anti-inflammatory properties. Additionally, previously isolated strains were combined in varying proportions to determine the optimal formulation that delivers the best functional and sensory characteristics for a fermented green tea beverage.

2. Materials and Methods

2.1. Samples Preparation

The water kefir grains were obtained from Tunisian household production and preserved by the Laboratory of Ecology and Microbial Technology (LETMi, INSAT, Tunisia). The water kefir grains were gradually adapted in a cultivation medium containing green tea and 30% honey through multiple cultures at 25 °C.

LAB and yeasts were isolated from a green tea–water kefir beverage sweetened with 42.85% honey and acidified with 1.657% lemon juice [11].

2.2. Fermentation of Green Tea Beverage Using Co-Cultures of Identified Strains

Fermented green tea beverages were produced using co-cultures of Leuconostoc citreum Lb4, Lactococcus lactis Lb5, and Saccharomyces cerevisiae Y9 in varying proportions: F1 (40% LB4, 40% LB5, and 20% Y9), F2 (35% LB4, 35% LB5, and 30% Y9), and F3 (25% LB4, 25% LB5, and 50% Y9). LAB strains were cultured separately in MRS broth (Merck Co., Darmstadt, Germany), while S. cerevisiae was grown in Sabouraud broth (Merck). After incubation, cultures were centrifuged (4000× g, 10 min, 4 °C), and pellets were washed twice with sterile saline (0.85% NaCl). Bacterial and yeast inocula were adjusted spectrophotometrically to 10⁷ CFU/mL and 10⁶ CFU/mL, respectively.

Co-cultures were inoculated (10% v/v) into a green tea infusion (0.8% w/v, steeped for 6 min), supplemented with 42.85% honey and 1.657% lemon juice [11], and fermented at 25 °C for 48 h.

2.3. LAB and Yeasts Isolation and Enumeration

LAB strains were isolated and enumerated using MRS agar (Merck), supplemented with 0.1% (v/v) cycloheximide (Merck) to inhibit yeast growth. The plates were incubated at 37 °C for 24–48 h before being stored at 4 °C for further analysis. For yeast isolation and enumeration, potato dextrose agar (PDA, Merck) was used, which was supplemented with 50 mg/L penicillin (Merck) to inhibit the growth of LAB [12]. The plates were incubated at 30 °C for 48 h. The isolated yeast strains were then stored for subsequent molecular and probiotic characterization.

2.4. Molecular Identification of Isolated Strains

2.4.1. LAB 16S Identification

LAB strains were incubated overnight in MRS broth until they reached the exponential phase. 16S rRNA sequencing was carried out using the Sanger method (Eurofins Genomics GmbH, Ebersberg, Germany). To extract the DNA, 100 µL of each isolate was crushed with 0.4 mm beads and then centrifuged for 30 s at 435× g. The 16S rDNA was amplified with universal primers 27F (5′-AGAGTTGATCMTGGCTCAG-3′) and R1492 (5′-TACGGYTACCTTGTTACGACTT-3′) (Eurogentec, Liège, Belgium). The PCR reaction mix included 22.2 µL of each primer, 88.8 µL of MgCl₂, 222 µL of buffer, 22.2 µL of each dNTP, 11.1 µL of GoTaq DNA Polymerase (Promega, Charbonnières-les-Bains, France), 2 µL of DNA, and 636.4 µL of sterile water.

The PCR cycling conditions were optimized as follows: an initial denaturation at 94 °C for 5 min, followed by 30 cycles of annealing at 55 °C for 10 min and elongation at 72 °C for 10 min. After amplification, the PCR products were analyzed on a 1% agarose gel with Tris-EDTA acetate buffer for optimal DNA resolution. The DNA was visualized under UV light after staining with ethidium bromide. The amplicons were then concentrated and purified using the QIAquick PCR Purification Kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer’s instructions. The purified PCR products were sequenced, and the sequence data were analyzed using the BLAST program from the NCBI website (http://www.ncbi.nlm.nih.gov/BLAST/, accessed on 23 February 2025).

2.4.2. 18S Yeasts Identification

The molecular identification of eight isolated yeasts was conducted by amplifying the 18S region performed by Sanger methodology (Eurofins Genomics). For this assay, isolated yeasts were incubated overnight in Sabouraud broth. Then, 10 µL of each yeast aliquot was added to 90 µL of pure distilled water. The mixture was crushed three times for 30 s each and then centrifuged for 30 s at 227× g. PCR reaction mixtures were prepared containing 120 µL of 5× buffer, 48 µL of MgCl₂, 12 µL of dNTPs, 12 µL of each universal primer Fung02F (5′-CAAGAGATCCGTTGTTGAAAGTT-3′) and Fun02R (5′-GGAAGTAAAAGTCGTAACAAGG-3′) (Eurogentec), 6 µL of BSA, 3.6 µL of Taq polymerase, and completed with 368 µL of sterilized water. Each mixture contained 1 µL of sample DNA. PCR cycling conditions were set at an initial denaturation at 94 °C for 5 min, followed by 34 cycles of annealing at 95 °C and a final elongation step at 72 °C for 10 min. Following PCR amplification, products were analyzed by 1% agarose gel electrophoresis to visualize the amplified DNA fragments. Amplified ITS fragments were then purified according to kit conditions and sequenced to identify the yeast species based on comparison with reference sequences in NCBI databases.

2.4.3. Phylogenetic Analysis

To examine the evolutionary relationships between the isolated LAB and yeast strains and known species, a phylogenetic analysis was performed. A phylogenetic tree was constructed by MEGA 12 software using the neighbor-joining method and a bootstrap analysis was performed to evaluate the reliability of the tree.

2.4.4. Phenotypic Characterization

The isolated strains were characterized based on colony morphology, including shape, color, and texture, followed by microscopic analysis to assess cell morphology and Gram stain properties.

Catalase Test

A 3% hydrogen peroxide (H₂O₂) solution was added to 24 h old cultures, with bubble formation indicating a positive catalase reaction [13].

Motility Test

A bacterial colony was inserted into sulfide indole motility (SIM, Sigma-Aldrich, St. Louis, MO, USA) medium in a test tube and incubated at 37 °C for 48 h. Growth restricted to the inoculation site indicated a negative result, while spreading growth indicated a positive result [13].

2.5. Characterization of Isolated Strains

2.5.1. Acidity Tolerance

This test followed the methodology of Zhang et al. [14], with some modifications. HCl and NaOH solutions were used to adjust the pH of Sabouraud broth (for yeasts) and MRS broth (for LAB) to 3 and 4. LAB strains were added to MRS broth and incubated at 37 °C, while yeast strains were added to Sabouraud broth and incubated at 30 °C, both at a standardized concentration to an optical density (OD) of 0.700 at 560 nm. To assess growth in an acidic environment, the OD at 560 nm was measured before and after the 24 h incubation period using a UV/Vis spectrophotometer (Model 63200; Jenway, Dunmow Essex, UK).

2.5.2. Bile Salt Resistance

The bile salt tolerance of both LAB and yeasts was evaluated using the method de-scribed by Yazidi et al. [15]. LAB isolates were cultured in 9 mL of MRS broth, with one set supplemented with 0.4% bile salt (Sigma-Aldrich) and the other set left unsupplemented as a control, then incubated at 37 °C for 24 h. Similarly, the yeast isolates were cultured in 9 mL of Sabouraud broth under identical conditions, but incubated at 30 °C for 24 h. After incubation, the growth of both LAB and yeast was determined by measuring the OD at 560 nm using a UV/Vis spectrophotometer (Model 63200; Jenway, UK). This allowed for a direct comparison between the bile-supplemented cultures and the control cultures.

2.5.3. Salinity Tolerance

The salinity tolerance of LAB and yeasts was conducted according to a modified method of Al Kotami et al. [13]. For this assay, MRS and Sabouraud broth were prepared and supplemented with 4% and 6% of NaCl (Merck). LAB and yeasts were standardized to an OD of 0.700 at 560 nm before inoculation into their respective media. LAB cultures were incubated at 37 °C, while yeast cultures were kept at 30 °C. To evaluate growth, the OD was measured at 560 nm using a UV/Vis spectrophotometer (Model 63200; Jenway, UK) at the beginning of incubation and after the 24 h incubation period.

2.5.4. Auto-Aggregation Test

The auto-aggregation test was carried out according to the method of Fadda et al. [16], with some modifications. Briefly, each cell culture of LAB in MRS broth and yeast in Sabouraud broth was incubated overnight at 37 °C and 30 °C, respectively. The cultures were then centrifuged at 5000× g for 10 min and resuspended in 3 mL of PBS. The OD at 560 nm was measured using a UV/Vis spectrophotometer (Model 63200; Jenway, UK) after vortexing each tube. The LAB and yeast cultures were then incubated under the appropriate conditions, and the OD of the tubes was measured after 2 h and 24 h of incubation.

The auto-aggregation percent was calculated using the formula:

Auto-aggregation (%) = [1 − (A_t/A₀)] × 100

where A₀ represents the absorbance at time 0 h and A_t represents the measured absorbance at 2 and 24 h.

2.5.5. Hydrophobicity Activity

Hydrophobicity of isolates was assessed according to the method of Fadda et al. [16]. Briefly, 5 mL of each culture (10⁷ CFU/mL) was added to 5 mL of phosphate-buffered saline. Then, 3 mL of each culture was mixed with 0.6 mL of hexadecane (Sigma-Aldrich) and vortexed thoroughly. The mixture was incubated for 2 h at 37 °C for LAB and 30 °C for yeasts. After incubation, the absorbance was measured at 560 nm using a UV/Vis spectrophotometer (Model 63200; Jenway, UK). The hydrophobicity of each strain was calculated using the following equation:

Hydrophobicity (%) = [1 − (A_F/A₀)] × 100

where A₀ is the absorbance of the sample before contact with hexadecane and A_F is the final absorbance (after incubation).

2.5.6. Antibiotic Resistance of LAB Strains

The antibiotic susceptibility of isolated LAB was assessed using the disk diffusion technique [17]. The assay includes five antibiotics (disk diameter = 6 mm): penicillin (10 µg/disk), streptomycin (10 µg/disk), gentamicin (30 µg/disk), ampicillin (10 µg/disk), and olfaxin (5 µg/disk) (Humeau, Couëron, France). MRS agar plates were prepared and inoculated with a standardized LAB suspension containing 10⁷ CFU/mL. These plates were incubated at 37 °C for 48 h. After the incubation period, the diameter of the inhibition zones around each antibiotic disk was measured. Based on these measurements and established criteria, the LAB strains were determined as either susceptible or resistant to the antibiotics tested.

2.5.7. Antimicrobial Activity of LAB Strains

The antimicrobial activity of LAB strains was evaluated using the well agar diffusion method against six pathogens; bacteria was investigated. The pathogens tested included Staphylococcus aureus subsp. aureus ATCC 6538, Escherichia coli ATCC 11229, Salmonella typhimurium ATCC 14028, Shigella sonnei ATCC 25931, Pseudomonas paraeruginosa ATCC 9027, and Bacillus cereus ATCC 10876 [18].

Pathogenic strains were first incubated in BHI broth at 37 °C for 24 h. They were then spread onto Mueller–Hinton agar (Merck) and incubated for another 24 h at 37 °C. The bacterial cells were suspended in sterile physiological saline to reach a final density of 10⁷ CFU/mL. One milliliter of the pathogen suspension was mixed with 19 mL of melted Mueller–Hinton agar and poured into Petri dishes. After solidification, the wells (5–6 mm in diameter) were punched into the agar. Each well was filled with 0.5 mL of either crude or NaOH-neutralized LAB supernatant.

Chloramphenicol (30 μg, Humeau, Couëron, France) was used as a reference control to assess the sensitivity of the tested pathogens. The plates were incubated at 37 °C for 24 h, after which the inhibition zones around the wells were measured in millimeters. The antimicrobial activity of LAB was recorded as the diameter of these inhibition zones.

2.5.8. Cytotoxicity Using Caco-2 Cell

The cytotoxicity of the identified isolates was evaluated using an in vitro model with Caco-2 cells [19]. Initially, the isolates were cultured overnight in the appropriate media until they reached the logarithmic growth phase. These cultures were then neutralized to pH 7 using a 0.1% NaOH solution. Meanwhile, Caco-2 cells (Clini Sciences, Nanterre, France) were maintained in Dulbecco’s modified Eagle’s minimal essential medium (DMEM, Sigma-Aldrich), supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich), 1% (v/v) L-glutamine, 1% (v/v) penicillin–streptomycin solution (Clini Sciences, Nanterre, France), and 7.5% NaHCO₃. The cells were kept at 37 °C in a humidified atmosphere with 5% CO₂.

For the assay, DMEM was removed from the wells, and 1 mL of each neutralized strain culture was added in triplicate. A negative control consisting of DMEM without Caco-2 cells or isolates was included. The inoculated plates were incubated for 2 h at 37 °C in a 5% CO₂ atmosphere. Following incubation, the contents of each well were transferred to tubes, and the number of viable cells was quantified using a cytometer (Attune NxT, AFC2, Life Technologies Holding, Singapore). This method, adapted from previous studies, provided a quantitative assessment of the potential cytotoxic effects of the isolated strains on intestinal epithelial cells.

2.5.9. Cumulative Probiotic Potential of Identified Strains

The cumulative probiotic potential (CPP) is a scoring system used to evaluate the overall suitability of isolated strains for probiotic applications [20]. It assesses multiple functional and safety properties—such as resistance to gastric acidity, bile tolerance, adhesion to the intestinal mucosa, antimicrobial activity, production of beneficial metabolites, and absence of harmful traits.

For each characteristic, a score is assigned based on predefined criteria. In a simple binary system, a score of 1 may be given if the characteristic meets or exceeds a threshold, and 0 if it does not. In other cases, a weighted or graded scoring system might be used. The overall CPP is determined by summing these individual scores.

The CPP percentage is then calculated using the following formula:

CPP (%) = (Observed score/maximum score) × 100

Observed score is the sum of the scores assigned to each evaluated characteristic and maximum score is the highest possible score that could be obtained if all characteristics met their optimal criteria.

2.6. Physico-Chemical Analysis of Fermented Beverage

2.6.1. Determination of pH and Total Titratable Acidity

The pH during the fermentation of beverages was monitored using a pH meter (pH/mV Meter 86502, AZ, Taiwan, China). Additionally, the total titratable acidity (TTA) was determined through a titration method [21]. In this process, a 10 mL sample was titrated with a 0.1 N sodium hydroxide (NaOH) solution. The results were calculated and expressed as a percentage of lactic acid.

2.6.2. Determination of Total Phenolic Compounds

The quantification of total phenolic compounds (TPCs) was conducted using a modified version of the Folin–Ciocalteu method, as optimized and validated by Musci and Yao [22]. To eliminate interference from oxidized molecules, phenolic compounds were extracted through liquid–liquid separation using ethyl acetate. Following the evaporation of ethyl acetate, the residue was resuspended in an equal volume of aqueous methanol (80% v/v). Subsequently, 250 µL of sample was mixed with 1 mL of Folin–Ciocalteu reagent (Sigma-Aldrich) and 1 mL of 10% Na₂CO₃. The prepared solution was kept in darkness at room temperature for 60 min to allow the reaction to proceed. After this incubation period, the ODwas measured at 765 nm against a blank sample using a UV/Vis spectrophotometer (Model 63200; Jenway, UK). The TPC content in tea samples was determined using a gallic acid standard curve ranging from 0.1 to 2 μg/mL (R² = 0.9941). The results were expressed as mg of gallic acid equivalents per mL of sample (mg GAE/mL).

2.6.3. Determination of Antioxidant Activity

2,2-Diphenyl-1-Picrylhydrazyl Radical Scavenging Assay

The antioxidant activity was evaluated using the DPPH (2,2-diphenyl-1-picrylhydrazyl) radical scavenging method [23]. A volume of 500 µL of the sample was mixed with 2.5 mL of the methanolic DPPH^• solution (0.5 mM) (Sigma-Aldrich). The mixture was incubated in the dark for 2 h before measuring the absorbance at 517 nm using a UV/Vis spectrophotometer (Model 63200; Jenway, UK). The absorbance of the methanolic DPPH^• solution served as a control. The percentage of DPPH^• scavenging activity was calculated using the formula:

% DPPH^• scavenging activity = [(A_control − A_sample)/A_control] × 100,

where A_control is the absorbance of the control reaction and A_sample is the absorbance of the sample.

2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulfonic Acid) Diammonium Salt Radical Scavenging Assay

The antioxidant activity was also assessed using the ABTS (2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt) method [24]. The ABTS solution was prepared by mixing 7 mM ABTS stock solution with 2.5 mM potassium persulfate (Sigma-Aldrich) and stored in darkness for 16 h. The absorbance of the ABTS solution was adjusted to 0.700 ± 0.2 at 734 nm by dilution in ethanol. For the analysis, 15 µL of the sample was mixed with 950 µL of freshly prepared ABTS solution. The mixture was allowed to react for 6 min, after which the absorbance was measured at 734 nm using a UV/Vis spectrophotometer (Model 63200; Jenway, UK). The percentage of ABTS^•+ scavenging activity was calculated using the formula:

% ABTS^•+ scavenging activity = [(A_control − A_sample)/A_control] × 100,

where A_control is the absorbance of the control reaction (ethanol mixed with ABTS) and A_sample is the absorbance of the sample.

Power-Reducing Activity

The power-reducing activity was assessed using the method outlined by Son et al. [25]. The procedure involved combining 0.2 mL of extract with 0.2 mL of 0.2 M sodium phosphate buffer (pH 6.6) and 0.2 mL of 1% potassium ferricyanide (Sigma-Aldrich). This mixture was then incubated at 50 °C for 20 min.

Following incubation, 0.2 mL of 10% trichloroacetic acid was added, and the sample was centrifuged at 8000× g for 10 min at 4 °C. The resulting supernatant was collected and mixed with 0.4 mL of distilled water and 0.1 mL of ferric chloride. After allowing the reaction to proceed for 10 min, the absorbance was measured spectrophotometrically at 700 nm using a spectrophotometer (model 63200 UV/Vis; Jenway, UK). HCl-cysteine served as the standard for this assay.

2.6.4. Sensory Analysis

The sensory analysis was conducted with 60 participants, including 45 students from the National Institute of Applied Sciences (TUNISIA) who had received semi-formal training in sensory analysis through practical laboratory sessions. Participants were screened for regular consumption of tea, kefir, or kombucha. Additionally, it was confirmed that they had no allergies to key ingredients such as honey, lemon, or green tea. The participants ranged in age from 20 to 58 years, with a nearly equal gender distribution (35 women and 25 men).

Consumers assessed color, aroma, sweetness, acidity, appearance, and overall acceptability (OA) using a 9-point hedonic scale, where 9 = “Like extremely” and 1 = “Dislike extremely”.

2.7. Anti-Inflammatory Activity of Strains and Fermented Beverage

The anti-inflammatory potential of the identified LAB and yeast strains, along with the fermented beverage, was assessed using a modified protocol based on Zheng et al. [26]. The strains were grown in appropriate media until reaching their optimal logarithmic phase. Simultaneously, a co-culture system of Caco-2 and RAW 264.7 cells (Clini Sciences, Nanterre, France) was established to serve as a model for evaluating anti-inflammatory activity.

Each microbial suspension or the fermented beverage was added to the co-culture system, with positive and negative controls included. The samples were incubated at 37 °C in a 5% CO₂ environment for 3 h. Inflammation was induced by introducing 1 mL of lipopolysaccharide (LPS) (Merck,) mixed with EMEM medium to all wells except the negative control, followed by an additional incubation of 3 h and 30 min.

After incubation, the supernatants were collected and stored for ELISA analysis, using a commercial kit (Pierce Endogen, Rockford, IL, USA). Briefly, 96-well plates were coated with 50 µL of capture antibody mixed with TNF-α and incubated at room temperature for 2 h. The plates were then washed three times with washing buffer, followed by the addition of 100 µL of streptavidin and a 30 min incubation. After another wash cycle, 100 µL of TMB substrate solution was added and incubated in the dark for 30 min. The reaction was stopped with 100 µL of stop solution, and absorbance was measured at 450–550 nm using an automated microplate ELISA reader (Lisa Plus, Darmstadt, Germany). TNF-α levels were determined using a standard curve. All experiments were performed in triplicate, with results reported as mean values ± SD.

2.8. Statistical Analysis

SPSS software Statistics 25 was used to determine the significant difference between the tested strains for different properties. Data were analyzed in triplicate by comparing means using ANOVA. Significant differences were considered at a p-value < 0.05.

3. Results

3.1. Phenotypic Characterization of LAB and Yeast Strains

Distinct morphological characteristics were observed in the selected colonies. The five LAB isolates (designated LB4, LB5, LB6, LB13, and LB25) exhibited a coccus shape and were confirmed to be Gram-positive, catalase-negative, and non-motile. In parallel, three yeast isolates (Y7, Y9, and Y10) showed round morphology on agar plates, with a white to creamy coloration and medium colony size.

3.2. Molecular Identification of Selected LAB and Yeasts

The 16S rDNA gene of LAB isolates from water kefir grains was successfully amplified, as confirmed by electrophoresis. The amplified sequences were then sequenced using the Sanger method and analyzed with BioEdit and BLAST against the NCBI database. Sequence alignment revealed a 97–100% similarity with reference strains, identifying 13 isolates as Leuconostoc citreum (similar to JCM9698) and 13 as Lactococcus lactis (similar to DSM20481) (

Table 1. Molecular identification and phylogenetic characterization of bacterial (16S rDNA) and yeast (18S rDNA) isolates from water kefir.

Strain Designation	Species (16S,18S r DNA Gene Analysis)	Nearest Phylogenetic Neighbor	Identity (%)
LB4	Lc. citreum	Lc. citreum (JCM9698)	99.65%
LB5	L. lactis	L. lactis (DSM20481)	99.58%
LB6	Lc. citreum	Lc. citreum (JCM9698)	99.51%
LB13	Lc. citreum	Lc. citreum (JCM9698)	99.58%
LB25	L. lactis	L. lactis (DSM20481)	99.41%
Y7	S. cerevisiae	S. cerevisiae (CBS1171)	97.29%
Y9	S. cerevisiae	S. cerevisiae (CBS1171)	99.01%
Y10	S. cerevisiae	S. cerevisiae (CBS1171)	97.63%

).

Phylogenetic analysis grouped the isolates into two distinct clusters corresponding to Lc. citreum and L. lactis (Figure 1 and Figure 2). Based on these findings, one representative isolate from each species—Lc. citreum LB4 and L. lactis LB5—was selected for further characterization.

For yeast identification, three isolates were selected based on distinctive morphological and biochemical criteria. The 18S rDNA genes of these isolates were amplified and sequenced, revealing 97–99% similarity to the S. cerevisiae CBS1171 strain (Makegolli) through BLAST analysis (Table 1). Phylogenetic analysis (Figure 3) confirmed that these isolates cluster within the Saccharomyces genus, grouping closely with other S. cerevisiae strains. Given their high genetic similarity and consistent phenotypic characteristics, one representative isolate—S. cerevisiae Y9—was selected for further characterization and subsequent analyses.

3.3. Probiotic Potential of Identified Strains

3.3.1. Tolerance to Acidity, Salinity, and Bile Salts

The two LAB strains showed moderate growth at pH 3, with significantly improved tolerance at pH 4, indicating better survival in mildly acidic conditions. In contrast, S. cerevisiae Y9 exhibited lower resistance to acidic pH (

Table 2. Tolerance to acidity, salinity, and bile salts for isolates strains from water kefir after 24 h of incubation at 37 °C for LAB and 30 °C for yeast.

Strain	∆OD600nm at pH 3	∆OD600nm at pH 4	∆OD600nm at 4% NaCl	∆OD600nm at 6% NaCl	∆OD600nm At 0.4%Bile Salt
Lc. citreum LB4	0.322 ± 0.13	0.532 ± 0.12	1.118 ± 0.098	−0.222 ± 0.05	0.418 ± 0.18
L. lactis LB5	0.416 ± 0.076	0.636 ± 0.013	1.131 ± 0.013	−0.181 ± 0.12	0.667 ± 0.32
S. cerevisiae Y9	0.161 ± 0.05	0.396 ± 0.042	0.614 ± 0.016	0.51 ± 0.081	0.131 ± 0.24

), suggesting a reduced ability to grow under such conditions compared to the LAB strains.

For salt tolerance, the LAB strains were more sensitive to increasing NaCl concentrations than S. cerevisiae Y9 (Table 2). While the yeast maintained consistent growth at both 4% and 6% NaCl, LAB strains showed growth inhibition at 6% NaCl, demonstrating a lower osmotolerance compared to the yeast.

In the presence of bile salts, all tested strains survived at 0.4% bile concentration (Table 2). However, L. lactis LB5 exhibited greater bile resistance than Lc. citreum LB4 under all conditions.

3.3.2. Auto-Aggregation of Identified Strains

The three strains exhibit a high ability for auto-aggregation, with no significant difference observed between the two LAB strains. S. cerevisiae Y9 demonstrated a similar auto-aggregation ability to the LAB strains at 4 h (Figure 4). However, after 24 h, S. cerevisiae Y9 showed a greater increase in auto-aggregation, surpassing the LAB strains.

3.3.3. Hydrophobicity of Identified Strains

The strains Lc. citreum LB4, L. lactis LB5, and S. cerevisiae Y9 exhibited hydrophobicity rates of 76.18 ± 0.042%, 70.25 ± 0.013%, and 69.89 ± 0.056%, respectively (Figure 5). Leuconostoc showed higher hydrophobicity than Lactococcus, while LAB strains overall displayed greater hydrophobicity than yeast.

3.3.4. Antimicrobial Activity of Identified LAB Strains

The results indicate that both Lc. citreum LB4 and L. lactis LB5 strains exhibit antimicrobial activity against the tested pathogens. However, the extent of this activity varies depending on the type of supernatant used (pure or neutralized) and the specific bacterium targeted (

Table 3. Antibacterial activity of isolated LAB from water kefir after incubation at 37 °C for 24 h (PS: pure supernatant, NS: neutralized supernatant).

Microorganism	Media Growth Inhibition Halos (in mm)				Chloramphenicol (30 µg)
	Lc. citreum LB4		Lc. lactis LB5
	PS	NS	PS	NS
Escherichia coli ATCC11229	14 ± 0.03 a	13 ± 0.011 b	16± 0.002 c	16 ± 0.01 c	15 ± 0.03 c
Shigella sonnei ATCC25931	15 ± 0.0015 a	14 ± 0.003 b	15± 0.09 a	12 ± 0.017 c	14 ± 0.072 b
Staphylococcus aureus subsp. aureus ATCC6538	14 ± 0.001 a	14 ± 0.002 a	14± 0.09 a	14 ± 0.017 a	12 ± 0.065 b
Salmonella thyphimirium ATCC14028	12 ± 0.0012 a	10 ± 0.001 b	13 ±0.014 a	11 ± 0.024 c	11 ± 0.021 c
Bacillus cereus ATCC 10876	15 ± 0.007 a	15 ± 0.004 a	15 ±0.012 a	15 ± 0.045 a	12 ± 0.04 c
Pseudomonas paraeroginosa ATCC9027	12 ± 0.023 a	11 ± 0.0012 b	13 ± 0.07 c	12 ± 0.068 a	13 ± 0.01 c

For a tested pathogenic bacterium, (^{a, b, c}) means with the same superscripts indicates no significant difference between the samples (p > 0.05).

).

Both strains demonstrated significant inhibitory activity against all tested pathogenic bacteria, regardless of whether the supernatant was pure or neutralized. Moderate activity was observed against Salmonella typhimurium ATCC14028 and Pseudomonas aeruginosa ATCC9027, with inhibition being more pronounced in an acidic environment.

Furthermore, the antibacterial activity of L. lactis LB5 was generally more pronounced than that of Lc. citreum LB4 (both pure and neutralized supernatants), except in the case of S. aureus subsp. aureus ATCC6538 and B. cereus ATCC10876, where both strains exhibited similar levels of activity.

3.3.5. Antibiotic Resistance of Identified LAB Strains

Antibiotic susceptibility of L. lactis LB5 and Lc. citreum LB4 to five antibiotics (penicillin, gentamicin, ampicillin, ofloxacin, and streptomycin) was assessed. The average diameters of the inhibition zones are presented in

Table 4. Evaluation of antibiotic sensitivity of LAB strains isolated from water kefir.

Antibiotic	Inhibition Halos (in mm)
	Lc. citreum LB4	L. lactis LB5
Penicillin	14 ± 0.01	Resistant
Streptomycin	resistant	13 ± 0.016
Gentamicin	16 ± 0.023	19 ± 0.012
Ampicillin	Resistant	14 ±0.05
Olfaxin	15 ± 0.033	Resistant

Diameter of the antibiotic disc = 6 mm.

. L. lactis LB5 showed susceptibility to streptomycin, gentamicin, and ampicillin, with inhibition zones of 13 mm, 19 mm, and 14 mm, respectively. However, it exhibited resistance to both penicillin and ofloxacin. In contrast, Lc. citreum LB4 displayed resistance to streptomycin and ampicillin.

3.3.6. Cytotoxicity of Identified Strains

The exposure of CaCo-2 cells to Lc. citreum LB4 resulted in approximately 52.568 ± 2.769% cell viability, indicating moderate cytotoxicity. Similarly, L. lactis LB5 exhibited a cell viability of around 50.975 ± 5.981%, suggesting a comparable level of moderate cytotoxicity. In contrast, S. cerevisiae Y9 showed a significantly higher cell viability of 95.142 ± 1.127%, indicating a very low level of cytotoxicity (Figure 6).

3.3.7. Anti-Inflammatory Activity of Identified Strains

The anti-inflammatory activity of the identified strains was assessed by their ability to modulate the immune response triggered by LPS and influence the levels of pro-inflammatory cytokines, particularly TNF-α (

Table 5. Evaluation of anti-inflammatory activity of identified strains following LPS exposure.

Strain	Anti-Inflammatory Activity
	TNF-α Before LPS	TNF-α After LPS
Lc. citreum LB4	563.22 a ± 465.21	1019.66 b ± 350.17
L. lactis Lb5	1026.4 a ± 84.55	1253 b ± 22.91
S. cerevisiae Y9	311.505 a ± 88.74	451.62 b ± 119.549

^{a, b}: means with the same superscripts in the same row indicates no significant difference between the samples (p > 0.05).

). The results indicate that these strains do not exhibit significant anti-inflammatory potential.

3.3.8. Cumulative Probiotic Potential of the Identified Strains

The CPP was calculated to evaluate the overall probiotic potential of these strains (

Table 6. Cumulative probiotic potential of identified strains.

Probiotic Characteristics	Probiotic Isolates (Scores)
	Lc. citreum LB4	L. lactis LB5	S. cerevisiae Y9
Acidity tolerance	1	1	1
Bile salt tolerance	1	1	1
Auto-aggregation activity	1	1	1
Hydrophobicity	1	1	1
Anti-microbial activity	1	1	nd
Antibiotic sensivity	0.6	0.6	nd
Total	5.6/6	5.6/6	4.0/4
CCP (%)	93.33	93.33	100

nd: not determined.

). Both Lc. citreum LB4 and L. lactis LB5 showed a high cumulative probiotic potential of 93.33%, while S. cerevisiae Y9 exhibited a calculated value of 100%.

3.4. Fermentation of Green Tea Infusion by Co-Culture of Isolated Strains

The probiotic strains previously isolated were used in co-culture at different proportions (F1, F2, F3) to ferment a green tea beverage sweetened with honey and acidified with lemon juice (42.85% honey and 1.657% lemon juice). The formulation of this beverage was optimized in a previous study using a central composite design (CCD) [11]. The results regarding the physicochemical, antioxidant, antimicrobial, and sensory properties of fermented tea-based beverages (BF1, BF2, BF3) were evaluated.

3.4.1. Microbial Growth and Acidification During Fermentation

Analysis of

Table 7. Changes in acidity, pH, and viable cell count in different fermented beverages of green tea by F1 (40% Lc. citreum LB4, 40% L. lactis LB5, 20% S. cerevisiae Y9), F2 (35% Lc. citreum LB4, 35% L. lactis LB5, 30% S. cerevisiae Y9), and F3 (25% Lc. citreum LB4, 25% L. lactis LB5, 50% S. cerevisiae Y9) after 48 h of fermentation at 25 °C.

Formulation	Time (h)	Viable Cell Count of LAB (log10₀ UFC/mL)	Viable Cell Count of Yeasts (log10₀ UFC/mL)	Acidity (%)	pH
BF1	0	6.75 ± 0.16 a	4.47 ± 0.12 a	0.5 ± 0.04 a	5.42 ± 0.05 a
	48	7.89 ± 0.15 b	6.35 ± 0.19 b	1.50 ± 0.05 b	4.25 ± 0.02 b
BF2	0	6.32 ± 0.32 c	4.75 ± 0.02 c	0.55 ± 0.01 a	5.5 ± 0.07 a
	48	7.52 ± 0.01 d	6.28 ± 0.01 d	1.60 ± 0.01 c	4.18 ± 0.01 c
BF3	0	5.24 ± 0.05 e	5.47 ± 0.01 e	0.50 ± 0.02 a	5.55 ± 0.03 a
	48	6.91 ± 0.06 f	7.02 ± 0.04 f	1.74 ± 0.03 d	4.02 ± 0.015 d

Means of the same column with the same superscripts (^a–f) indicate that there is no significant difference between the two samples according to Tukey’s test (p > 0.05).

, which presents pH, acidity levels, and final microbial concentration in different fermented beverages (BF1, BF2, BF3), reveals significant differences among the formulations (p < 0.05). BF1 and BF2, which contained moderate proportions of S. cerevisiae Y9 (20% and 30%, respectively), showed higher bacterial growth and acidity compared to BF3. These results suggest that a balanced ratio of yeast promotes a synergistic coexistence with LAB, supporting effective fermentation. In contrast, BF3, which had a higher proportion of S. cerevisiae Y9 (50%), exhibited significantly reduced bacterial growth, higher yeast proliferation, a decrease in pH, and an increase in total acidity (1.74%).

3.4.2. Total Phenolic Content and Antioxidant Activity of Fermented Beverages

Results in

Table 8. Changes in total phenolic compounds in different green tea formulations fermented by co-culture of isolated strains (F1 (40% Lc. citreum LB4, 40% L. lactis LB5, 20% S. cerevisiae Y9), F2 (35% Lc. citreum LB4, 35% L. lactis LB5, 30% S. cerevisiae Y9), and F3 (25% Lc. citreum LB4, 25% L. lactis LB5, 50% S. cerevisiae Y9) after 48 h of fermentation at 25 °C.

Time (h)	Total Phenolic Content (mg GAE/mL)
	BF1	BF2	BF3
0	48.32 ± 0.011 a	48.13 ± 0.18 a	48.37 ± 0.022 a
24	51.39 ± 0.014 b	49.74 ± 0.08 c	48.91 ± 0.011 a
48	55.13 ± 0.01 d	52.91 ± 0.023 e	50.082 ± 0.05 c

Means with the same superscripts (^a–e) indicate that there is no significant difference between the two samples according to Tukey’s test (p > 0.05).

indicate that microbial composition significantly influenced the TPC of fermented beverages. After 24 h of fermentation, all formulations exhibited increased TPC, but with notable differences depending on the microbial strain proportions. BF1 showed the highest increase (+12.2%, reaching 55.13 mg GAE/mL at 48 h), and BF2 (+9.5%, 52.91 mg GAE/mL). In contrast, BF3 (yeast-rich) displayed the lowest increase (+4%).

Among the tested formulations, BF1 exhibited the highest antioxidant activity. BF2 also showed strong antioxidant potential, though it was slightly lower in power-reducing activity (64.15 μmol HCL-Cystein/L). In contrast, BF3 displayed the lowest antioxidant activity across all assays (Figure 7).

3.4.3. Sensory Analysis of the Different Fermented Beverages

BF1, BF2, and BF3 exhibited distinct sensory profiles influenced by their microbial composition. While all formulations shared a visually uniform appearance, differences emerged in other sensory attributes (Figure 8).

Aroma analysis revealed an imbalance in BF3, marked by a pronounced alcoholic note due to its higher yeast content. In contrast, BF1 and BF2 displayed a well-balanced aroma characteristic of mixed lactic–alcoholic fermentations.

Sweetness was most pronounced in BF1 and BF2, whereas BF3 had a noticeably reduced sweetness. Overall acceptability scores ranked BF1 highest (7.7), while BF3 received the lowest score (6.0).

3.4.4. Anti-Inflammatory Activity

After evaluating the sensory acceptability of the three formulations, BF1, the most favorable, was selected for anti-inflammatory analysis. Before LPS exposure, its TNF-α concentration was 695.18 ± 372.24 pg/mL. After exposure, a slight but statistically significant decrease (688.28 ± 324.27 pg/mL, p < 0.05) was observed.

4. Discussion

4.1. Probiotic Properties of Isolated Strains

The microbial composition of the green tea water kefir beverage, containing 4 × 10⁷ CFU/mL of LAB and 2 × 10⁷ CFU/mL of yeasts [11], differed from previously reported findings. Lc. citreum was identified in our sample, although it is less commonly found in studies, where Lc. mesenteroides is typically predominant [27]. Additionally, L. lactis, which is primarily associated with milk kefir, was also detected. This aligns with previous reports identifying L. lactis in Brazilian water kefir grains [2]. The presence of S. cerevisiae in our sample is consistent with existing literature, as it is one of the most abundant yeasts in water kefir grains [28].

Water kefir typically contains LAB, acetic acid bacteria, and yeasts. However, no acetic acid bacteria were detected in our sample. All isolated bacteria were Gram-positive and catalase-negative, a characteristic feature of LAB, which lack the catalase enzyme responsible for converting hydrogen peroxide into water and oxygen [13].

Several studies have reported that LAB can survive in environments with pH values ranging between 2 and 4 [29,30]. Rallu et al. [31] demonstrated that L. lactis adapts to acid stress through active protein synthesis, pH regulation via ahrC, and the (p)ppGpp pathway, which also enhances acid resistance. However, our findings do not align with the study by Pundir et al. [32], which showed that LAB isolated from food samples could not survive at pH 3 but were able to grow at pH 4. The limited survival of LAB at pH 3 may be explained by disruptions in their cellular processes, membrane integrity, and metabolic activity. Similarly, Alegría et al. [33] reported that none of the Lc. citreum strains they tested were able to grow at pH 4.5 or lower, indicating a low tolerance to acidic conditions.

Contrary to previous studies, S. cerevisiae Y9 showed lower tolerance to acidic conditions than LAB. Diosma et al. [34] reported that several S. cerevisiae strains demonstrated significant resistance to acidic environments, with survival rates ranging from 50% to 95% after 3 h of incubation. However, in our case, S. cerevisiae Y9 was more sensitive to acid stress. This can be explained by changes in membrane fluidity and ion gradient destabilization, which may also contribute to growth inhibition [35].

Regarding salt tolerance, a clear relationship was observed between the growth of LAB strains and NaCl concentration. High salt concentrations negatively impacted the growth of these strains. For instance, Reale et al. [36] demonstrated that the growth behavior of different LAB strains was comparable between 2% and 4% NaCl after 24 and 48 h of incubation. However, their survival significantly decreased when the salt concentration exceeded 5%, suggesting that LAB strains have a limited tolerance to high salt concentrations, beyond which their growth is inhibited.

In contrast, Lee et al. [37] reported that Lc. citreum strains 4AC15, 7A7b, and 7G2c achieved higher cell densities than other strains at salt concentrations up to 6.5%. This variability in salt tolerance among LAB strains could be attributed to differences in their adaptive mechanisms. The resistance of S. cerevisiae to salt stress has also been demonstrated in other studies. Yeasts improve their salt tolerance by synthesizing osmotic protectants like glycerol [38]. Additionally, yeasts can mitigate cytotoxicity by actively exchanging intracellular sodium ions with extracellular potassium ions [39].

Bile salt resistance is another important property to consider when studying the potential probiotic qualities of microorganisms. Their ability to survive in bile salts indicates their potential for survival in the small intestine, where they can positively contribute to host health. Bile salts act as antimicrobial agents, capable of damaging and inhibiting bacterial growth. Therefore, probiotic microorganisms must demonstrate resistance to these compounds to successfully traverse the gastrointestinal tract. L. lactis LB5 and Lc. citreum LB4 showed good survival in bile salt environments, which may be due to their ability to produce hydrolytic enzymes that degrade conjugated bile salts, thereby mitigating bile’s inhibitory effects [40]. However, Likotrafiti et al. [41] noted that L. lactis strains were unable to survive at a bile salt concentration of 0.08%. Additionally, the bile salt tolerance of S. cerevisiae may be attributed to an increase in lipid content following bile salt exposure, which helps protect the cells from bile salt–induced toxicity [42].

The ability of strains to adhere to each other and form aggregates, known as auto-aggregation, is essential for pathogen suppression, colonization, and persistence within the gastrointestinal tract. Studies have shown that strains of Lc. citreum and L. lactis exhibit biofilm-forming abilities and adhesion to intestinal cells, which are essential for persistent colonization [43,44]. The high auto-aggregation observed in S. cerevisiae may be attributed to surface structures such as exopolysaccharides [45].

These strains also display strong hydrophobicity, a key factor for adhesion to intestinal epithelial cells, promoting their colonization and long-term persistence in the gastrointestinal tract. Their interaction with hydrophobic surfaces, particularly in this environment, suggests significant potential for adhering to intestinal walls, facilitating their implantation and persistence. This adhesion is a prerequisite for exerting beneficial effects, such as modulating the microbiota, competing with pathogens, or stimulating the immune system. For instance, Muthusamy et al. [40] demonstrated significant hydrophobicity in Lc. citreum strains KCC-57 and KCC-58, with values of 63.37% and 71.60% in solvents like chloroform and xylene after 180 min of incubation. Similarly, Tarazanova et al. [46] observed increased hydrophobicity in L. lactis strains during their exponential phase. Regarding S. cerevisiae, studies have revealed a diversity in hydrophobicity levels among strains, influenced by cell surface composition and growth conditions. Some strains exhibited high hydrophobicity, while others showed moderate to low levels [47].

In terms of antimicrobial activity, both L. lactis LB5 and Lc. citreum LB4 exhibit inhibitory effects against various pathogenic bacteria, underscoring their potential to modulate the intestinal microbiota. Numerous studies have confirmed the antimicrobial potential of L. lactis, primarily attributed to the production of bacteriocins, particularly nisin [48]. Rodriguez et al. [49] further demonstrated the antimicrobial activity of L. lactis strains against Listeria monocytogenes, linked to the production of pediocin. However, De Chiara et al. [50] reported that L. lactis strains inhibited the growth of Shigella sonnei ATCC25931, suggesting that organic acids, rather than bacteriocins, were responsible for this inhibition. They observed that the bactericidal effect was neutralized when the pH of the supernatants was adjusted. Comparatively, the antibacterial activity of Lc. citreum has also been demonstrated in other studies. Muthusamy et al. [40] have shown that Lc. citreum strains showed an antibacterial activity against pathogens. This activity is attributed to the production of various antimicrobial substances, including organic acids, CO₂, H₂O₂, diacetyl, phenyllactic acid, and small molecular antimicrobial metabolites. These compounds can damage cell membranes, inhibit enzyme systems, and disrupt the metabolic pathways of pathogenic bacteria [51]. However, Lc. citreum exhibited slightly weaker antimicrobial activity than L. lactis, a finding consistent with Grześkowiak et al. [43], who also observed lower antimicrobial effects in Lc. citreum compared to other LAB strains.

Beyond their antimicrobial properties, the antibiotic resistance of LAB is an essential aspect to consider, particularly due to the potential risk of horizontal gene transfer of resistance genes to pathogenic bacteria. Khemariya et al. [52] conducted antibiotic sensitivity testing on L. lactis strains, revealing that all were sensitive to ampicillin and spectinomycin. However, L. lactis LB5 has been reported to exhibit resistance to certain antibiotics. Florez et al. [53] found that only 5.4% of L. lactis strains were resistant to streptomycin. This resistance is primarily mediated by multidrug transporters such as LmrA and LmrP. LmrA, an ATP-binding cassette (ABC) transporter, expels a wide range of antibiotics, while LmrP, a proton/drug antiporter, utilizes the proton motive force to remove toxic substances from the cell [54,55].

Regarding Lc. citreum, this species was found to be sensitive to penicillin, gentamicin, and ofloxacin. Similarly, Morandi et al. [56] confirmed the sensitivity of all Leuconostoc strains isolated from Italian cheeses to penicillin and ampicillin. However, resistance to streptomycin and ampicillin has been observed in Lc. citreum, which could be attributed to intrinsic mechanisms or the acquisition of resistance genes through horizontal gene transfer [57]. The study by Flórez et al. [58] examined the antibiotic sensitivity of Leuconostoc isolates and reported that several strains exhibited resistance to three or more antimicrobial classes, identifying them as multi-drug resistant (MDR). All MDR Leuconostoc strains were resistant to at least four antimicrobials, and one strain was found to be resistant to nine. More specifically, an MDR phenotype was observed in Lc. citreum LE46 and in four strains of Lc. mesenteroides.

Regarding cytotoxicity, both Lc. citreum LB4 and L. lactis LB5 exhibited moderate cytotoxicity, in agreement with the findings of Han et al. [59] and Al-Shaibani et al. [60], who also reported low to moderate cytotoxicity in these strains. The very low cytotoxicity of S. cerevisiae (95.142% viability) further supports its established safety in food fermentation processes. Previous research has indicated that S. cerevisiae can mitigate mycotoxin-induced cytotoxicity, reinforcing its potential as a safe probiotic [61].

With respect to anti-inflammatory activity, none of the individual strains demonstrated anti-inflammatory potential, contrary to some previous findings. Studies have suggested that Lc. citreum and L. lactis can exhibit anti-inflammatory effects under specific conditions. For instance, Lc. citreum isolated from kimchi has been reported to suppress inflammation by inhibiting nitric oxide production [62,63]. Additionally, L. lactis has been shown to exert anti-inflammatory effects in animal models of colitis [64], while certain S. cerevisiae strains can modulate pro-inflammatory responses and inhibit NF-κB activation [65].

4.2. Physico-Chemical and Sensory Properties of Fermented Beverages

The use of specific microbial strains directs fermentation toward the production of target bioactive molecules, optimizes yields, and reduces the formation of undesirable by-products. Not only does the nature of the strains play an essential role, but their relative proportion in the ferment also significantly influences the nutritional and sensory properties of the final product. Significant differences were observed in all physicochemical parameters of the three fermented beverages (BF1, BF2, and BF3), with these variations being directly linked to the microbial ratios.

A moderate proportion of yeasts promotes a balanced fermentation between LAB and yeasts, enabling a synergistic coexistence of both microbial populations. In contrast, a high yeast concentration can lead to a reduction in bacterial growth, likely due to competition for essential nutrients. Indeed, yeast cells, being approximately 20 to 50 times larger than bacterial cells, absorb proportionally more resources [66]. Additionally, Carvalho et al. [67] demonstrated that yeasts can inhibit bacterial growth through the production of ethanol and volatile metabolites.

Regarding acidity, LAB primarily produce lactic acid, which is less acidic than the acetic acid generated by S. cerevisiae. As a result, even with high fermentative activity, the total acidity may appear lower in LAB-dominated samples. Moreover, in the presence of oxygen, yeasts can produce acetic acid via the acetification pathway, which not only increases total acidity but also imparts a sharper taste [68].

Similarly, antioxidant activity plays a significant role in mitigating oxidative stress–related diseases by neutralizing free radicals and slowing biomolecular oxidation. To accurately assess this activity, at least two complementary methods should be used alongside TPC quantification, an important indirect marker of antioxidant potential. The TPC content of fermented beverages is largely influenced by the composition and synergistic interactions of microorganisms, with prolonged fermentation and high microbial biomass optimizing TPC release. Beverages dominated by LAB demonstrated superior TPC and antioxidant activities compared to this with a higher yeast composition. This difference may stem from LAB’s ability to produce bioactive metabolites, such as EPS, antioxidant enzymes, bioactive peptides, and manganese ions, which directly enhance antioxidant capacity [69].

L. lactis, for example, produces phosphorylated EPS with significant in vivo and in vitro antioxidant activity, attributed to its ability to scavenge hydroxyl and superoxide anion radicals while boosting catalase and superoxide dismutase activity [70]. Likewise, Lc. citreum demonstrates strong antioxidant effects linked to EPS production during fermentation [71]. In contrast, fermentations with a higher yeast proportion, while effective in ethanol production, may prioritize metabolic pathways less associated with phenolic compound stabilization. Although, S. cerevisiae also contributes to antioxidant activity through mechanisms such as glutathione synthesis and enzymatic antioxidants (e.g., superoxide dismutase and catalase) [72].

An increase in TPC and antioxidant activity has been observed in water kefir, where polyphenol levels more than doubled after extended fermentation [73]. These findings underscore the significance of strain selection and precise control of fermentation conditions in maximizing the functional quality of fermented beverages, particularly their antioxidant potential [74,75].

The three fermented beverages showed distinct sensory profiles, strongly influenced by their microbial composition. BF1 showed no significant difference from green tea water kefir [11], validating its close similarity to water kefir, likely due to a LAB/yeast ratio similar to that found in water kefir grains. Conversely, the dominance of S. cerevisiae in BF3 disrupted the acidity–aroma balance, reducing its overall sensory appeal. The more pronounced aroma in BF3 created an imbalance, further affecting its sensory profile.

The evaluation of BF1’s anti-inflammatory activity revealed weaker effects compared to those observed in a green tea water kefir beverage [11]. These results suggest that kefir grain–fermented beverages exhibit superior anti-inflammatory activity, potentially due to the microbial diversity and complex bioactive metabolites produced by these grains.

The anti-inflammatory potential of water kefir beverages may also be attributed to kefiran, a bioactive polysaccharide with well-documented anti-inflammatory properties in various scientific studies [76,77]. In vivo and in vitro research has demonstrated that kefiran reduces inflammatory cell infiltration and mitigates oxidative stress markers, which are often involved in chronic inflammatory processes [78]. Additionally, its prebiotic effect promotes gut microbiota balance, indirectly contributing to the modulation of systemic inflammation [79].

5. Conclusions

The identified strains (Lc. citreum LB4, L. lactis LB5, and S. cerevisiae Y9) showed potential probiotic properties and a strong CCP. Furthermore, using specific strains instead of traditional water kefir grains enhances the scalability and industrial applicability of fermented beverage production, potentially boosting commercial acceptance. Adjusting fermentation with these strains optimizes the production of desired molecules while minimizing unwanted by-products. The balance between LAB and yeast populations plays an essential role in influencing the physicochemical and sensory qualities of fermented beverages. Additionally, when used in combination, these strains showed anti-inflammatory properties. However, further research, particularly in vivo studies, is essential to confirm the long-term benefits of this beverage consumption and assess the bioactivity of these microorganisms.