Non-targeted metabolomics and postbiotics survey of an Iranian Koupa cheese Lactiplantibacillus plantarum isolate Authors: Somayeh Aghabeyk, Mahdi Moridi Farimani, Alireza Sadeghi, Mostafa Alilou, Atousa Aliahmadi Published: 2026 ABSTRACT Bacterial postbiotics, bioactive compounds produced by probiotic bacteria, offer enhanced stability and safety compared with live microorganisms, making them valuable candidates for functional foods and therapeutic applications. In the present study, the probiotic properties and postbiotic compounds of a Lactiplantibacillus plantarum strain isolated from a traditional Iranian cheese were investigated. The strain exhibited a 71.28% survival rate under simulated gastrointestinal (SGI) conditions and demonstrated cell surface hydrophobicity and auto-aggregation capabilities of 60.39% and 54.30%, respectively. The cell-free supernatant (CFS) of bacterial culture was extracted with methanol and subsequently fractionated using n-hexane and ethyl acetate. A bioassay-guided fractionation approach was applied to the CFS, involving the identification of the most effective antibacterial extract/fraction against selected microorganisms, the isolation of potent antibacterial fractions via Sephadex LH-20 column, and identification of the active compounds through liquid chromatography-mass spectrometry. EtOAc proved to be the most effective solvent for extracting antibacterial compounds. Fractions 4 and 5 exhibited the highest inhibitory effect against Staphylococcus aureus and Escherichia coli, respectively. Identification of nocardamine and cyclic dipeptides, specifically diketopiperazines such as cyclo(Hyp-Leu) and cyclo(Tyrosyl-Prolyl), as major bioactive constituents in these fractions underscores their notable potential for enhancing health benefits and combating pathogenic bacteria.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-51098-6. Introduction Probiotic bacteria, particularly certain lactic acid bacteria (LAB), are widely recognized to benefit human health by modulating the immune system, improving gut barrier function, and inhibiting pathogenic microorganisms1,2. Probiotics also offer protection against several disorders, including allergies, diarrhea, inflammatory bowel disease (IBD), irritable bowel syndrome (IBS), infections, infant colic, and even some cancers1,3,4. Furthermore, some strains can alleviate side effects of antibiotic treatment by reducing antibiotic-associated diarrhea and supporting microbial balance1,5. However, concerns about administering live microorganisms, such as the risk of infection and the potential transfer of antibiotic resistance genes, have prompted growing interest in postbiotics. Postbiotics are defined as non-viable bacterial products or metabolic byproducts with beneficial biological activity6,7. They provide a safer and more stable alternative for harnessing the health-promoting properties of probiotics, especially when the viability of probiotics cannot be guaranteed6,8. Beyond their immunomodulatory and health benefits, postbiotic metabolites also exhibit potent antimicrobial activities against foodborne pathogens and spoilage organisms, making them valuable for food safety and preservation9,10. These bioactive compounds, including organic acids, bacteriocins, peptides, enzymes, and siderophores, suppress bacterial growth through multiple mechanisms such as cell membrane disruption, competitive exclusion, and iron sequestration10,11. The use of postbiotics as natural preservatives has attracted considerable interest in using them as natural preservatives to reduce reliance on synthetic additives12. Research has showed postbiotic compounds from Lactiplantibacillus plantarum and other LAB strains possess promising antimicrobial properties. Yang et al. reported L. plantarum AF1, from kimchi produces δ-dodecalactone, an antifungal compound with strong inhibitory activity against mold species13. Similarly, postbiotics from Lactobacillus casei ATCC 39,392 inhibited Escherichia coli in dairy products, and their addition to sterilized milk extended shelf life under refrigerated conditions14. These findings support previous reports indicating that LAB-derived postbiotics can act as multifunctional bio-preservatives through various mechanisms including organic acid production, cell membrane disruption, and competitive exclusion15–17. Cell-free supernatants (CFS) of LAB cultures are particularly rich sources of antimicrobial metabolites, including organic acids (such as lactic and acetic acids), hydrogen peroxide, and bacteriocins peptides18,19. Among these metabolites, cyclic dipeptides (especially diketopiperazines) have garnered interest due to their broad-spectrum efficacy against both bacteria and fungi. Diketopiperazines inhibit pathogens through disrupting microbial cell membranes, inhibiting biofilm formation, and interfering with quorum sensing, making them promising candidates for natural food preservatives and therapeutic applications18,20. The identification of such metabolites often relies on liquid chromatography-mass spectrometry (LC-MS), which provides high sensitivity and selectivity, for profiling complex microbial extracts21,22. More recently, feature-based molecular networking (FBMN) has enhanced metabolite annotation and dereplication in metabolomics studies by clustering similar MS/MS spectra, facilitating efficient dereplication and structural characterization23. Koupa cheese, a traditional Iranian dairy product fermented in clay pots and buried underground for approximately three months, harbors a diverse microbiota rich in probiotic LAB. This unique fermentation environment makes Koupa cheese a valuable source of strains with potential digestive and immunomodulatory benefits24,25. In this study, we aimed to isolate and characterize postbiotic compounds from L. plantarum SA-8, a probiotic bacterium isolated from Koupa cheese using chromatographic separation techniques. Following the preparation and fractionation of CFS samples by liquid-liquid fractionation, the antimicrobial and anticancer activities of the extracts and fractions were assessed. Due to promising antibacterial activity, bioassay-guided fractionation, including determination of minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC), an agar-overlay bioautography, was performed to identify antibacterial postbiotic compounds. LC-MS combined with FBMN was then used for metabolite profiling. Materials and methods Koupa cheese was obtained from East Azerbaijan Province, located in the northwest of Iran. Calcium chloride, bile acids sodium salt, lyophilized (salt-free) pepsin from pig gastric mucosa, and pancreatin from porcine were purchased from Sigma-Aldrich (Burlington, VT, USA). Xylen, NaCl, KCl and NaHCO3 were supplied by Merck (Darmstadt, Hesse, Germany). Analytical grade solvents, including n-hexane, EtOAc, methanol (MeOH), acetonitrile (ACN), chloroform (CHCl3), ethanol, and n-butanol, were purchased from Merck. RPMI-1640 and trypsin/EDTA were obtained from Gibco. MCF-7 (a human breast cancer cell line) was purchased from National Cell Bank of Iran (Pasteur Institute, Iran). MTT [3-(4,5-dimethylthiazol-2-yl)−2,5-diphenyltetrazolium Bromide] and aflatoxins (B1, B2, G1, and G2) were obtained from Sigma-Aldrich. Sephadex LH-20 (Pharmacia Biotech AB, Uppsala, Sweden) was used for column chromatography. High performance liquid chromatography (HPLC) was performed on a Waters 2475 system (Marshall Scientific, USA) consisting of a quaternary low-pressure mixing pump with a degasser module, column oven, UV and fluorescence detectors, and an auto sampler. Isolation of the predominant cheese LAB Each cheese sample (10 g) was mixed with 90 mL of Ringer’s solution, and 100 µl of the serially tenfold diluted samples were spread on De Man-Rogosa-Sharpe (MRS) agar. Plates were incubated under anaerobic conditions for 24-h at 37 °C. Subsequently, Gram-staining and catalase test were performed on purified colonies. Preliminary microscopic examination of the pure colonies revealed various morphologies, including spherical or oval-shapes and short rod-shaped chains. Finally, all pure stock cultures obtained from rod-shaped Gram-positive bacteria were selected and preserved at −70 °C in 2X broth medium containing up to 25% glycerol, for further analysis. DNA extraction and molecular identification of selected LAB isolates Genomic DNA was extracted from LAB isolates using the Gene Transfer Pioneers (GTP, Iran) Kit following the manufacturer’s instruction. The polymerase chain reaction (PCR) was used to amplify the 16S rRNA gene of the LAB isolates using universal primers (27F 5’-AGAGTTTGATCCTGGCTCAG-3’ and 1492R 5’-TACGGTTACCTTGTTACGACTT-3’). DNA sequencing was then performed to identify the LAB species using a DNA database. The PCR products were subsequently examined through agarose gel electrophoresis, and sequencing was performed using Bioneer (South Korea). Species-level identification was achieved using the basic local alignment search tool (BLAST) algorithm. A phylogenetic evolutionary tree was constructed MEGA 6 software based on the 16 S rRNA gene sequences obtained from the Gene-Bank. Probiotic properties of L. plantarum strain Survivability in simulated gastric and intestinal (SGI) conditions The viability of the predominant LAB isolate was assessed under simulated gastric and intestinal conditions as summarized in Table 1, according to the procedure described by Zarali et al.26 and Shahryari et al.27 Table 1Simulated in vitro gastrointestinal conditions.Chemicalssimulated gastricsimulated intestine NaCl 6.21.28 g/L KCl 2.2 g/L0.239 g/L CaCl 2 0.22 g/L- NaHCO 3 1.2 g/L6.4 g/L bile salts -0.3% pH 2.06.5 Incubate 120 min at 37 °C180 min at 37 °C Pepsin 2000 U/mL- Pancreatin -200 mL Table 1Simulated in vitro gastrointestinal conditions.Chemicalssimulated gastricsimulated intestine NaCl 6.21.28 g/L KCl 2.2 g/L0.239 g/L CaCl 2 0.22 g/L- NaHCO 3 1.2 g/L6.4 g/L bile salts -0.3% pH 2.06.5 Incubate 120 min at 37 °C180 min at 37 °C Pepsin 2000 U/mL- Pancreatin -200 mL Simulated in vitro gastrointestinal conditions. Adhesion properties Cell-surface hydrophobicity The hydrophobicity of the isolate was assessed based on the ability of cells to adhere to xylene26. Bacterial cells were washed and then resuspended in phosphate buffer saline (PBS) to an absorbance (OD600) of about 0.25 ± 0.05 to achieve a standardized bacterial concentration (107−108 CFU/mL). Xylene was then added to the sample, and the mixture was vortexed for 5 min. Following incubation for 1-h, the aqueous phase was separated, and the absorption at 600 nm was measured. The percentage of hydrophobicity was determined using the following formula: Hydrophobicity = [1 – (Aₓ/A₀)] × 100. Where AX and A0 are the absorbance (600 nm) before and after mixing with xylene, respectively. Auto-aggregation. The spectrophotometry method was used to evaluate auto-aggregation28. Overnight bacterial cultures were centrifuged at 6000 × g for 20 min at 4 °C, washed three times with PBS, and adjusted to a bacterial concentration of 108 CFU/mL. The LAB isolate was then incubated at 37 °C for 24-h, after which the absorbance was measured at 600 nm. The auto-aggregation percentage was calculated by the formula: Auto-aggregation = [1 – (A₂₄/A₀)] × 100. Where A0 and A24 are the absorbance (600 nm) before and after incubation for 24-h, respectively. Co-aggregation. The co-aggregation ability of the L. plantarum strain with S. aureus (ATCC12600), E. coli (ATCC11775), B. cereus and C. albicans (ATCC10321) was assessed following the method described by Collado et al.29. Preparation of the CFS To prepare the CFS, 100 mL of MRS broth was inoculated with a single colony of the bacterium and incubated at 37 °C with shaking (100 rpm) for 24 or 48-h. The cultures were centrifuged at 12,000 rpm for 20 min at 4 °C and filtered through 0.22 μm syringe filters to obtain clarified supernatant. The resulting CFS was concentrated under reduced pressure at 40 °C using a rotary evaporator. The dried CFS residues were kept at 4 °C until further use. A control extract was prepared from sterile MRSA broth subjected to identical conditions. Extraction of CFS metabolites The dried CFS material (10 g) was suspended in 100 mL MeOH and sonicated (20 s) in a water bath. The resulting methanolic extract was evaporated (yield 9.0 g), resuspended in distilled water (90 mL), and sequentially partitioned by n-hexane (3 × 100 mL) and EtOAc (3 × 100 mL) at room temperature30. These fractions were concentrated in vacuo to afford dark gummy residues of n-hexane (0.9 g), EtOAc (6.28 g), and H2O (2.02 g). All fractions were stored at 4 °C for further chemical and in vitro biological analyses. Biological activity Antimicrobial and cytotoxic activities of the total extract and fractions were evaluated. Based on promising results of the antimicrobial assays, the antibacterial compounds in CFS from the 48-h culture of the isolated strain were further investigated. Antimicrobial activity The MIC values were determined using the broth microdilution technique in sterile 96-well plates, as recommended by Balouiri et al., with some modifications31. Samples were dissolved in sterile distilled water. Inoculums of S. aureus ATCC12600, E. coli ATCC11775, Salmonella typhi PTCC1609, Pseudomonas aeruginosa ATCC27853, Candida albicans ATCC10321, Propionibacterium acnes 98–6912, and Bacillus subtilis ATCC6051 were prepared in Mueller-Hinton broth at 0.5–1.5 × 106 CFU/mL. Plates were incubated for 20-h at 37 °C. MBCs were defined as the lowest concentration capable of killing 99.9% of the test strains, determined by culturing wells without visible growth onto nutrient agar plates. Anticancer activity The anticancer potential of various bacterial culture fractions was evaluated using the MCF-7 human breast cancer cell line. Cells were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum and antibiotics. For the MTT assay, cells were seeded in 96-well plates and treated with different concentrations (31.25 to 1000 µg/mL) of 24-h and 48-h CFS, as well as EtOAc and aqueous fractions. After incubation, MTT reduction was measured at 570 nm. The half-maximal inhibitory concentration (IC50) values were calculated from dose-response curves. Aflatoxin inhibition Aflatoxin (B1, B2, G1, and G2) reduction was evaluated based on a modified protocol from Haskard et al.32 LAB cultures were grown overnight in MRS broth at 37 °C, adjusted to 2 × 109 CFU/mL in PBS, centrifuged at 2500 × g for 10 min at < 10 °C, and washed twice with PBS (pH 7). The bacterial pellet was then resuspended in PBS (900 µL). The supernatant and pellet suspensions were incubated with aflatoxins (100 µL) at 37 °C for 4-h. Following centrifugation and filtration (0.45 μm), samples were purified through immunoaffinity columns for aflatoxin purification. Aflatoxins were eluted using methanol and analyzed by HPLC, equipped with a C18 column (5 μm, 125 × 4.0 mm) and a Kobra Cell for fluorescence detection (excitation 365 nm, emission 435 nm). The mobile phase consisted of water/MeOH/ACN (65:25:10 v/v/v), with a flow rate of 1 mL/min at 25 °C. Injection volume was 100 µL with a retention time of 25 min. Quantification was performed using calibration curves (0.4–10 ng/mL). Fractionation of EtOAc extract Based on antimicrobial activity, EtOAc fraction was selected for bioassay-guided purification. The extract (1 g) was dissolved in MeOH/CHCl3 and subjected to a Sephadex LH-20 column using MeOH-CHCl3 (70:30) as the eluent at a flow rate of 1.0 mL/min. Fractions (5 mL each) were collected, analyzed by TLC, pooled into yield 14 combined fractions. TLC bioautography Each pooled fraction (1 mg/spot) was applied to TLC plates33. Sterile molten Mueller–Hinton agar (about 45 °C) containing about 107 CFU/mL of tested microorganisms (S. aureus ATCC12600 and E. coli ATCC11775) and resazurin, as growth indicator, was transferred to plates. TLC plates were incubated at 37 °C in a sealed container for about 16-h. Chemical analysis of effective fractions LC-MS analysis Fractions with the strongest antibacterial activity (fractions 2, 4 and 5) were further analyzed by a Vanquish UPLC system coupled with an Exploris-120 mass spectrometer (Thermo Scientific, USA). Chromatographic separation was conducted on an Accucore C18 column (150 × 2.1 mm, 2.6 μm) at 35 °C using a gradient of H2O + 0.1% FA (A) and ACN + 0.1% FA (B) from 5% to 99% B in 20 min, at a constant flow rate of 0.4 mL/min. The injection volume was 1 µL, and samples were prepared at a concentration of 5 mg/mL. MS data were acquired in positive electrospray ionization mode using the following parameters; spray voltage: 3200 V, sheath gas (N2): 45 arbitrary unit, auxiliary gas (N2): 17 arbitrary units, ion transfer tube temperature: 350 °C, vaporizer temperature: 400 °C, and RF Lense: 70%. Full MS scans (MS1) covered a mass range of 150–1500 m/z at a resolution of 60,000 (FWHM) in centroid mode, with one microscan per spectrum. For structural elucidation, data dependent MS² spectrometry (dd-MS2) was performed, automatically selecting the top three most abundant precursors per cycle for fragmentation. Higher-energy collisional dissociation (HCD) was applied with stepped normalized collision energies of 15, 30, and 45%. Product ion spectra were acquired at 15,000 resolutions (FWHM) in centroid mode. Additional settings included an intensity threshold (1.5 × 10⁵), dynamic exclusion after single trigger (duration: 3 s, mass tolerance ± 10 ppm), apex trigger at 30%, isotopic pattern exclusion filter (assigned), restriction to singly charged ions (z = 1). Data Processing and Molecular Networking Raw MS/MS data of the biologically active subfractions were processed via MZmine 4.6 software34. The resulting aligned feature list were uploaded to the GNPS Web platform (http://gnps.ucsd.edu) for generation of an Ion Identity Molecular Network35. The network was built with the following settings: precursor mass tolerance of 0.02 Da, MS/MS fragment ion tolerance of 0.02 Da, minimum cosine score of 0.7, at least five matched fragment ions, minimum cluster size of 2, and a TopK value of 10 for network edges. The resulting molecular network was exported and visualized in Cytoscape 3.8.2 version for further interpretation36. The GNPS job is publicly accessible at: https://gnps.ucsd.edu/ProteoSAFe/status.jsp?task=3cecf1e267e8433da451783f99255f11. To further characterize the structural classes of the potential bioactive compounds, the aligned feature list was additionally analyzed using SIRIUS37. For annotation refinement, the data were filtered using following thresholds: minimum ZodiacScore of 0.9, minimum ConfidenceScore of 0.1, and minimum class-level confidence of 0.8. The filtered results, including molecular formulas and predicted compound classes, were exported as a.tsv file and used to generate compound classes via a custom Python script (Fig. 7B). Statistical analysis The diameters of the inhibition zones in each image were measured using Image J software (Image Processing and Analysis in Java software, version 1.51, 2018, from NIH, Bethesda, MD, USA). Each fraction was measured multiple times to ensure accuracy. Differences among fractions were evaluated using one-way analysis of variance (ANOVA). Results and discussion Identification of the predominant LAB Isolate Sequencing of the 16 S rRNA gene revealed that the predominant isolate (SA-8) belongs to the species L. plantarum. The corresponding sequence has been deposited in the NCBI database under the accession number PV776685.1. The phylogenic tree illustrating the taxonomic position of SA-8 is presented in Fig. 1. Additionally, 16 S rRNA gene sequencing was utilized to identify other bacterial isolates obtained and purified under anaerobic conditions on MRS plates. These isolates were assigned the GenBank accession numbers of OR157959, OR157958, OR157957, and OR145146. Fig. 1Phylogenetic tree derived from 16 S rRNA gene sequences, giving the position of the L. plantarum SA-8 isolate among the cheese LAB. (Visualization of Neighbor joining tree from NCBI BLAST alignments using BLAST Tree View tool.). Phylogenetic tree derived from 16 S rRNA gene sequences, giving the position of the L. plantarum SA-8 isolate among the cheese LAB. (Visualization of Neighbor joining tree from NCBI BLAST alignments using BLAST Tree View tool.). Probiotic characterization of L. plantarum SA-8 The survival rate of L. plantarum SA-8 under SGI conditions was 71.28% ± 0.78. This value exceeds the commonly accepted threshold of more than 70% survival for candidate probiotic strains. The survival rate through the harsh physicochemical conditions of the gastrointestinal tract is a critical requirement for the selection of probiotic strains, as tolerance to bile salts and acidic pH strongly influences bacterial colonization and metabolic activity in the host intestine38. These conditions are particularly challenging due to the combination of gastric acidity and bile salt–mediated stress. Several studies have documented comparable or higher tolerance levels in LAB isolated from fermented foods. For example, L. plantarum BR4, isolated from various Egyptian cheeses showed the survival rate of 85% under simulated gastrointestinal conditions (0.3% bile salts after 3-h)39. Likewise, among 45 strains of LAB isolated from Feta-type cheese, the Lactobacillus paracasei K5 showed enhanced resistance to low pH, pepsin, pancreatin, and bile salts, outperforming the commercial probiotic L. Plantarum40. Moreover, five LAB strains isolated from fermented rose jam demonstrated excellent survival at pH 2 and 3. These studies support the observation that food-derived LAB frequently exhibit significant resistance to digestive system conditions41. The cell-surface hydrophobicity of the LAB isolate was measured at 60.39% ± 2.84. Zommara et al. reported a comparable hydrophobicity value (66.23%) for isolate L. plantarum BR4 obtained from cheese39. Significant hydrophobic potential (31.7–95.5%) have also been observed in the probiotic L. plantarum isolated from naturally fermented tangerine vinegar42. Cell-surface hydrophobicity reflects non-specific interactions between microbial cells and the host epithelial surfaces. Strains with high hydrophobicity generally adhere more effectively to mucosal cells43,44, enhancing their capacity to grow and survive in the human gastrointestinal tract. Auto-aggregation capability of the LAB isolate was determined to be 54.30% ± 4.13. This result aligns closely with the percentage of auto aggregation values reported for LAB strains from Algerian dairy products (53% to 72%), indicating a good self-aggregation capacity45. Auto-aggregation is closely linked to colonization potential of LAB, as it enhances the organism’s capacity to establish and maintain its presence in the gastrointestinal tract46. The results obtained indicated that SA-8 strain could aggregate effectively on the surface of the host cell. This property may help prevent pathogen colonization through competitive exclusion. Co-aggregation assays revealed that L. plantarum SA-8 effectively interacted with several studied foodborne pathogens. The highest percentage of co-aggregation was observed with C. albicans (48.24 ± 5.80), followed by E. coli (42.36 ± 3.55) and B. cereus (39.78 ± 6.12) (Fig. 2). The lowest percentage of aggregation was observed with S. aureus bacteria (11.5 ± 7.09), which differed significantly (p < 0.05) from those of the other tested pathogens. The strong tendency of strain LAB to co-aggregate with Candida species may play a role in their probiotic potential by limiting the colonization of pathogenic yeasts. Notably, SA-8 strain exhibited stronger co-aggregation activity compared to LAB isolates obtained from other fermented products39,47. Co-aggregation is recognized as an important probiotic inhibitory mechanism, as it enables beneficial bacteria to physically interact with pathogens and limit their access to epithelial cells and nutrients26. This interaction can thereby reduce pathogen invasion and may contribute to the overall antimicrobial mechanisms in potential probiotic bacteria. Fig. 2Comparison of the co-aggregation activity percentage of the L. plantarum SA-8 isolate with assessed foodborne pathogenic bacteria and yeast. The different letters indicate significant difference (p < 0.05) among the studied foodborne pathogens in terms of co-aggregation with the isolate, which was determined using ANOVA statistical analysis and LSD post-hoc with three replicates. Comparison of the co-aggregation activity percentage of the L. plantarum SA-8 isolate with assessed foodborne pathogenic bacteria and yeast. The different letters indicate significant difference (p < 0.05) among the studied foodborne pathogens in terms of co-aggregation with the isolate, which was determined using ANOVA statistical analysis and LSD post-hoc with three replicates. Biological activities The antibacterial activity of the bacterial culture was evaluated via three consecutive steps. First, the most active extract was identified based on the MIC and MBC values against selected bacterial pathogens and one yeast strain. Among the tested samples, the EtOAc extract obtained from the 48-h CFS showed the strongest antibacterial activity. This extract was subsequently subjected to column chromatography, and the resulting fractions were screened using an agar-overlay assay to find the most active fractions. As shown in Table 2, the lowest MIC values were observed for methanolic extract of the 48-h culture supernatant and the EtOAc fraction. Accordingly, the EtOAc fraction was selected for further characterization of proposed antibacterial compound(s). Among tested microbial strains, the EtOAc fraction showed the greatest susceptibility against S. aureus, P. aeruginosa, B. subtilis, and P. acne, each displaying MIC values of 4 mg/mL. However, MBC values of this fraction were approximately 8 mg/mL against P. aeruginosa, B. subtilis and P. acne. Table 2MIC and MBC values (mg/mL) of CFS extracted/fractioned materials, against selected human pathogens.StrainsMethanolic extract of CFSn-HexanefractionEtOAcfractionAqueous fractionStandard antimicrobial agents*24-h48-h MIC MBCMICMBC MIC MBC MICMBCMICMBCMICMBCS. aureus ATCC126008> 6486432> 6441664> 64416E. coli ATCC1177532> 6416> 6464> 6416> 6464> 640.51S. typhi PTCC160916> 648> 6464> 6486464> 64816P. aeruginosa ATCC278531616816> 64> 644864> 6432128P. acne 98–6912163241664> 64481664432B. subtilis ATCC6051163281632> 644864> 64132C. albicans ATCC1032164643264> 64> 64163264> 6464256*Cefixime for bacteria and Nystatin for C. albicans, all in µg/mL. Table 2MIC and MBC values (mg/mL) of CFS extracted/fractioned materials, against selected human pathogens.StrainsMethanolic extract of CFSn-HexanefractionEtOAcfractionAqueous fractionStandard antimicrobial agents*24-h48-h MIC MBCMICMBC MIC MBC MICMBCMICMBCMICMBCS. aureus ATCC126008> 6486432> 6441664> 64416E. coli ATCC1177532> 6416> 6464> 6416> 6464> 640.51S. typhi PTCC160916> 648> 6464> 6486464> 64816P. aeruginosa ATCC278531616816> 64> 644864> 6432128P. acne 98–6912163241664> 64481664432B. subtilis ATCC6051163281632> 644864> 64132C. albicans ATCC1032164643264> 64> 64163264> 6464256*Cefixime for bacteria and Nystatin for C. albicans, all in µg/mL. MIC and MBC values (mg/mL) of CFS extracted/fractioned materials, against selected human pathogens. *Cefixime for bacteria and Nystatin for C. albicans, all in µg/mL. Previous studies have demonstrated that LAB derived from various fermented dairy products frequently possess notable antibacterial activity. For example, an investigation of LAB isolated from 11 types of Brazilian cheeses showed that 75% of isolates inhibited the growth of pathogenic bacteria such as S. aureus48. Additionally, five LAB strains isolated from artisanal Colombian cheese exhibited strong probiotic properties and broad-spectrum antibacterial effects against foodborne pathogens49. The results of the present study highlight the efficiency of EtOAc as an effective solvent for extracting antibacterial compounds from the CFS of bacterial cultures. EtOAc is known to preferentially extract semi-polar to moderately non-polar compounds, many of which possess significant antibacterial properties. Comparison of extracts from 24-h and 48-h cultures indicates that the EtOAc fraction from the 48-h CFS displayed stronger antibacterial properties, suggesting that longer incubation periods promote higher production of secondary metabolites with antimicrobial potential. The MRS broth control showed no antimicrobial activity, with MIC values exceeding 64 mg/mL for all tested microorganisms. According to the results of the preliminary antimicrobial assays (Table 2), the EtOAc extract was selected for fractionation on a Sephadex LH-20 column. All resulting fractions were evaluated for their inhibitory effect against S. aureus and E. coli (Fig. 3). ANOVA analysis showed significant differences among the mean inhibition zones produced by the different fractions. Clear inhibition zones (blue zones indicating complete absence of bacterial growth) were observed around seven fractions (F1-F7). As shown in Fig. 3, fraction 4 showed the strongest antibacterial effect against S. aureus, significantly (p < 0.05) surpassing all other fractions. Fraction 2 also demonstrated notably strong inhibition against S. aureus, while fraction 5 showed the highest inhibitory effect against E. coli. Fig. 3Antibacterial effects of Sephadex column fractions against two model bacterial strains using agar-overlay method. (A) Growth inhibition results of seven fractions (B) selected fractions (1 mg/spot) showing different level of bacterial growth inhibition on silica plate. Probiotics might exert both preventive and therapeutic influences on breast cancer cells, largely through modulation of gut microbiota and host immune system50. Previous studies have indicated that L. plantarum can directly inhibit the proliferation of MCF-7 breast cancer cells without causing damage to healthy cells51. In the present study, cytotoxicity assays revealed that the 24-h CFS showed slightly higher cytotoxicity against MCF-7 cells (IC50: 8.48 mg/mL) than the 48-h CFS (9.39 mg/mL), indicating moderate antiproliferative activity (Fig. 4). Antibacterial effects of Sephadex column fractions against two model bacterial strains using agar-overlay method. (A) Growth inhibition results of seven fractions (B) selected fractions (1 mg/spot) showing different level of bacterial growth inhibition on silica plate. Probiotics might exert both preventive and therapeutic influences on breast cancer cells, largely through modulation of gut microbiota and host immune system50. Previous studies have indicated that L. plantarum can directly inhibit the proliferation of MCF-7 breast cancer cells without causing damage to healthy cells51. In the present study, cytotoxicity assays revealed that the 24-h CFS showed slightly higher cytotoxicity against MCF-7 cells (IC50: 8.48 mg/mL) than the 48-h CFS (9.39 mg/mL), indicating moderate antiproliferative activity (Fig. 4). Fig. 4Cell viability of MCF-7cells following 48-h treatment with 31.25, 62.5, 125, 250, 500, 1000 µg/mL solutions of 24-h and 48-h CFS. One effective biological method to reduce aflatoxin exposure is the incorporation of probiotic yeasts and bacteria into the diet. These microorganisms can interact with aflatoxins in the gastrointestinal tract, binding the toxins and limiting their absorption, thereby reducing their negative effects on animal and human health52. In this study, the anti-aflatoxigenic capacity of the bacterial culture, pellet, and supernatant was investigated. Aflatoxin concentrations were quantified using a standard HPLC method and toxin reduction in each sample was determined at ppb level according to the regression equation. HPLC analyses indicated that the LAB isolate lowered the initial concentrations of aflatoxins G2, G1, B2, and B1 (Fig. 5). Cell viability of MCF-7cells following 48-h treatment with 31.25, 62.5, 125, 250, 500, 1000 µg/mL solutions of 24-h and 48-h CFS. One effective biological method to reduce aflatoxin exposure is the incorporation of probiotic yeasts and bacteria into the diet. These microorganisms can interact with aflatoxins in the gastrointestinal tract, binding the toxins and limiting their absorption, thereby reducing their negative effects on animal and human health52. In this study, the anti-aflatoxigenic capacity of the bacterial culture, pellet, and supernatant was investigated. Aflatoxin concentrations were quantified using a standard HPLC method and toxin reduction in each sample was determined at ppb level according to the regression equation. HPLC analyses indicated that the LAB isolate lowered the initial concentrations of aflatoxins G2, G1, B2, and B1 (Fig. 5). Fig. 5HPLC chromatograms comparing standard pure aflatoxins (A), untreated control (B: PBS+aflatoxins), and treated samples aflatoxins incubated with whole culture (C), supernatant (D), pellet (E). HPLC chromatograms comparing standard pure aflatoxins (A), untreated control (B: PBS+aflatoxins), and treated samples aflatoxins incubated with whole culture (C), supernatant (D), pellet (E). Among all tested samples, the bacterial pellet exhibited the highest aflatoxin-reducing capability. It showed particularly significant effect on aflatoxins B2 (96.1% reduction) and G2 (94.5% reduction) (Fig. 6). Fig. 6Comparison of reduction percentage of aflatoxins (G2, G1, B2, B1) by bacterial culture, supernatant and pellet of LAB isolate.Numerous investigations on probiotic bacteria isolated from traditional fermented dairy products have demonstrated their capacity to significantly attenuate the concentration of aflatoxin B1. The observed variability in toxin-binding efficiency among the strains further suggests the binding sites among different probiotic strains53. Comparison of reduction percentage of aflatoxins (G2, G1, B2, B1) by bacterial culture, supernatant and pellet of LAB isolate.Numerous investigations on probiotic bacteria isolated from traditional fermented dairy products have demonstrated their capacity to significantly attenuate the concentration of aflatoxin B1. The observed variability in toxin-binding efficiency among the strains further suggests the binding sites among different probiotic strains53. Preliminary analysis of the secondary metabolites content To analyze the chemical diversity of LAB isolate and preliminarily dereplicate its potentially active constituents, Feature-Based Molecular Networking (FBMN) was performed using the GNPS platform54. FBMN groups metabolites into clusters based on the similarities in their MS/MS fragmentation patterns, providing insights into the structures of unknown compounds through annotated molecular features55. Based on the TLC bioautography results against S. aureus and E. coli (Fig. 3), three fractions (F2, F4, and F5) were selected for LC-MS analysis followed by FBMN- and SIRIUS-based dereplication studies. SIRIUS predictions indicated that most detected metabolites are from alkaloids and peptides classes (Fig. 7B). These results were consistent with the metabolite dereplication results from GNPS database. Several metabolite clusters containing alkaloids and peptides were identified, including clusters annotated as 2,5-diketopiperazine derivatives and nocardamine-related compounds (Fig. 7A and Fig. S1), which may contribute potentially to the observed bioactivities. The corresponding MS/MS spectra of the identified compounds (query vs. library spectrum) were included in the supplementary, along with cosine scores and retention times (Fig. S2 [A-C]). The high cosine scores (more than 0.7) indicate a great match between acquired and library MS/MS spectra. Moreover, a relative abundance of each of the potential active compounds were calculated across the active fractions (Fig. S2 [A-C]). Fig. 7FBMN of three most active fractions F2 (Green), F4 (Orange) and F5 (Blue). A: Examples of the FBMN clusters obtained from GNPS and annotated node, B: Compound class analysis summary obtained from SIRIUS for all three fractions. FBMN of three most active fractions F2 (Green), F4 (Orange) and F5 (Blue). A: Examples of the FBMN clusters obtained from GNPS and annotated node, B: Compound class analysis summary obtained from SIRIUS for all three fractions. 2,5-Diketopiperazines (DKPs), the smallest cyclic peptides found in nature, consist of two α-amino acids linked through a head-to-tail peptide bond. Due to their structural simplicity, chemical stability, and biological properties, these compounds have attracted significant attention as scaffolds for developing new therapeutic agents56. Several DKPs previously reported from LAB, such as cyclo(Gly-Leu), cyclo(Phe-Pro), and cyclo(Phe-4-hydroxy-Pro), are known for their antifungal effects57. Both cyclic dipeptides (DKPs) detected in the present study are consistent with these earlier findings and likely contribute to the observed antibacterial activity. Nocardamine is a cyclic, non-ribosomal peptide antibiotic that was originally isolated from a marine sponge. Based on the dereplication data, a cluster of nocardamine, including various derivatives such as dehydroxy and desmethylenyl, were identified58. This type of compounds are particularly notable for their strong ferric ions (Fe+³) chelating capacity (siderophore), which enhances its biological efficacy under iron-deficient conditions. Nocardamine has been isolated from various bacterial strains, including Pseudomonas stutzeri, Streptomyces hygroscopicus and Nocardia spp., and showed considerable antibacterial activity, especially against Mycobacterium species. Its dual antibacterial and iron-binding properties make nocardamine promising candidate for the development of new drugs and a valuable research tool for investigating various biological mechanisms and host–pathogen interactions59. Considering these findings, both DKPs and nocardamine derivatives can potentially explain the antibacterial activity observed in the LAB extract against S. aureus and E. coli. Nonetheless, detailed structural and functional investigations are required to fully elucidate the specific metabolites responsible for these bioactivities. Conclusion In this study L. plantarum SA-8 strain was isolated from a traditional dairy source, identified and characterized as a probiotic bacterial strain. The antibacterial substances of the cell-free supernatant; CFS of bacterial culture were investigated through bioassay-guided fractionation. Initially, MIC and MBC assays identified the most active extract, with the EtOAc fraction of the 48-h CFS showing strongest antibacterial activity. This fraction was further separated by Sephadex LH-20 chromatography. Agar-overlay bioautography revealed the most active fractions, which were analyzed by LC-MS/MS analysis, followed by non-targeted metabolomics. Preliminary dereplication studies using FBMN and SIRIUS revealed diketopiperazine and nocardamine derivatives as key bioactive constituents, highlighting their potential in mediating the observed antibacterial effects. These findings underscore the valuable role of traditional dairy products as sources of beneficial postbiotic compounds with application in health-promoting formulations. While the extracts exhibited no cytotoxicity against the MCF-7 human breast cancer cell line, the bacterial biomass demonstrated significant anti-aflatoxigenic activity, particularly against aflatoxins B2 and G2, indicating its potential as a postbiotic agent for mycotoxin detoxification. Collectively, these results highlight the richness of traditional fermented dairy products as reservoirs of beneficial microorganisms and bioactive metabolites. Strain SA-8 and its postbiotic fractions merit further development for applications in functional foods, probiotic/postbiotic supplements, and natural antimicrobial strategies. However, further investigations utilizing robust methodologies, biotechnological approaches to ensure bioequivalence, and adherence to good manufacturing practices are essential to accurately identify the specific compounds responsible for these effects, elucidate their mechanisms of action, and investigate their safety. Such future endeavors will deepen our understanding of L. plantarum’s applications and enhance its potential as a safe postbiotic with antimicrobial and anti-aflatoxigenic properties. Electronic Supplementary Material Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1 Supplementary Material 2 Supplementary Material 2