Metabolic and Safety Characterization of Lactiplantibacillus plantarum Strains Isolated from Traditional Rye Sourdough

Abstract

Lactiplantibacillus plantarum is a versatile lactic acid bacterium (LAB) with broad ecological and metabolic adaptability, contributing to both technological and probiotic functions. The prevalence and functional diversity of locally adapted L. plantarum strains in traditional sourdough fermentations remain poorly understood. This study aimed to characterize ten L. plantarum strains isolated from traditional Estonian rye sourdoughs, focusing on safety, enzymatic and carbohydrate metabolism, fermentation performance, exopolysaccharide (EPS) production, and genotype-associated functional diversity, including interaction with aflatoxin B₁ (AFB₁). Strains were γ-hemolytic and susceptible to major antibiotics. Strong aminopeptidase and β-glucosidase activities were observed, whereas α-glucosidase and α-galactosidase activities varied among strains and genotypes. Strains efficiently utilized mono- and disaccharides, with genotype-specific patterns for complex carbohydrates. During sourdough fermentation, all strains acidified the dough (pH < 4.5) and produced lactic and acetic acids in optimal ratios, while fermentation kinetics differed in a genotype-dependent manner. EPS yields (131–225 mg/L) were stable across genotypes. All strains retained fermentative activity in the presence of AFB₁ and exhibited high binding capacity (~100%). These findings demonstrate the safety, metabolic versatility, and genotype-structured functional diversity of L. plantarum from traditional Estonian rye sourdoughs, supporting their application as robust, multifunctional starter cultures for sustainable food fermentations.

1. Introduction

Lactiplantibacillus plantarum (formerly Lactobacillus plantarum) is a Gram-positive lactic acid bacterium (LAB) renowned for its exceptional ecological and metabolic versatility [1]. It exhibits a nomadic lifestyle, acting as a dominant member of microbial communities in spontaneous plant-based fermentations and frequently occurring in fermented dairy and meat products [2,3]. Its ecological flexibility also enables colonization of the gastrointestinal tract in both humans and animals [4].

L. plantarum is facultatively heterofermentative, employing both the Embden–Meyerhof–Parnas and phosphoketolase pathways for carbohydrate metabolism. This metabolic adaptability allows the bacterium to shift between homolactic and heterolactic fermentation depending on substrate availability and environmental conditions [5]. Beyond lactic acid and ethanol, L. plantarum synthesizes a diverse array of secondary metabolites whose qualitative and quantitative profiles are strain-dependent and modulated by growth conditions, including oxygen availability [6]. Compared with many other LAB, L. plantarum can utilize a remarkably broad spectrum of carbohydrates, including fructo- and galactooligosaccharides, thereby enhancing its probiotic functionality and technological versatility [4]. This extensive metabolic capacity is supported by a relatively large genome of approximately 3.3 million base pairs, reflecting substantial intraspecific diversity [3,7]. Carbohydrate utilization patterns differ among strains owing to variations in transport systems, metabolic pathways, and strain-specific genes [8]. Even isolates originating from the same ecological niche may exhibit pronounced genotypic and phenotypic heterogeneity, metabolizing distinct substrate spectra under comparable environmental conditions [9].

Among sourdough-associated LAB, L. plantarum is one of the most frequently detected species, together with Fructilactobacillus sanfranciscensis and Levilactobacillus brevis [10], underscoring its pivotal role in traditional fermentations and sustainable food systems. However, the increasing industrial application of standardized, non-local starter cultures, coupled with the global decline in microbial diversity, has reduced the prevalence of locally adapted strains [11,12,13]. This trend threatens the authenticity, sensory integrity, and microbial safety of heritage fermented foods. Locally adapted LAB populations, shaped by long-term coevolution with specific raw materials and fermentation environments, often harbor unique metabolic and functional traits that are essential for preserving the distinctive characteristics and microbial resilience of traditional foods.

Despite the extensive characterization of L. plantarum across diverse food ecosystems, limited attention has been paid to how environmental contaminants, particularly mycotoxins, influence its metabolic activity. Beyond its fermentative performance, the functional potential of L. plantarum strains depends on their ability to withstand environmental stressors and to synthesize metabolites that contribute to product quality and technological functionality. For instance, exopolysaccharide (EPS) production enhances the texture, stability, and overall technological performance of sourdoughs [14], whereas interactions with mycotoxins such as aflatoxins are critical for ensuring food safety and maintaining the ecological resilience of fermentative microbiota. Although the detoxification capacity of LAB toward aflatoxins has been described [15], little is known about how these toxins affect the metabolic activity of the bacteria themselves.

Therefore, this study aims to characterize the safety, enzymatic, fermentative, EPS-producing, and mycotoxin-binding properties of L. plantarum strains isolated from traditional Estonian rye sourdoughs, providing insights into their functional diversity and potential for application as robust starter cultures and biotechnological agents in sustainable food systems.

2. Materials and Methods

2.1. Bacterial Strains

Ten Lactiplantibacillus plantarum strains (

Table 1. Lactiplantibacillus plantarum strains used in this study and their corresponding genotypes.

Genotype	Isolate
G1	L. plantarum A38
	L. plantarum A57
G5	L. plantarum A45
	L. plantarum A52
G6	L. plantarum A44
	L. plantarum A51
G9	L. plantarum A26
	L. plantarum A37
G12	L. plantarum A10
	L. plantarum A71

) used in this study were previously isolated from industrial rye bread sourdoughs and genotypically characterized by ITS-PCR and ITS-PCR/TaqI restriction analysis as described by Lutter et al. [16]. Based on the size and combination of ITS-PCR fragments and their TaqI restriction patterns, the strains were assigned to five genotypes (G1, G5, G6, G9, and G12), each representing a well-resolved and reproducible molecular profile. The cultures were maintained in de Man–Rogosa–Sharpe (MRS) broth (Biolife Italiana, Monza, Italy) containing 50% (v/v) sterile glycerol and stored at −80 °C until analysis.

2.2. Safety Assessment

2.2.1. Hemolytic Activity

The hemolytic activity of L. plantarum strains was assessed by streaking the cultures onto 5% defibrinated horse blood agar (Oxoid Ltd., Thermo Fisher Scientific, Basingstoke, UK) and incubating the plates at 35 °C for 48 h under microaerophilic conditions (10% CO₂). Colonies were classified as β-hemolytic (clear zone), α-hemolytic (greenish zone), or γ-hemolytic (no zone). All assays were performed in triplicate.

2.2.2. Antibiotic Susceptibility

The antibiotic susceptibility of each L. plantarum strain was determined on Mueller-Hinton agar (Biolife Italiana, Monza, Italy) using Etest^® Strips (bioMérieux SA, Marcy-l’Étoile, France) according to the manufacturer’s protocol. Bacterial suspensions were adjusted to McFarland standard 2.0 (≈6.0 × 10⁸ cfu/mL), and 10 µL aliquots were inoculated in triplicate onto agar surfaces. Plates with Etest^® strips were incubated at 35 °C for 24 h under microaerobic conditions (10% CO₂). Susceptibility to ampicillin, chloramphenicol, clindamycin, tetracycline, erythromycin, vancomycin, gentamicin, kanamycin, and streptomycin was evaluated within a range of 0.002 to 32 µg/mL. Minimum inhibitory concentrations (MICs) were read directly from the Etest^® scale and interpreted using the cut-off values established by the European Food Safety Authority [17].

2.3. Enzymatic, Biochemical, and Fermentative Properties of L. plantarum Strains

2.3.1. Enzymatic Activity

Enzymatic activities of L. plantarum strains were determined using the semi-quantitative API ZYM system (bioMérieux SA, Marcy-l’Étoile, France) according to the manufacturer’s instructions. The test strips were incubated at 37 °C for 4 h under aerobic conditions, and enzymatic activities were scored on a colorimetric scale from 0 to 5, where 0 indicated no activity, 1–2 weak activity (5–20 nmol of substrate hydrolyzed), and 3–5 strong activity (>20 nmol of substrate hydrolyzed).

2.3.2. Carbohydrate Utilization

The carbohydrate fermentation profiles of L. plantarum strains were determined using the API 50 CHL system (bioMérieux SA, Montalieu-Vercieu, France) according to the manufacturer’s instructions. The inoculated strips were incubated at 30 °C for 48 h, and fermentation reactions were recorded as positive (yellow) or negative (no color change).

2.3.3. Rye Sourdough Fermentation and Analyses

For fermentation assays, rye dough medium was prepared by mixing 400 g of rye flour (KALEW, Tartu Mill AS, Tartu, Estonia) with 800 mL of distilled water. One milliliter of each bacterial suspension (3.0 × 10⁸ cfu/mL) was added to 1200 mL of medium, resulting in an initial bacterial density of approximately 2.5 × 10⁵ cfu/mL. Fermentations were carried out in an Applikon my-Control fermenter (Applikon Biotechnology, Delft, The Netherlands) at 30 °C for 48 h with continuous stirring (10 rpm). Enumeration of lactic acid bacteria in the fermented sourdoughs was performed using the pour plate technique. One milliliter of sample was transferred into 9 mL of sterile 0.9% (w/v) NaCl solution to obtain a tenfold dilution. The suspension was homogenized using a vortex mixer (Vortex Genie 2 Mixer, Scientific Industries Inc., Bohemia, NY, USA), and serial tenfold dilutions were subsequently prepared. One milliliter of each dilution was plated in duplicate into sterile Petri dishes and overlaid with MRS agar at 50 °C (Biolife Italiana, Monza, Italy). Plates were incubated for 48 h at 35 °C, after which colonies were enumerated and expressed as log cfu/mL of sample. pH, temperature, and CO₂ levels were automatically monitored, and all fermentations were performed in triplicate. Physicochemical parameters of fermented and unfermented doughs were analyzed in triplicate, including soluble solids (°Brix) and titratable acidity (TTA). TTA was determined by titrating 5 g of sample diluted in 40 mL of distilled water and expressed in Thörner degrees (°Th), where 1 mL of 0.1 M NaOH required to neutralize 100 mL of sample corresponds to 1 °Th [18]. Soluble solids were measured using a refractometer (DR-A1, Atago Co. Ltd., Tokyo, Japan) according to ISO 2173:2003 [19]. Concentrations of lactic acid, acetic acid, and ethanol were determined by gas chromatography (Agilent 7890A, Agilent Technologies, Santa Clara, CA, USA) with a packed glass column (1.83 m × 2 mm ID) containing 4% Carbowax 20 M on 80/120 Carbopack B-DA (Supelco Inc., Bellefonte, PA, USA) and a flame ionization detector. For analysis, 50 g of each homogenized sourdough sample was mixed with 100 g of distilled water, incubated at room temperature for 1 h, and filtered. One milliliter of filtrate was combined with 100 µL 1% pivalic acid (internal standard) and 100 µL 0.36 M oxalic acid (to denature proteins), centrifuged at 5000 rpm for 10 min, and 1.5 mL of supernatant was transferred to vials. Metabolite concentrations were quantified based on peak areas at the corresponding retention times.

2.4. Quantification of Exopolysaccharides

EPS extraction was performed according to a modified method described by Kavitake et al. [20]. Cell suspensions of L. plantarum strains were adjusted to McFarland standard 2.0. An aliquot of 5 mL from each suspension was inoculated into 45 mL of MRS broth supplemented with 2% (w/v) sucrose and incubated for 48 h at 35 °C under 10% CO₂. After incubation, samples were heated for 10 min at 100 °C, cooled in cold water for 5 min, and centrifuged (4000× g, 15 min, 4 °C; Sigma 3-18KS, Germany). Supernatants were collected and trichloroacetic acid (TCA) was added to 2% (w/v) final concentration, incubated for 24 h at 4 °C, and centrifuged again (4000× g, 30 min, 4 °C). Three volumes of cold ethanol (−20 °C) were added to the supernatants to precipitate exopolysaccharides (EPS), followed by 24 h incubation at 4 °C and centrifugation (4000× g, 20 min, 4 °C). The resulting EPS pellets were dissolved in distilled water and dialyzed using 12–14 kDa cut-off membrane (SpectraPor^® Regenerated Cellulose Dialysis Membrane, Repligen, Waltham, MA, USA) for 48 h at room temperature with three water changes. Dialyzed samples were frozen (−20 °C), lyophilized (CHRIST alpha 3-4LSCbasic, Osterode am Harz, Germany; −108 °C, 0.3 bar, 48 h), and weighed to determine dry EPS mass.

The sugar content of EPS was quantified spectrophotometrically using the phenol–sulfuric acid method [21] and expressed as µg of glucose equivalents per mg of EPS. Lyophilized EPS (5 mg) was dissolved in 5 mL of distilled water to prepare a 1 mg/mL stock solution. For the assay, 250 µL of this solution was diluted to 250 µg/mL, mixed with 1 mL of concentrated sulfuric acid, and incubated at 95 °C for 15 min. After cooling, 250 µL of 5% (w/v) phenol was added, and the mixture was left to stand for 30 min. Absorbance was measured at 490 nm in triplicate using a 96-well microplate reader (Bio Tek Epoch 2 Microplate Spectrophotometer, Agilent Technologies, CA, USA).

2.5. Effect of Aflatoxin B₁ on Fermentative Properties of L. plantarum Strains

The effect of aflatoxin B₁ (AFB₁) on the fermentative activity of L. plantarum strains was evaluated according to the method of Sutic and Banina [22]. Culture media were prepared using BBL™ Phenol Red Broth Base (Becton, Dickinson, France) supplemented with 1% of the respective carbohydrate (glucose, galactose, lactose, or sucrose). Media containing glucose or galactose were autoclaved at 121 °C for 15 min, and those with lactose or sucrose for 3 min. Each tube containing 5 mL of medium and an inverted Durham tube was inoculated with 100 µL of bacterial suspension adjusted to McFarland standard 2.0. AFB₁ (Aflatoxin B₁ crystalline, Acros Organics, Geel, Israel) was added at final concentrations of 1, 10, or 20 µg/mL. A control without AFB₁ was included, and all treatments were performed in triplicate. Escherichia coli WDCM 0020 was used in duplicate as a reference strain to verify fermentation performance in media without AFB₁. All samples were incubated at 30 °C for 12 days. Acid and gas production were monitored after 24, 48, and 72 h, and on days 5 and 12; acid production was indicated by a color change of the phenol red indicator (red to yellow), and gas formation by bubble accumulation in the inverted Durham tube.

2.6. Aflatoxin B₁ Binding Ability

The AFB₁ binding capacity of L. plantarum strains was evaluated after two subcultures in MRS broth. Each strain was cultivated individually in triplicate. Following the second subculture, cells were harvested by centrifugation (2000× g, 10 min, 4 °C), washed three times with phosphate-buffered saline (PBS; pH 7.2; Sodium Chloride, Honeywell, Seelze, Germany), and adjusted to 10⁸ cells/mL in PBS. A portion of the cells was heat-killed by autoclaving at 121 °C for 30 min, while the rest remained viable.

For binding assays, 990 µL of either viable or heat-killed cell suspension was mixed with 10 µL of AFB₁ (100 µg/mL; Aflatoxin B₁ crystalline, Acros Organics, Geel, Israel) in 1.5 mL tubes. The mixtures were incubated at 37 °C for 30 min with shaking (200 rpm), centrifuged, and the supernatants were collected for ELISA analysis. AFB₁ concentrations were quantified using an Aflatoxin B₁ ELISA Test Kit (EuroProxima, Arnhem, The Netherlands) according to the manufacturer’s instructions. Each strain was analyzed in six replicates, and absorbance was measured at 450 nm using a spectrophotometer (BioTek Epoch 2, Agilent Technologies, Santa Clara, CA, USA).

The percentage of AFB₁ bound by viable and heat-killed cells was calculated using the following equation:

$${\mathrm{A}\mathrm{F}\mathrm{B}}_1 \mathrm{b}\mathrm{i}\mathrm{n}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{g} (\mathrm{\%})=\left(1-{\displaystyle \frac{A_1}{A_0}}\right)\times 100$$

where A₁ represents the concentration of AFB₁ in the supernatant after incubation, and A₀ represents the initial AFB₁ concentration in the control sample (1 µg/mL).

2.7. Statistical Analysis

Experimental results were expressed as arithmetic means ± standard deviations of three replicates. Hemolytic activity, enzymatic activity, and carbohydrate utilization profiles were analyzed descriptively. Exopolysaccharide (EPS) quantification data were analyzed using one-way analysis of variance (ANOVA, single factor), followed by pairwise comparisons with Student’s t-tests. Differences in aflatoxin B₁ binding between viable and non-viable cells were assessed using Student’s t-test. Statistical significance was accepted at p < 0.05. For multiple comparisons of fermentative parameters of L. plantarum strains, a Bonferroni correction was applied (p/n, where n = 45) following Student’s t-tests, resulting in an adjusted significance threshold of p < 0.0011. All statistical analyses were performed using Microsoft Excel 2021 (Microsoft Corporation, Redmond, WA, USA).

3. Results and Discussion

3.1. Hemolytic Activity and Antibiotic Susceptibility

Hemolytic activity is recognized as a microbial virulence factor [23]. According to EFSA guidelines, this property must be assessed in all bacterial strains intended for food applications, regardless of their GRAS (Generally Recognized as Safe) or QPS (Qualified Presumption of Safety) status [24]. None of the ten L. plantarum strains tested showed discoloration or clearing on blood agar, indicating γ-hemolysis (no hemolytic activity) and confirming their safety. These findings are consistent with previous reports describing L. plantarum as non-hemolytic [25,26,27,28,29], although rare exceptions have been reported [30,31].

LAB may harbor antibiotic resistance (AMR) genes that are either intrinsic or acquired through horizontal gene transfer [32]. EFSA [17] and FAO/WHO [24] recommend that food-grade strains should not carry acquired or transferable AMR determinants due to the potential risk of gene transfer to pathogens. All L. plantarum strains in this study were susceptible to ampicillin, chloramphenicol, clindamycin, kanamycin, and gentamicin (

Table 2. Minimum inhibitory concentrations (µg/mL) of antibiotics for Lactiplantibacillus plantarum strains.

Genotype	Strain	Antibiotic Class
		β-Lactam	Amphenicol	Lincos-amide	Macrolide	Aminoglycoside			Tetracycline	Glycopeptide
		AMP	CHL	CLI	ERY	GEN	KAN	STR	TET	VAN
G1	A38	0.50 ± 0.00	2.00 ± 0.00	0.02 ± 0.00	0.12 ± 0.07	1.06 ± 1.68	28.00 ± 5.66	10.00 ± 8.49	2.00 ± 0.00	>32.00
	A57	0.67 ± 0.29	1.67 ± 0.58	0.02 ± 0.00	0.14 ± 0.10	3.00 ± 0.00	>32.00	16.00 ± 0.00	10.00 ± 2.83	>32.00
G5	A45	0.07 ± 0.02	1.83 ± 0.29	1.75 ± 0.35	0.35 ± 0.00	1.29 ± 1.48	>32.00	24.00 ± 0.00	6.00 ± 2.00	>32.00
	A52	0.11 ± 0.07	2.17 ± 0.76	0.02 ± 0.00	0.04 ± 0.02	0.75 ± 0.25	7.00 ± 1.41	14.00 ± 2.83	1.50 ± 0.50	>32.00
G6	A44	0.27 ± 0.10	1.83 ± 0.29	0.02 ± 0.00	0.10 ± 0.02	1.83 ± 0.29	29.33 ± 4.62	21.33 ± 4.62	2.50 ± 0.87	>32.00
	A51	0.32 ± 0.11	0.92 ± 0.14	0.02 ± 0.00	0.19 ± 0.06	2.50 ± 1.32	32.00 ± 0.00	13.33 ± 4.62	1.08 ± 0.38	>32.00
G9	A26	0.13 ± 0.06	1.17 ± 0.76	0.02 ± 0.00	0.10 ± 0.02	0.92 ± 0.14	32.00 ± 0.00	12.00 ± 4.00	2.33 ± 1.53	>32.00
	A37	0.27 ± 0.20	1.33 ± 0.58	0.02 ± 0.00	1.36 ± 1.10	0.83 ± 0.14	10.00 ± 2.83	11.33 ± 5.03	1.58 ± 0.72	>32.00
G12	A10	0.42 ± 0.14	2.50 ± 1.32	0.02 ± 0.00	0.17 ± 0.04	1.83 ± 0.29	28.00 ± 5.66	20.00 ± 5.66	3.67 ± 0.58	>32.00
	A71	0.34 ± 0.28	3.00 ± 0.00	0.02 ± 0.00	0.13 ± 0.00	0.83 ± 0.14	32.00 ± 0.00	24.00 ± 0.00	2.67 ± 1.15	>32.00
Cut-off values 1		≤2	≤8	≤4	≤1	≤16	≤64	n.r.	≤32	n.r.

Abbreviations: AMP ampicillin, CHL chloramphenicol, CLI clindamycin, ERY erythromycin, GEN gentamicin, KAN kanamycin, STR streptomycin, TET tetracycline, VAN vancomycin, n.r. not required. ¹ Antibiotic cut-off values by EFSA [17].

), consistent with previous findings showing sensitivity of L. plantarum to β-lactams and protein synthesis inhibitors [33,34], suggesting the absence of transferable AMR genes.

Some LAB genera possess intrinsic resistance to bacitracin, vancomycin, kanamycin, teicoplanin, and quinolones [35]. In this study, kanamycin MICs among L. plantarum strains varied notably, with some isolates (A45, A57) exceeding 32 µg/mL. Variation also occurred within genotypes: for G1 (A38, A57) values were 28 and >32 µg/mL, while for G5 (A45, A52) kanamycin and streptomycin MICs ranged from 7 to >32 µg/mL and 14–24 µg/mL, respectively. These differences indicate strain-specific phenotypic variation despite shared genotypes, likely due to minor genomic or regulatory differences. In contrast, G6 (A44, A51) showed uniform MICs (29–32 µg/mL), suggesting stable intrinsic resistance.

Kanamycin, gentamicin, and streptomycin belong to the aminoglycoside class of antibiotics that bind to the bacterial ribosome and inhibit protein synthesis [36]. Although L. plantarum susceptibility to aminoglycosides varies among strains [37,38], all isolates in the present study were resistant to vancomycin (MIC > 32 µg/mL), in agreement with previous reports describing the high intrinsic resistance of this species [37,39]. This resistance is attributed to the specific structure of the L. plantarum peptidoglycan layer, which terminates with D-alanyl-D-lactate (D-Ala-D-Lac) or, less frequently, D-alanyl-D-alanine, thereby preventing vancomycin binding [40].

Streptomycin MICs for L. plantarum strains ranged from 10 to 24 µg/mL, indicating low-level intrinsic resistance. Previous studies have confirmed that L. plantarum inherently resists streptomycin [41,42,43,44], whereas MICs in other Lactobacillus species may exceed 64 µg/mL. The observed variability likely reflects strain-specific phenotypic differences rather than the absence of intrinsic resistance. Since this resistance is chromosomally encoded and non-transferable [17], it is regarded as safe with respect to horizontal gene transfer.

One strain (A37) showed an erythromycin MIC of 1.36 ± 1.10 µg/mL, slightly above the EFSA cut-off value of 1 µg/mL. The high replicate variability suggests that this deviation was due to experimental fluctuation rather than genuine resistance.

3.2. Enzymatic Activity Profiles

The proteolytic system of LAB includes a cell-envelope proteinase, peptide transport mechanisms, and a range of intracellular peptidases that break down (poly)peptide chains into smaller molecules such as di- and tripeptides and free amino acids [45,46,47]. In this study, all L. plantarum strains exhibited strong aminopeptidase activities, particularly leucine and valine arylamidase, which were consistent across all five genotypes (

Table 3. Enzymatic activity of Lactiplantibacillus plantarum strains.

Enzyme	Genotypes and Strains
	G1		G5		G6		G9		G12
	A38	A57	A45	A52	A44	A51	A26	A37	A10	A71
Alkaline phosphatase	1	0	0	1	0	1	0	0	0	0
Esterase	3	0	0	1	0	1	1	3	1	1
Esterase lipase	3	1	1	1	1	1	1	3	1	1
Lipase	0	1	0	0	0	0	0	0	0	0
Leucine arylamidase	5	5	5	5	5	5	5	5	5	5
Valine arylamidase	4	4	3	4	5	4	5	3	5	5
Cystine arylamidase	2	2	2	2	3	3	2	2	3	3
Trypsin	0	1	0	0	0	0	0	0	0	0
α-chymotrypsin	0	0	0	0	0	0	0	0	0	0
Acid phosphatase	1	3	1	2	3	3	3	2	3	3
Naphthol-AS-BI phosphohydrolase	2	3	3	3	3	3	3	3	2	2
α-galactosidase	3	0	0	0	0	0	0	5	0	0
β-galactosidase	5	4	3	4	4	4	4	5	4	4
β-glucuronidase	1	0	0	0	0	0	0	0	0	0
α-glucosidase	5	0	3	0	0	0	0	5	0	0
β-glucosidase	4	5	5	5	5	5	5	5	5	5
N-Acetyl-β-glucosaminidase	3	3	3	5	5	5	5	3	5	5
α-mannosidase	0	0	0	0	0	0	0	0	0	0
α-fucosidase	0	0	0	0	0	0	0	0	0	0

Value 0 indicated a negative reaction, 1–2 represented weak activity (5 to <20 nmol of substrate hydrolyzed), and 3–5 denoted strong activity (>20 nmol of substrate hydrolyzed). Background colour indicates the intensity of the values, with darker colours representing higher activity levels.

). Moderate cystine arylamidase activity was detected in strains A44, A51 (G6) and A10, A71 (G12), suggesting some intra-genotypic variability. In contrast, trypsin and α-chymotrypsin were inactive in all strains, in agreement with earlier reports indicating the absence of these endopeptidases in L. plantarum [48,49]. High aminopeptidase activity across all L. plantarum strains underscores the functional importance of intracellular peptide hydrolysis, efficient nitrogen utilization, and the release of bioactive peptides that enhance both sensory and health-promoting properties of fermented foods [50,51,52]. These peptides may exert antioxidant, anti-inflammatory, and gut-protective effects [53], consistent with the presence of multiple genes encoding intracellular peptidases in the L. plantarum genome [5].

In addition to proteolytic activity, glycolytic enzyme profiles were assessed to evaluate carbohydrate metabolism. All L. plantarum strains exhibited strong β-glucosidase activity, consistent with reports for LAB from the Lactobacillaceae and Bifidobacteriaceae families [54,55,56]. β-glucosidases hydrolyze β-1,4-glycosidic bonds in plant-derived glucosides, oligosaccharides, and phenolic compounds, releasing β-D-glucose and bioactive aglycones [55,56,57,58]. This activity enhances the flavor, aroma, and nutritional quality of fermented foods and supports the bioconversion of ginsenosides, isoflavones, and phenolics during fermentation [57,58,59].

α-glucosidase activity varied between strains, being strong in A38 (G1) and A37 (G9), moderate in A45 (G5), and absent in other isolates, indicating both inter- and intra-genotypic variation. A similar pattern was observed for α-galactosidase, which was highest in A37 and moderate in A38, while α-mannosidase and α-fucosidase were inactive in all strains.

β-galactosidase and N-acetyl-β-glucosaminidase were consistently active across nearly all strains, indicating efficient carbohydrate utilization and hydrolysis of plant-derived glycosidic bonds. In contrast, β-glucuronidase activity was absent, which is desirable given its potential carcinogenicity [60]. β-Galactosidase catalyzes the breakdown of lactose into glucose and galactose, potentially alleviating symptoms of lactose intolerance [49].

Overall, the carbohydrate utilization profiles revealed both inter- and intra-genotypic diversity. While β-glucosidase and β-galactosidase activities were uniform, α-glucosidase and α-galactosidase varied substantially between genotypes and strains, suggesting strain-specific metabolic contributions during food fermentation and possible differences in health-promoting properties [61,62]. Several L. plantarum strains have been reported to inhibit α-glucosidase and α-amylase, thereby reducing carbohydrate digestion and improving glycemic control through mechanisms involving exopolysaccharide and short-chain fatty acid production [63,64,65,66,67,68,69].

Enzymes related to phosphatase activity, including acid phosphatase and naphthol-AS-BI phosphohydrolase, showed moderate to high activity across most strains, suggesting adaptability to phosphate-limited environments. Esterase and esterase-lipase activities were variable but generally moderate, whereas lipase was weak or absent. Strong N-acetyl-β-glucosaminidase activity, as also noted by Gökmen et al. [70] and Lee et al. [49], indicates an active role in polysaccharide degradation and may contribute to antifungal properties through the hydrolysis of chitin-like structures present in the cell wall of Aspergillus niger [71].

Together, these findings highlight the broad enzymatic capacity of L. plantarum strains, reflecting their adaptability to nutrient-limited environments and their potential antifungal activity mediated through diverse hydrolytic mechanisms.

3.3. Carbohydrate Metabolism Profiles

All L. plantarum strains efficiently utilized major monosaccharides, including D-glucose, D-galactose, and D-fructose, as well as several disaccharides such as D-maltose, D-lactose, D-melibiose, and D-sucrose (

Table 4. Utilization of carbohydrates by Lactiplantibacillus plantarum strains.

Carbon Source	Genotypes and Strains
	G1		G5		G6		G9		G12
	A38	A57	A45	A52	A44	A51	A26	A37	A10	A71
L-Arabinose	+	+	-	-	-	-	+	+	+	-
D-Ribose	+	+	+	+	+	+	+	+	+	+
D-Xylose	+	-	-	-	-	-	+	+	-	-
D-Galactose	+	+	+	+	+	+	+	+	+	+
D-Glucose	+	+	+	+	+	+	+	+	+	+
D-Fructose	+	+	+	+	+	+	+	+	+	+
D-Mannose	+	+	+	+	+	+	-	+	+	+
D-Mannitol	+	+	+	+	+	+	-	+	+	+
D-Sorbitol	+	+	+	+	+	+	-	+	+	+
Methyl-α-D-mannopyranoside	+	+	+	+	+	+	-	+	+	-
N-Acetylglucosamine	+	+	+	-	+	+	-	+	+	+
Amygdalin	+	+	+	+	+	+	-	+	+	+
Arbutin	+	+	+	+	+	+	-	+	+	+
Esculin	+	+	-	+	-	+	+	+	+	+
Salicin	+	+	+	+	+	+	-	+	+	+
D-Cellobiose	+	+	+	+	+	+	-	+	+	+
D-Maltose	+	+	+	+	+	+	+	+	+	+
D-Lactose	+	+	+	+	+	+	+	+	+	+
D-Melibiose	+	+	+	+	+	+	+	+	+	+
D-Saccharose	+	+	+	+	+	+	+	+	+	+
D-Trehalose	+	+	+	+	+	+	-	+	+	+
Inulin	+	-	+	-	+	-	-	+	+	-
D-Melezitose	+	+	+	+	+	-	-	+	+	+
D-Raffinose	+	+	+	+	+	-	+	+	+	+
Starch	-	-	-	-	-	+	-	-	-	-
Gentiobiose	+	+	+	+	+	-	-	+	+	+
D-Turanose	-	-	-	-	-	-	-	-	+	+
D-Lyxose	-	-	-	-	-	-	-	-	+	-
D-Tagatose	-	-	-	-	-	-	-	-	+	-
D-Fucose	-	-	-	-	-	-	-	-	+	-
L-Fucose	-	-	-	-	-	-	-	-	+	-
D-Arbitol	-	-	+	-	+	-	-	-	+	-
L-Arbitol	-	-	-	-	-	-	-	-	+	-
Potassium gluconate	+	+	+	+	+	+	+	-	+	+
Potassium 5-ketogluconate	-	-	-	-	-	-	+	-	-	-

Symbols: “+” indicates utilized; “-“ indicates not utilized.

). This broad utilization pattern reflects the species’ metabolic versatility and aligns with previous findings [3,4,8,72]. Of the 49 tested carbon sources, most were metabolized by over 80% of strains, whereas some substrates—such as L-arabinose and D-xylose—showed genotype-specific utilization, indicative of ecological adaptation to plant-associated environments rich in complex polysaccharides.

L-arabinose metabolism was observed in genotype G1 (A38, A57) and selected G9 and G12 strains (A26, A37, A10), while genotypes G5 and G6 lacked this ability. D-xylose utilization followed a similar pattern, occurring in genotypes G1 and G9 but not G5 or G6. Additional variation in the fermentation of inulin, D-melezitose, and D-arabitol further demonstrates strain-specific metabolic diversity. Such variability is consistent with the presence of genomic “lifestyle islands” containing carbohydrate utilization cassettes [3], which enable niche-specific metabolic flexibility and support targeted prebiotic or synbiotic applications [7].

While genotypic grouping reflected certain similarities, substantial strain-to-strain variability was evident—particularly within G5, G6, and G9. Strains of G1 (A38, A57) displayed nearly identical fermentation profiles, suggesting a strong genotype-phenotype relationship, whereas considerable intra-genotypic diversity appeared in G9 and G12. Moreover, strains from distinct genotypes occasionally exhibited similar carbohydrate profiles, indicating that metabolic traits do not always align strictly with genotype.

Comparison of carbohydrate fermentation and enzymatic activity (Table 3 and Table 4) revealed both concordance and divergence. High β-galactosidase and β-glucosidase activities correlated with the ability to metabolize lactose, cellobiose, and other β-linked carbohydrates, confirming the conserved functional role of these enzymes in L. plantarum. This enzymatic repertoire likely underpins the species’ capacity to exploit diverse dietary substrates. However, exceptions were noted: strain A26, despite strong β-glucosidase activity, failed to utilize amygdalin, arbutin, salicin, and D-cellobiose, indicating that substrate utilization may depend on enzyme specificity, expression level, or transport efficiency rather than enzyme presence alone.

α-glucosidase activity was detected only in strains A38, A45, and A37, corresponding to their ability to ferment maltose and trehalose. Although α-fucosidase and α-mannosidase activities were absent in all strains, discrepancies emerged between enzyme profiles and substrate use. Strain A10 metabolized both D- and L-fucose despite lacking α-fucosidase activity, while all strains except A26 fermented D-mannose, and A26 together with A71 failed to utilize methyl-α-D-mannopyranoside. These results suggest that alternative transport or metabolic pathways may compensate for the absence of the canonical enzymes.

These results underscore that enzymatic potential does not always correspond to actual metabolic activity, as enzyme expression may be substrate-induced, strain-specific, and dependent on physiological state. The findings emphasize the metabolic plasticity of L. plantarum and its ability to adapt to diverse carbohydrate environments, reflecting its genomic diversity and ecological versatility.

3.4. Fermentative Properties of L. plantarum Strains

During all sourdough fermentation experiments, the initial abundance of lactic acid bacteria (LAB) in the rye dough medium was 5.40 log cfu/mL. The final abundance of the studied L. plantarum strains was similar for most isolates (9/10; p > 0.05), ranging from 8.50 to 10.84 log cfu/mL. Intra-genotypic variation was observed within genotype G9: strain A26 exhibited markedly higher cell counts than strain A37 (p < 0.05) and also exceeded the abundance of strains belonging to genotype G1. In contrast, strain A37 (G9) showed the lowest final abundance among all tested strains. Inter-genotypic comparisons revealed statistically significant differences in LAB abundance (p < 0.05), with pairwise differences detected between G1 (A38, A57) and G6 (A44, A51), as well as between G1 and G5 (A45, A52).

The LAB counts observed in the sourdough samples fall within the range reported in previous studies. Žugić-Petrović et al. [73] reported cell densities of 7–8 log cfu/g after 24 h of fermentation, while Müller et al. [74] observed values around 9.8 log cfu/mL under comparable conditions. Variability among studies can be attributed to differences in fermentation time, temperature, inoculum type (mother starter versus defined inoculation), and refreshment practices [75,76]. Such intra- and inter-genotypic differences may stem from variations in carbohydrate utilization pathways, acid tolerance, and growth capacity under the acidic and osmotically stressful conditions characteristic of sourdough fermentation.

The fermentative performance of ten L. plantarum strains was evaluated in a rye flour-water medium over 48 h. The pH of all sourdoughs decreased below 4.5 by the 16th hour, reaching final values between 3.56 and 3.77. Strains A52, A45, A37, and A26 acidified most rapidly (pH ≤ 4.5 within 13 h), whereas A38, A57, and A44 reached this threshold at 16 h. The rapid decrease in pH reflects organic acid production typical of L. plantarum-based sourdoughs [77], which inhibits flour amylases below pH 4.5 while supporting optimal proteolytic activity (pH 3.5–4.5) [78]. Although no significant differences in pH were detected among individual strains or within genotypes, inter-genotypic analysis revealed statistically significant variation in LAB abundance (p < 0.05), with significant pairwise differences primarily involving genotype G5, particularly in comparisons with G1, G6, and G12, as well as a notable difference between G1 and G9.

Total titratable acidity (TTA) after 48 h ranged from 152.3 to 190.2 °Th (Figure 1). Genotype-based comparisons of TTA revealed no consistent genotype-associated pattern; however, significant intra-genotypic variation was observed within genotype G6 (A51 and A44). The highest TTA (190.2 °Th) was observed for A51 (G6), which was significantly higher than strains belonging to genotypes G1 (A38, A57), G9 (A37), and also the other G6 strain, A44 (p < 0.05). The lowest TTA (152.3 °Th) was observed for A38, which differed significantly only from A51. These values correspond well with those previously reported for rye sourdoughs (145–216 °Th) [79]. Overall, both inter-genotypic and intra-genotypic comparisons indicate that TTA variation is primarily strain-specific rather than genotype-driven.

Brix measurements increased significantly (p < 0.05) in all sourdoughs relative to the initial slurry (7.3–8.5 °Bx; Figure 2), indicating carbohydrate hydrolysis and solute release [77,80,81]. Inter-genotypic comparisons revealed statistically significant differences (p < 0.05) in final Brix values between G1 (A38, A57) and G6 (A44, A51), G1 and G9 (A26, A37), G1 and G5 (A45, A52), and between G12 (A10, A71) and G5. Significant intra-genotypic variation was also observed within G5 and G12. The highest final Brix value was observed for with A45 (12.63 °Bx), which differed significantly from A38 and A71. The lowest value was recoded for A38 (11 °Bx). Following Bonferroni correction, only the pairwise differences between A45 and A38, and between A45 and A71, remained statistically significant. Intra-genotypic variation may reflect minor genetic differences, strain-specific enzyme expression, or prior adaptation to environmental conditions, highlighting the contributions of both genotype and individual strain characteristics to sourdough fermentation performance.

CO₂ evolution during fermentation paralleled the pH decrease (Figure 3). All L. plantarum strains initiated gas production after 8–10 h, reaching maximum levels between 14 and 16 h (22–100%) before declining. Strains A38 and A45 peaked at 15–17 h (100% and 56%, respectively), decreased to 6–7% by 24–26 h, and later partially recovered (23–36%). During the first 24 h, A38 and A44 were the most active CO₂ producers, whereas A26, A52, and A71 exhibited higher activity in the later stages. These late-stage increases likely reflect secondary metabolic activity, while early-phase differences may arise from strain-specific variation in carbohydrate metabolism and adaptation to the sourdough environment. This biphasic pattern aligns with observations by Wehrle and Arendt [82], who reported earlier CO₂ production in inoculated sourdoughs (~10 h) compared with spontaneous fermentations (~20 h).

Significant (p < 0.05) inter-genotypic differences in CO₂ production were detected for all genotype comparisons, except between G1 (A38, A57) and G6 (A44, A51), and between G9 (A26, A37) and G12 (A10, A71). Intra-genotypic analyses revealed significant variation (p < 0.05) within genotypes G5 (A45, A52) and G12, demonstrating that strains belonging to the same genotype can differ markedly in their CO₂ production profiles.

Lactic acid concentrations in sourdoughs fermented with L. plantarum strains ranged from 11.4 to 21.6 g/kg (

Table 5. Content of lactic acid, acetic acid, and ethanol (g/kg) in rye sourdoughs prepared with L. plantarum strains.

Parameter	Strains and Genotypes
	G1		G5		G6		G9		G12
	A38	A57	A45	A52	A44	A51	A26	A37	A10	A71
Lactic acid	12.7 ± 3.1 ab	14.0 ± 1.6 ab	13.3 ± 2.8 ab	21.6 ± 4.1 ab	13.3 ± 1.8 ab	11.4 ± 0.9 a	12.7 ± 3.2 ab	15.4 ± 0.4 b	15.1 ± 4.8 ab	16.3 ± 2.6 ab
Acetic acid	2.8 ± 0.3 a	2.0 ± 0.6 ab	1.7 ± 0.2 b	1.9 ± 0.8 ab	2.2 ± 0.7 ab	2.2 ± 0.7 ab	1.8 ± 0.2 b	2.3 ± 0.2 ab	2.3 ± 0.9 ab	2.5 ± 0.5 ab
Total acids	15.5 ± 3.1	16.0 ± 1.7	15.0 ± 2.8	23.5 ± 4.2	15.5 ± 1.9	13.6 ± 1.1	14.5 ± 3.2	17.7 ± 0.5	17.4 ± 4.9	18.8 ± 2.6
Ethanol	5.2 ± 1.2 b	4.1 ± 0.8 b	4.2 ± 4.2 a	11.8 ± 1.5 a	3.9 ± 0.5 c	3.7 ± 1.5 b	6.7 ± 2.2 a	5.7 ± 1.2 a	4.6 ± 1.9 a	13.8 ± 2.7 a

Notes: Different lowercase letters following the mean values within the same row indicate statistically significant differences among strains (p < 0.05). Values sharing the same letter do not differ significantly.

). The highest level was produced by strain A52 (p < 0.05), exceeding most other strains except A37, A10, and A71. The lowest value occurred in A51 (11.4 g/kg), which differed significantly from A57, A71, A52, and A37. After Bonferroni correction, only the difference between A51 and A37 remained significant. Comparable studies reported 9.1 g/kg lactic acid in rye sourdoughs fermented for 18 h [83], while wheat sourdoughs typically contain 1–7 g/kg depending on strain and conditions [84,85,86]. These data confirm the higher lactic acid accumulation in rye sourdough, reflecting both substrate composition and microbial metabolism. Within individual genotypes, no significant differences (p > 0.05) were observed among strains; however, inter-genotypic comparisons revealed statistically significant differences (p < 0.05), specifically between G6 and G12 as well as between G9 and G12.

L. plantarum is capable of synthesizing acetic acid from pyruvate, which, together with lactic acid, represents a major end-product of heterofermentative metabolism [6]. Acetic acid concentrations ranged from 1.7 to 2.8 g/kg (Table 5). Strain A38 produced the highest amount (2.8 g/kg, p < 0.05), significantly exceeding A45 and A26, while A45 showed the lowest level (1.7 g/kg). After Bonferroni correction, A38 differed significantly from both A45 and A26. These values are comparable to those in wheat sourdoughs (1.8–3.3 g/kg) [84] and exceed those reported for short fermentations (0.67 g/kg) [86], confirming strain- and condition-dependent variability in acetic acid formation. Within genotype G1, significant intra-genotypic variation was observed, whereas inter-genotypic comparisons revealed statistically significant differences (p < 0.05) between G1 and G9, G1 and G5, and between G5 and G12.

Comparison of total acid content showed that L. plantarum strain A52 produced the highest combined lactic and acetic acid levels (23.5 g/kg), significantly exceeding most other strains except A10, A71, A37, and A26 (p < 0.05). The lowest total acid concentration was found in A51 (13.6 g/kg), differing significantly only from A71 and A52. After Bonferroni correction, no significant differences remained. The lactic-to-acetic acid ratio, a key indicator of sourdough quality, ranged from 87.2:12.8 in this study, well within the optimal 90:10 to 70:30 range [87]. Comparable results have been reported for rye sourdoughs (77.7:22.2) [88], with higher acetic acid levels linked to pentose utilization by heterofermentative LAB [89,90,91]. Within genotype G5, statistically significant intra-genotypic variation was detected, while inter-genotypic comparisons revealed significant differences involving G6, specifically in comparisons with G12 and with G5.

L. plantarum produces several secondary metabolites during heterofermentative metabolism, including ethanol [6], with additional contributions from flour-derived yeasts [92]. After 48 h fermentation, ethanol concentrations averaged 6.4 g/kg (Table 5). The highest levels were observed in sourdoughs fermented with strains A71 (13.8 g/kg) and A52 (11.8 g/kg) (p < 0.05), while the lowest occurred in A51 (3.7 g/kg) and A44 (3.9 g/kg). After Bonferroni correction, A52 differed significantly from A51, A38, A57, and A44, whereas A71 differed only from A44. These ethanol values are lower than those reported by Buksa [88] (18.6 g/kg in 24 h rye sourdough), likely due to yeast supplementation in that study. Within genotypes, significant intra-genotypic variation in ethanol content was observed for strains belonging to G5 and G12. Inter-genotypic comparisons revealed statistically significant differences involving G6 (in comparisons with G12, G1, G9, and G5) as well as between G1 and G12 and between G1 and G5, suggesting that both genotype- and strain-specific traits influence ethanol production.

Collectively, these results demonstrate that L. plantarum exhibits considerable biochemical versatility, with both inter-genotypic and intra-genotypic variation significantly influencing acidification kinetics, CO₂ production, and the synthesis of key metabolites such as lactic acid, acetic acid, and ethanol. Such variability highlights the importance of strain- and genotype-specific characteristics in shaping sourdough fermentation dynamics. These findings emphasize that careful selection and combination of particular strains or genotypes can be strategically employed to design tailored starter cultures, thereby optimizing the technological performance, metabolic outputs, and sensory attributes of sourdough products.

3.5. Exopolysaccharide Yield and Quantitative Sugar Content

The glucose-equivalent sugar content of exopolysaccharides (EPS) extracted from L. plantarum strains ranged from 278.3 to 362.8 µg/mg (Figure 4), with no statistically significant differences between strains or genotypes. The dry mass of lyophilized EPS, ranged from 22.1 ± 4.9 mg to 32.2 ± 3.2 mg, also without significant differences. Within genotype G5 (A52 and A45), EPS dry mass was significantly lower (p < 0.05) compared with most other genotypes, except G6 (A51 and A44).

EPS yields based on dry mass ranged from 131.4 to 225.2 mg/L, consistent with previously reported [93] values for L. plantarum (71–279 mg/L). No consistent strain- or genotype-dependent trends were observed, although minor pairwise differences were statistically significant (p < 0.05; significant differences were detected only for EPS yield in some pairwise comparisons, e.g., A51 vs. A10, A71, A37). EPS production depends on multiple factors, including strain genetics, medium composition, and nutrient availability [94,95]. Several studies report markedly higher yields under optimized conditions [96,97,98,99]. In the present study, MRS medium supplemented with 2% sucrose was used; however, previous research has shown that EPS production can be enhanced by optimizing the carbon source. Elmansy et al. [96] identified sucrose, lactose, and fructose as the most effective substrates, whereas Zhang et al. [93] reported glucose as the preferred carbon source for most L. plantarum strains, with some exceptions (e.g., L. plantarum 70810), which exhibited enhanced EPS production with sucrose. Similarly, Wang et al. [98] and Afreen et al. [100] observed increased yields with lactose. These findings likely reflect strain-specific metabolic preferences and emphasize the need to optimize carbon sources individually to maximize EPS biosynthesis.

Notably, the optimal conditions for EPS synthesis do not always coincide with those for microbial growth. In some cases, microorganisms produce EPS only under stress conditions [101]. For instance, Imran et al. [97] reported exceptionally high EPS concentrations (956 mg/L), likely resulting from optimized culture conditions and differences in purification procedures, as trichloroacetic acid (TCA) was not used during EPS recovery. Consistently, Rimada et al. [102] demonstrated that TCA treatment can reduce EPS yield by up to 50%. Nitrogen source selection also influences EPS production: yeast extract supports maximal synthesis due to its amino acid and peptide content, while ammonium sulfate is the most suitable inorganic source. However, some nitrogen-containing compounds may interfere with the phenol–sulfuric acid assay, leading to overestimation of sugar content [103].

The relationship between EPS sugar content and dry mass was evaluated. The low coefficient of determination (R² = 0.0177) indicated no linear correlation, suggesting compositional variability among strains. For example, strain A37 exhibited both high sugar content and dry mass, whereas A57 had high dry mass but the lowest sugar content. Conversely, A45 showed relatively low dry mass but above-average sugar content.

Carbohydrates constitute the predominant component of L. plantarum EPS, accounting for approximately 90–96% of total mass [97,104]. Minor fractions include sulfated groups, proteins, nucleic acids, and uronic acids [104,105]. For example, Xiao et al. [106] found that purified EPS contained 88.3% carbohydrates, 3.9% proteins, 3.0% uronic acids, and 1.5% sulfates, while Tsuda et al. [107] reported 81% carbohydrates and 3.1% proteins. These data confirm that although EPS are mainly carbohydrate polymers, total dry mass does not necessarily correlate with sugar content, as non-carbohydrate components can substantially contribute to the total EPS mass.

3.6. Influence of Aflatoxin B₁ on the Carbohydrate Fermentation Profiles of L. plantarum

All L. plantarum strains fermented all four sugars tested (glucose, galactose, lactose, and sucrose), and this ability was unaffected by AFB₁, except in strain A71. Fermentation profiles remained stable throughout incubation (48 h–12 days), with minor variations during the first 24 h. No gas production was observed during the 12-day incubation. Among the five genotypes, only G1 (A38, A57) and G5 (A52, A45) showed identical fermentation profiles, whereas intra-genotypic variation occurred in others, indicating six distinct phenotypes. Fermentation was most efficient with glucose and weakest with lactose. These results align with reports describing L. plantarum as a metabolically versatile species capable of utilizing diverse carbohydrates through its extensive repertoire of sugar transport and metabolic genes [5,9]. The observed strain-specific differences likely reflect functional variation in transport systems and intracellular metabolic pathways [8].

During the first 24 h in lactose-containing media, isolates A38 and A57 (G1) and A37 (G9) produced little acid, whereas strain A26 (same genotype as A37) actively fermented all sugars throughout incubation. Similarly, within genotype G6, strain A44 showed limited lactose fermentation and weak galactose utilization, in contrast to the more active profile of A51. Strain A10 fermented glucose and galactose rapidly, but required 48 h for lactose and sucrose. In the same genotype, A71 exhibited similarly reduced activity, producing minimal acid from lactose and galactose during the first 24 h. Interestingly, for galactose, acid formation was absent in toxin-free and low-AFB₁ media (1 µg/mL) but slightly increased at 10–20 µg/mL, suggesting that higher toxin levels may transiently stimulate acid production—an effect not previously described.

Research on the influence of AFB₁ on L. plantarum fermentation is limited. Sutic and Banina [22] reported that L. plantarum retained its carbohydrate fermentation capacity in the presence of AFB₁ and, notably, produced gas only under toxin exposure. These results suggest that AFB₁ does not inhibit sugar utilization but may affect fermentation kinetics in a strain- and substrate-specific manner, highlighting the metabolic adaptability of L. plantarum and its potential role in stabilizing and detoxifying mycotoxin-contaminated cereal matrices.

3.7. Aflatoxin B₁ Binding Capacity

All L. plantarum strains, regardless of viability, exhibited strong AFB₁-binding capacity, with efficiencies approaching 100%. The highest binding was observed for viable cells of L. plantarum A26, whereas the lowest occurred in non-viable cells of A10 (Figure 5). Previous studies have reported highly variable AFB₁-binding efficiencies among LAB, ranging from 0.6% to 100% [108,109,110,111,112]. Several LAB species, including L. plantarum, Levilactobacillus brevis, Limosilactobacillus fermentum, Lacticaseibacillus rhamnosus, Lactobacillus acidophilus, and Lactococcus lactis, have shown high efficacy in AFB₁ removal, sometimes achieving complete degradation [113]. Co-cultivation of probiotic strains may further enhance aflatoxin removal [114]. Factors such as incubation temperature, exposure time, microbial load, toxin concentration, and pH significantly influence AFB₁ reduction [113,115].

Despite extensive studies on L. plantarum, strain-level differences in AFB₁ binding remain less characterized. Møller et al. [116] reported wide variability among L. plantarum isolates: at pH 6.5, binding ranged from <20% to >70%. Similarly, Park et al. [117] observed strain-dependent adsorption efficiencies (3.3–48.9%) among three L. plantarum isolates, confirming strong intra-species variability.

Viable L. plantarum cells generally bound more AFB₁ than heat-inactivated ones, consistent with Khanafari et al. [112] and Asurmendi et al. [118], who observed toxin desorption from heat-killed cells during incubation. Conversely, several reports indicate higher AFB₁ binding by non-viable cells [119,120,121,122]. Enhanced adsorption by inactivated cells is attributed to heat-induced protein denaturation and structural modifications in peptidoglycan, which increase surface hydrophobicity and expose new binding sites [123,124,125].

Although numerical differences between viable and non-viable cells were small, statistical analysis revealed a significant overall effect of heat treatment (p < 0.001). Within-genotype comparisons showed no significant differences in most cases (p > 0.05), except for genotype G5, where non-viable strains A52 and A45 differed significantly (p < 0.05). Across genotypes, binding efficiency was comparable, though minor differences were noted between G1 (A38, A57) and G12 (A10, A71) non-viable cells, with G1 slightly outperforming G12 (99.9552% vs. 99.9546%).

Heat-killed cells exhibited more intra-strain variability (8 strains with significant differences) than viable cells (2 strains), suggesting differential structural responses to inactivation. Strains A71 and A44 showed significant viability-dependent differences (p < 0.05), confirming genotypic distinctness. In contrast, strains within G12, G6, and G9, despite differing in carbohydrate profiles, exhibited similar adsorption patterns, indicating that AFB₁ binding is less influenced by these metabolic traits.

Differences between studies likely reflect not only inter-strain variability (e.g., cell wall composition) but also methodological factors such as the cell-to-toxin ratio, incubation conditions, pH, and analytical methods used for quantification [109,126,127,128,129]. Moreover, the method of cell inactivation—heat, acid, or irradiation—can markedly alter adsorption efficiency [119,122]. Despite such variability, numerous studies consistently confirm LAB as effective AFB₁ binders, underscoring their potential in improving food safety and mitigating mycotoxin risks.

4. Conclusions

All examined Lactiplantibacillus plantarum strains isolated from traditional rye sourdough were non-hemolytic and exhibited antibiotic susceptibility profiles consistent with food-grade safety standards. The isolates displayed broad enzymatic and metabolic diversity, with pronounced aminopeptidase and β-glucosidase activities, indicating strong proteolytic and glycolytic potential. Carbohydrate utilization profiles revealed extensive substrate adaptability, while inter- and intra-genotypic differences confirmed notable metabolic plasticity.

During rye sourdough fermentation, all strains rapidly acidified the dough and produced lactic-to-acetic acid ratios consistent with optimal sourdough quality, confirming their technological suitability. Exposure to aflatoxin B₁ did not impair fermentation performance, and all strains showed exceptionally high toxin-binding efficiency, indicating potential for biological detoxification of contaminated cereals.

Exopolysaccharide yields and sugar contents were stable across genotypes, suggesting consistent biosynthetic capacity under standardized conditions. When the genotype- and strain-level data were integrated, clear functional patterns emerged: although strain-specific traits accounted for most of the observed variation, reproducible genotype-associated tendencies were evident. Genotype G6 demonstrated the strongest overall fermentative performance, G5 exhibited enhanced sugar utilization and ethanol formation, and G1 showed consistently weaker metabolic activity, whereas G9 and G12 displayed more strain-dependent behavior. Within these genotype backgrounds, marked strain-level differences persisted, with A52 expressing the most versatile and robust phenotype, A26 and A71 exhibiting strong but more specialized metabolic profiles, and A38 and A51 consistently demonstrating weak fermentative performance.

Collectively, the studied L. plantarum strains exhibited functional robustness, combining efficient fermentation, stress tolerance, and mycotoxin-binding ability. Their genotype-specific tendencies and complementary strain-level capabilities underscore their promise as multifunctional starter cultures for safe, high-quality, and nutritionally enhanced fermented foods.