Sampling and isolation
In this study, ten LAB isolates from two traditional cheeses of the Djelfa province (Jben and Klila) were collected from the localities of Aïn Oussera, Djelfa and Birine. The purified isolates were initially screened for Gram-positive and catalase-negative traits to confirm their identity as LAB (Fig 1).
Fig 1 shows the various stages of LAB observation. Colonies grown on MRS and M17 media were circular, smooth and ranged from whitish to cream in color, typical of LAB, indicating their adaptation to mesophilic conditions. Microscopic examination revealed cocci- or rod-shaped cells arranged in clusters or chains, characteristic of LAB commonly present in fermented cheeses. These results confirm the dominance of LAB and demonstrate their morphological diversity, supporting their potential for use in developing starter cultures.
Genotypic characterization
The isolation and identification results presented in Table 1 show that traditional Jben and Klila cheeses from the Djelfa region contain a diverse indigenous LAB community, primarily composed of
Enterococcus and lactobacilli from the
plantarum/brevis group. Ten isolates were identified through 16S rRNA gene sequencing as members of
Enterococcus,
Lactiplantibacillus and
Levilactobacillus. This combined phenotypic (Fig 1) and molecular approach (Table 1) is commonly employed to characterize indigenous LAB in traditional cheeses and the assignment of accession numbers (PX736099–PX736107, PX737322) ensures traceability and facilitates their future use, as reported by
Azzouz et al., (2025b) and
Bouchibane et al., (2023).
The species distribution shows cheese-specific patterns, with
E. faecium and
L. plantarum dominating in Jben and
L. pentosus and
Lev. brevis primarily associated with Klila. This pattern aligns with previous studies on traditional Mediterranean and Algerian cheeses, as well as dairy products from the Black Sea region and Morocco. These results confirm that Jben and Klila host characteristic LAB species, underscoring their potential for the development of indigenous starter cultures with both technological and probiotic applications
(Azzouz et al., 2025b; Bouchibane et al., 2023).
The phylogenetic tree (Fig 2) displays distinct clustering of isolates by species, showing high similarity to reference strains and confirming their identification via 16S rRNA gene analysis as members of the genera
Enterococcus,
Lactiplantibacillus and
Levilactobacillus. Such concordance between 16S rRNA-based phylogenetic trees and reference taxonomy is widely used to validate bacterial identification, including LAB. The short branch lengths and tight clustering indicate low genetic divergence and strong phylogenetic relatedness, reflecting relatively homogeneous populations (
Al-shammary et al., 2025).
Technological characterization
pH kinetics and acidifying activity
All isolates showed a gradual decrease in pH accompanied by a corresponding increase in titratable acidity (°D) over 24 h, confirming their fermentative activity (Fig 3). Initial measurements were consistent (pH ≈6.70; acidity ≈19°D), reflecting standardized experimental conditions. By 6 h, a moderate acidification phase was observed, with pH values between approximately 5.75 and 6.15 and acidity rising to 24-36°D. After 24 h of incubation, pH values ranged from 4.02 to 5.80, while titratable acidity reached 33-83°D (Table 2).
Statistical analysis revealed highly significant differences among isolates (p<0.01), confirming variability in acidifying performance. Strains IB4 (JDj03) and IB2 (JDj02) showed the highest acidification capacity (pH ≈4.02-4.16; acidity ≈80-83°D), followed by IB7 (JDj04), IB10 (JDj12) and IB5 (LO23), which also exhibited strong acidifying activity (acidity ≥66°D). In contrast, IB3 (JO09) and IB1 (JO01) displayed significantly weaker acidification (p<0.05), with pH values above 5.2 and acidity below 40°D, while IB6 (JB01), IB8 (CDz09) and IB9 (LB10) showed intermediate profiles.
Based on these results, isolates can be classified into three functional groups: strongly acidifying strains (pH drop ≥ 2 units; acidity ≥66°D), moderately acidifying strains and weakly acidifying strains. A clear inverse relationship between pH and acidity was observed, reflecting lactose conversion into lactic acid.
The acidification kinetics observed followed typical lactic fermentation patterns, characterized by a decrease in pH and an increase in titratable acidity within 24 h. Isolates such as JDj03, JDj02 and LO23 exhibited strong acidifying capacity, confirming their suitability as starter cultures, as rapid acidification enhances milk coagulation and microbial safety. Moderately acidifying strains (CDz09, LB10, JB01) may serve as adjunct cultures, while weak acidifiers (JO01, JO09) were more likely involved in flavor development and maturation, as reported by
Durango Zuleta et al. (2023) and
Sesín et al. (2023).
These findings are supported by the significant pH reduction and acidity values exceeding 60°D observed in active LAB. Overall, the functional diversity among isolates supports the development of multi-strain starter cultures combining complementary properties to improve cheese quality and standardization while preserving traditional characteristics, in agreement with previous studies reported by
Grujović et al. (2024),
Sesín et al. (2023) and
Coelho et al., (2022).
Proteolytic activity
Proteolytic activity, measured by halo diameter (mm), reflects protein hydrolysis and protease production (Table 2). Halo sizes ranged from 0 to 36 mm, indicating substantial variability among isolates. Statistical analysis showed highly significant differences (p<0.001), confirming pronounced functional heterogeneity. While the high coefficient of variation highlights inter-strain differences, low standard deviations indicate good reproducibility of the measurements. Based on halo diameters, isolates were categorized into three functional groups:
• Super-proteolytic (≥ 30 mm): JO01 (31 mm), JDj03 (36 mm) and LO23 (36 mm), corresponding to the highest statistical group a (p<0.05).
• Moderately proteolytic (10-29 mm): JDj02 (20 mm), JB01 (23 mm) and CDz09 (10 mm), falling into intermediate groups b and c.
• Weakly or non-proteolytic (< 10 mm or 0 mm): JO09, JDj04, LB10 and JDj12, forming a homogeneous group
d.
The high activity observed in super-proteolytic strains reflects a strong capacity for casein degradation, promoting peptide formation and flavor development. These strains therefore represent excellent candidates for accelerating cheese ripening and enhancing their bioactive properties, particularly in fresh or ripened cheeses, as reported by
Novak et al., (2021). Isolates with moderate activity play an essential intermediate role by ensuring controlled protein hydrolysis. They help limit bitterness and stabilize the rheological properties of the curd, acting as “metabolic bridges” within the microbial ecosystem. In contrast, weakly or non-proteolytic strains were mainly involved in rapid acidification of the medium, contributing to microbiological safety without directly participating in protein degradation, as described by
Coelho et al., (2022).
This functional stratification, ranging from highly proteolytic to primarily acidifying strains, reflects a rational approach to starter culture selection. Combining these complementary profiles makes it possible to design optimized mixed cultures capable of synchronizing acidification, proteolysis and flavor development, thereby meeting the technological and sensory requirements of traditional cheeses, according to
Coelho et al., (2022) and
Novak et al., (2021).
Assessment of texturizing ability
The ability of the ten LAB isolates from Jben and Klila to produce exopolysaccharides was first screened qualitatively on sucrose-enriched modified BHI agar. Only three isolates, JO01 (IB1), LB10 (IB9) and JDj12 (IB10), displayed a clear viscous (“slimy”) phenotype, indicating active EPS synthesis. This observation is consistent with the common use of colony viscosity as a rapid indicator of EPS-producing strains.
Quantitative analysis performed under controlled conditions (30°C, 48 h, anaerobiosis) confirmed these results. After extraction, purification and lyophilization (Fig 4), EPS yields reached 0.69±0.25 g/L for IB1, 0.86±0.30 g/L for IB9 and 0.56±0.20 g/L for IB10 (Table 2), whereas no detectable production was observed for the remaining isolates. The overall production range (0-0.86 g/L) and the relatively high standard deviations highlight pronounced inter-strain variability, which was statistically significant (p<0.05). This heterogeneity reflects the presence of a limited number of efficient producers within a predominantly non-producing population.
The calculated sugar-to-EPS conversion efficiency (based on 7.5 g sucrose/250 mL) ranged from 1.85% to 2.86%, indicating an active but non-optimized metabolic capacity for polysaccharide synthesis. Despite the absence of process optimization, these values suggest that the selected strains possess a favorable baseline for EPS production.
From a technological perspective, EPS-producing isolates are of particular interest due to their ability to improve the rheological properties of fermented dairy products. These polymers contribute to viscosity enhancement, water retention and structural stabilization of the matrix. In contrast, non-producing strains may still play complementary roles, such as acidification and microbial balance, supporting the concept of mixed starter cultures combining EPS-positive and EPS-negative strains to achieve both textural and microbiological benefits.
When compared with literature data, the yields obtained in this study can be considered relatively high for non-optimized systems. For example, the marine isolate MSD8 produced only 0.20 g/L under similar conditions (
Abdel-monem et al., 2024), which is markedly lower than the values reported here. This difference suggests that isolates from traditional dairy environments may be naturally better adapted for EPS biosynthesis.
Nevertheless, higher production levels have been described for certain
Enterococcus faecium strains. Under rich but non-optimized conditions, strains such as R114 and T52 (isolated from Kishk) reached 2.68 and 2.39 g/L, respectively (
Rahnama Vosough et al., 2022), illustrating the strong influence of strain-specific metabolic potential. Furthermore, optimization strategies based on statistical designs (
e.g., RSM or CCD) have enabled yields of approximately 2.5-3.2 g/L for strains such as R114, F58 and KT990028 (
De Brito et al., 2024;
Zanzan et al., 2023; Rahnama Vosough et al., 2021). In some cases, extreme production levels have been achieved, particularly for fish-derived strains such as MC13 and MC-5, reaching up to 11.6 and 16.5 g/L after optimization
(Tilwani et al., 2021; Kanmani et al., 2013).
Overall, the EPS yields obtained for IB1, IB9 and IB10 fall within the upper range typically reported for LAB cultivated under non-optimized conditions. Although lower than those achieved after process optimization, these results demonstrate a strong intrinsic production capacity. This baseline performance provides a solid foundation for future improvement through bioprocess optimization and supports the potential application of these strains in food and biotechnological fields.