This study involved the collection of freshly voided fecal samples from captive Asian elephants and in vitro laboratory experiments on bacterial isolates. Animals were not restrained, sedated, or subjected to invasive procedures. Samples were collected immediately after defecation during routine husbandry at four tourist elephant camps in Krabi Province, Thailand. Written permission was obtained from all participating elephant camps, and all procedures complied with institutional and international animal welfare guidelines.

The study was conducted between January 2022 and May 2024. Fecal samples were collected from captive Asian elephants at four tourist elephant camps in Krabi Province, Thailand. All laboratory procedures and analyses were performed at the Faculty of Veterinary Science, Rajamangala University of Technology, Srivijaya, Thungyai, Nakhon Si Thammarat, Thailand.

Twenty-five Asian elephants, comprising 23 females and two males, aged 30–50 years and weighing approximately 2,000–4,000 kg, housed at a tourist elephant camp in Krabi Province, Thailand, were included in this study. All elephants were clinically healthy and had not received any medications, dietary supplements, or probiotics prior to the collection of the samples. The elephant camp maintained high hygiene standards and was situated in a forested environment that allowed free roaming, access to natural water streams for bathing, and daily physical activity. The elephants were managed under a free-ranging, non-captive system without cage confinement, which supports good physical and psychological well-being. Low stress and normal behavior indicators, including regular ear flapping and trunk swinging, were observed. The elephants were fed a diet consisting of pineapple stems, Napier grass (Pennisetum purpureum), and Bana grass (P. purpureum × Pennisetum americanum).

Fecal samples were obtained from 25 clinically healthy captive Asian elephants aged between 30 and 50 years, housed at a tourist elephant camp in Krabi Province, Thailand. Freshly voided feces were collected within 4 h of defecation in the morning. All 25 elephants were individually sampled, and fecal samples were collected once from each animal. Fecal specimens were collected in sterile plastic zip bags and immediately placed on ice for transport to the laboratory within 3 h to preserve microbial viability.

For LAB isolation, 10 g of each fecal sample was diluted 10-fold in sterile 0.85% normal saline solution. Subsequently, 0.1 mL aliquots of the 10^-3 to 10^-6 dilutions were spread onto de Man, Rogosa, and Sharpe (MRS) agar (pH 6.5; bioMérieux, Marcy l’Etoile, France). Plates were incubated at 37°C for 48 h under aerobic conditions. The number of LAB isolates recovered per elephant was comparable, with approximately 7–8 isolates obtained from each animal.

Colonies exhibiting distinct morphology were selected and repeatedly subcultured to obtain pure isolates. Preliminary phenotypic characterization was conducted by incubating the pure cultures on MRS agar at 37°C for 48 h under anaerobic conditions, followed by Gram staining, spore staining, and standard biochemical tests, including catalase test, indole test, and oxidase test to confirm their identity as LAB [33].

The antimicrobial activity of LAB isolated against common pathogenic bacteria was evaluated using the disk diffusion assay. In brief, LAB isolates were cultured in MRS broth (bioMérieux, Marcy l’Etoile, France) and incubated at 37°C for 72 h under anaerobic conditions. After incubation, the cultures were centrifuged at 8,000 × g for 10 min to separate the bacterial cells. The supernatant was collected and further sterilized by filtration through a 0.22 µm membrane filter (Sartorius, USA) to obtain the cell-free supernatant, which was used to assess antimicrobial activity.

Pathogenic indicator bacteria, including Escherichia coli (American Type Culture Collection [ATCC] 25922), Salmonella enterica serovar Typhimurium (ATCC 13311), Pseudomonas aeruginosa (ATCC 27853), Staphylococcus aureus (ATCC 25923), and Klebsiella pneumoniae (ATCC 700603), were cultured in tryptic soy broth (bioMérieux, Marcy l’Etoile, France) at 37°C for 24 h under aerobic conditions. The cultures were centrifuged at 8,000 × g for 10 min, and the bacterial pellets were washed twice with phosphate-buffered saline (PBS; pH 7.4). PBS was prepared by dissolving 8.0 g sodium chloride (NaCl; HiMedia, India), 0.2 g potassium chloride (KCl; HiMedia), 1.44 g disodium hydrogen phosphate (Na₂HPO₄; HiMedia), and 0.24 g potassium dihydrogen phosphate (KH₂PO₄; HiMedia) in distilled water, with the final volume adjusted to 1 L and the pH adjusted to 7.4 using hydrochloric acid (HCl; HiMedia) or sodium hydroxide (NaOH; HiMedia). The buffer was sterilized by autoclaving before use.

The washed cell pellets were resuspended in sterile 0.85% NaCl and adjusted to a turbidity equivalent to a 0.5 McFarland standard (approximately 10⁸ Colony-forming units [CFU]/mL) using a suspension turbidity detector (Den-1B, Biosan, Latvia). The bacterial suspensions were swabbed onto Mueller-Hinton agar plates (HiMedia, India). Sterile MRS broth without LAB was included as a negative control for the antimicrobial assay. Sterile paper disks were placed on the agar surface, and 50 µL of cell-free supernatant from each LAB isolate was applied to the disks [34]. Plates were incubated at 37°C for 24 h under aerobic conditions, after which the inhibition zone diameters surrounding the disks were measured. These measurements were used to evaluate the LAB isolates’ inhibitory potential against the tested pathogens.

Cell surface hydrophobicity is a key characteristic of LAB associated with their ability to adhere to the intestinal mucosa, which is considered a critical criterion for the selection of probiotics. The hydrophobicity of LAB isolates was assessed using the MATH method described by Rahman et al. [37]. Briefly, LAB cells were harvested by centrifugation at 8,000 × g for 10 min, and the optical density at 600 nm (OD₆₀₀) of the bacterial suspension was measured and recorded as A¹.A 3.0 mL aliquot of the suspension was then transferred into a clean test tube, and 1.0 mL of hydrocarbon (xylene: Cat. No. 9490-03; J. T. Baker, Avantor, United Kingdom) was added. The ratio of xylene to the bacterial suspension was 3:1 (v/v). The mixture was vigorously vortexed for 5 min and then left undisturbed for 10–15 min to facilitate phase separation. The aqueous (lower) phase was carefully collected, and its OD₆₀₀ was measured and recorded as A₂. The hydrophobicity index (HPBI) was calculated using the following equation:

HPBI (%) = [(A₁– A₂) / A₁] × 100

According to commonly adopted criteria, probiotic candidates exhibiting hydrophobicity values of ≥ 40% are considered acceptable, whereas values of ≥ 60% and ≥ 70% are indicative of good and excellent hydrophobicity, respectively [38].

The autoaggregation ability of the LAB isolates was evaluated following the method described by Kos et al. [39]. Briefly, the LAB isolates were grown in MRS broth at 37°C for 24 h under anaerobic conditions. Cultures were centrifuged at 8,000 × g for 10 min, and the bacterial pellets were washed twice with PBS (pH 7.4). The cell pellets were resuspended in PBS OD₆₀₀ of approximately 0.6 (approximately 10⁸ CFU/mL). Bacterial suspensions were incubated at room temperature (25°C ± 2°C) without agitation. The optical density of the upper suspension was measured at 0 and 4 h. The autoaggregation was calculated using the following equation:

Autoaggregation (%) = [1 − (OD₆₀₀ at 4 h / OD₆₀₀ at 0 h)] × 100

According to commonly adopted criteria, probiotic candidates exhibiting autoaggregation values of ≥ 40% are considered acceptable, whereas values of ≥ 60% and ≥ 70% are indicative of good and excellent autoaggregation ability, respectively [38]. Higher autoaggregation percentages indicated a strong ability of the isolates to self-associate, a property considered advantageous for intestinal colonization and probiotic functionality.

All measurements were performed in 11 biological replicates (n = 11), and each biological replicate was analyzed in triplicate. The enumeration of LAB was conducted in accordance with ISO 15214:2015, and probiotic characteristics were evaluated following the guidelines recommended by the Food and Agriculture Organization/World Health Organization [40, 41].

Bacterial species were identified using molecular techniques targeting the 16S rRNA gene. Genomic DNA was extracted from a bacterial suspension (approximately 10⁹ cells) using a commercial genomic DNA isolation kit (Bio-Helix, Taiwan) according to the manufacturer’s instructions. Total DNA quality was evaluated using a NanoDrop spectrophotometer (Maestrogen Inc., China), yielding an A260/280 absorbance ratio of 1.9, indicating high-quality DNA. Two microliters of the extracted DNA (10 ng/µL) served as a template for polymerase chain reaction (PCR) amplification using the universal bacterial primers 27F and 1492R [43]. The 16S rRNA gene was amplified using these universal primers, generating an amplicon of approximately 1,500 bp. Briefly, 2 µL of 27F (10 ng/µL) and 2 µL of 1492R (10 ng/µL) in 25 µL of PCR master mix (Cat. No. KMM-101; Toyobo, Japan). The PCR cycling conditions were as follows: initial denaturation at 95°C for 2 min, followed by 35 cycles of denaturation at 95°C for 1 min, annealing at 55°C for 1 min, and extension at 72°C for 1 min, with a final extension at 72°C for 10 min. PCR products were purified using a MultiScreen filter plate (Millipore Corp., Bedford, MA, USA). The PRISM BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) was used to perform sequencing reactions. The extension products were mixed with Hi-Di formamide (Applied Biosystems, Foster City, CA, USA), denatured at 95°C for 5 min, chilled on ice for 5 min, and analyzed using an ABI PRISM 3730XL DNA Analyzer (Applied Biosystems, Foster City, CA, USA).

Phylogenetic relationships were inferred using 16S rRNA gene sequences with the maximum likelihood method and the Kimura 2-parameter model implemented in MEGA version 11 software (www.megasoftware. net). The best-fit nucleotide substitution model was selected based on the Akaike Information Criterion and Bayesian Information Criterion values implemented in MEGA, indicating that the K2P model provided the optimal fit for the dataset. The robustness of the tree topology was assessed using bootstrap analysis with 1,000 replicates. The percentage of replicate trees supporting each node is shown next to the corresponding branches. The Neighbor-Joining and BioNJ algorithms were used to automatically generate initial trees for the heuristic search based on a matrix of pairwise distances estimated by the Maximum Composite Likelihood approach, and the topology with the highest log-likelihood value was selected.

Data normality was assessed using the Kolmogorov–Smirnov test before analysis. As the data were normally distributed, differences among groups were analyzed using one-way analysis of variance. When significant differences were detected, Duncan’s multiple range test was used to perform post hoc comparisons. Outlier detection was conducted using the Z-score method Z = (x − mean)/SD, and no outliers were identified. A p-value of <0.05 was considered statistically significant. All statistical analyses were performed using the Statistical Package for the Social Sciences software version 22 for Windows (SPSS Inc., Chicago, IL, USA). Data visualization and figure preparation were performed using Microsoft Excel and Microsoft Paint, respectively.

A total of 195 LAB isolates were recovered from fecal samples collected from 25 captive Asian elephants. All isolates were identified as Gram-positive, non-spore-forming bacteria and tested negative for catalase, oxidase, and indole activity. Morphological examination revealed that 133 isolates (68.20%) were Gram-positive rod-shaped bacteria, whereas the remaining 62 isolates (31.80%) exhibited a Gram-positive cocci-shaped morphology.

Among the 195 LAB isolates obtained from elephant feces, 52 isolates (26.67%) exhibited inhibitory activity against all five tested pathogenic bacterial strains. A substantial proportion of isolates demonstrated antagonistic activity against at least one indicator organism. Specifically, 124 isolates (63.59%) inhibited E. coli, 115 (58.97%) inhibited P. aeruginosa, 117 (60.00%) inhibited S. Typhimurium, 130 (66.67%) inhibited K. pneumoniae, and 108 isolates (55.38%) inhibited S. aureus. These findings indicate the broad-spectrum antimicrobial potential of LAB isolates, with inhibition zone diameters ≥12 mm.

The 52 isolates that inhibited all five pathogenic bacteria were selected for further evaluation of probiotic properties. Among these, 11 isolates (I1, I9, I12, I56, I90, I91, I115, I131, I145, I146, and I182) exhibited favorable characteristics, including acid and bile salt tolerance, cell surface hydrophobicity, and autoaggregation capacity, all exceeding 50%. Further analysis showed that isolates I9, I12, I115, and I145 exhibited significantly higher inhibitory activity (p < 0.05) against S. Typhimurium, while isolate I9 demonstrated the strongest inhibition of E. coli (Figure 1). The antimicrobial activity and probiotic properties of the selected isolates are summarized in Table 1.

(a) Colony morphology of isolate I9; (b) Gram-stained micrograph of isolate I9 showing Gram-positive rod-shaped bacteria under light microscopy (100×); (c) disk diffusion assay illustrating the antimicrobial activity of the cell-free supernatant from isolate I9 cultured in de Man, Rogosa, and Sharpe (MRS) broth against Escherichia coli, with the largest inhibition zone indicated by an arrow; (d) negative control (sterile MRS broth without lactic acid bacteria) showing no inhibition zone (arrow). Inhibition zone diameters include the 6-mm paper disc. Scale bar = 20 µm.

Results are presented as mean inhibition zone diameters (mm) ± standard deviation from three independent replicates (n = 11). Mean values within the same column followed by different superscript letters indicate statistically significant differences (p < 0.05). Inhibition zone diameters include the diameter of the paper disc (6 mm). E. coli, S. Typhimurium, P. aeruginosa, K. pneumoniae, and S. aureus were used as indicator pathogenic microorganisms.

The acid and bile salt tolerance of the 11 selected LAB isolates is summarized in Table 2. All isolates exhibited survival rates greater than 50% under acidic conditions (pH 3.0) and in the presence of 1% bile salts, indicating physiological resilience compatible with probiotic application. Isolate I9 showed the highest tolerance to both acid and bile salt exposure. Notably, four isolates (I1, I9, I56, and I145) exhibited acid tolerance rates exceeding 90%, while two isolates (I9 and I145) demonstrated bile salt tolerance rates above 95%.

As shown in Figure 2, isolates I9, I56, I115, I131, and I145 exhibited the highest levels of cell surface hydrophobicity among all tested strains. In contrast, isolates I146 and I182 displayed the lowest hydrophobicity values.

High autoaggregation capacity was observed in isolates I9, I56, I115, I131, I145, I146, and I182 (Figure 3), indicating a strong ability for cell-to-cell adhesion, a trait associated with enhanced colonization potential within the host gastrointestinal tract. Conversely, isolates I1, I12, I90, and I91 exhibited comparatively lower autoaggregation values.

All 11 selected LAB isolates were susceptible to five commonly used antibiotics, namely ampicillin, chloramphenicol, streptomycin, tetracycline, and enrofloxacin (Table 3). In addition, hemolysis assays performed on Columbia agar supplemented with 5% sheep RBC confirmed that all isolates exhibited non-hemolytic activity, corresponding to gamma-hemolysis.

Isolates were classified as susceptible (S) when inhibition zone diameters were ≥17.5 mm for ampicillin, chloramphenicol, streptomycin, and tetracycline, and ≥18.0 mm for enrofloxacin. Resistance (R) was defined as inhibition zone diameters ≤14.5 mm for ampicillin, chloramphenicol, streptomycin, and tetracycline, and ≤14.0 mm for enrofloxacin.

The 11 selected LAB isolates were subjected to molecular identification using 16S rRNA gene sequencing followed by phylogenetic analysis (Figure 4). Sequence alignment results identified four isolates as L. plantarum, with sequence identity values ranging from 99.54% to 99.87%. The remaining seven isolates were identified as Enterococcus faecalis, with sequence identities ranging from 99.58% to 99.80%. Detailed molecular identification data are presented in Table 4. This study represents the first molecularly validated identification of L. plantarum strains isolated from captive Asian elephants based on full-length 16S rRNA gene sequencing and phylogenetic confirmation.

Phylogenetic tree of 11 lactic acid bacteria isolates exhibiting probiotic characteristics, constructed using 16S rRNA gene sequences and the maximum likelihood method with the Kimura two-parameter model implemented in MEGA version 11 software. Tree topology robustness was evaluated by bootstrap analysis based on 1,000 replicates. The tree with the highest log-likelihood value is presented. Bootstrap support values, expressed as percentages, are shown next to the corresponding branches and indicate the frequency with which associated taxa clustered together among replicate trees. Initial trees for the heuristic search were generated using the Neighbor-Joining and BioNJ algorithms based on a matrix of pairwise distances estimated by the Maximum Composite Likelihood approach, and the topology with the highest log-likelihood value was selected.

This study successfully isolated 195 LAB strains from fecal samples of 25 healthy Asian elephants. To the best of our knowledge, this is the first study to conduct a comprehensive, multidimensional evaluation of elephant-derived LAB by integrating antimicrobial activity, acid and bile tolerance, cell surface hydrophobicity, autoaggregation, antibiotic susceptibility, and hemolytic safety testing. Among these, 52 isolates (26.67%) demonstrated antimicrobial activity against five clinically pathogenic bacteria, namely E. coli, P. aeruginosa, S. Typhimurium, K. pneumoniae, and S. aureus, which are common causative agents of gastrointestinal and opportunistic infections in elephants, other animals, and humans [27–31].

Isolate I9 showed the highest inhibition of E. coli (Figure 1) and K. pneumoniae. Isolates I9, I12, I115, and I145 exhibited significantly higher inhibition of S. Typhimurium, whereas isolate I145 showed the strongest inhibition of P. aeruginosa. In addition, isolates I9, I56, I145, and I146 exhibited significant inhibition of S. aureus. This study represents the first report demonstrating strong broad-spectrum antimicrobial activity of elephant-derived L. plantarum against clinically relevant enteric and opportunistic pathogens. These findings highlight strain-specific variation in antimicrobial efficacy among the isolates, consistent with previous reports [18].

The selection of E. coli (ATCC 25922), S. Typhimurium (ATCC 13311), P. aeruginosa (ATCC 27853), S. aureus (ATCC 25923), and K. pneumoniae (ATCC 700603) as indicator pathogens followed internationally accepted probiotic screening strategies, ensuring methodological robustness and enabling comparison with existing studies. As environmentally ubiquitous bacteria, their inclusion provides ecologically relevant preliminary evidence of the capacity of probiotic strains to inhibit common enteric and opportunistic pathogens encountered by elephants through dietary and environmental exposure.

The inhibition zones produced by the cell-free supernatant of LAB isolates ranged from 11 to 13 mm, classifying them as strong antimicrobial activity according to established criteria [32]. These findings are consistent with previous studies in which L. plantarum isolated from swine feces inhibited swine pathogens [44], and L. plantarum strains derived from sheep feces suppressed foodborne pathogens [45]. In addition, LAB species such as Lactobacillus pentosus, Lactococcus garvieae, and Enterococcus hirae have been isolated from wild elephant feces in Thailand [46]. The antimicrobial activity of probiotics is attributed to the production of bioactive compounds, including bacteriocins, lactic acid, acetic acid, propionic acid, ethanol, and antimicrobial peptides [47]. Variations in antimicrobial efficacy among isolates may reflect differences in microbial species and growth conditions that influence the type and quantity of antimicrobial compounds produced [48].

Acid and bile salt tolerance are critical attributes for probiotic bacteria to survive gastrointestinal transit and establish persistence in the host. In this study, acid tolerance was assessed at pH 3.0, reflecting the typical gastric pH range observed in elephants [49]. Previous studies have shown that the robust cell wall structure of Lactobacilli, composed primarily of thick peptidoglycan and teichoic acids, contributes to resistance under acidic conditions [50].

Bile salts exert antimicrobial effects by disrupting bacterial phospholipid bilayers and cell membranes, leading to cell lysis [51], and higher bile salt concentrations are associated with increased bacterial mortality [52]. The 1% bile salt concentration used in this study represents a physiologically relevant level in the intestinal tract of animals [53]. Exopolysaccharide production by L. plantarum has been reported to enhance resistance to acidic and bile stress conditions [54]. All 11 LAB isolates demonstrated survival rates exceeding 50% under these stress conditions, with isolate I9 showing the highest tolerance to both acid and bile salts, indicating strong potential for gastrointestinal survival.

Cell surface hydrophobicity and autoaggregation are key attributes contributing to bacterial adhesion and intestinal mucosal colonization. High hydrophobicity is generally associated with increased adherence to epithelial cells and is influenced by bacterial cell envelope components, including lipoteichoic acids, teichoic acids, S-layer proteins [55], and mannose-specific lectins [56]. Together with autoaggregation, these properties contribute to microbial homeostasis and competitive exclusion of pathogenic microorganisms [57].

Previous studies have shown that L. plantarum, L. rhamnosus, L. acidophilus [58], and Lactobacillus salivarius [59] can initiate autoaggregation within minutes to several hours, depending on environmental and growth conditions. Hydrophobicity and autoaggregation are strain-specific traits that vary among Lactobacillus isolates [60]. In this study, isolates I9, I56, I115, I131, and I145 exhibited significantly higher hydrophobicity and autoaggregation than other tested strains, indicating enhanced mucosal adhesion and colonization potential. The L. plantarum strains were isolated directly from the elephant gut ecosystem and exhibited traits consistent with adaptation to hindgut fermentation physiology, supporting the concept of host-specific probiotics.

Antibiotic susceptibility and hemolytic activity are critical parameters for probiotic safety assessment. Hemolytic activity facilitates host tissue invasion by pathogenic bacteria; therefore, the absence of hemolysis is a key safety requirement [61]. All 11 LAB isolates were susceptible to ampicillin, chloramphenicol, streptomycin, tetracycline, and enrofloxacin, suggesting a low risk of transferable antibiotic resistance.

Furthermore, none of the isolates exhibited hemolytic activity when cultured on blood agar supplemented with sheep RBC, and all were classified as gamma-hemolytic. These findings provide the first verified safety profile of elephant-derived LAB, supporting their suitability for probiotic development in animals and humans.

Molecular identification based on 16S rRNA gene sequencing revealed that four isolates belonged to L. plantarum and seven isolates were identified as E. faecalis. Both species are commonly present in the gastrointestinal tract and have been widely studied for probiotic potential. L. plantarum is recognized for its safety and efficacy as a probiotic in humans and animals [55]. Although E. faecalis is considered a commensal organism, its opportunistic pathogenic potential necessitates careful strain-level safety evaluation. Non-hemolytic and antibiotic-sensitive E. faecalis strains, such as those identified in this study, may be considered acceptable probiotic candidates [62]. However, concerns regarding virulence factors and antimicrobial resistance limit broader application of Enterococcus-based probiotics, further supporting L. plantarum as the primary candidate for continued investigation.

Among the identified isolates, I9, I56, I115, I131, and I145 exhibited significantly higher hydrophobicity and autoaggregation. Isolates I9, I56, and I145 demonstrated the most promising probiotic potential overall, based on superior performance across antimicrobial activity, acid and bile salt tolerance, hydrophobicity, and autoaggregation. All three isolates were identified as L. plantarum based on 16S rRNA gene analysis. This study represents the first report identifying elite LAB strains from elephants with consistently strong probiotic profiles.

The probiotic potential of L. plantarum has also been reported in other megaherbivores, including rhinoceros (Rhinoceros spp.) [63], hippopotamus (Hippopotamus amphibius) [64], and giraffes (Giraffa camelopardalis) [65]. These findings indicate that L. plantarum is a recurrent component of the gastrointestinal microbiota of large herbivores. Although L. plantarum is not a primary producer of short-chain fatty acids, LAB can support short-chain fatty acid metabolism through cross-feeding interactions [66–69]. The strong tolerance, hydrophobicity, and autoaggregation observed in elephant-derived strains suggest persistence and functional relevance in the elephant hindgut. Host-derived strains may offer improved adaptation compared with non-host commercial probiotics, warranting further in vivo validation. From a One Health perspective, host-specific probiotics for elephants may reduce antibiotic use, limit antimicrobial resistance selection, and decrease pathogen dissemination across animals, the environment, and humans [70].