Milk samples, ~1 mL each, were obtained from two healthy domestic Beagle dogs that had not been administered antibiotics or probiotics for at least 1 month prior to sample collection. The samples were immediately transported to the laboratory under cold chain conditions and stored at 4°C until further analysis.

The Isolation, Purification and Species Identification was conducted based on Zhang's report with minor adjustments (24). To isolate LAB strains, 0.1 mL of canine milk was inoculated onto de Man, Rogosa, and Sharpe (MRS) agar plates (Hopebio, Qingdao, China) and incubated aerobically at 37°C for 24–72 h until distinct colonies appeared. Isolated colonies were subjected to three rounds of purification by re-culturing on MRS agar plates, followed by enrichment in MRS broth. All subsequent bacterial counting work is completed by Halo counter HD-4 Cell Counter (Hiscore Inc., China). The 16S rRNA genes of all isolates were amplified using universal primers 27F (5′-AGGTTTTTGATCCTGGCCAG-3′) and 1492R (5′-TACGACTTAACCCCAATCGC-3′). The resulting amplicons were sequenced by Sangon Bioengineering (Chengdu) Co., Ltd., and the sequences were aligned using BLAST against GenBank bacterial nucleic acid databases. Species identification was performed using MEGA6 software (Mega Limited, Auckland, New Zealand), employing the neighbor-joining method and the Kimura 2-parameter model to construct a phylogenetic tree. A sequence similarity threshold of >99% was used to define species.

Growth performance was evaluated by measuring the growth curve of LAB isolates as previously described by Liu et al. (25). A 50 μL aliquot (1%) of LAB culture in the mid-exponential phase was inoculated into 50 mL of fresh MRS broth and incubated at 37°C for 48 h. Optical density (OD600) measurements were taken every 2 h during the first 12 h and subsequently every 4 h up to 48 h.

Tolerance of the LAB isolates to simulated gastrointestinal conditions was tested according to the method of Zhang et al. (24), with slight adjustments to simulate the specific gastrointestinal fluid conditions in dogs. The LAB isolates were incubated in MRS broth for 12 h. Then, 10 ml of the fresh cultures were centrifuged at 8,000 × g for 10 min at 25°C to obtain bacterial cells. Bacterial cells were suspended in 10 mL of simulated gastric juice and incubated at 37°C for 3 h. Cells were then harvested by centrifugation and transferred into 10 mL of simulated intestinal juice for incubation at 37°C for an additional 4 h. Bacterial survival at 0, 3, and 7 h was assessed using colony-forming units (CFU) determined by plate counting. Survival rates were calculated as the ratio of surviving CFUs at each time point to the previous time point count.

The Antibacterial activity detection was conducted based on Zhang's report with minor adjustments (24). The antibacterial potential of LAB isolates was tested using the Oxford cup method against four common enteropathogens: Staphylococcus aureus ATCC 25923, Escherichia coli ATCC 25922, Salmonella enterica H9812, and Pseudomonas aeruginosa PAO1. Cell-free supernatants (CFS) were obtained by centrifuging LAB cultures at 4,500 × g for 10 min and filtering the supernatant through a 0.22 μm sterile filter. The pathogenic bacteria were cultured in LB broth, diluted to about 10⁷ CFU/mL, and 100 μL of each suspension was spread onto LB agar plates. Three sterile Oxford cups were placed on each plate and filled with 200 μL of CFS. Plates were incubated at 37°C for 24 h, and inhibition zones were measured using a caliper to determine antibacterial activity.

Co-aggregation of LAB with the four pathogens was assessed by mixing equal volumes (2 mL) of LAB and pathogen suspensions, followed by incubation at 37°C for 2 h. Absorbance (A₆₀₀) was measured before and after incubation for both individual and mixed cultures. The co-aggregation rate was calculated using the formula: Co-aggregation rate (%) = 1 – A_mix / [ (A_LAB + A_pathogen) / 2] × 100.

According to the report by Zhang et al. (24). Auto-aggregation activity was measured by adjusting the LAB culture to 1 × 10⁸ CFU/mL in phosphate-buffered saline (PBS). After incubation at 37°C for 6 h, the OD600 of the upper layer of the bacterial suspension was measured, and auto-aggregation was calculated as: Auto-aggregation rate (%) = 1 – (A₁/A₀) × 100, where A₀ is the initial OD and A₁ is the OD at 6 h.

The hydrophobicity of LAB strains in various organic solvents was determined using a modified version of the method reported by Kos (26). The LAB strains that were activated were introduced into MRS liquid medium at a concentration of 1% (v/v) and incubated overnight. Afterward, they were centrifuged at 8,000 × g for 10 min at 4°C to collect the microorganisms. The microorganisms were then washed three times with sterile phosphate-buffered saline (PBS, pH = 7.4) and resuspended in PBS. The absorbance of the lactobacilli suspension was adjusted to OD600 = 0.60 ± 0.05 (A₀). LAB strains suspension (3 mL) was mixed with 1 mL of different organic solvents (ethyl acetate, xylol, and trichloromethane), vortexed and shaken for 2 min, and then allowed to stand for 20 min. The OD value of the aqueous phase was measured at 600 nm using a UV spectrophotometer (A₁). The experiment was repeated three times. The hydrophobicity of LAB was calculated using the following formula: Hydrophobicity rate (%) = (1-A₁/A₀) × 100.

The Adhesion to Caco-2 cells was conducted based on Wang's report with minor adjustments (27). Caco-2 cells were cultured in a flask until they reached a sub-confluent state of 80–90%. They were then digested with 0.25% trypsin and counted using a hemocytometer. The concentration of viable cells was adjusted to 1 × 10⁵ cells/mL (V_C) using DMEM medium. The cell suspension was added to a 12-well cell culture plate at a volume of 1 mL per well. The plate was then incubated at 37°C and 5% CO₂ until the cells formed a confluent monolayer, typically within 24–48 h. Prior to the adhesion assay, the cells were cultured for 1 day, and the medium was replaced with high-glucose DMEM (without antibiotics) to avoid any influence of antibiotic resistance on the adhesion assay. At the beginning of the assay, the cells were washed three times with sterile PBS and 1 mL of a LAB suspension at a concentration of 1 × 10⁸ CFU/mL (V₀) was added to each well. The cells were then incubated at 37°C and 5% CO₂ in a constant-temperature cell culture incubator for 2 h. After the incubation period, the cells were washed three times with sterile PBS to remove any unattached LAB. The cells were digested using 0.25% trypsin, which should be handled with care due to its potential hazards. After complete digestion, the cells were collected and subjected to 10-fold gradient dilution. The number of viable adherent LAB after dilution on MRS solid media was determined using plate colony counting (V₁) after incubating the plates at 37°C for 48 h in an anaerobic jar. The experiment was conducted three times. The adhesion rate and adhesion index of LAB to Caco-2 cells were calculated using the following formula: Adhesion rate (%) = (V₁ / V₀) × 100; Adhesion index (CFU/cell) = V₁ / V_C.

Research indicates that LAB with a strong ability to form biofilm exhibit better resistance to heat and freezing (28). The biofilm-forming ability of LAB strains was determined using crystal violet staining (29). The concentration of LAB was adjusted to 1 × 10⁷ CFU/mL. The LAB suspension was inoculated into 96-well cell culture plates at a volume of 200 μL per well. The plates were then incubated in a 37°C incubator for 24 h to form a stable biofilm. Blank MRS liquid medium was used as a control. The bacteria were washed three times with sterile PBS to remove the planktonic bacteria, and then dried at room temperature for 15 min. Next, they were fixed in 200 μL of methanol solution for 15 min and dried again at room temperature for 10 min. The sample was immersed in a 1% crystal violet solution (200 μL) for 20 min, rinsed three times with distilled water, and air-dried for 10 min. It was then eluted in a 33% acetic acid solution (200 μL) for 10 min, and the resulting decolorized solution was analyzed at a wavelength of 595 nm using an enzyme marker. The OD value of the decolorized solution at 595 nm was measured using an enzyme counter (the control was recorded as A₀ and the LAB were recorded as A). The strength of LAB biofilm formation ability was evaluated based on the following criteria: no biofilm formation ability (–): A <A₀; weak biofilm forming ability (+): A₀ <A ≤ 2A₀; moderate biofilm forming ability (++): 2A₀ <A ≤ 4A₀; and strong biofilm forming ability (+++): A > 4A₀. The experiments were repeated three times.

The Safety assessment test was conducted based on Zhang's report with minor adjustments (24). LAB isolates were streaked onto blood agar plates and incubated at 37°C for 48 h. Staphylococcus aureus ATCC 25923 was used as a positive control. Hemolysis was assessed by the presence of clear zones (β-hemolysis), green zones (α-hemolysis), or no zones (γ-hemolysis).

The susceptibility of the selected LAB strains to antibiotics was assessed using the disc-diffusion test. A total of 14 antimicrobials (Shunyoubio, Shanghai, China) were tested, including penicillin G (P, 10 μg), ampicillin (AMP, 10 μg), amoxicillin (AML, 25 μg), erythromycin (E, 15 μg), Cefuroxim (CXM, 30 μg), and cefotaxime (CTX). Oxacillin (OX, 5 μg), Cefazolin (KZ, 30 μg), Norfloxacin (NOR, 5 μg), Rifampicin (RD, 5 μg), clindamycin (DA, 10 μg), chloramphenicol (C, 30 μg), tetracycline (TE, 30 μg), and vancomycin (VA, 30 μg). Fresh overnight cultures of each LAB strain were diluted to a concentration of 10⁸ CFU/ml. Subsequently, 100 μl of the diluted cultures were spread on MRS agar plates and dried. Three uniform antibiotic discs were manually placed on the surface of the dried MRS plates. The plates were then inverted and incubated for 48 h under anaerobic conditions at 37°C. Antibiotic susceptibility was classified as resistant (R), moderately susceptible (M), or sensitive (S) based on the diameter of the zone of inhibition (mm) according to the parameters of the Clinical and Laboratory Standards Institute (65).

The tolerance of LAB to H₂O₂ was measured using the method reported by Xiong et al. (30). A liquid culture of LAB with a concentration of 1 × 10⁸ CFU/mL was inoculated into MRS liquid culture medium containing 0, 0.5, 1.0, 1.5, and 2.0 mmol/L H₂O₂ at an inoculum volume of 2% (v/v). The mixture was incubated for 8 h at 37°C in a constant-temperature incubator, and the OD value of the culture medium was measured at a wavelength of 600 nm using a UV spectrophotometer. The experiment was repeated three times.

The ability of LAB to scavenge the 1,1-diphenyl-2-picrylhydrazyl (DPPH) free radical was determined using the method described in the literature (31). Two milliliter of 0.2 mmol/L DPPH absolute ethanol solution was added to a centrifuge tube containing 1 mL of cell-free supernatant or bacterial suspension of lactic acid bacteria. The mixture was vortexed and left to react for 30 min at room temperature in the dark at 4°C. The supernatant was collected by centrifuging at 8,000 × g for 10 min. The OD value of the supernatant was measured at a wavelength of 517 nm using a UV spectrophotometer (OD_sample). Anhydrous ethanol was used as the blank group instead of DPPH absolute ethanol solution (OD_blank). Distilled water was used as the control group instead of the reaction sample (OD_control). The experiment was conducted three times. Afterward, we calculated the DPPH free radical scavenging rate of LAB using the following formula: DPPH free radical scavenging rate (%) = [1– (OD_sample – OD_blank) / OD_control] × 100.

The OH⁻ scavenging capability was determined following the protocol outlined by Alam et al. with certain adjustments (32). Five hundred microliters of LAB cell-free supernatant or suspension were added to a centrifuge tube, followed by 1 mL of 0.1% 1,10-phenanthroline, 1 mL of PBS, 1 mL of 2.5 mmol/L FeSO4, and 1 mL of 20 mmol/L H₂O₂. After incubating for 1.5 h in a water bath set at 37°C, we measured the OD536 of the reaction mixture (OD_sample). In the blank group, we substituted a consistent volume of absolute ethanol for H₂O₂ (OD_blank). Similarly, in the control group, we replaced the sample solution with an equivalent volume of distilled water (OD_control). We determined the OH⁻ radical scavenging rate using the following formula: OH⁻ free radical scavenging rate (%) = [(OD_sample − OD_control)/ (OD_blank− OD_control)] × 100.

The ability of LAB to scavenge O²⁻ free radicals was evaluated using the method described by Liu et al. (33). LAB cell-free supernatant or bacterial suspension (100 μL) was mixed with 2.8 mL of 0.05 mol/L Tris-HCl (pH 8.2) and 100 μL of 0.05 mol/L pyrogallol. The mixture was vortexed and incubated at 25°C in the dark. After 4 min, the reaction was stopped by adding 1 mL of 8 mol/L HCl. To measure the OD value of the reaction solution at a wavelength of 320 nm (OD_sample), use a UV spectrophotometer. Adjust to zero with distilled water, which replaces the sample for reaction as a control group (OD_control). The experiment was repeated three times. Then, calculate the O²⁻ free radical scavenging rate of LAB using the following formula: O²⁻ free radical scavenging rate (%) = [1 − OD_sample / OD_control] × 100.

The ability of LAB to produce EPS was determined using the method described by Ren et al. (34). To prepare the cell-free LAB supernatant, follow the same procedure as in the antioxidant test, as previously described. Mix the supernatant with 800 mg/mL trichloroacetic acid to obtain a final concentration of 40 mg/mL. Incubate the mixture at 4°C in a temperature-controlled incubator overnight. Centrifuge the mixture at 4°C at 8,000 × g for 10 min to collect the supernatant. Add 250 μL of 6% phenol solution and 1 mL of concentrated sulfuric acid to 250 μL of supernatant. Mix well in an ice-water bath and cool to room temperature. Next, add 200 μL of the reaction solution to a 96-well cell culture plate and measure the OD value of the reaction solution at a wavelength of 490 nm using a microplate reader, with the blank set to the reagent mixture without the supernatant. To measure the concentration of EPS produced by LAB, prepare a standard curve using glucose solutions with concentrations of 3.125, 6.25, 12.5, 25, 50, and 100 mg/L in the presence of trichloroacetic acid, phenol, and sulfuric acid. Repeat the experiment three times to ensure accuracy.

The ability of LAB to produce GABA was determined using the Berthelot colorimetric method as reported by Zhang et al. (35). The LAB strains were activated and then introduced into a fermentation medium containing glucose, yeast extract, and peptone (GYP), with an inoculation volume of 2% (v/v). After being cultured overnight, the LAB were centrifuged at 8,000 × g for 10 min at 4°C. The supernatant should be collected and mixed with 200 μL of 0.2 mol/L borate buffer (pH 9.0), 1 mL of 6% phenol, and 0.4 mL of sodium hypochlorite solution with an available chlorine content of 5.5%. After adding these components, vortex the mixture. Place the mixture in a boiling water bath for 10 min, followed by an ice water bath for 20 min. Then, add 2 mL of a 60% ethanol solution and mix thoroughly using a vortex. Measure the OD value of the reaction solution at a wavelength of 645 nm using a UV spectrophotometer and adjust to zero with distilled water. To calculate the concentration of GABA produced by LAB, draw a standard curve using GABA standards with concentrations of 0, 0.2, 0.4, 0.6, 0.8, and 1.0 g/L. Repeat the experiment three times.

The production of BSH by LAB was determined following the method described in Wang et al. (36). The resulting supernatant was collected to obtain a cell-free extract. To prepare the antioxidant test samples, the bacterial suspension was mixed with 10 mmol/L dithiothreitol, sonicated for 10 min in an ice bath, and then centrifuged at 4°C and 8,000 × g for 10 min. To 10 μL of LAB cell-free supernatant or cell-free extract, add 180 μL of 0.1 mol/L PBS (pH 6.0) and 10 μL of 0.1 mol/L sodium taurocholate solution. Then, add 200 μL of 15% trichloroacetic acid and react for 1 min. Centrifuge the mixture at 4°C, 8,000 × g for 10 min and collect the supernatant. Add 100 μL of the supernatant to 1.9 mL of ninhydrin chromogenic solution and vortex. Vortex the mixture and place it in a 37°C water bath for 30 min. Put the mixture in a boiling water bath for 15 min and then in an ice water bath for 3 min. The OD value of the reaction solution was measured at a wavelength of 570 nm using a UV spectrophotometer. Trichloroacetic acid was added to the sample, followed by the addition of sodium taurocholate solution as a control group for the reaction. A standard curve was constructed using glycine standards at concentrations of 0, 0.1, 0.2, 0.3, 0.4, and 0.5 μmol/L. Enzyme activity was defined as the production of 1 μg of glycine per minute at 37°C, with 1 unit of enzyme activity being equivalent to this amount. The experiment was repeated three times.

The 16S rRNA sequences of the 6 LAB isolates obtained were compared with the sequences in GenBank. The strains were identified as members of the Lactobacillus and Enterococcus genera, which are commonly recognized as lactic acid bacteria. Specifically, the strains were related to Lactobacillus johnsonii (L218), Lactobacillus reuteri (L219), Enterococcus faecium (L220), Lactobacillus acidophilus (L221), Lactobacillus animalis (L222), and Enterococcus faecalis (L223), with sequence similarity exceeding 98%. The phylogenetic tree, constructed based on the 16S rRNA gene sequences and shown in Figure 1, provides a visual representation of the genetic relatedness among these isolates.

The growth conditions of the six lactic acid bacteria strains and the control strain LGG are shown in Figure 2a. It was obvious that the growth cycles of all strains were relatively similar, entering the exponential growth phase after about 4 h, and the OD600 value increased exponentially. After 24 h of culture, the growth rate decreased and entered a stable phase. Among them, Lactobacillus rhamnosus LGG showed a significantly stronger growth speed and performance compared to the other strains.

Figure 2b shows the survival rate results of LAB strains under simulated gastrointestinal tract (GIT) conditions. After being treated with artificial gastric juice for 3 h, except for Enterococcus faecium L220, other isolated strains showed higher resistance to simulated gastric juice, with survival rates exceeding 60%. The specific average survival rates for these strains ranged from 43.91 to 73.65%, and there was no significant difference compared with the control group, which was treated with the same artificial gastric juice (P > 0.05). Seven hours after incubating the surviving cells in artificial intestinal fluid, strains L221 (93.59%) and L219 (83.39%) showed better resistance to artificial intestinal fluid, with no significant difference compared with the control group (P > 0.05).

Table 1 shows that the six strains exhibit varying levels of antagonistic activity against four common intestinal pathogenic bacteria. The control strain LGG exhibited the strongest antagonistic activity against Escherichia coli and Staphylococcus aureus. Its antagonistic activity against Escherichia coli was significantly better than that of the other isolated strains (P < 0.05). L221 and L222 were the isolates with the strongest antagonistic activity against Salmonella and Pseudomonas aeruginosa, respectively. Both were superior to the control strain LGG (P < 0.05).

Table 2 shows that the co-aggregation rates of the six LAB strains against four common pathogens. The strains with the highest co-aggregation ability against all four pathogens was Lactobacillus reuteri strain L219, including Escherichia coli (68.03%), Staphylococcus aureus (69.70%), Salmonella (68.10%), and Pseudomonas aeruginosa (69.60%).

To evaluate the adhesion ability of the six isolates, we performed the following experiments: auto-aggregation assay, adhesion ability assay of Caco-2 cell line, cell surface hydrophobicity assay, and biofilm formation ability assay.

The auto-aggregation ability is shown in Figure 3a. L221 (88.93%) has the highest self-agglutination rate, followed by LGG (78.31%) and L219 (74.40%).

Figures 3b, c show the adhesion ability of the six LAB strains to the Caco-2 cell line. Among them, Strain L221 showed the strongest adhesion ability (adhesion rate 3.00%, 30.00 CFU/cell), followed by control strain LGG (adhesion rate 2.88%, 28.77 CFU/cell).

Figure 3d displays the results of cell surface hydrophobicity. Significant differences were observed in the hydrophobicity of various LAB strains when exposed to different solutions. L218 exhibited the highest hydrophobicity against two organic solvents: Ethyl acetate (87.63%) and Xylol (49.03%). L222 exhibited the highest hydrophobicity toward the organic solvent Trichloromethane (81.76%).

Supplementary Table 2 displays the biofilm formation ability of the six isolates. Strains L218, L219, L220, and L223 exhibited moderate biofilm formation ability (++), while strain L221, L222 and LGG demonstrated strong biofilm formation ability (+++).

The hemolytic activity test results showed that these six isolates were non-hemolytic (Supplementary Figure 1).

Table 3 shows the susceptibility results of six isolates to 14 antibiotics. The resistance rates (includes resistance and intermediate) were 0% (0/6) to penicillin G, 0% (0/6) to amoxicillin, 0% (0/6) to amoxicillin, 16.67% (1/6) to erythromycin, 0% (0/6) to Cefuroxim, 0% (0/6) to cefotaxime, 83.33% (5/6) to Oxacillin, 0% (0/6) to Cefazolin, 100% (6/6) to Norfloxacin, 33.33% (2/6) to Rifampicin, 16.67% (1/6) to clindamycin, 0% (0/6) to chloramphenicol, 33.33% (2/6) to tetracycline, 100% (6/6) to vancomycin. Isolates L219, L221, and L222 had the highest sensitivity to 14 antibiotics (78.57%), followed by L218 (71.43%), L220 (64.29%), and L223 (64.29%).

Table 4 displays the tolerance of 6 LAB strains to varying concentrations of H₂O₂. The results showed that all six isolates could survive well in 0.5 mmol/L H₂O₂, with survival rates ranging from 86.01 to 98.96%. Except for isolates L220 and L221, the survival rates of the other strains were found to be statistically significantly higher than the control strain LGG (P < 0.05). When the concentration increased to 1 mmol/L H₂O₂, the survival rate of strains L220, L221, and LGG exceeded 28%. When the concentration increased to 1.5 and 2 mmol/L H₂O₂, the overall survival rate was lower than 10%.

The results of the free radical scavenging experiment are shown in Table 5. The cell-free supernatant of the six LAB isolates was always better than the bacterial suspension. The clearance rate of DPPH by bacterial suspension is 18.05% ± 0.15 to 24.40% ± 0.35, all were higher than the control strain LGG (17.65 ± 0.36). The highest clearance rate (24.40 ± 0.35) was observed in strain L219. Interestingly, the overall DPPH clearance rate of the supernatant for the isolated strains remained within the range of 77.91% ± 1.09 to 88.44% ± 0.99, all lower than the control strain LGG (93.43 ± 1.16). The clearance rate of OH⁻ in bacterial suspension is 11.27% ± 1.05 to 22.38% ± 3.73, all lower than the control strain LGG (39.67 ± 1.35). The removal rate of OH- in the supernatant liquid is 58.00–71.40%. Strain L223 showed the highest removal rate (71.40 ± 1.57). The isolated strains showed no ability to clear O2⁻ from bacterial suspensions or cell-free supernatants, which is consistent with the control strain LGG.