Healthy juvenile tilapia (Oreochromis niloticus) samples were purchased from the food market located in Chenggong District, Kunming, Yunnan, China, and sent to the laboratory immediately within 3–5 h. Later, six individuals (mean body weight: 12 ± 1.5 g, mean body length: 7 ± 1.2 cm) were anesthetized with ethyl 3-aminobenzoate methane sulfonate (concentration: 15 mg/L, MS-222, Sigma, MO, United States). Then, the tilapia individuals were washed twice with 75% ethanol and 5 times with sterile water, and the intestines of tilapia were aseptically cut on a super-clean worktable (Optec, Chongqing, China) and placed in a 5.0-mL sterile Eppendorf tube with 2.0-mL sterile water, and thoroughly ground with a grinding rod. Afterward, 100 μL dilutions of four concentrations (i.e., 10^–1, 10^–3, 10^–5, and 10^–7) obtained by gradient dilution with sterile water were dispersed on the surface of MRS Culture Medium (Solarbio, Beijing, China) (37°C, 24 h). Furthermore, single colonies were picked out by sterile inoculation loop. Finally, these single colonies were preliminarily identified and stored at 4°C according to their phenotypic characteristics.

S. flexneri strain BDS14 (referred to as S. flexneri_14) isolated from chicken was obtained in our previous studies (He et al., 2021), and it is being stored at Faculty of Life Science and Technology, Kunming University of Science and Technology. Species of LAB strains were confirmed by morphological and microscopic observation under a light microscopy (400 × magnification) (Zeiss Primo Star, Jena, Germany) according to the “manual for systematic identification of common bacteria” (Goodfellow et al., 2012). LAB strains possess several morphological characteristics such as rod-shaped appearance, gram-positive stain, no flagella, and no capsule.

The candidate LAB strains were cultured in MRS liquid medium (37°C, 24 h) and the culture solution was centrifuged (7,104 × g, 10 min) to discard the precipitate. Subsequently, a 0.22-μm filter membrane was employed to obtain a cell-free suspension. The obtained cell-free supernatant was adjusted to neutral pH (6.5) to exclude the interference of organic acids (including lactic acid). Screening of LAB strains against S. flexneri_14 was performed by the Oxford cup (diameter 8.0 mm, sample volume 200 μL) double-layer plate method (Voulgari et al., 2010). In brief, 100 μL of a suspension of S. flexneri_14 (10⁷ CFU/mL) cells in exponential phase was spread on the surface of LB semisolid medium (Solarbio). The Oxford cup was placed onto the surface of the LB semisolid medium and 200 μL of the cell-free supernatant was added, cultivated at 37°C for 24 h. Finally, a vernier caliper was used to determine the clearing zone around the well which was a result of the antibacterial activity of the LAB supernatant. Sterile water was used as control and three replicates were performed.

Based on the above obtained inhibitory zone, the LAB strain showing the best antibacterial effects against S. flexneri_14 was selected for the downstream experiments. The target strain was added to MRS liquid medium and then incubated at 37°C for 24 h. The taxonomic status of the target strain was verified using molecular markers. Briefly, DNA extraction of the target LAB strain was performed according to the directions of DNA Purification Kit (Tiangen, Beijing, China). PCR reaction system: 2 × Mix 12.5 μL, DNA template 1 μL, each universal primer (8F and 1492R for the bacterial 16S rRNA gene) 0.5 μL. PCR reaction conditions: 94°C for 3 min, 94°C for 32 s, 52°C for 32 s, 72°C for 50 s, final extension at 72°C for 5 min after 30 cycles. The PCR amplicons were detected by 1.5% agarose electrophoresis, and Sanger sequencing was performed by Geneseed Biotech Co., Ltd. (Guangzhou, China). The obtained 16S DNA sequences were aligned into the NCBI database using the online BLAST. MEGA6 software was used to reconstruct a phylogenetic tree (Tamura et al., 2013).

The target strain (L. plantarum) against S. flexneri_14 was added to MRS liquid medium (37°C, 24 h), and then centrifuged (7,104 × g, 10 min) to remove cell precipitate. The obtained cell-free supernatant was purified by an ÄKTA purifier automatic chromatograph in tandem coupled with a Superdex™ 30 Increase 10/300 GL Exclusion Chromatography columns (GE Healthcare, Marlborough, United States), as previously described (Li Z.R. et al., 2020; Li et al., 2021). The corresponding substances under each eluted peak were separately collected into a 25-mL centrifuge tube, and their antibacterial activity was determined by the Oxford cup double-plate method. The collection that showed the highest antibacterial activity was then loaded and purified again to obtain a purified antibacterial active substance (named as LFX01). Finally, the LFX01 fraction was preserved by freeze-drying in a FD-2A freeze dryer (Biocoll, Beijing, China), for downstream experiments.

The molecular weight and amino acid sequence of the purified LFX01 were determined by a Nano LC-MS/MS system, as previously described (Li Z.R. et al., 2020; Li et al., 2021). Briefly, LFX01 were eluted with 0.1% formic acid (Macklin, Shanghai, China), 100% water and 0.1% formic acid, and 100% acetonitrile. The LC linear gradient elution was implemented: from 6 to 9% B for 6 min; from 9 to 50% B for 30 min; and from 50 to 95% B for 3 min and 95–95% B for 5 min, with a flow rate at 0.3 μL/min. The primary mass spectrum was obtained by a single full scan (MS), and the secondary mass spectrum data is obtained by step-normalized collision energy. The obtained mass spectra signals by MS scan were further assessed for protein identification in Peaks Studio X (Bioinformatics, Waterloo, Canada). Antimicrobial peptide databases (i.e., NCBI, UniProt, and Swiss-Prot) were searched amino acid sequence of bacteriocin using the Blast online server. Finally, the antibacterial effects of LFX01 were verified by the Oxford cup double-plate method.

Bacteriocin concentration was estimated by the Pierce^® BCA Protein quantification Kit (Thermo Fisher Scientific, Waltham, United States), according to user instruction. Briefly, bovine serum albumin (BSA) standards (2.0 mg/mL) were diluted in ampoules to prepare the required concentration of the standard solution. Subsequently, the standard solution (25 μL) and the equivalent amount of LFX01 were pipetted into a 96-well plate. A total of 200 μL BCA working reagent was added to each well, the content was thoroughly mixed for 1 min, incubated room temperature for 25 min. Absorbance of the solution was measured at 565 nm at room temperature using a microplate reader (MR-96A, Mindray, Shenzhen, China). Free-BSA samples were set as control. The concentration of LFX01 in the samples was measured by plotting a standard curve according to y = 1.058x+0.1164, R² = 0.9962.

LFX01 (1.0 mg/mL) was added separately to catalase and different protease solutions [i.e., pepsin (20 mM in a in PBS, pH 2), trypsin, proteinase K, papain (20 mM in PBS, pH 7.5), and alkaline protease (20 mM in PBS, pH 9.0)] with a final concentration of 1.0 mg/mL (Solarbio). The mixture of LFX01 and enzyme was treated (37°C, 2 h), followed by deactivation (80°C, 5 min). Enzyme without LFX01 was set as control. The effect of each different enzyme on the antibacterial activity of LFX01 was confirmed by the Oxford cup double-plate method. Each experimental sample was independently tested in triplicate.

The following tests were performed to detect the acid-base and thermal stability of LFX01 (1.0 mg/mL), as previously described (Yi et al., 2018; Zhang et al., 2019): (1) treatment at 37, 60, 80, 100, and 121°C for 30 min; (2) pH of LFX01 was adjusted to 2.0, 4.0, 6.0, 8.0, 10.0, and 12.0 by 1.0 moL/L HCl or NaOH solution and then treated at 37°C for 2 h. Finally, pH of LFX01 solution was adjusted back to neutral pH (6.5). The antibacterial activity of LFX01 against S. flexneri_14 under different pH values and temperature was separately assessed by the Oxford cup double-plate method. LFX01 under room temperature (25°C) and neutral pH value was set as control. Each experimental sample was independently tested in triplicate.

Determination of the MIC values of LFX01 against S. flexneri_14 was conducted, as previously described (Zhang et al., 2019; Li et al., 2021). Briefly, S. flexneri_14 was precultured in LB liquid medium by incubation (37°C for 12 h). Then, LFX01 (1.0 mg/mL) was serially diluted into PBS. These dilutions (10 μL) were added separately into 90 μL of LB liquid medium with S. flexneri_14 (final concentration: 10⁷ CFU/mL), incubated at 37°C in a 96-well plate for 24 h. Growth of S. flexneri_14 was assessed in triplicate using a microplate reader at 595 nm. MIC was the lowest concentration of antibacterial substances that completely inhibited growth of S. flexneri_14, as previously described (Yi et al., 2018; Li et al., 2021). 2 × MIC was used as the effective concentration for bacteriostatic treatment, as previously reported (Yi et al., 2018; Lanhua et al., 2020).

The time-killing kinetics test was conducted to assess the dynamic changes in killing effects of LFX01 against the indicator strain across various treatment times. Namely, LFX01 was mixed with S. flexneri_14 (10⁷ CFU/mL) to reach a final concentration (2 × MIC) of LFX01 and then incubated for 0, 0.5, 1, 1.5, and 2 h. Subsequently, 100 μL of LFX01 at each incubation time points were spread on LB solid medium by 10 fold serial dilution method. After incubating at 37°C for 24 h, 30∼300 single colonies of S. flexneri_14 were selected for counting. LFX01-free samples were set as control groups. Each experimental sample was independently tested in triplicate.

The metabolic activity of S. flexneri_14 cells was determined by the XTT (i.e., 2,3-BIS (2-methoxy-4-nitro-5-sulfonyl)-2H-tetrazo-5-carboxyl) assay kit (Abcam, Cambridge, United Kingdom), according to use instruction strictly. Briefly, planktonic S. flexneri_14 samples and 10 μL of LFX01 (final concentrations: 2 × MIC) were added into a 96-well plate containing 80 μL of LB liquid medium, followed by the mixing and incubated (37°C, 1 h). Subsequently, 10 μL of XTT solution was added, mixed for 30 s and incubated (37°C, 2 h). Following a gentle homogenization of the microplate, absorbance of samples was measured using a microplate reader at 450 nm. Each experimental sample was independently tested in triplicate.

The survival rate of planktonic S. flexneri_14 cells after exposure to LFX01 was assessed by the Cell-Check™ viability/cytotoxicity kit (ABP Biosciences, Rockville, United States). Briefly, planktonic S. flexneri_14 cells were treated with LFX01 (2 × MIC) at 37°C for 10 min. Later, 3 μL of mixture composed of fluorescent dyes NucView Green and propidium iodide (PI) was added to the bacterial suspension, followed by incubated at 37°C for 10 min. LFX01-free planktonic S. flexneri_14 cells were set as control. The experimental samples were detected on the slides using fluorescence microscopy (Boschida, Shenzhen, China) (200 magnification). Each experimental sample was independently tested in triplicate.

Three milliliters of S. flexneri_14 (10⁷ CFU/mL) were centrifuged (7,104 × g, 5 min) to collect the precipitated bacterial cells. The bacterial cells were washed thrice with PBS and incubated (37°C, 2 h) with 1 mL of LFX01 (2 × MIC). Subsequently, the LFX01 treated bacterial cells were then fixed with 2.5% glutaraldehyde at 4°C for 8 h and dehydrated in the concentration gradient of ethanol (20, 40, 60, 80, and 95%) for 30 min each. Afterward, all samples were transferred to polished silicon wafers (10 × 10 mm) and dried at room temperature (25°C). The samples covered by gold powder were imaged with an SEM (S-3000N, Hitachi, Tokyo, Japan). LFX01-free S. flexneri_14 was set as control, and each treatment was performed in three replicates.

LFX01 was mixed with LB liquid medium and transferred to a 96-well plate to assess the effects of LFX01 exposure on S. flexneri_14 biofilm formation at final concentrations of: LFX01-free controls, 1/2 × MIC, 1 × MIC, 2 × MIC. Later, 100 μL of cell suspension of S. flexneri_14 was added to these LFX01 dilutions at different concentrations. After incubation at 37°C for 24 h (the optimal temperature and time for biofilm formation of S. flexneri_14), the liquid medium was discarded to remove planktonic S. flexneri_14 cells and expose the biofilm. The plates washing wells using PBS buffer, a MycoLight™ Live Bacteria Fluorescence Imaging Kit (AAT Bioquest, CA, United States) was employed to stain the exposed biofilm following the product manuals. Excessive dye was removed using PBS buffer. The formed biofilms were detected with fluorescence microscopy (100 magnification). PBS buffer was added to each well followed by homogenization. Absorbance (OD₅₉₅) of the mixture was determined in a microplate reader. Samples not treated with LFX01 were used as control.

Data analysis was conducted, as previously described (Yi et al., 2018; Li et al., 2021). In brief, results were exhibited as mean ± standard deviation (SD) of independent replicates. IBM^® SPSS^® Statistics version 22 was used to calculate the significance between two groups employing a non-paired two-tailed t-test. One-way ANOVA test was conducted to evaluate significance levels among groups. P-values < 0.05 indicated significant statistical difference.

Among pure cultures of LAB strains isolated from the tilapia intestine, the inhibitory zone diameter of the LF-8 strain against S. flexneri_14 was the largest (up to 26.69 ± 0.32 mm). Thus, this strain was selected as the object for further study. Colony morphology and micro-structure of LF-8 revealed white globular, smooth surface, short chain globular (Figures 1A,B). Furthermore, 16S rRNA amplicon of LF-8 with a nucleotide sequence of 1,465 bp (Genbank accession No. MW979597.1) was obtained. BLAST alignment results elucidated a 100.0% similarity of the sequence with L. plantarum species (Genbank No. NR117813.1), and the phylogenetic tree presented a cluster of LF-8 strain with L. plantarum (Figure 1C). Based on the above-obtained phenotypic and molecular evidence, LF-8 was identified as L. plantarum.

Among the four peaks identified in the chromatogram of the preliminary purified extract, peak A4 showed the greatest antibacterial activity (Figure 2A). The corresponding product in B1 peak was collected after purifying A4 and the antibacterial activity of B1 substances (LFX01) against S. flexneri_14 was confirmed by the formation of an inhibitory zone (25.12 ± 0.16 mm) (Figure 2B).

Mass spectrometry analysis determined a molecular weight of 1,049.56 Da (Figure 3) and amino acid composition of I-T-G-G-P-A-V-V-H-Q-A of LFX01 (Supplementary Figure 1). BLAST alignment did not search any similar protein sequence with LFX01 in protein storage databases.

LFX01 treated with catalase retained 97.02% activity compared to the controls (no significance, p > 0.05). This result excluded hydrogen peroxide-based antibacterial activity of LFX01. After treated with different proteases, the antibacterial activity of LFX01 significantly reduced (p < 0.05), particularly proteinase K, trypsin, and pepsin (reduced by more than 80%), indicating the sensitivity of LFX01 to various proteases (Figure 4). Additionally, the antibacterial activity of LFX01 gradually decreased with increase in pH values (Figure 5A). The antibacterial activity of LFX01 reached the lowest at pH 12.0 (reduced by 73.74%). The fastest decrease of antibacterial activity of LFX01 occurred at pH 8∼12 (reduced from 79.16 to 26.26%), and this change in antibacterial activity was significant compared with the controls (p < 0.05). Furthermore, the antibacterial activity of LFX01 gradually decreased with the increase of temperature (37∼121°C) (Figure 5B). Compared with the control, LFX01 retained 56.06% antibacterial activity at 121°C. The fastest decline in antibacterial activity of LFX01 occurred at 80∼121°C (reduced from 81.83 to 56.06%).

MIC of LFX01 against S. flexneri_14 was 12.65 μg/mL. The colony number of S. flexneri_14 gradually decreased with the treatment time and significantly decreased after 0.5 h (p < 0.05) when treated with 2 × MIC for 0∼2 h (Figure 6). The colony number of S. flexneri_14 reaches to the lowest value (lg4.65 CFU/mL) at 2 h, confirming the effectiveness of LFX01 to terminate planktonic S. flexneri_14 survival.

The XTT assays showed that absorbance values of planktonic S. flexneri_14 cells was decreased to 45.32% upon exposure to LFX01 (2 × MIC) for 2 h (p < 0.01) (Figure 7A). Compared with the untreated controls (live cells are stained green) (Figure 7B), S. flexneri_14 cells died after treatment with LFX01 (2 × MIC) (red-stained) (Figure 7C). These results further confirmed the effective killing of LFX01 against planktonic S. flexneri_14 cells.

Biofilm density of S. flexneri_14 appeared to be significantly reduced (p < 0.05) after treating with LFX01 at concentrations of 1/2 × MIC, 1 × MIC, and 2 × MIC. Furthermore, quantitative analysis determined that the biofilm exposed to the control (0 × MIC; Figure 8A), 1/2 × MIC (Figure 8B), 1 × MIC (Figure 8C), and 2 × MIC (Figure 8D) of LFX01 yielded OD595 values of 1.35 ± 0.15, 0.96 ± 0.09, 0.56 ± 0.05, and 0.29 ± 0.03, respectively. P-values of all comparison pairs between controls and each experimental group were less than 0.05. The SEM analysis further presented LFX01-free S. flexneri_14 cells (control) with uniform rod-shaped, smooth surface, neat edge, clear outline, and complete cell structure (Figures 9A,B). However, compared with the control, S. flexneri_14 cells incubated with LFX01 presented uneven shape, wrinkled surface, disruptive cells, content leakage, and cell surface perforations (Figures 9C,D).