Clostridium perfringens strain CP4 was obtained from D.A. Barnum, Department of Pathobiology, University of Guelph (18). Trypticase Sulfite Neomycin Agar (TSN) and Fluid Thioglycollate medium (FT) were used for culturing C. perfringens. Tryptic Soy Broth medium (TSB) was used for biofilm assay of C. perfringens. De Man, Rogosa and Sharpe medium (MRS) was used for screening and cultivating lactic acid bacteria. Medium and reagents were purchased from Qingdao Hi-Tech Industrial Park Hope Bio-Technology Co., Ltd. (Qingdao, China), and Solarbio Science & Technology Co., Ltd. (Beijing, China).

Fecal samples were collected from healthy wild birds in a natural scenic area in Jilin Province, China. The samples were serially diluted to 10⁻⁷ and spread onto MRS agar plates, which were then incubated at 37°C under aerobic conditions for 24 h. Bacteria colonies were selected and streaked onto fresh MRS agar plates for further activation. The activated colonies were then inoculated into 5 mL MRS broth and cultivated at 37°C for 24 h.

The agar well diffusion method, with some modifications based on the previous description (19), was employed to evaluate the antibacterial activity. Briefly, C. perfringens was used as the indicator bacterium. Both the isolated strains and the indicator bacterium were separately inoculated into MRS broth and FT medium, respectively, and cultivated at 37°C under anaerobic conditions for 16 h to reach the stationary phase. The indicator bacterium was then adjusted to an OD₆₀₀ of 0.8 (around 1 × 10⁸ CFU/mL). The isolated strains were centrifuged at 4000 rpm for 5 min to harvest the supernatant, which was subsequently filtered through a 0.22 μm membrane to obtain sterile cell-free supernatant (CFS). The adjusted indicator bacterium was evenly spread onto TSN agar plates, and place an Oxford cups on a TSN agar plates, the Oxford cups were filled with the CFS. After incubation at 37°C under anaerobic conditions, the diameters of the inhibition zones were recorded. The strains exhibiting the significant antibacterial activity was selected for subsequent experiments.

The top 4 strains with the high antibacterial activity were selected and genomic DNA were extracted using an Ezup column-based bacterial genomic DNA extraction kit (Sangon Biotech Co., Ltd., Shanghai, China), with the procedure performed according to the manufacturer’s instructions. PCR of the 16S rDNA with primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-TACGGCTACCTTGTTACGACTT-3′) (20) used 0.5 μL of DNA template (20–50 ng/μL), 2.5 μL of 10 × PCR buffer (with 50 mol/L Mg²⁺), 1.0 μL of dNTP (2.5 mmol/L each), 0.2 μL of DreamTaq™ DNA Polymerase (5 U/μL) (Thermo Scientific, Waltham, MA, USA), and 0.5 μL of primers (10 μmol/L each), and ddH2O was added to achieve a final volume of 25 μL. PCR was performed on a 2720 thermal cycler (Applied Biosystems, Foster City, CA, USA) based on the following procedure: 5 min of denaturation at 95°C, 30 cycles of denaturation at 94°C for 30 s, annealing at 57°C for 30 s, extension at 72°C for 90 s, and a final extension at 72°C for 10 min. The amplified fragments were sequenced using a 3730XL sequencer (Applied Biosystems) by Sangon. The nucleotide sequences were aligned with the available sequences in the GenBank database through NCBI blasting,¹ and a phylogenetic tree was constructed using MEGA software (version 11) (21) in order to determine the species of the bacterial strains.

The 4 strains of E. faecalis with the strongest antagonistic activity against C. perfringens, designated as YL-EF7, YL-EF25, YL-EF28, and YL-EF32, were streaked onto plates containing 5% sheep blood agar medium and incubated at 37°C to observe their hemolytic ability. Additionally, these strains were inoculated into 100 mL of MRS broth and cultivated at 37°C with shaking at 120 rpm for 36 h. Samples of 5 mL were taken at the 2nd, 4th, 6th, and every 6th hour thereafter to measure the OD₆₀₀ value and pH of the bacterial suspension. The data were recorded and used to plot growth curves and acid production curves.

The 4 strains were streaked onto MRS agar plates and incubated at 37°C. Colonies were then picked and inoculated into MRS broth, which was cultivated at 37°C overnight. The bacterial suspension concentration was adjusted to 1 × 10⁸ CFU/mL. To determine the initial bacterial count, 1 mL of the adjusted suspension was serially diluted in PBS and spread onto MRS agar plates, which were then incubated at 37°C for 24 h to calculate the initial bacterial concentration [log₁₀ CFU/mL (I)]. The remaining suspension was centrifuged at 5000 rpm for 10 min, and the bacterial pellet was resuspended in corresponding bile salt solutions (0.2, 0.25%) and artificial gastric acid solution (0.32% pepsin, pH = 2) prepared in PBS. The suspensions were then incubated at 37°C for 3 h. After incubation, the suspensions were serially diluted in PBS and spread onto MRS agar plates, which were incubated at 37°C for 24 h to calculate the bacterial count in the treated groups [log₁₀ CFU/mL (L)] and the survival rate. The survival rate was calculated using the following formula:

The phenol tolerance test for E. faecalis was conducted based on the previous description method (22), with some modifications. The overnight cultures of the E. faecalis strains were inoculated into MRS broth containing different concentrations of phenol (0.1, 0.2, 0.3, and 0.4%) and cultivated at 37°C for 24 h. The OD₆₀₀ values of the bacterial suspensions in each group were measured.

The biofilm inhibition test was performed according to the previous described method (23), with some modifications. Overnight cultures of C. perfringens (CP) and E. faecalis (YL-EF7/25/28/32) were washed once with PBS and resuspended in TSB medium to an OD₆₀₀ of 0.1. A 100 μL aliquot of the CP suspension was inoculated into each well of a 96-well polystyrene tissue culture plate, followed by the addition of 100 μL or 200 μL of the EF suspension or filter-sterilized CFS. Each well was supplemented with TSB medium to a total volume of 300 μL per well. Control groups included CP alone (300 μL CP) and a negative control (300 μL TSB). The experimental groups were designed as follows: CP, CP + EF, CP + TSB + EF, CP + CFS, and CP + CFS + TSB. The cultures were incubated in sealed culture plate to prevent evaporation and incubate at 25°C for 5 days. The OD₆₀₀ values of the supernatants from each group were measured, and the viable counts of CP in the planktonic phase were determined using selective medium (TSN). The wells containing biofilms were gently washed twice with PBS, and 300 μL of a 0.1% crystal violet solution was added to each well and stained at room temperature for 30 min. The crystal violet solution was carefully removed, and the wells were washed twice more with PBS. The biofilm-bound crystal violet was extracted with 300 μL of methanol for 30 min, and the methanol solution was transferred to a new 96-well plate. The OD570 values were measured using a multifunctional microplate reader (TECAN, Switzerland). The ratio of A570 to OD₆₀₀ was used as a relative measure of biofilm production.

The 71 carbon sources utilization and sensitivity to 23 chemicals of E. faecalis YL-EF25 and YL-EF32 were assessed using Biolog GEN III microplates (Biolog, USA) to provide a phenotypic profile of their carbon source metabolism. Briefly, the strains were streaked onto MRS agar medium for activation and then inoculated into special inoculating fluid (IF) to prepare bacterial suspensions with the recommended optical density. A 100 μL aliquot of each suspension was inoculated into a GEN III microplate and incubated in an OmniLog instrument for 30 h. During incubation, bacterial respiration reduces tetrazolium redox dye, turning it purple. The OmniLog automated microbial identification system automatically collects phenotypic data.

The animal experiments were conducted on 90 one-day-old broilers purchased from a poultry farm in Changchun, China, and all animal experiments were approved by the Ethics Committee for Animal Protection and Use of Jilin Agricultural University. The rearing temperature was controlled at 33°C for the first 7 days, and then decreasing cumulatively 3°C in every week until the temperature of 24°C was reached. The humidity was maintained at around 60%, and the animal room was illuminated for 24 h during the first 3 days, followed by 20 h of light per day. The animals were allowed free access to water and feed. A total of 90 male broiler chickens were randomly divided into 5 treatments groups, with 6 cages in each group and 3 chickens per cage.

The treatments were as follows: (1) negative control group birds fed a basal diet and not challenged (NC); (2) CP-infected group birds fed a basal diet and orally challenged with 2 × 10⁴ coccidia (Eimeria tenella) oocysts and 4 × 10⁸ cfu CP (PC); (3) antibiotic group fed a basal diet containing 10 mg of the Enramycin/kg of feed and then orally challenged with 2 × 10⁴ coccidia oocysts and 4 × 10⁸ cfu CP (PT); (4) E. faecalis YL-EF25 group fed a basal diet containing E. faecalis YL-EF25 (10⁹ cfu/kg of feed) and then orally challenged with 2 × 10⁴ coccidia oocysts and 4 × 10⁸ cfu CP (EF25); and (5) E. faecalis YL-EF32 group fed a basal diet containing E. faecalis YL-EF32 (10⁹ cfu/kg of feed) and then orally challenged with 2 × 10⁴ coccidia oocysts and 4 × 10⁸ cfu CP (EF32). The basal diets were formulated according to the Chinese Feeding Standard for Chickens. The nutrient ingredient compositions of the feed are shown in Supplementary Table S1. From the 15th to the 25th day, adjust the crude protein content in the diet to 28% using fish meal to increase the effect of CP infection. The specific details of the animal experimental design are shown in Figure 1. The body weights of the chickens were measured on days 0, 14, and the final day (day 25) of the experiment. Feed intake (FI), body weight gain (BWG), and feed conversion ratio (FCR) were calculated and recorded for further analysis. The FCR value was calculated using the following formula:

Peripheral blood lymphocytes were isolated from chickens using a lymphocyte separation kit (Solarbio, China). On day 14, lymphocytes (1 × 10⁶ cells/well in 96-well plates) were stimulated with 20 μg/mL Concanavalin A for 48 h. Proliferative activity was then quantified using a CCK-8 assay (Melunbio, China), with the stimulation index (SI) calculated according to the manufacturer’s protocol.

On day 25, the lesion scores of the small intestine were evaluated. Chickens from each group were euthanized, and their small intestines were dissected for objective lesion scoring. The scoring details were based on the method described by Shojadoost et al., where intestinal lesions were rated on a scale of 0 to 5 (24).

Jejunum samples were collected from broiler chickens on day 25. Total RNA was isolated using TRIzol reagent (Transgen, China) and quantified. Subsequently, cDNA synthesis was performed using the BeyoRT™ II cDNA First Strand Synthesis Kit (RNase H-) (Beyotime, China). The expression levels of the respective genes were detected using BlasTaq™ 2 × qPCR MasterMix (Beyotime, China) and quantified on an Applied Biosystems 7500 Instrument. GAPDH was utilized as an endogenous control in the comparative CT method. The primers used are listed in the Supplementary Table S2.

Total genomic DNA from ileum samples was extracted using the E.Z.N.A™ Mag-Bind Soil DNA Kit (Omega, USA) according to the manufacturer’s instructions. On Day 26, broiler ileal digesta were sampled. DNA was extracted from ileal digesta using the E.Z.N.A™ Mag-Bind Soil DNA Kit (Omega, USA) according to the guidelines. According to the specifications outlined by Illumina, all DNA samples were pretreated for MiSeq compositional sequencing. The V3-V4 region of the 16S rRNA gene was amplified, and Illumina index primers were attached in two separate PCRs. FLASH software (v1.2.7) was used to generate raw tags (25). Effective tags were obtained by the UCHIME algorithm (26) and QIIME (v1.7.0) analysis (27) UPARSE software (v7.0.1001) was used to analyze sequences, and the sequences were clustered at 97% similarity as operational taxa (OTUs). The GreenGene database was used to compare sequences and classify the different classification levels of these sequences. Microbial diversity was detected through QIIME software and Python scripts.

All of these data were analyzed with SPSS Version 20.0 for Windows (SPSS Inc., Chicago, IL, USA). One-way ANOVA was performed if there was a significant difference, and Duncan’s method was used for multiple comparisons. The values were expressed as means ± SD and differences were considered significant at p < 0.05.

A total of 34 isolates were screened from wild bird feces using MRS medium, and their cell-free supernatants (CFS) were tested for inhibitory activity against bird-derived CP strain using the agar well diffusion method. The results showed that isolates #7, #25, #28, and #32 exhibited strong inhibitory activity against CP (Table 1). Therefore, these 4 isolates were selected for further study. Based on 16S rRNA gene sequencing, the 16S sequences of the 4 isolates were obtained and compared with known sequences. The results revealed a 100% sequence similarity to E. faecalis. A phylogenetic tree was constructed based on the neighbor-joining method using MEGA 11 software, shown in Supplementary Figure S1. Thus, these 4 isolates were identified as E. faecalis and named YL-EF7, YL-EF25, YL-EF28, and YL-EF32.

The probiotic characteristics of the E. faecalis (E.F) strains were analyzed in vitro. The results showed that none of the strains produced hemolytic circles, indicating their safety in terms of hemolysis (Figure 2a). The growth curves and acid production capabilities of the strains in MRS medium were also evaluated (Figure 2b). When exposed to simulated gastric juice for 3 h, the strains maintained a survival rate of 40% (Figure 2c) and showed survival rates of around 90 and 65% in 0.2 and 0.25% bile salt solutions, respectively (Figure 2d). These strains can metabolize phenol, which has strong inhibitory activity, in the gastrointestinal tract. Experiments demonstrated that the E. faecalis strains did not show significant inhibition when cultured in MRS medium containing 0.1 and 0.2% phenol (Figure 2e). These findings indicate that the E. faecalis strains have strong resistance to gastric acid, high concentrations of bile salts, and phenol. They are safe, environmentally tolerant, and candidate strains for anti-CP applications.

Clostridium perfringens can utilize biofilm formation as a strategy for persistent survival in various environments or hosts (23). To investigate whether the E. faecalis strains can inhibit CP biofilm formation, a biofilm assay was established. All CP strains were inoculated into 96-well plates, and except for the CP control group, other groups were supplemented with E. faecalis suspensions or CFS at ratios of 1:1 or 1:2. After 5 days of incubation, biofilm formation was observed by calculating the ratio of A570 to OD₆₀₀. Compared with the CP control group, both the E. faecalis suspensions and CFS significantly reduced CP biofilm formation, especially in the EF + CFS group (Figure 2f). Additionally, the number of CP in the supernatants of each experimental group with E. faecalis was determined. The results showed that the presence of E. faecalis reduced the number of CP, particularly in the EF32 group (Figure 2g). These data indicate that the presence of the E. faecalis strains can effectively inhibit CP biofilm formation during the biofilm formation process, especially YL-EF25 and YL-EF32.

In this study, microbial metabolic analysis was used to investigate the metabolic capabilities of E. faecalis YL-EF25 and YL-EF32 toward 71 carbon sources and their sensitivity to 23 chemicals. The results showed that the two strains had similar metabolic capabilities for the 71 carbon sources and sensitivity to the 23 chemicals (Figure 3). Notably, both strains can utilize dietary fibers such as dextran, pectin, and cellobiose (Full line marker) and substrates for propionic acid synthesis pathways (Dashed line marker) as energy sources. Dietary fibers can be metabolized by probiotics to produce short-chain fatty acids (SCFA), which contribute to maintaining intestinal health.

Growth performance results are presented in Table 2. Compared to the NC group, CP challenge (PC group) significantly reduced body weight gain (BWG) and increased the feed conversion ratio (FCR) in broilers during the challenge phase (14–25 days) (p < 0.05), but had no significant effect on feed intake (FI) (p > 0.05). During the challenge phase, the BWG of the PC group was significantly lower than that of the NC group, PT group, EF25 group, and EF32 group (p < 0.05), indicating that supplementation with YL-EF25, YL-EF32, or enramycin (PT group) partially alleviated the CP-induced reduction in weight gain. The FCR of the PC group was significantly higher than that of the NC group, PT group, EF25 group, and EF32 group (p < 0.05), while no significant differences in FCR were observed among the PT, EF25, and EF32 groups (p > 0.05). The proliferative capacity of lymphocytes is a crucial indicator for assessing immune performance. We examined the proliferative ability of peripheral blood lymphocytes in broilers raised for 14 days. Compared to the NC group, the lymphocyte proliferative capacity was significantly enhanced in the EF25 group (p < 0.05), while there was no significant difference in the EF32 group (Figure 4a).

Intestinal lesion scores were assessed on day 25 in the jejunum (Table 3). The intestinal macroscopic lesions are shown in Supplementary Figure S2. The CP infection group (PC) significantly increased intestinal lesion scores (p < 0.01). Dietary supplementation with E. faecalis YL-EF25 (EF25) and YL-EF32 (EF32) significantly reduced intestinal lesion scores induced by CP infection (p < 0.05), particularly in the EF32 group (p < 0.01). As shown in Figure 4b, the PC group led to a decrease in tight junction protein (TJPs)-related genes such as OCLN and ZO-1. However, the EF25 and EF32 groups increased the expression of OLCN and ZO-1 genes, with significant elevations in the EF32 group (p < 0.05). On the other hand, the PC and PT groups increased the expression of genes related to increased intestinal barrier permeability, such as MLCK. Dietary supplementation with E. faecalis YL-EF25 and YL-EF32 have a trend to alleviated the increase in MLCK, although there were no significant differences between them. This trend may indicate a modulatory role of the probiotics, but further validation with larger sample sizes or complementary functional assays is warranted. These data suggest that dietary supplementation with E. faecalis YL-EF25 and YL-EF32 improves intestinal barrier function. PC group and the PT group significantly increased the expression of inflammatory cytokine genes, such as TNF-α, IFN-γ, and IL-10. The EF25 and EF32 groups alleviated the upregulation of inflammatory cytokines induced by PC infection (Figure 4c). These data indicate that dietary supplementation with E. faecalis YL-EF25 and YL-EF32 exerts anti-inflammatory effects during CP infection.

Through 16S rRNA sequencing analysis, we compared the composition and differences in ileal microbiota among different treatment groups on days 14 and 25. The results showed that on day 14, the EF25 and EF32 groups demonstrated distinct microbiota regulatory effects (Figure 5a). The EF25 group increased the relative abundance of Lactobacillus, while the EF32 group showed an increased relative abundance of segmented filamentous bacteria (SPF) in the ileal microbiota of chickens. On day 25, compared to the normal group (NC), the necrotic enteritis group (PC) and the antibiotic group (PT) exhibited a decrease in Lactobacillus abundance in the ileal microbiota and a certain degree of microbiota disruption. However, the EF25 and EF32 groups alleviated the microbiota disruption caused by necrotic enteritis in chickens. Compared to the PC group, the EF25 and EF32 groups had higher Lactobacillus abundance (Figure 5b), and their microbiota abundance and composition were more similar to the NC group (Figure 5c). Further comparison of the Shannon index and principal component analysis (PCA) among groups further demonstrated that the microbiota abundance and composition structure of the EF25 and EF32 groups were closer to those of the NC group (Figures 5d,e).