Prior to initiating the formal trial, 41 one-day-old neonatal lambs were subjected to a 14-day fecal scoring assessment to evaluate the prevalence of diarrhea on the experimental farm. Fecal scores were recorded twice daily by two independent observers using a five-point grading system based on criteria previously established (3). The average fecal score was calculated to quantify diarrhea severity. Diarrhea occurrence was calculated daily by identifying lambs with fecal score exceeding 3 (3). The diarrhea occurrence (%) was computed using the following formula: (number of lambs exhibiting diarrhea × number of diarrheic days) / (total number of lambs × total trial days) × 100% (3). The result indicated a diarrhea occurrence of 35.93% and an average fecal score of 2.41, providing essential baseline data for the subsequent experimental phases.

Thirty-two neonatal Hu lambs (4.29 ± 0.22 kg, 1-day-old) of comparable body weight were randomly assigned to four treatments, with eight lambs per treatment. All lambs were suckled by their dams throughout the trial. And the dams in this trial were ensured to be healthy. The treatment groups received yeast peptides supplementation at 0 (CON), 500 (YP500), 1,000 (YP1000), and 2,000 (YP2000) mg/d, respectively. The designated dosage of yeast peptides was uniformly mixed with 10 mL of milk, maintained at 40°C, and administered orally via syringe once daily at 8:00 h. Lambs in the CON group received an equivalent volume of milk without supplementation. The 14-day experimental duration was justified by the brief postpartum period (first two weeks) during which lambs are particularly susceptible to diarrhea and no supplementary feeding is provided, eliminating the need for an adaptation period (3). The pens underwent weekly sanitation to maintain optimal hygienic conditions for both ewes and lambs. Lamb health management and ewe feeding protocols followed standard sheep farm practices. Two lambs from the CON group and one lamb from the YP1000 group were excluded from final analysis due to mortality attributable to insufficient milk intake or diarrhea.

Yeast peptides represent a kind of antibacterial peptide characterized by their unique lasso-like topological structure. This 19-amino-acid-residue peptide, with the specific sequence GGVGKIIEYFIGGGVGRYG, possesses a molecular weight of 1.9 kilodaltons (kD). Notably, yeast peptides maintain a highly stable structural configuration, featuring a distinctive circular form that is intricately folded and contains a lasso running through its core.

The trial lambs were weighed on the initial and final days of the trial, after which the average daily gain (ADG) was calculated. Diarrhea incidence was computed as previously described (3).

On the fifteenth day of the trial, six lambs from each of the four treatments, each selected to approximate the average body weight of its respective group, were designated for slaughter. On the morning of the final day of the trial period, approximately 5 mL of venous blood was collected via jugular venipuncture and subsequently centrifuged at 4°C (3,000 × g for 10 min) to obtain serum for analysis. After slaughter, a 10 cm segment of colonic tissue was excised from the midsection of the colon and cut longitudinally. The colonic luminal contents were rinsed out with phosphate-buffered saline (PBS) and the mucosal layer of the inner colonic wall was scrapped off and transferred into freezing tubes for preservation.

Serum biochemical indicators, including diamine oxidase (DAO) and D-lactase (DLA), as well as biochemical indices of the colonic mucosa, including interleukin 1β (IL-1β), interleukin 4 (IL-4), interleukin 6 (IL-6), interleukin 10 (IL-10), total antioxidant capacity (T-AOC), glutathione peroxidase (GSH-Px), catalase (CAT), superoxide dismutase (SOD) and malondialdehyde (MDA), were assayed using enzyme-linked immunosorbent assay (ELISA) and biochemical kits provided by Nanjing Jiancheng Bioengineering Institute Co., Ltd., Nanjing, China.

After grinding the colon mucosa into powder using a tissue grinder, the tissues were fully lysed by the addition of Trizol. Total RNA from the colon mucosa was then obtained through centrifugation and filtration, and its concentration and purity were then determined with a UV–visible spectrophotometer. After confirming the integrity, complementary DNA (cDNA) was obtained by reverse transcription of the RNA using a cDNA synthesis kit. The resulting cDNA served as the template for quantitative PCR (qPCR) with the primer sequences specified in Table 1. The expression levels of mucoprotein 2 (MUC2), Claudin-1, Claudin-4, Occuludin and Zonula occludens-1 (ZO-1) were quantified using the 2^−ΔΔct method, with β-actin as the internal reference gene. Each biological replicate was validated in triplicate by qPCR to ensure reliability.

After slaughter, the proximal colonic contents were collected and transferred into freezing tubes, which were subsequently stored at −80°C. DNA extraction was performed on the colonic contents using DNA stool kits, resulting in the isolation of genomic DNA. The concentration and purity of the genomic DNA were determined using a Thermo NanoDrop 2000 Ultraviolet Microspectrophotometer (Thermo Fisher Scientific, Waltham, MA), and 1% agarose gel were employed for integrity of quality assessment. Then, targeted amplification of the V3-V4 region of the bacterial 16S rDNA gene was conducted using primers (338F: ACTCCTACGGGAGGCAGCAG; 806 R: GGACTACHVGGGTWTCTAAT). The PCR amplification procedure consisted of an initial denaturation step at 94°C for 5 min, followed by 28 amplification cycles, which included denaturation at 94°C for 45 s, annealing at 55°C for 30 s, and extension at 72°C for 45 s. The process concluded with a final extension step at 72°C for 10 min. The resulting PCR products were purified using AxyPrep DNA gel recovery kits. The constructed library was quantitatively assessed using Qubit and sequencing was performed on the Illumina Miseq PE250 platform, yielding sequenced fragments of 425 bp.

The Fastq sequencing data obtained from Miseq platform were subjected to quality control processing. Subsequently, Qiime2 software was utilized to conduct statistical bioinformatics analysis of Amplicon Sequence Variants (ASVs) at a 100% similarity level. ASV-based analyses determined the diversity and relative abundance of the microflora through summed normalization. The Venn diagram and Mantel test analysis were made with R software (version 4.3.2). The LEfSe analysis was employed to identify the microbial biomarkers across the four treatments. Visualization of co-occurrence network analysis using Cytoscape software (version 3.7.1, Bethesda, MD, USA). Additionally, the relationship between microbial biomarkers and phenotypic parameters was analyzed using a Spearman correlation heatmap in R software (version 4.3.2). PICRUSt2 was utilized for prediction of function compared in comparison with the KEGG database.

Data regarding growth performance, fecal scores, serum indices, and colon biochemical indices were statistically analyzed using the one-way ANOVA procedure by R software (v4.3.2). The incidence of diarrhea was analyzed using the Chi-square test. And the dose-dependent effect of yeast peptides in diet was analyzed by both linear and quadratic regression models. Duncan’s method was employed for multiple comparisons among the four treatments. Microbial abundances were analyzed using the Kruskal-Wallis test to assess differences between the four treatments. Trend plots and column bar charts were visualized by using R software (v4.3.2). A p-value of less than 0.05 was considered statistically significant.

As shown in Table 2, there was no significant difference in initial BW among the treatments, satisfying the prerequisites of the trial design. The administration of yeast peptides at varying doses did not significantly affect the final BW or average daily gain (ADG) of the lambs. Nevertheless, the supplementation of yeast peptides to diets significantly reduced (p < 0.05) the diarrhea occurrence and the fecal scores of the lambs (Table 2). Specifically, the occurrence of diarrhea was significantly lower (p < 0.05) in the YP1000 and YP2000 groups compared to the YP500 group (Table 2).

Supplemental feeding of yeast peptides significantly increased (p < 0.05) the levels of IL-1β and IL-6 of the colonic mucosa, while it decreased (p < 0.05) the levels of IL-4 and IL-10 (Table 3). No significant differences were observed in the levels of colonic mucosal inflammatory factors between the YP500 and YP1000 lambs. Notably, compared to the YP500 and YP1000 groups, the YP2000 group exhibited higher (p < 0.05) levels of anti-inflammatory factors (IL4 and IL10) and lower (p < 0.05) levels of pro-inflammatory factors (IL-1β and IL-6).

Table 4 presents the levels of antioxidant indices in the colonic mucosa of lambs. The addition of yeast peptides significantly increased (p < 0.05) T-AOC, SOD, CAT and GSH-Px levels, while concurrently decreasing (p < 0.05) MDA levels in the colonic mucosa. No significant differences were observed in antioxidant indices of the colonic mucosa between the YP500 and YP1000 groups. Among the four treatment groups, the YP2000 group exhibited the highest (p < 0.05) levels of T-AOC, SOD, CAT and GSH-Px, while the lowest (p < 0.05) level of MDA was also recorded in the YP2000 group.

Feeding different doses of yeast peptides significantly increased the levels of DAO in serum (p < 0.05), and the DAO levels in the serum of YP1000 and YP2000 groups were lower than those in YP500 group (p < 0.05). However, yeast peptides did not affect the serum DLA levels of lambs (Table 5). Supplementary yeast peptides significantly upregulated the expressions of MUC2, claudin-4, and Occludin mRNA in colonic mucosa of lambs (p < 0.05), whereas the expressions of claudin1 and ZO-1 mRNA were similar across all four groups (Table 5). Moreover, the expression levels of claudin-4 and Occludin genes in YP1000 and YP2000 groups were significantly higher than those in YP500 group (p < 0.05), while no differences were observed in the expression levels of MUC2 among the three yeast peptides treatment groups.

The sequencing of the 24 samples yielded a total of 358,436,379 clean reads from the colon contents, resulting in the identification of 1,150 ASVs. The four groups exhibited no significant differences in terms of alpha diversity, as measured by Richness, Shannon, Simpson, and Chao1 indices (Figure 1A). Specifically, a total of 288, 247, 320, and 285 ASVs were detected in the CON, YP500, YP1000, and YP2000 groups, respectively (Figure 1B). Furthermore, 142 ASVs were shared among the four groups, while unique ASVs included 48 in CON, 43 in YP500, 29 in YP1000, and 17 in YP2000 (Figure 1B). The Weighted Unifrac Anosim analysis revealed significant differences (p < 0.05) between the four groups (Figure 1C). Additionally, the Weighted Unifrac PCoA analysis indicated significant separation between pairs of the four groups (Figure 1D). The colonic microbial compositions among four groups are shown in Figure 2. The ASVs were identified as 8 phyla, and the relative abundance of Firmicutes, Bacteroidetes, Verrucomicrobia, Proteobacteria greater than 1% were considered to be predominant phyla across groups, accounting for 99.32% of the relative abundance (Figure 2A). There were 80 genera detected in all samples. And the predominant genus (relative abundance > 0.5%) included 20 genera across groups, such as Bacteroides, Lactobacillus, Parabacteroides, Faecalibacterium, Escherichia_Shigella, and Butyricicoccus, accounting for 74.36% of the relative abundance (Figure 2B). The LEfSe analysis suggested that the Escherichia_Shigella, Bacteroides and Staphylococcus were significantly enriched in CON group, the Prevotella was enriched in YP1000 group, and the Pseudoflavonifractor, Lactobacillus, Megasphaera, Romboutsia and Roseburia were enriched in YP2000 group (Figure 2C). Based on the mantel test analysis, the abundance of Roseburia is negatively correlated with both Escherichia_Shigella and Staphylococcus, while a positive correlation between the relative abundance of Escherichia_Shigella and Staphylococcus was observed (Figure 2D). These three genera were significantly associated with diarrhea occurrence and fecal scores in the colon microbiota (Figure 2D). Then, co-occurrence network analysis was conducted to identify the core bacterial genera (R > 0.6; degree cutoff: 1; K-Core: 2) in each treatment group, and the results showed that Escherichia_Shigella and Staphylococcus were the core bacterial genera in the CON group, while the Roseburia was the common core bacterial genera in the YP500, YP1000 and YP2000 groups (Figures 3A–E). Then, the relative abundance of Staphylococcus and Escherichia_Shigella in YP500, YP1000 and YP2000 groups decreased compared with CON group, while the relative abundance of Roseburia increased (Figure 3F).

Next, the main functional pathways of colonic microbiota were identified based on PICRUSt2 across different treatments (Figure 4). Notably, no significant differences were observed in KEGG level 1 pathways, with the exception of the human diseases pathway. The results indicated that the abundance of human diseases pathway in YP500, YP1000 and YP2000 groups was significantly lower (p < 0.05) than that in CON group (Figure 4A). Further analysis of KEGG level 2 pathways of human diseases revealed that only three pathways were detected: drug resistance: antimicrobial, infectious disease: parasitic and infectious disease: bacterial pathways (Figure 4B). Among these, only the infection disease: bacterial pathway in YP500, YP1000 and YP2000 was significantly lower (p < 0.05) than the CON group at the second functional level (Figure 4B). The results of the KEGG level 3 pathway analysis indicated that the pathways for Shigellosis and Staphylococcus aureus infection were reduced (p < 0.05) in YP500, YP1000 and YP2000 groups compared to CON group (Figure 4C). Spearman correlation analysis was conducted for the signature genera and phenotype parameters, which included intestinal immune and antioxidant function (Figure 5). As expected, the microbes Staphylococcus and Escherichia-Shigella mainly enriched in the CON group exhibited a positive relationship with IL-1β, IL-6, MDA, and DAO. In contrast, these bacteria demonstrated a negative correlation with IL-4, IL-10, CAT, SOD, T-AOC, GSH-Px, and claudin 4 in the colon (Figure 5). Furthermore, Escherichia_Shigella displayed a negative correlation with the MUC2, while Staphylococcus was negatively correlated with Occludin and ZO-1 (Figure 5). Conversely, Roseburia exhibited a positive correlation with CAT, SOD, T-AOC and claudin 1, whereas a negative correlation was observed with MDA, IL-6, and IL-1β (Figure 5).