Thirty weaned Xiangcun black pigs (28 days of age; half male and half female; male piglets were castrated) were selected for the feeding trial. After 7 days of the adaptation period, piglets with an average body weight (BW) of 11.47 ± 1.13 kg were randomly allocated into three groups with 10 pigs (replicates) per group. Three groups included the control group (CON, fed a corn-soybean meal-based basal diet), C. vulgaris group (CHV, fed a corn-soybean meal-based basal diet and soybean meal replaced with 5% C. vulgaris), and lysozyme group (LYSO, fed a corn-soybean meal-based basal diet and soybean meal replaced with 5% C. vulgaris and 100 mg/kg lysozyme). The feeding trial lasted 56 days. Dietary C. vulgaris and lysozyme were uniformly mixed with the basal diets. Dietary C. vulgaris was obtained from Xian Saiyang Biotechnology Co. Ltd. (Xian, China), and lysozyme was provided by the Sunson Bioscience Technology Development Co. Ltd. (Beijing, China). The supplementing dose of C. vulgaris and lysozyme was based on the previous findings (9, 10). The experimental diet’s ingredients and proximal composition were formulated (Table 1) to meet the Chinese nutrient requirements for domestic pigs in China (NY/T65-2004).

Experimental pigs were housed in individual pens (1.10 × 0.60 m) equipped with an adjusted drinking water nipple, single-hole feeder, and plastic-slatted floored room. The piggery was facilitated with forced ventilation and environmentally controlled temperature (23–26°C) and humidity (60 ± 5%). All pigs had free access to water and food at all times. All experimental animals enrolled in this study were in good health conditions, and had no antibiotic exposure or any gastrointestinal diseases before the trial.

At the end of the feeding trial and 12 h after fasting, all pigs were selected for sampling after weighing. Blood samples (10 mL) were collected from the anterior vena cava of each pig into anticoagulated vacuum blood collection tubes, mixed upside down, and then centrifuged for 10 min at 3,500 × g and 4°C to obtain the plasma. The plasma samples were immediately stored at −80°C for further analyses of immunoglobulins and immuno-cytokines. Afterward, all pigs were anesthetized with an intramuscular injection of Zoletil^® 50 (Tiletamine and zolazepam; Beijing Lab Anim. Tech. Dev. Co. Ltd., Beijing, China) and exsanguinated to collect jejunum, ileum, and colon samples. The samples (2–3 cm) of the jejunum (10 cm below the flexure of duodenum-jejunum) and ileum (10 cm above the ileo-cecal junction) were excised, flushed in phosphate buffer solution, and then scrapped by glass slides. The mucosa scrapings (~2 g) were sampled, snap-frozen in liquid nitrogen, and immediately stored at −80°C to determine intestinal immuno-cytokines. Colonic contents (mid-section) were harvested into 1.5 mL sterile tubes and stored at −80°C for colonic microbiota and metabolite analyses.

The initial and final BW of each pig was measured to calculate the average daily gain (ADG). Feed consumption and feed refusal of each pig were recorded daily to calculate the average daily feed intake (ADFI) of pigs. The ratio of feed intake to weight gain (F/G) was calculated. The diarrhea rate of experimental pigs was evaluated daily using Hart and Dobb’s method as described previously (11, 12). Briefly, the diarrheal conditions of pigs were recorded daily as a fecal score of 0 for normal, 1 for mild diarrhea, 2 for moderate diarrhea, and 3 for severe diarrhea, respectively. The diarrhea rate (%) was calculated as follows:

Total diarrhea times/(total number of piglets × experimental days) × 100.

Plasma biochemical parameters, including total protein (TP), albumin (ALB), ammonia (AMM), triglycerides (TG), total cholesterol (TC), high-density lipoprotein-cholesterol (HDL-C), low-density lipoprotein-cholesterol (LDL-C), and glucose (GLU) concentrations, as well as alpha-amylase (α-AMY), alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), lactate dehydrogenase (LDH), and cholinesterase (CHE) activities were analyzed using their corresponding available kits (F. Hoffmann-La Roche Ltd., Basel, Switzerland) and instruments (Cobas c311, Basel, Switzerland). The levels of plasma immunoglobulins, including immunoglobulin A (IgA, RC-01340P2), IgG (RC-01237P2), and IgM (RC-01236P2) were determined using the commercially available ELISA kits (Haiyang Biological Co. Ltd., Changsha, China) and following the guidelines provided by the manufacturer.

The concentrations of plasma, jejunal, and ileal interleukin (IL)-1β (RC-01256P1), IL-2 (RC-01255P1), IL-6 (RC-01252P1), IL-10 (RC-01255P1), IL-17 (RC-01211P1), tumor necrosis factor (TNF)-α (RC-01217P1), and interferon (IFN)-γ (RC-01246P1) were evaluated with commercially available porcine-specific ELISA kits (Haiyang Biological Co. Ltd., Changsha, China) according the guidelines provided by the company. Absorbance values were read on a Multiscan Spectrophotometer (Infinite M200 PRO; TECAN, Männedorf, Switzerland).

Approximately 0.30 g of colonic contents of each sample were thawed and homogeneously mixed for bacterial genomic DNA extraction with a Mag-Bind^® Stool DNA kit (Omega, Guangzhou, China). The concentration and quality of the extracted DNA were determined with the NanoDrop OneC miniature Volume UV–Vis spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and agarose gel electrophoresis was used for resulting PCR products. The hypervariable V4–V5 region of the colonic bacterial 16S rRNA genes was amplified with universal forward primer 338F (5′-ACTCCTACGGGGAGGCAGCA-3′) and reverse primer 806R (5′-GGACTACHVGGGGTWTCTAAT-3′). The PCR reaction thermal cycle conditions were followed by the manufacturer’s standard protocols (New England Biolab. Inc., Ipswich, MA, USA). The obtained PCR products were purified with Vazyme VAHTSTM DNA Clean Beads to remove non-specific products, and then the constructed DNA libraries were quantitatively analyzed using the Quant-iT PicoGreen dsDNA Detection Kit (Invitrogen, Carlsbad, CA, USA). Based on the standard protocols of Shanghai Personal Biotech. Co. Ltd. (Shanghai, China), purified amplicons were subjected to equimolar and pair-end (2 × 300) sequencing by the MiSeq Reagent kit on an Illumina MiSeq platform (Illumina, San Diego, CA, USA).

The raw sequence data obtained from the Illumina Miseq were processed using the DADA2 plugin and QIIME2 for sequence and subsequent data analysis as previously described (13). After that, using the default parameter, a representative sequence was selected from each operational taxonomic unit (OTU) and classified based on BLAST searching against the Greengenes database. The alpha-diversity indices, including Chao 1, Shannon index, Simpson index, and Observed_species, were calculated using the OTU table and QIIM2. Beta-diversity analysis was performed using principal coordinate analysis (PCoA) and non-measured multidimensional scaling (NMDS) plots based on the Bray-Curtis distance metric to identify the structural differences in microbial communities among samples. The differences in the taxonomic composition of colonic microbiota at the phylum and genus levels were evaluated using the Kruskal-Wallis test. The linear discriminant analysis size effect (LEfSe) was performed to analyze the taxonomic hierarchical distribution of marker species in each group of samples using the histograms of the distribution of linear regression analysis (LDA) values of different species and species taxonomic branching diagrams (Cladograms) to present the significantly enriched species and their levels of importance in each group of samples.

The short-chain fatty acid (SCFA), including straight-chain fatty acids (acetate, propionate, butyrate, and valerate) and branched-chain fatty acids (isobutyrate and isovalerate) concentrations in the colonic contents of weaned pigs were measured with the gas chromatography (7890A; Agilent Technologies Inc., Santa Clara, CA, USA), following the previously described protocols (14). Briefly, colonic contents (1.00 g) were homogeneously mixed with ultrapure water (5 mL), and supernatants were obtained by centrifuging at 1000 × g and 4°C for 10 min. The obtained supernatants were mixed with 25% metaphosphoric acid at a 9:1 (v/v) ratio in 2 mL sterile centrifuge tubes, centrifuged at 2000 × g and 4°C for 10 min, and then filtered through a 0.45-μm polysulfone filter for gas chromatography analysis.

Data analyses of growth performance, plasma biochemical parameters, immunoglobulins, and plasma and intestinal immuno-cytokines were analyzed by one-way analysis of variance (ANOVA). Multiple comparisons among different groups were evaluated using Tukey’s post-hoc test with the SPSS 26.0 software package (IBM Inc., Chicago, IL, USA). In addition, two pigs were removed from the microbiome analysis due to experimental outliers, and the sample size was adjusted to n = 8 per group, as indicated where applicable. Colonic microbiome data were analyzed using the Kruskal-Wallis test. The individual pigs were considered the experimental unit for all analyses. All data are expressed as means with their pooled standard error of the means (SEM). Differences among groups were considered statistically significant when p < 0.05, and a trend at 0.05 ≤ p < 0.10. Correlations between abundant top 20 bacterial taxa at the genus level and plasma biochemical parameters, immunoglobulins, immuno-cytokines, and colonic SCFA concentrations were analyzed using Spearman’s rank correlation test.

The effects of dietary C. vulgaris and lysozyme supplementation on the growth performance of weaned Xiangcun black pigs are presented in Table 2. The diarrhea rate was decreased (p < 0.001) in the CHV group compared with the CON and LYSO groups. Additionally, the diarrhea rate in the LYSO group was increased regarding the CON and CHV groups (p < 0.001). However, dietary C. vulgaris and lysozyme had no significant impacts (p > 0.10) on the growth performance parameters, including BW, ADG, ADFI, and F/G of weaned pigs.

The changes in plasma biochemical parameters of weaned Xiangcun black pigs are listed in Table 3. The plasma ALB (p = 0.052) concentration and ALP (p = 0.090) activity displayed increasing trends in the CHV group compared with the CON and LYSO groups. The plasma TC concentration was higher (p < 0.05) in the LYSO group compared with the CON and CHV groups. Moreover, plasma HDL-C concentration was higher (p < 0.05) in the CHV and LYSO groups compared with the CON group. No significant differences (p > 0.10) were observed in other plasma biochemical parameters among the three groups.

The level of plasma IgM was decreased (p < 0.05) in the CHV group compared with the CON and LYSO groups. However, supplementation of C. vulgaris and lysozyme did not affect (p > 0.10) the plasma levels of IgA and IgG in weaned pigs (Table 3).

As shown in Table 4, the IL-6 content was lower while IL-10 content was higher in the jejunum of the LYSO group (p < 0.05) compared with the CON and CHV groups. In the ileum, the IL-1β content was higher (p < 0.05) in the LYSO group compared with the CON group. Moreover, the TNF-α content was lower (p < 0.05) in the ileum of the CHV group compared with the CON and LYSO groups. However, supplementation of C. vulgaris and lysozyme did not affect (p > 0.10) the plasma immno-cytokines levels.

Colonic contents from 24 samples (n = 8) were assessed to evaluate the effects of C. vulgaris and lysozyme on the microbiota diversity of weaned pigs. The Venn analysis (Figure 1a) showed that there were 3,530, 3,444, and 3,040 unique OTUs in the colonic contents of the CON, CHV, and LYSO groups, respectively. Among them, there were 953 common OTUs among the three groups, of which 226, 261, and 412 common OTUs between the CON vs. CHV, CON vs. LYSO, and CHV and LYSO groups, respectively (Figure 1a).

The alpha-diversity analysis was performed to identify the diversity and structure of the colonic microbiota among different groups. The Simpson index displayed an increasing trend (p = 0.07) in the CHV group compared with the CON and LYSO groups. However, there were no significant changes (p > 0.10) in the Chao1, Shannon, and Observed_species among the three groups (Figure 1b). Furthermore, the beta-diversity analysis was performed to visualize the differences in the colonic microbiota community among the three groups. However, the PCoA showed no distinct separation among the three groups (Figure 1c).

At the phylum level, Firmicutes, Bacteroidota, and Proteobacteria were the top three bacterial phyla (>90%) in the colon of weaned pigs in all groups, while other phyla were shown at very low relative abundance (Figure 2a). The top three phyla (including Firmicutes, Bacteroidota, and Proteobacteria) accounted for 81.16, 15.74, and 1.69% in the CON group, 71.16, 21.88, and 4.11% in the CHV group, and 75.90, 16.66, and 3.82% in the LYSO group, respectively (Figure 2a). The relative abundance of Firmicutes was decreased, while Desulfobacterota was increased in the CHV and LYSO groups compared with the CON group (p < 0.05). Moreover, the relative abundance of Actinobacteriota was higher (p < 0.05) in the LYSO group compared with the CON and CHV groups (Figure 2b).

Colonic bacterial community composition at the genus level is shown in Figure 3. Lactobacillus, Prevotella, and Megasphaera_A were the top three bacterial genera in the colon of all pigs. Lactobacillus (21.71%), Megasphaera_A (11.25%), Prevotella (9.10%), Limosilactobacillus (4.88%), Gemmiger_A (3.43%), Agathobacter (3.47%), Faecalibacterium (2.28%), and Phascolarctobacterium_A (1.74%) were the most abundant bacterial genera in the CON group. The most abundant bacterial genera in the CHV group were Lactobacillus (10.27%), Prevotella (7.75%), Sodaliphilus (4.87%), Limosilactobacillus (3.85%), Megasphaera_A (3.72%), Phascolarctobacterium_A (3.50%), and Faecalibacterium (1.45%). In addition, Lactobacillus (10.72%), Prevotella (7.75%), Megasphaera_A (5.99%), Limiosilactobacillus (3.90%), and Phascolarctobacterium_A (3.50%) were the most abundant bacterial genera in the LYSO group (Figure 3a). The relative abundance of Phascolarctobacterium_A was higher, whereas Megasphaera_A was lower in the CHV and LYSO groups compared with the CON group (p < 0.05). The relative abundance of Faecalibacterium was higher (p < 0.05) in the LYSO group compared with the CON and CHV groups, while Anaerovibrio was higher (p < 0.05) in the LYSO group compared with the CHV group (Figure 3b).

To identify the differences in colonic microbial function among different groups, LEfSe analysis (LDA threshold score ≥ 3.0) was performed at the phylum and genus levels (Figure 4a). There was a significant enrichment of Firmicutes in the CON group and Desulfobacterota in the LYSO group at the phylum level. At the genus level, Succinivibrio was enriched in the CON group, while Eubacterium, Faecousia, CAG_238, and Peptococcus were enriched in the CHV group. Additionally, Dusulfovibrio_R, Bariatricus, Facccalibacterium, Phascolarctobacterium_A, Anaerovibrio, Bulleidia, and Holdemanella were enriched in the LYSO group (Figure 4a). Moreover, Firmicutes_C in the CON group, Peptococcia in the CHV group, and Desulfobacterota_I in the LYSO group were the most dominant microbiota (Figure 4b).

The effects of dietary C. vulgaris and lysozyme supplementation on the colonic SCFA concentrations of weaned pigs are presented in Table 5. Colonic acetate, propionate, butyrate, and valerate concentrations were lower (p < 0.001) in the CHV and LYSO groups compared with the CON group. However, acetate and butyrate concentrations were higher (p < 0.001) in the LYSO group compared with the CHV group. Moreover, colonic isobutyrate and isovalerate concentrations were higher (p < 0.05) in the CHV group compared with the CON and LYSO groups.

Spearman’s correlation analysis was performed to reveal the associations between the levels of plasma biochemical parameters, immunoglobulins, immuno-cytokines, colonic SCFA, and the top 20 colonic bacterial taxa at the genus level (Figure 5). Plasma TP was negatively (p < 0.05) correlated with Lactobacillus, whereas plasma ALB was negatively correlated (p < 0.01) with Lactobacillus and Megasphaera_A and positively correlated (p < 0.05) with Faceousia, Streptococcus, and Cryptobacteroides. Plasma ALP was negatively correlated (p < 0.05) with Lactobacillus, Megasphaera_A, and Gemmiger_A, while positively correlated (p < 0.01) with Sodaliphilus and Cryptobacteroides. Plasma LDH was positively correlated with Sodaliphilus, while negatively correlated with Megasphaera_A and Agathobacter (p < 0.05). There were positive correlations between plasma HDL-C with Phascoloractobacterium_A and Alloprevotella and plasma LDL-C with Lachnospira and Schwartzia (p < 0.05). Plasma AMM was positively correlated with Dysosmobacter and negatively correlated with Limosilactobacillus (p < 0.05). Moreover, plasma IgM was positively correlated (p < 0.05) with Lactobacillus and Schwartzia and negatively correlated (p < 0.01) with Sodaliphilus (Figure 5a).

The correlations between plasma and intestinal immuno-cytokines and colonic bacterial taxa at the genus level are shown in Figures 5b,c. The positive correlation included between plasma IL-1β with Germmiger_A and plasma IL-17 with Escherichia, while the negative correlation included between plasma IL-2 with Mitsuokella, plasma IL-10 with Limosilactobacillus, and plasma IFN-γ with Germmiger_A (Figure 5b). The IL-2 in the jejunum was positively correlated with Agathobacter, while IL-17 in the jejunum was positively correlated with Cryptobacteroides and negatively correlated with Phascoloractobacterium_A (p < 0.05). In addition, IL-1β in the ileum was positively correlated with Phascoloractobacterium_A, while IL-2 in the ileum was negatively correlated with Anaerovibrio and Mitsuokella (p < 0.05). The TNF-α in the ileum was positively correlated with Lactobacillus, Megasphaera_A, and Blautia_A, while it was negatively correlated with Faecousia, Escherichia, Alloprevotella, Streptococcus, Cryptobacteroides, and Dysosmobacter (p < 0.05). Furthermore, IFN-γ in the ileum was negatively correlated (p < 0.05) with Alloprevotella (Figure 5c).

The correlations between SCFA and colonic bacterial taxa at the genus level are shown in Figure 5d. The positive correlation included between acetate with Gemmiger_A and Agathobacter (p < 0.05), propionate with Lactobacillus (p < 0.01), Megasphaera_A (p < 0.05), Limosilactobacillus (p < 0.05),, and Mitsuokella (p < 0.05), butyrate with Lactobacillus and Gemmiger_A (p < 0.05), valerate with Lactobacillus (p < 0.05) and Megasphaera_A (p < 0.01), isobutyrate with Sodaliphilus, Faecousia, Cryptobacteroides, and Dysosmobacter (p < 0.05), and isovalerate with Prevotella, Sodaliphilus, Faecousia, Cryptobacteroides, and Dysosmobacter (p < 0.05). The negative correlation (p < 0.05) included between acetate with Phascoloractobacterium_A, Sodaliphilus, and Streptococcus, propionate (p < 0.05) with Phascoloractobacterium_A, Sodaliphilus, Faecousia (p < 0.001), Alloprevotella, Streptococcus, Cryptobacteroides, and Dysosmobacter, butyrate with Phascoloractobacterium_A (p < 0.001) and Faecousia (p < 0.05), valerate (p < 0.01) with Phascoloractobacterium_A, Faecousia, and Streptococcus (p < 0.05), isobutyrate with Germmiger_A (p < 0.01) and Faecalibacterium (p < 0.05), isovalerate with Lactobacillus (p < 0.05), Germmiger_A (p < 0.01), and Faecalibacterium (p < 0.05).