Broilers (male, 1-day-old) of the Arbor Acres (AA) breed were acquired from a nearby poultry farm and reared in hygienic housing with consistent light/dark cycles, provided with water and feed ad libitum. All the broilers underwent testing, and none of them were found to have MG. When the trial was over, all chickens were euthanized by cervical dislocation.

A total of 96 one-day-old male AA broilers with an initial average body weight (39.29 ± 0.54 g) were selected and divided into 4 groups: (1) broilers were fed a control diet (Con); (2) broilers were fed a control diet supplemented with 10⁸ CFU/kg of B. subtilis KC1 (KC); (3) broilers were fed a control diet and challenged with MG at day 14 (MG); and (4) broilers were fed a control diet supplemented with 10⁸ CFU/kg of B. subtilis KC1, and further challenged with MG at day 14 (MG_KC). Each treatment was replicated 3 times, with 8 broilers per replicate. All diets were formulated to meet the National Research Council and Chinese chicken feeding standards (NY/T-33–2004) (17, 18). A single-layer, three-dimensional chicken coop was used. Temperature was gradually reduced by 2°C per week from 33 to 23°C and then kept constant.

The MG strain R_low was obtained from the China Veterinary Culture Collection Center and cultured in a Mycoplasma medium (H910KJ, BasalMedia, Shanghai, China), supplemented with 1% ampicillin (ST008, Beyotime, Shanghai, China) and 10% fetal bovine serum (S660JJ, BasalMedia, Shanghai, China). The procedure for infecting the broilers with MG involved the use of a syringe to inoculate a 200 μL volume of solution, containing 2 × 10⁹ color change units (CCUs) of MG, into the left air sacs of the broilers.

The strain of B. subtilis KC1 (GenBank accession number OL721931) was previously isolated from the fecal matter of healthy chickens (16). This B. subtilis KC1 strain was cultured in a modified liquid medium, composed of 1.25 g of glucose, 1.25 g of yeast extract, 1.25 g of peptone, 1.25 g of beef paste, and 1.25 g of NaCl dissolved in 250 mL of sterilized water, adjusted to a pH of 7.2, to achieve a bacterial suspension of 10⁸ CFU/mL. Subsequently, the B. subtilis KC1 was mixed fully with the control diet, achieving a final concentration of 10⁸ CFU per kilogram of feed.

The broilers of each replicate were weighed at the beginning and day 42. Feed intake and the number of broilers were recorded to calculate the average daily gain (ADG), average daily feed intake (ADFI) and feed conversion ratio (FCR) of every replicate. ADG = [total body gain (g)/test days]/number of broilers in each replicate, ADFI = [total feed intake (g)/test days]/number of broilers in each replicate, FCR = ADFI/ADG.

Blood samples were collected (6 samples per treatment, 2 per replicate) at day 42. Serum was obtained by centrifuging the blood at 3,000 × g for 10 min at 4°C. Following the manufacturer’s instructions, serum superoxide dismutase (SOD, A001-3-2, Nanjing Jiancheng Biotechnology Co., Ltd., Nanjing, China), malondialdehyde (MDA, A003-1-2, Nanjing Jiancheng Biotechnology Co., Ltd., Nanjing, China), interleukin-1 (IL-1β, ml002787, Shanghai Enzyme-linked Biotechnology Co., Ltd., Shanghai, China) and tumor necrosis factor-a (TNF-α, ml002790, Shanghai Enzyme-linked Biotechnology Co., Ltd., Shanghai, China) levels were detected using commercially ELISA kits.

Serum was obtained as described above. Untargeted metabolomics was completed by Beijing Genomics Institute (BGI, Shenzhen, China). To prepare the samples for metabolomic analysis, add 100 μL of serum from each sample to a tube. Subsequently, add 700 μL of an extractant containing internal standard (prepared at a ratio of methanol: acetonitrile: water = 4:2:1, V/V/V), shake vigorously for 60 s, and then place the tubes in a −20°C refrigerator for 2 h. After the incubation, centrifuge the samples at 25,000 × g for 15 min at 4°C. Carefully remove the samples from the centrifuge and transfer 600 μL of the supernatant into a new tube. Proceed with drying using a drying machine until complete. Once dry, add 180 μL of a methanol: pure water mixture (1,1, v/v), vortex for 10 min to ensure complete dissolution in the reconstituted solution. Centrifuge the samples again at 25,000 × g for 15 min at 4°C. Collect the supernatant into a fresh tube. Take 20 μL of each sample and combine them to create quality control (QC) samples. Finally, transfer the prepared supernatants to the LC–MS/MS analysis steps. The sample extracts were analyzed using Waters UPLC I-Class Plus (Waters, United States) equipped with QTRAP 6500 Plus (SCIEX, United States). After importing the off-line data of mass spectrometry into compound discoverer 3.3 (Thermo Fisher Scientific, United States) software and analyzing the mass spectrometry data in combination with BMDB (BGI metabolome database), mzcloud database and chemspider online database, a data matrix containing information such as metabolite peak area and identification results will be obtained. These metabolites were annotated using the KEGG database and HMDB database. The principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA) were performed at meta X. Metabolites were considered to be significantly different if they had a variable important for the projection (VIP) value greater than 1, a p-value less than 0.05, and a fold change of at least 1.2 or not more than 0.83. Volcano plots were then generated based on the log2 of the fold change and the negative log10 of the p-value for these metabolites. The functions of these differential metabolites and their associated metabolic pathways were investigated using the KEGG database.

Lung, liver and ileum samples were collected (3 samples per treatment, 1 per replicate) at day 42. Total RNA was extracted using TRIzol reagent (ET101-01-V2, Beijing TransGen Biotech Co., Ltd., Beijing, China). The RNA concentration and integrity were detected using a Fragment Analyzer (Agilent, California, United States). RNA samples that complied with the criteria of having a concentration of more than 40 ng/μL, an integrity value greater than 7.0, and a total content of 1 μg or more were chosen for further experimentation. Transcriptomic sequencing was completed by BGI. The procedures for library construction, sequencing and bioinformatics analysis were as follows. Oligo dT beads were employed to enrich mRNA carrying a poly(A) tail. Subsequently, the RNA was fragmented, and first-strand cDNA production was initiated through random N6-primed reverse transcription. This was followed by the synthesis of second-strand cDNA, where dUTP was utilized instead of dTTP. The resulting cDNA then underwent end-repair and subsequent 3′ adenylation. Afterward, adaptors were attached to the termini of these 3′ adenylated cDNA segments. Before initiating PCR amplification, the strand marked with dUTP was selectively degraded using Uracil-DNA-Glycosylase (UDG). The remaining strand was then amplified to create a cDNA library suitable for sequencing. To enrich the purified cDNA template, multiple rounds of PCR amplification were carried out using PCR primers. Heat was used to denature the PCR product, and subsequently, the single-stranded DNA was circularized with the aid of a splint oligo and DNA ligase. Ultimately, DNA nanoball synthesis and sequencing were executed on the DNBSEQ platform in paired-end 150 sequencing mode. Sequencing data were called raw reads, and quality control was then performed on the raw reads to obtain clean reads. After quality control, the filtered clean reads were aligned to the reference sequence. Then, gene expression quantification, gene difference analysis, and enrichment analysis were performed. Data quality control was completed using SOAPnuke (v1.5.2). Genome alignment was completed using HISAT2 (2.0.4). Transcript quantification and differential comparison were completed via RSEM and DESeq2, respectively. Genes were considered to be significantly different if they had a q-value (adjusted p-value) less than 0.05 (19). Volcano plots were then generated based on the log2 of the fold change and the negative log10 of the q-value. The functions of these differential genes and their associated metabolic pathways were investigated using the KEGG database.¹

The results were presented as mean ± standard deviation (SD) and were analyzed utilizing GraphPad Prism 8.0 software (GraphPad Software, Inc.). The comparisons among multiple groups were conducted using one-way analysis of variance (ANOVA) accompanied by Tukey’s multiple comparison test. A p-value less than 0.05 was deemed to indicate statistical significance.

Compared with the Con group, the KC group exhibited significantly increased final body weight (p < 0.0001) and ADG (p = 0.0023) of broilers, as well as significantly decreased FCR (p = 0.0019) (Figures 1A,B,D). Compared with the Con group, the MG group exhibited significantly reduced final body weight (p < 0.0001) and ADG (p < 0.0001) of broilers, and significantly increased FCR (p < 0.0001) (Figures 1A,B,D). Compared with the MG group, the MG_KC group exhibited significantly increased final body weight (p < 0.0001) and ADG (p = 0.0034) of broilers, and significantly reduced FCR (p < 0.0001) (Figures 1A,B,D). There was no significant difference in ADFI among all groups (Figure 1C). These results indicated that dietary B. subtilis KC1 supplementation improved the growth performance of MG-infected chickens.

Compared with the Con group, the KC group exhibited a increased trend in SOD activity (p = 0.0770) and significantly reduced MDA content (p = 0.0030) in serum (Figures 2A,B). Compared with the Con group, the MG group exhibited significantly reduced SOD activity (p < 0.0001) and significantly increased MDA content (p < 0.0001) in serum (Figures 2A,B). Compared with the MG group, the MG_KC group exhibited significantly increased SOD activity (p = 0.0006) and significantly reduced MDA content (p < 0.0001) in serum (Figures 2A,B).

Compared with the Con group, the KC group did not affect the levels of IL-1β (p = 0.5922) and TNF-α (p = 0.9985) (Figures 2C,D). Compared with the Con group, the MG group exhibited significantly increased serum IL-1β (p < 0.0001) and TNF-α levels (p < 0.0001) (Figures 2C,D). Compared with the MG group, the MG_KC group exhibited significantly reduced serum IL-1β (p < 0.0001) and TNF-α levels (p = 0.0164) (Figures 2C,D). These results indicated that dietary B. subtilis KC1 supplementation alleviated the oxidative stress and pro-inflammatory response caused by MG.

A total of 4,579 annotated metabolites were identified (Supplementary File S1). The PCA indicated that the serum samples from the MG group were significantly separated from the Con group (Figure 3A). Moreover, distinct clusters from the MG group compared with the Con group were confirmed by the OPLS-DA (Figure 3B). To assess the quality of the model without the risk of overfitting, 200 response permutation tests (RPT) were performed on the OPLS-DA model. The results indicated that the model was reliable and did not suffer from overfitting, as evidenced by the fact that the intercept of the goodness-of-fit measure (R2) was greater than the intercept of the goodness-of-prediction measure (Q2), and furthermore, the intercept of Q2 was less than zero (Figure 3C). Total 531 significantly differential metabolites were identified in the MG group compared with the Con group (Supplementary File S2). The volcano plots illustrated the global distribution of metabolites, revealing distinct differences in metabolite levels between the Con and MG groups (Figure 3D). KEGG pathway enrichment analysis indicated that the differential metabolites mainly enriched in the amino acids metabolism related pathways (Figure 3E).

The PCA indicated that the serum samples from the MG_KC group were significantly separated from the MG group (Figure 3F). Moreover, distinct clusters from the MG_KC group compared with the MG group were confirmed by the OPLS-DA (Figure 3G). In order to judge the quality of the model without fitting risk, 200 response permutation tests (RPT) were performed on the OPLS-DA model. The results indicated that the model was reliable and did not suffer from overfitting, as evidenced by the fact that the R2 was greater than the Q2, and furthermore, the intercept of Q2 was less than zero (Figure 3H). Total 318 significantly differential metabolites were identified in the MG_KC group compared with the MG group (Supplementary File S3). The volcano plots illustrated the global distribution of metabolites, revealing distinct differences in metabolite levels between the MG and MG_KC groups (Figure 3I). KEGG pathway enrichment analysis indicated that the differential metabolites mainly enriched in the amino acids metabolism related pathways (Figure 3J).

Compared with the Con group, the MG group exhibited significantly reduced l-proline, l-threonine, l-histidine, l-lysine, l-serine, l-aspartic acid, l-arginine, l-glutamine, Leucine, l-tryptophan, glycine, asparagine, O-acetyl-l-serine, l-Homoserine and l-allo-threonine levels (Table 1). Compared with the MG group, the MG_KC group exhibited significantly increased l-proline, l-histidine, l-lysine, l-aspartic acid, l-arginine, leucine, l-tryptophan, asparagine and l-homoserine (Table 1). Above all, these results indicated that MG infection induced abnormal amino acid metabolism, while B. subtilis KC1 alleviated the abnormal amino acid metabolism caused by MG.

A total of 1,570 significantly differential genes in the lung were identified in the MG group compared to the Con group (Supplementary File S4). The volcano plots illustrated the global distribution of genes, revealing distinct differences in gene expression levels in the lung between the Con and MG groups (Figure 4A). KEGG pathway enrichment analysis indicated that 12 pathways related to amino acid metabolism were significantly enriched (Figure 4B). Furthermore, a total of 458 significantly differential genes in the lung were identified in the MG_KC group compared with the MG group (Supplementary File S5). The volcano plots illustrated the global distribution of genes, revealing distinct differences in gene expression levels in the lung between the MG and MG_KC groups (Figure 4C). KEGG pathway enrichment analysis indicated that five pathways related to amino acid metabolism were significantly enriched (Figure 4D). It is worth noting that four specific metabolic pathways, namely Cysteine and methionine metabolism, Arginine and proline metabolism, Valine, leucine and isoleucine degradation, as well as Alanine, aspartate and glutamate metabolism, are significantly enriched in both of the above enrichment analysis results (Figures 4B,D).

A total of 2,738 significantly differential genes in the liver were identified in the MG group compared with the Con group (Supplementary File S6). The volcano plots illustrated the global distribution of genes, revealing distinct differences in gene expression levels in the liver between the Con and MG groups (Figure 5A). KEGG pathway enrichment analysis indicated that 14 pathways related to amino acid metabolism were significantly enriched (Figure 5B). Furthermore, a total of 2,967 significantly differential genes in the liver were identified in the MG_KC group compared with the MG group (Supplementary File S7). The volcano plots illustrated the global distribution of genes, revealing distinct differences in gene expression levels in liver between the MG and MG_KC groups (Figure 5C). KEGG pathway enrichment analysis indicated that 14 pathways related to amino acid metabolism were significantly enriched (Figure 5D). It is worth noting that 14 metabolic pathways, including valine, leucine, and isoleucine degradation; cysteine and methionine metabolism; glycine, serine, and threonine metabolism; tryptophan metabolism; arginine and proline metabolism; biosynthesis of amino acids; lysine degradation; alanine, aspartate, and glutamate metabolism; histidine metabolism; beta-alanine metabolism; tyrosine metabolism; arginine biosynthesis; phenylalanine metabolism; and phenylalanine, tyrosine, and tryptophan biosynthesis, are significantly enriched in both of the above enrichment analysis results (Figures 5B,D).

A total of 3,553 significantly differential genes in the ileum were identified in the MG group compared with the Con group (Supplementary File S8). The volcano plots illustrated the global distribution of genes, revealing distinct differences in gene expression levels in the ileum between the Con and MG groups (Figure 6A). KEGG pathway enrichment analysis indicated that 14 pathways related to amino acid metabolism were significantly enriched (Figure 6B). Furthermore, a total of 579 significantly differential genes in the ileum were identified in the MG_KC group compared with the MG group (Supplementary File S9). The volcano plots illustrated the global distribution of genes, revealing distinct differences in gene expression levels in ileum between the MG and MG_KC groups (Figure 6C). KEGG pathway enrichment analysis indicated that nine pathways related to amino acid metabolism were significantly enriched (Figure 6D). It is noteworthy that nine metabolic pathways, including lysine degradation, arginine and proline metabolism, valine, leucine, and isoleucine degradation, glycine, serine, and threonine metabolism, tryptophan metabolism, biosynthesis of amino acids, tyrosine metabolism, alanine, aspartate, and glutamate metabolism, and arginine biosynthesis, are significantly enriched in both of the aforementioned enrichment analysis results (Figures 6B,D). In summary, these results indicated that MG infection induced amino acids metabolic disturbance in chickens, while B. subtilis KC1 alleviated the amino acid metabolism disorder caused by MG.