Ten high-yielding Holstein dairy cows (695.2 ± 33.5 kg) from a dairy farm in Jiusan (Heilongjiang Province, China) were selected for ruminal fluid collection using a nasopharyngeal tube before morning feeding. After bottling, the vessels were wrapped with a sealing membrane and quickly transported to the laboratory in an insulation foam box. Sterile medical gauze was folded into four layers, poured into the rumen fluid, and filtered. The filter was then placed in a new sterile blue cap bottle and centrifuged at 100 × g for 10 min. The supernatant of rumen fluid from all cows was spread on tryptose soy agar (TSA) plates (Qingdao Hope Bio-Technology Co. Ltd., Qingdao City, China) to isolate the bacteria. All strains grown on TSA plates for 24 h at 37°C under anaerobic and static conditions were cultured by repeating this step 4–5 times to obtain the purified strain. Bacillus subtilis was initially isolated from rumen fluid samples, and bacterial DNA was extracted using the Tian Gen kit and amplified by polymerase chain reaction (PCR). Amplified products were sent to Shanghai Bioengineering Co., Ltd. for sequencing.

The obtained bacterial solution was diluted with saline, and 100 μl of the diluted bacterial solution was extracted and evenly coated onto TSA for 24 h at 37°C. The number of colonies on the plate was counted and the concentration of the original bacterial solution was calculated and diluted to different concentrations (10⁴, 10⁵, 10⁶, 10⁷,10⁸, and 10⁹ CFU/ml) using Luria–Bertani (LB) broth for the rumen explant assay. The inactivated bacterial broth was obtained by placing the living bacteria broth in a sterilizer at 121°C for 15 min. The bacterial culture supernatant was collected by centrifugation (100 × g at 4°C for 10 min). The biological characteristics of the target bacteria were determined. The tolerance of the bacterial solution under different culture conditions was determined. Different temperatures at 32, 34, 36, 37, 38, and 40°C; different pH at 6.0, 6.5, 7.0, 7.5, 8.0; different bile salt concentrations of 0.5, 1.0, 1.5, and 2.0%; all cultures were incubated in a shaking table at 150 revolutions per min (rpm/min) for 24 h, and absorbance was determined at 600 nm using a microplate reader.

The Ethics Committee of Heilongjiang Bayi Agricultural University (Daqing, China) approved the study protocol for the use and care of animals. Animal studies were performed in accordance with the guiding principles adopted by the Chinese Association for Laboratory Animal Sciences. Three lactating Holstein cows were selected from a farm slaughterhouse in Daqing, Heilongjiang province. Average parity, body weight, days in milk, and milk yield prior to slaughter of the cows were 3.6 ± 0.6, 686 ± 25.2, 228.6 ± 22.2, and 30.3 ± 2.1 kg/day [mean ± standard deviation (SD)], respectively. The collection and culture of rumen explants were carried out according to the method described by Zhang et al. (20) and Fathi et al. (26). Briefly, the rumen from each cow was emptied and surface debris was rinsed off with saline. A subsample of tissue was placed in a beaker containing a mixture of phosphate buffer saline (PBS, contained 2.5 mg/ml amphotericin B, 100 mg/ml streptomycin, 50 mg/ml penicillin, and 50 mg/ml gentamicin, pH 7.4, 37°C). After rinsing the rumen tissue with PBS to remove residual feed, the rumen mucosal layer was separated and discarded using sterile forceps. The mucosal tissue was cut into 4 mm² sections with a surgical scalpel and incubated in duplicate with the medium. The explant surface was rinsed several times with PBS to wash off the blood and other debris. During the isolation process, tissue from three cows was washed several times with PBS at 37°C to slow-down tissue activity. Three separate tissue blocks from each cow were then laid out nipple-up with sterile forceps in 24-well plates containing 2.5 mL of culture medium A or B and incubated at 37°C in a 5% CO₂ incubator.

To detect the activity of rumen explants at different times, explants from 0, 2, 8, 12, 16, 20, 24, 48, and 72 h were rinsed 4–5 times with the PBS mixture and placed in culture plates with 2 ml of medium containing 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) components and incubated at 37°C with 5% CO₂ for 4 h. Blue-purple methanamine (formazan) was allowed to use in tissue cells to be fully deposited. At the end of the incubation, the MTT-containing medium on the surface of the explants was washed off with PBS and immersed in a methanol solution to allow complete precipitation of methane. The explants were removed and placed in a drying oven for 30 min, the dry weight of the explants was recorded, and the optical density of the solution containing methanate at 570 nm was measured using an enzyme marker to calculate the explant viability. The formula used was explant viability = optical density/explant dry weight × 100% (20).

To evaluate the optimal culture time and medium for subsequent experiments, two series of cultures were used. First, medium A contained 80% Dulbecco's modified eagle medium: nutrient mixture F-12 (DMEM/F12 medium), 4% insulin, transferrin, and selenium additive, 100 mg/ml penicillin, 2.5 mg/ml amphotericin B, 100 mg/ml streptomycin, 50 mg/ml gentamicin, 2 mM/L glutamine, and 10% heat-inactivated serum. The second series involved using medium B containing 90% DMEM/F12 medium, 4% insulin, transferrin, and selenium additive (100 ×), 100 mg/ml penicillin, 2.5 mg/ml amphotericin B, 100 mg/mL streptomycin, 50 mg/ml gentamicin, and 2 mM/ml L-glutamine. The culture time encompassed 10-time points from 0 to 72 h (0, 2, 4, 8, 12, 16, 20, 24, 48, and 72 h), and samples from each culture series were collected at each time point to assess the activity of the explants. After sample collection, a portion was fixed in 4% formaldehyde, paraffin-embedded, sectioned and subjected to hematoxylin and eosin (HE) staining to observe structural integrity under a microscope. The remaining portion of each sample was frozen in liquid nitrogen. Subsequently, the tissue was ground and crushed, and TRIzol reagent (Catalog No. 15596-018, LIFE) was added to extract the RNA. After reverse transcription, CDH1 gene expression levels in the explants were measured using quantitative fluorescence PCR. The reaction conditions were 95.0°C for 3 min, 95.0°C for 10 s, 56.0°C for 30 s, 72.0°C for 45 s, and melting for 15 s over 40 cycles.

Freshly-isolated explants from each of the three cows were incubated in triplicates for 1 h at 37°C with 5% CO₂. Briefly, explants were removed and rinsed 4 to 5 times with PBS at 37°C, and then were placed in a new 24-well plate containing 1.8 ml of medium A. Then, 200 μl of 2.0 × 10⁴, 10⁵, 10⁶, 10⁷, 10⁸, and 10⁹ CFU/ml of living and inactivated B. subtilis. The bacterial liquid was placed in an autoclave at 121°C for 30 min, prolonged high temperatures could denature the proteins in bacteria, causing them to lose their biological role. After that, the inactivated bacterial solution was coated on the nutrient agar medium and was cultured at 37°C for 24 h. If no colonies were generated, the inactivated bacterial solution was used for subsequent tests. About 200 μl of B. subtilis solution was added to the culture plates containing rumen explants and 1.8 ml of medium. After 2 h of stimulation, all liquid in the plates was discarded, and the rumen explants were rinsed with PBS 4 to 4–5 times and placed into new plates with 2.5 ml of medium A in each well. Incubation was continued in a CO₂ incubator at 37°C with 5% CO₂ for 12 h. LB broth was used as a control group.

Total RNA was extracted from rumen explants and reverse-transcribed into cDNA using an RNAiso Plus kit (TaKaRa Biotechnology Co. Ltd., Beijing, China) according to the manufacturer's protocol. Subsequently, a reverse transcription kit (RR047A; TaKaRa Biotechnology Co., Ltd., Beijing, China) was used to synthesize cDNA with 1 μg of total RNA according to the supplier's protocols. The SYBR Green Plus Reagent Kit (Roche, Shanghai, China) was used for qRT-PCR assays with a 7500 Real-Time PCR System (Applied Biosystems). Six key factors in the PIK3CB–AKT–mTORC1 pathway, 4EBP1, AKT, MTOR, P70S6K, PIK3CB, and-PDCD, were selected as target genes for qPCR (Applied Biosystems, Analytik Jena, Germany). Raw data were analyzed using the comparative quantification method, and the 2^−ΔΔCt method was used to calculate the relative expression of the target genes, with actin serving as an internal control. Primers were designed using Primer Premier 5 Tool (Premier; Supplementary Table 1; https://doi.org/10.6084/m9.figshare.14660184.v1). The reaction conditions for fluorescence quantitative PCR were: 95.0°C for 3 min, 95.0°C for 10 s, 56.0°C for 30 s, 72.0°C for 45 s, and melting for 15 s over 40 cycles. All samples were analyzed in triplicate.

Western blotting was performed to determine the abundance of six target proteins in the PIK3CB–AKT–mTORC1 pathway in rumen explants cultured at 37°C for 12 h. Total tissue proteins were extracted using a protein extraction kit (C510003; Sangon Biotech Co.Ltd., Shanghai, China) according to the manufacturer's instructions. The protein concentration was measured using the Enhanced BCA Protein Assay Kit (P0010, Beyotime Co. Ltd., Shanghai, China). Protein samples were denatured for 5 min at 100°C, separated on an 8% SDS-PAGE gel (30 μg of protein per sample), and electro-transferred onto polyvinylidene difluoride (PVDF) membranes (Bio-Rad, Shanghai, China). After blocking with blocking buffer Tris Buffered saline with Tween-20 (TBST) (Tris 50 mM for pH 7.6, NaCl 150 mM, and Tween-20 0.1%) containing 5% BSA (4240, Saiguo Biotech Co. Ltd., Guangzhou, China) for 2 h at room temperature, the membranes were then incubated with primary antibodies against PIK3CB (1:1,000, Ab40776, Abcam), PDCD4 (1:1,000, 9535S, Cell Signaling), site-specific phosphorylated AKT (Ser⁴⁷³, 1:1,000, 4060S, Cell Signaling), 4EBP1 (Thr³⁷, 1:1,000, YP001, Immunoway), P70S6K (Thr³⁸⁹, 1:1000, YP1427, Immunoway), mTOR (Ser²⁴⁴⁸,1:2,000, YP0176, Immunoway), and β-Actin (1:10,000, 3700S, Cell Signaling Technology) overnight at 4°C.

The membranes were rinsed with TBST 3 times (10 min each time) and incubated with HRP-conjugated secondary antibodies (anti-mouse IgG1, 96714S, Cell Signaling; anti-rabbit IgG, 14708S, Cell Signaling) diluted at 3:5,000 in TBST for 30 min at room temperature. After brief washing, an enhanced chemiluminescence solution detection kit (ECL, Millipore) was used to visualize the specific bands. Image J2x 2.1.4.7 Analyzer (Rawak Software Inc., Stuttgart, Germany) was used to quantify the intensity of each band. Refer to (27–29); LY294002 (50 μmol/L) and rapamycin (10 nmol/L; 9904) levels were determined with reference to previous studies and combined with the results of previous toxicity tests (30).

After stable incubation of rumen explants for 1 h, LY294002 (50 μmol/L; 9901; Cell Signaling Technology) and rapamycin (10 nmol/L; 9904; Cell Signaling Technology) was added for 30 min, after which explants were removed and rinsed 4–5 times with PBS solution. Subsequently, they were stimulated with 2.0 × 10⁷ CFU/ml of living or inactivated bacteria for 12 h. Subsequently, the explants were washed three times before RNA extraction. Rumen explants treated with inhibitors served as the control group. As in the previous experiment, protein and RNA were extracted from each group of explants, and the tissue culture supernatant was collected.

An automatic amino acid analyzer s-433d (Sykam, Germany) and a vacuum tube concentrator TVE-1100 (EYELA, Tokyo) was used for amino acid profiling. Hierarchical clustering analysis was performed on all significantly different free amino acids, and the results are reported in a heat map. Correlation analysis was then used to measure the closeness of the relationship between the statistically significant amino acids. The degree of linear correlation between two free amino acids was measured using the Pearson correlation coefficient, and the free amino acids with significantly different contents between groups were selected for visualization. Lastly, pathway enrichment analysis of significant free amino acids was performed using the Kyoto Encyclopedia of Genes and Genomes database for a better appreciation of the pathways contributing to amino acid and protein metabolism.

Data were analyzed using SPSS (version 22.0; SPSS Inc., Chicago, IL, USA) and GraphPad Prism 7.00 (GraphPadSoftware, San Diego, CA, USA). Results are expressed as the mean ± SD. All experiments were independently repeated at least three times. Statistical significance was determined using a two-tailed unpaired Student's t-test or one-way ANOVA followed by Tukey's post-hoc test for comparisons between two or more two groups, respectively. Differences were considered significant at ^*p < 0.05 and highly-significant at ^**p < 0.01. For the screening test of optimal culture conditions of explants (experiment II), differences in CDH1 expression were used (“^*” indicates a significant difference compared to the 0 h group, p < 0.05. “^**” indicates highly significant difference compared to 0 h, p < 0.01. “#” indicates a significant difference between the MB group and the MA group at the same incubation time (p < 0.05).

For the addition of B. subtilis (experiment III), ^*p < 0.05 and ^**p < 0.01 were used to indicate the effects of different concentrations and different states of B. subtilis on the expression levels of each key factor in the PIK3CB–AKT–mTORC1 signaling pathway in rumen explants (compared to the control). ^#p < 0.05 was used to indicate the effects of living or inactivated bacteria on gene expression in the pathway.

For the addition of LY294002 and rapamycin, ^*p < 0.05 and ^**p < 0.01 were used to indicate the effects of different inhibitors on the expression levels of each key factor in the PIK3CB–AKT–mTORC1 signaling pathway in rumen explants (compared to the control).

Microscopic observations revealed that the isolated strain B. subtilis IIVE-4 was a gram-positive rod-shaped bacterium with an overall oval or rod-like morphology. Soft, slightly yellowish colonies were formed after incubation on Mannitol-Egg-Yolk-Polymyxin Agar Base at 37°C for 24 h (Figures 1A,B). An amplified gene fragment of 1,500 bp was obtained through bacterial 16S rDNA electrophoresis, and sequence analysis via GenBank indicated that the isolated bacteria were isogenous with B. subtilis IIIVE-4 (Figures 1C,D). Bacillus subtilis IIIVE-4 grew the fastest at 38°C, with growth declining at higher temperatures (Supplementary Figure 1A). After culturing in 0.5, 1.0, and 1.5% bile salts for 24 h, the survival rate of B. subtilis IIIVE-4 was more than 50% of the standard culture state (Supplementary Figure 1B). This suggests that B. subtilis IIIVE-4 could survive and grow in conditions similar to those of the small intestine-like intestinal bile concentration in healthy animals to 0.3% (31), pH range of the gastrointestinal tract in healthy animals (6.15 < pH <6.80) (32). The survival rate of B. subtilis IIIVE-4 increased rapidly when the culture pH was increased from 3.0 to 6.0 (Supplementary Figure 1C). When the pH values were 6.0 and 7.0, the bacterial survival rate leveled-off.

Bacillus subtilis IIIVE-4 had some tolerance to pH between 3.0 and 4.0, suggesting that it could survive in the animal gastrointestinal tract. Similarly, when the pH of the medium was >7.0, B. subtilis IIVE-4 tended to first increase and then decrease to a stable point, indicating resistance to a strong alkali environment. The survival rate was highest at pH 7.5.

Thiazolyl blue tetrazolium bromide (MTT) was used in this study to evaluate the proliferation and cellular activity of ruminal explants. The MTT curve showed that the viability of rumen tissue cells was relatively stable when the incubation time ranged from 4–12 h (Figure 2). By 16 h, there was a small decrease in cell viability. After 28 h, there was a substantial decrease in the cell viability. Cell viability was maintained at a relatively low level as incubation time continued to increase.

No structural differences were detected at any time with medium A (M.A.) or medium B (M. B.) during the first 12 h of culture (Figure 3). Starting from 16 h, regardless of M. A. or M. B. culture media, the tissue organizational structure began to disintegrate, and longer incubations led to further negative effects on tissue structure after 72 h. Thus, a 12 h incubation was deemed suitable for subsequent experiments.

As shown in Figure 4, for 0–12 h, the expression of CDH1, which represents the ability of epithelial cell differentiation in explant tissue, gradually increased and then leveled-off. After 12 h of culture, expression of CDH1 began to decrease along with the apparent viability of the tissue. Both results were consistent with the histological results. The overall trend of CDH1 mRNA abundance in tissue was the same in culture media A or B, with or without fetal bovine serum. However, at each time point, CDH1 was more highly expressed in medium A than in medium B. Thus, culture medium A was used in subsequent experiments.

As shown in Figure 5, different concentrations of living and inactivated bacteria had no significant effect on explant viability (MTT assay). As shown in Figure 6, the addition of living bacteria and inactivated bacteria had varying degrees of influence on the expression levels of PIK3CB, AKT, mTOR, P70S6K1, and 4EBP1 when the bacterial concentration was 2.0 × 10⁷ CFU/ml, while the expression of PDCD4 was lowest. Moreover, the expression levels of PIK3CB, AKT, mTOR, and P70S6K1 genes were more pronounced with the addition of living bacteria than with the addition of inactivated bacteria. As shown in Figure 7, the expression of target proteins at different concentrations of living (Figure 7A) and inactivated bacteria (Figure 7B) was consistent with gene expression. For instance, PIK3CB, p-AKT, p-mTOR, p-P70S6K1, and p-4EBP1 protein expression was the highest at a bacterial concentration of 10⁷ CFU/ml, and PDCD4 was the lowest, suggesting an opposite relationship between protein synthesis and apoptosis. The highest protein and gene expression were observed at a bacterial concentration of 10⁷ CFU/ml.

To further clarify the differences in the effects of B. subtilis and inactivated bacteria on protein synthesis, we compared the results of the effects of living and inactivated bacteria at a concentration of 2.0 × 10⁷ CFU/ml on key proteins in the PI3K–AKT–mTORC1 signaling pathway, and the results are shown in Figure 8. The results from Figure 8 showed that both living and heat-inactivated B. subtilis at 2.0 × 10⁷ CFU/ml were able to exert a more pronounced effect on key factors in the pathway, and there was a significant difference in the effect on the expression levels of key proteins between living and heat-inactivated bacteria (except for 4EBP1), while there was a significant difference in the expression levels of PI3K, AKT, and mTOR genes. The effects of key proteins and genes in the pathway were greater in the living bacteria group than in the inactivated bacteria group (p < 0.05). The results indicated that protein and gene expression in response to inactivated bacteria was no less than that of living bacteria. The effect of adding living bacteria on proteins and genes was significantly higher than that of the control.

The addition of supernatant from cultures receiving inactivated bacteria upregulated the expression of PIK3CB and AKT (Figure 9A). Only P70S6K and 4EBP1 were significantly upregulated, whereas PDCD4 was downregulated. With the exception of PDCD4, the expression of all other proteins evaluated was upregulated by the addition of supernatant from inactivated bacteria during the 12 h incubations (Figure 9B).

MTT data indicated that the addition of LY294002 or rapamycin did not affect viability or had toxic effects on rumen explants (Figure 10). A significant decrease in protein expression was detected after the addition of LY294002 and rapamycin. Similarly, the expression of PIK3CB, mTOR, AKT, and P70S6K was markedly inhibited (Figure 11).

As shown in Table 1, glutamic acid, glycine, phenylalanine, histidine, arginine, and dl-proline had the highest content in the control group and were significantly different from the other two groups. The highest levels of tyrosine and cysteine were detected in the supernatants of the living bacterial culture group. The aspartic acid, threonine, serine, d-alanine, valine, methionine, isoleucine, n-carbobenzoxy-dl-leucine, and lysine contents in the culture supernatant of inactivated bacteria were the highest and differed significantly from the other two groups.

Hierarchical clustering of the top 17 amino acids differentially affected metabolites in the bacterial culture supernatant (Figure 12A) indicated that the contents of glycine, methionine, n-carbobenzoxy-dl-leucine, isoleucine, valine, lysine, serine, threonine, alanine, and aspartic acid in the supernatant of inactivated bacteria were greater than those in the living bacteria and LB control group.

The lysine and serine contents in the living bacterial supernatant and LB control group were similar. In contrast, the content of cysteine in the supernatant of living bacteria was the highest among the three groups, followed by that in the supernatant of inactivated bacteria, and was lowest in the control group. Among the 17 free amino acids, the content of glycine, arginine, tyrosine, dl-proline, phenylalanine, histidine, and glutamic acid in the control group was the highest, with the content of arginine and tyrosine in the supernatant of living and inactivated bacteria being similar. The contents of glycine, dl-proline, phenylalanine, histidine, and glutamic acid were the lowest in the supernatant of inactivated bacteria.

Correlation analysis (Figure 12B) showed that methionine was positively correlated with n-carbobenzoxy-dl-leucine, isoleucine, valine, lysine, serine, threonine, d-alanine, and aspartic acid, and had the strongest positive correlation with n-carbobenzoxy-dl-leucine and the strongest negative correlation with histidine. n-Carbobenzoxy-dl-leucine was positively correlated with isoleucine, valine, lysine, serine, threonine, d-alanine, and aspartic acid, and had the strongest negative correlation with phenprobamate. Isoleucine was positively correlated with valine, lysine, serine, threonine, D-alanine, and aspartic acid, and had the strongest positive correlation with valine and the strongest negative correlation with tyrosine.

Enrichment analysis of KEGG pathways (Figure 12C) showed that aminoacyl tRNA biosynthesis, cyanoamino acid metabolism, monobactam biosynthesis, glycine, serine, and threonine metabolism, valine, n-carbobenzoxy-dl-leucine, and isoleucine biosynthesis were among the most represented within the list of affected metabolites. Among the free amino acids that were significantly affected, the aminoacyl tRNA biosynthesis pathway had the highest coverage. Cyanoamino acid metabolism; monobactam biosynthesis; and glycine, serine, and threonine metabolism were among the top pathways enriched with amino acids. Valine, n-carbobenzoxy-dl-leucine, and isoleucine biosynthesis were among the most represented within the list of affected metabolites.