All animal experimental procedures were approved by the Ethics Committee of Jilin Agricultural University (Changchun, Jilin, China).

Forty-eight 35-day-old “Yorkshire-Landrace-Duroc” weaned piglets (average 11.59 kg) were randomly divided into four diet groups (n = 4 each, i = 3 each). Every three piglets are kept in a cage. After a 7-day acclimation period, a 28-day animal experiment was started. The probiotics were cultured in MRS Medium at 37°C for 18h. The control group was fed the basal diet of the National Research Council (2012) nutritional requirements, and each piglet was fed 10 mL of MRS without bacteria per day. The experimental group was fed the basal diet with MRS containing probiotics, and the dose of probiotics per pig per day was included 10 mL containing L. plantarum JL01 (1 × 10⁹ CFU/mL Lac group), L. rhamnosus GG ATCC53103 (1 × 10⁹ CFU/mL, Rha group) or both probiotics (0.5 × 10⁹ CFU/mL L. plantarum JL01 and 0.5 × 10⁹ CFU/mL L. rhamnosus GG ATCC53103, mix group). Probiotics or MRS were mixed into a small amount of the basic diet and then fed to the piglets separately with a spoon to ensure that all weaned piglets would eat the whole dose. All piglets can get free water and feed. The composition of the diet is shown in Table 1.

On the 23rd day of the rearing period. The metabolic experiments were conducted for 5 days. Using the method of collecting feces and urine, we randomly selected three piglets from each group for the nitrogen metabolism test, fecal weight and urine volume were measured using a separate collection device for feces and urine in metabolic cages, and feces and urine were collected at 8:00 and 16:00 every day (Feyera et al., 2019). Urine samples were added with 10% sulfuric acid for nitrogen fixation and stored at −80°C for testing. We mixed the fecal samples collected every day, kept them in zip lock bags, recorded the weight, took out 200 g of fecal samples, quickly added 10% sulfuric acid to fix nitrogen, baked them at 60 °C to constant weight, then pulverized, and stored at −80 °C refrigerator for testing. On the morning of the 28th day, three weaned piglets were randomly selected from each group to collect intestinal samples, and a 3% sodium barbital solution was used as anesthesia for weaned piglets. The jejunum and ileum were dissected, and the contents of the jejunum and ileum were carefully squeezed out, and then, chyme was collected in sterile cryogenic vials. Afterward, 10 mL of hepatic portal vein blood was collected and placed in a blood collection tube containing heparin sodium. After centrifugation, plasma layers were extracted and distributed into 1.5 mL cryopreservation tubes. Afterward, the skeletal muscle (gluteus maximus) of piglets was quickly collected with a sterile EP tube and quickly frozen in liquid nitrogen. The samples were stored in a −80°C refrigerator.

The nitrogen content in feed and feces was measured using Nitrogen determination apparatus (FOSS Kjeltec 8420 Kjeldahl +nitrogen determination apparatus, FOSS [Technology and Trade Co. Ltd., Beijing, China]). The formulas used were as follows: nitrogen emission (g/d) = fecal nitrogen (FN) (g/d) + urine nitrogen (UN) (g/d), nitrogen apparent digestibility (%) = [(nitrogen intake (NI) - FN)/NI] × 100%, nitrogen deposition rate (%) = [(NI-FN-UN)/NI] × 100%, and nitrogen utilization rate (%) = [(NI-FN-UN) / (NI-FN)] × 100%.

Total RNA from the muscle tissue was extracted using TRIzol. The steps for RNA-reverse transcription and RT-qPCR were the same as before (Ding et al., 2020). PrimeScript™ RT reagent kit and SYBR^® Premix Ex T aq™ II were used for reverse transcription and RT-qPCR processes, respectively. RT-qPCR analysis was conducted using the Agilent Stratagene Mx3005P qPCR system (Agilent Technologies, Inc.). These RT-qPCR assays were run in triplicate for each sample, and the results were normalized to the expression of β-actin. Expression was analyzed using 2–ΔΔCt comparison method. Primer sequences are provided in Table 2.

Gut chyme microbial DNA was extracted using the TIANamp Stool DNA kit (Tiangen, Beijing, China).

The sequences of 16S rDNA genes were analyzed using the Illumina HiSeq™ 2500 sequencing platform. The raw data were filtered using the FASTP software. All samples were clustered, and the effective sequence similarity ≥ 97% was clustered into operation classification units (OTUs). The OTU sparsity curve and rank abundance curve were plotted in QIIME. An inter-group Venn analysis was performed in R (version 3.4.1) to identify unique and common otus. We use QIIME to calculate Chao1, Simpson, etc. α Diversity Index.

Portal vein plasma samples weighing 50 mg were carefully measured and transferred into an EP tube. Subsequently, the EP tube was filled with the solvent (acetylene-methane-water, 2: 2: 1, including internal standards). After rotating for 30s at an average speed of 45 Hz, the sample was subjected to a 4-min ice water bath. Then, it was ultrasonicated for 5min. The average ultrasound is repeated three times and then stored at −20 °C for 1 h and for 15 min for 12, 000 RPM and 4 °C. The obtained Shangqing was transferred to the LC-MC bottle, analyzed in UHPLC-QE or orBITRAP MS, and mixed with the equivalent of all samples to clear the quality control (QC) sample.

We used the T-test as a variable analysis to screen the different metabolites, and the VIP threshold was set to 1, and the differences between the two groups of the metabolites of a P-value of < 0.05 and a VIP of ≥ 1 were selected. The metabolites were enriched in the KEGG metabolic pathway for enrichment analysis. Compared with the entire background, the enrichment analysis determined the significant enrichment of metabolites or signal transition channels in the difference.

We used the Nextera XT DNA Library to prepare the library for the macro-base group and sequence the Illumina Hiseq platform. We also used Metaphlan (V.2.2) (Truong et al., 2015) and Humann2 to analyze macroscopic group data. Then, we use the Unigene to query the enzyme committee (EC) database through Diamond software. NCBI's RefSeq database was used to build a microbial index with significant differences in pigs annotated by Metaothello.

The results were expressed as the mean ± SD. Statistical analysis was performed using Prism software (GraphPad Software, San Diego, CA) to carry out one-way ANOVA followed by Bonferroni post-hoc analysis. Values of p < 0.05 were considered significant. Different microbial, genetic functions, metabolites, and metabolic channels were drawn through Adobe Illustrator software (USA).

The effects of L. plantarum JL01 and L. rhamnosus GG ATCC53103 on the nitrogen metabolism of weaned piglets are shown in Table 3. Compared to the Con group, the ingested nitrogen of the piglets in the Rha group, Lac group, and mix group was significantly increased (P < 0.05), fecal nitrogen, urinary nitrogen, and total nitrogen emissions showed a decreasing trend, the apparent nitrogen digestibility of Rha and Lac groups tended to increase; the apparent nitrogen digestibility of the mix group increased by 1.88% (P < 0.05), and the nitrogen deposition rate increased respectively increased by 4.17, 4.49, and 5.65% (P < 0.05), and the nitrogen utilization rate increased by 2.82, 3.23, and 3.70%, respectively (P < 0.05).

As shown in Figure 1, the relative expression of mTOR and S6K mRNA in the skeletal muscle of piglets in the Rha, Lac, and mix groups was significantly increased compared with the Con group (P < 0.05).

A total of 2, 800, 880 sequences were obtained in this experiment, and the original sequences were filtered, and finally, 2, 591, 938 valid sequences were obtained, with an average of 128, 160 valid sequences per sample for subsequent analysis. Microbial diversity is the diversity in a particular habitat or ecosystem and uses two parameters, species richness, and evenness, to reflect the diversity and abundance of microorganisms (Grice et al., 2009). Chao1 and ACE indices were used to estimate species richness information of samples, and Simpson and Shannon were used to estimate species diversity information (Segata et al., 2011). The effects of L. plantarum JL01 and L. rhamnosus GG ATCC53103 on the microbial diversity of the small intestine of weaned piglets are shown in Figure 2. Compared with the Con group, the Shannon and Simpson indices of the Rha, Lac, and mix groups showed an increasing trend in the jejunum but were insignificant. In the ileum, the ACE, Chao1, Shannon, and Simpson indexes of piglets in the mix group were significantly increased compared with the Con group (P < 0.05), while the uniformity and abundance of other intestinal contents were not significant.

The microbial composition of the gut contents of weaned piglets was analyzed. Figure 3A shows the 10 most abundant phyla among all phyla. Firmicutes, Proteobacteria, and Bacteroidetes are the common dominant phyla in each treatment group, and their relative abundances together account for the bacterial abundances in each treatment group. They accounted for more than 93% of the total bacteria. As shown in Figure 3B, compared with the Con group, in the jejunum, the Firmicutes in the mix group were significantly decreased (P < 0.05). The relative abundance of Bacteroidetes significantly increased (P < 0.05), and the relative abundance of Proteus significantly decreased (P < 0.05) in the Lac and mix groups. Firmicutes/Bacteroidetes were significantly increased in piglets in the Lac and mix groups (P < 0.05). In the ileum, the relative abundance of Firmicutes in the Rha and mix groups was significantly lower than those in the Con group (P < 0.05). The relative abundance of Bacteroidetes increased significantly in Rha, Lac, and mix groups (P < 0.05). Firmicutes/Bacteroidetes were significantly increased in Rha, Lac, and mix groups (P < 0.05).

At the genus level, 181 genera were identified from all samples. The relative abundances of the top 25 genera are shown in Figure 3C. As shown in Figure 2D, compared with the control group, the relative abundance of Lactobacillus, Actinobacillus, and Prevotella_9 in the mix group increased significantly (P < 0.05), and the relative abundance of Ruminococcaceae_UCG-005 decreased significantly (P < 0.05). The relative abundances of Ruminococcaceae_UCG-005, Actinobacillus, Streptococcus, Blautia, and Prevotella_9 increased significantly (P < 0.05), while the relative abundances of Clostridium_sensu_stricto_1 and Terrisporobacter decreased significantly (P < 0.05). Rha group in the jejunum, including Clostridium_sensu_stricto_1, Terrisporobacter, and Actinobacillus, were found to be significantly different. The relative abundance of Lactobacillus, Ruminococcaceae_UCG-005, and Streptococcus decreased significantly (P < 0.05). Conversely, in the ileum, the relative abundance of the Rha group, Prevotellaceae_NK3B31_group, Lactobacillus, Ruminococcaceae_UCG-005, and Akkermansia, increased significantly (P < 0.05). Additionally, the relative abundance of Clostridium_sensu_stricto_1 and Terrisporobacter decreased significantly (P < 0.05). The relative abundance of the Lac group, including Prevotellaceae_NK3B31_group, Lactobacillus, Ruminococcaceae_UCG-005, Prevotella_9, showed a significant increase (P < 0.05). Similarly, in the ileum, the relative abundance of the Lac group, comprising Prevotellaceae_NK3B31_group, Lactobacillus, Ruminococcaceae_UCG-005, and Streptococcus, also increased significantly (P < 0.05). In contrast, the relative abundance of Clostridium_sensu_stricto_1 and Terrisporobacter exhibited a significant decrease (P < 0.05).

This experiment only focuses on the differentiated microorganisms in the small intestine and the functional genes encoding amino acid metabolism and protein synthesis. As shown in Figure 4, Clostridium_ sensu_ stricto_ 1 and ruminocaceae_ Ucg-005 have a gene encoding d-3-phosphoglycerate dehydrogenase; Actinobacillus, Clostridium, Lactobacillus, and Streptococcus have genes encoding carbamate kinase and Carbamoyl phosphate synthase; Akkermansia and Ruminococcus have genes encoding Carbamoyl phosphate synthase; Actinomycetes, Clostridium, Prevotellaceae_ NK3B31_ Group, Prevotella_ 9, Rumencocci and Streptococcus have genes encoding aconitate hydratase; Actinomycetes, Klebsiella, Streptococcus, Rumencocci, Prevotellaceae_ NK3B31_ Group, Prevotella_ 9 and Clostridium has a gene encoding tryptophan synthase.

In this test, no significant peaks were detected in the blank sample, indicating that there was no cross-contamination between samples in this test. Substance identification was conducted using a self-written R program package and a self-built secondary mass spectrometry database. The second-order mass spectrometry matching degree cutoff = 0.2. A total of 11, 875 metabolites were obtained from the detection results of this study, and 617 substances were identified by the secondary spectrum.

As shown in Figure 5A, each treatment group had apparent separation compared with the control group, indicating that feeding L. rhamnosus and L. plantarum had a significant effect on the plasma metabolites of the hepatic portal vein of weaned piglets. In this study, the differential metabolites related to nitrogen metabolism were selected for analysis. As shown in Table 4, in the portal vein metabolites, compared with Con, the differential metabolites produced by the Rha group were 3-Phosphonooxypyruvate (P < 0.05), the differential metabolites produced by the Lac group were L-Tryptophan (P < 0.05), and the differential metabolites produced by the mix group were L-Tryptophan, 3-Phosphonooxypyruvate, cis-Aconitate and Carbamoyl phosphate (P < 0.05). As shown in Figures 4A–D, after the metabolites were enriched, the first 20 paths with the lowest enrichment q value were selected to draw the map. This experiment only focused on the metabolic pathway of nitrogen nutrients. The results showed that the differential metabolite 3-phosphodeoxypyruvate enriched in the portal vein of weaned piglets produced by feeding the mixed bacteria of L. rhamnosus and L. plantarum was glycine, serine, and threonine metabolism pathway; the metabolic pathway enriched by the differential metabolite L-Tryptophan is the tryptophan metabolic pathway; the metabolic pathway enriched by the differential metabolite cis-Aconitate is the Citrate cycle (TCA cycle) pathway; The metabolic pathway for the differential metabolite Carbamoyl phosphate is the Arginine biosynthesis pathway.

To further study the relationship between intestinal microflora and metabolites, we visualized the correlation between jejunum and ileum microflora and metabolites and carried out Spearman correlation analysis on the relative abundance of different microbial species and metabolites. As shown in Figure 6A, in the jejunum, L-Tryptophan and cis-Aconitate were positively correlated with Streptococcus, Lactobacillus, and Actinobacillus and negatively correlated with Akkermansia. Carbamoyl phosphate was positively correlated with Akkermansia, Lactobacillus, and Actinobacillus and negatively correlated with Streptococcus. 3-Phosphonoxypyruvate was positively correlated with Lactobacillus and Actinobacillus and negatively correlated with Akkermansia and Streptococcus. As shown in Figure 6B, in the ileum, we found 3-Phosphonooxypyruvate, Carbamoyl phosphate, cis-Aconitate, L-Tryptophan and Actinobacillus, Blautia, Lactobacillus, Streptococcus, Ruminococcaceae_UCG-005, Prevotella_9, Prevotellaceae_NK3B31_group, and Akkermansia were positively correlated and negatively correlated with Clostridium_sensu_stricto_1.