A comprehensive collection of lactic acid bacteria from sow milk
Breast milk is one of the common sources of probiotic strains[15]. We hypothesized that these lactic acid bacteria (LAB) from sow milk could facilitate adaptive, functional changes for the optimal weaning transition of piglets, which are vulnerable to various stressors. However, the diversity and function of probiotics in porcine milk remain relatively understudied[16]. To establish a sow milk bacterial collection (smBC), the large-scale cultivation and identification of the sow milk microbiota (SMM) were performed by three procedures (Fig. 1, steps 1–3).
After the first three steps, we obtained 1240 isolates derived from the sow milk microbiota by the culturomics of continuous culture and interval sampling. These isolates were grouped into 271 bacterial taxa in the CD-HIT analysis based on a nonredundant set of sequences of the V1-V5 region of the 16S rDNA gene with a cutoff value of 99% identity for classification. The Silva version 132 16S rRNA database, NCBI nucleotide collection (nr/nt) database and DAIRYdb reference database were used to classify all isolated prokaryotes into two categories: suspected new species and previously identified prokaryote species (including known species from dairy products). A phylogenetic tree was built based on the calculated distances between pairs of sequences (Table S1,2). The results revealed that 151 taxa were assigned to previously described species (black, Fig. S1), while the other 120 taxa could not be assigned to any known species (blue, Fig. S1), which were defined as suspected new species. The alignment against the DAIRYdb reference database revealed that 107 out of the 271 taxa were assigned to species in dairy products (red dots, Fig. S1).
Specifically, 23 species were assigned to previously described species found in both the DAIRYdb reference database and Silva version 132 16S rRNA sequences database or the NCBI nucleotide collection (nr/nt) database, including A. lwoffii, Acinetobacter sp., C. perfringens, Pelomonas, Enterococcus sp., E. durans, L. amylovorus, L. taiwanensis, L. garvieae, L. mesenteroides, Enterobacteriaceae bacterium DHL-32, Lactobacillales bacterium, Lactobacillaceae bacterium, Streptococcaceae bacterium, S. enterica and Salmonella sp. Nine species were suspected new species and could not be assigned to any known species in the DAIRYdb reference database. In addition, five species were assigned to a previously described species in the Silva version 132 16S rRNA database or NCBI nucleotide collection (nr/nt) database, but not in the DAIRYdb database, including Acidovorax sp. SEPRH9, S. hyovaginalis, S. mitis, Streptococcus sp. S2 and S. thoraltensis (Fig. 2 and Fig. S1).
A total of 922 out of 1240 isolates belong to Lactobacillales (Suppl. Data 2), and this group was dominated by Lactococcus lactis, which was roughly consistent with the microbiota composition of sow milk during lactation according to a recent report [16]. The genera Staphylococci and Streptococci represented 5.81% and 4.03% of the total bacterial isolates, respectively. These results supported the view that commensal Staphylococci and Streptococci commonly occur in breast milk[17, 18], which may originate from the maternal skin[19]. The collection of 1240 isolates obtained through these efforts provides insight into the diversity of sow milk microbiota and enables more studies on their function associated with mammalian health and diseases.
Screening of candidate probiotic P. pentosaceus strains
To explore health-promoting bacterial genera, which are not limited to traditional probiotic of Lactobacillus and Bifidobacterium, we focused on Pediococcus. Antagonistic activity against pathogens to control their spread is a prerequisite for a potential probiotic[20]. Piglets commonly encounter pathogens on farms at increasing frequencies[21], including Salmonella typhimurium (S. typhimurium)[22], enterohemorrhagic Escherichia coli (EHEC)[23], enterotoxigenic Escherichia coli (ETEC)[24], Klebsiella pneumoniae (K. pneumoniae)[25], Aeromonas punctate (A. punctate)[26], Staphylococcus aureus (S. aureus)[27], Listeria monocytogenes (L. monocytogenes)[2] and Clostridium perfringens (C. perfringens) [28]. The fluctuating size of the inhibition zone against pathogens indicates the inhibitory activities of different P. pentosaceus against both gram-negative and gram-positive pathogenic bacteria (Fig. 3a), revealing strain-specific antimicrobial activity against different bacteria[29]. The morphology of the top ten strains was observed (Fig. S2), and they were then used in subsequent tests.
For these P. pentosaceus strains with strong antimicrobial activity, we next used a simple animal model to rapidly screen bacteria that show potent antioxidant activity in vivo (Fig. 3b). Considering the similarities of the intestinal development with mammals and the cost of the mouse model[30, 31], Drosophila could be an appropriate model to evaluate the ability of our bacteria in protection of the host from reactive oxygen species (ROS), whose accumulation causes damage to the health of both Drosophila and mammals[30, 32]. Resistance to paraquat can be used as a measure of free radical scavenging activity in the Drosophila system[32, 33]. After paraquat treatment for 45 hours, the flies colonized with P. pentosaceus SMM914 showed a significantly elevated survival rate in response to paraquat challenge (p < 0.05, log-rank test) (Fig. 3c).
SMM914 showed a strong ability to inhibit the proliferation of S. aureus, while almost one-third of P. pentosaceus strains exhibited no antimicrobial ability toward this species. Through the sodium hydroxide neutralization reaction, the antibacterial effect of SMM914 against S. aureus and ETEC was abolished, which proved that its bacteriostatic effect is mainly due to the presence of organic acids in either anaerobic or aerobic conditions (Fig. S3). We measured the time curves of growth and pH for SMM914. The results revealed that the strain entered the stationary phase after 12 h of fermentation, while the pH was stabilized at approximately 3.9 after 36 h of fermentation under either anaerobic or aerobic conditions (Fig. S4 a, b). Additionally, after sequencing verification of the 16S rDNA, L-lactate dehydrogenase 1 and L-lactate dehydrogenase 2 genes, it was revealed that SMM914 is closest to P. pentosaceus SRCM100194 (Table S1).
Growth performance and serum biochemical parameters in the pig-feeding trial
Because SMM914 showed bacteriostatic activity and the strongest resistance to paraquat-induced stress in Drosophila, it was selected to treat the piglets in the low-dose (LD) group or high-dose (HD) groups prior to early weaning (Fig. 4a). There were no significant differences among the three groups regarding growth performance (Table S2). However, the HD group showed a tendency toward an increased length of the small intestine compared to the control (p = 0.07, n = 7). The consumption of SMM914 in the HD group increased the final body weight by 7.23% (p > 0.05, n = 18). Regarding visceral indices, a higher heart coefficient was observed in piglets treated with SMM914 (p < 0.05, n = 7) (Table S2). The heart coefficient has been reported to be negatively associated with oxidative stress via changes in angiotensin II-aldosterone-brain natriuretic peptide[34]. We speculated that the increased heart coefficient observed in this study could be an indicator of alleviated stress.
Blood biochemical parameters demonstrated that treatment with SMM914 reduced ALT, LDH, TP and ALB levels, which are closely related to protein metabolism and reflect liver damage (Table S3). Other serum biochemical traits were not influenced by oral gavage. Weaning is frequently associated with liver injury and alters blood chemistry related to liver function[35]. The elevation of ALT and LDH activities in serum can be used as a biomarker of hepatic disorders and the loss of functional integrity under oxidative stress[36–38]. The shift observed in our study was in accordance with another report that the administration of P. pentosaceus LI05 significantly prevented increases in TP and ALT in the context of acute liver failure[39]. Thus, compared with the control in our study, liver injuries in the treated groups were ameliorated.
Pretreatment of piglets with SMM914 induces the Nrf2/Keap1 antioxidant signaling pathway
The liver is a target organ of stress in vertebrates and is commonly accepted to be involved in the secretion of bile salts, the phagocytosis of residual materials and the metabolism of proteins as well as detoxification[40]. When constantly challenged by various endogenous or exogenous free radicals, the liver is susceptible to damages [22, 41]. To provide a theoretical basis for the clinical use of SMM914, this study investigated the alteration of the Nrf2/Keap1 signaling pathway in the liver by western blotting analysis and enzyme activity assays. The Nrf2/Keap1 signaling pathway is conserved across metazoans[42, 43]. Keap1 is a specific repressor of Nrf2 via tight binding. Antioxidant metabolites can cause the dissociation of Keap1 and Nrf2 complex, promoting Nrf2 movement into the nucleus[44]. Nrf2 transfers from the cytosol to the nucleus, resulting in the coordinated transcriptional upregulation of a battery of antioxidant enzymes and detoxifying proteins[45]. CAT and SOD1 are widely recognized as important endogenous antioxidant enzymes that scavenge hydroxyl and superoxide anion radicals[46, 47].
As expected, in the western blotting analysis of this study, the protein level of Keap1 was remarkably suppressed in piglets receiving SMM914 at both high and low doses. We found that SMM914 not only markedly increased the intranuclear protein expression level of Nrf2 but also led to elevated protein levels of NQO-1, HO-1, CAT and SOD1 in a concentration-dependent manner (Fig. 4b-e).
In the enzyme activity assays, the HD group simultaneously increased glutathione peroxidase (GSH-Px) activity, CAT activity and SOD activity (p < 0.05) in the livers of the high-dose group (Fig. 4f). Additionally, the HD group showed a significant decrease in MDA, a lipid peroxidation end product, in the liver compared with the control group (Fig. 4f). The western blotting data combined with enzyme activity tests suggested that SMM914 functions as a promising probiotic conferring antioxidant capacity by activating the Nrf2/Keap1 antioxidant signaling pathway in weaning piglets.
The altered pathways of amino acid metabolism and lipid metabolism in plasma
To provide a better understanding of the antioxidant mechanism of SMM914, we further examined the metabolic profiles of blood plasma from the three groups (n = 7). PLS-DA plots showed separated clusters with an optimal goodness of fit (R2 = 0.996, Q2 = 0.681 (Fig. 5a); R2 = 0.994, Q2 = 0.479 (Fig. 5b), indicating that the models were suitable and reliable for prediction. The KEGG enrichment of differential metabolites between the HD group and the control group revealed that the pathways of amino acid metabolism and lipid metabolism were the main perturbed metabolic pathways (Fig. 5c).
In regard to amino acid metabolism, several critical antioxidant metabolites (cysteine-S-sulfate, DL-methionine sulfoxide, L-methionine) (Fig. 5d) closely related to cysteine and methionine metabolism were significantly increased by 1.41-2.03-fold in the HD group compared with the control group (p < 0.05). D-proline, L-proline and L-glutamate, (Fig. 5e) which are involved in arginine and proline metabolism, were increased in the LD group compared with the control group by 1.43-fold (p < 0.05), 1.30-fold (p < 0.05) and 1.34-fold (p = 0.056), respectively. In the glycine, serine and threonine pathway, choline was also significantly increased in the HD group (Fig. 5f).
However, the intensities of cholic acid, taurochenodeoxycholate and glycochenodeoxycholate (Fig. 5g), which are involved in the biosynthesis pathway of primary bile acid, were decreased in the LD group to 0.43-fold (p < 0.05), 0.58-fold (p = 0.08) and 0.47-fold (p < 0.05), respectively, compared to the levels in the control. Decreased levels of corticosterone and cortisol (Fig. 5h) were also observed in plasma, which are related to steroid hormone biosynthesis.
The possible protective effect of SMM914 on the weaned piglets is depicted in Fig. 5c. L-methionine is a limiting amino acid in early lactation[48] associated with various key physiologic events, and the increased availability of L-methionine in early lactation could have positive effects on plasma lipid metabolism and overall antioxidant status[49]. Methionine sulphoxide is also biologically available as a methionine source. High methionine bioavailability is likely to increase the entry of L-methionine into the one-carbon metabolism cycle, where S-Adenosyl-L-methionine is then used to generate S-Methyl-5’-thioadenosine and 1-Aminocycloproane-1-carboxylic acid. Through the transsulfuration and transmethylation pathway for the synthesis of the amino acid L-cysteine, L-methionine also serves as a substrate for glutathione, an endogenous sulfur-containing antioxidant[50, 51]. Glutathione is required for regulating the cell redox state and detoxification in all cell types through a direct reaction with free radicals.
Cysteine and methionine metabolism is overlapped with choline metabolism tightly because choline can serve as the substrate for L-methionine synthesis. Choline is an essential vitamin for humans and other mammals to regulate amino acid metabolism[52], particularly when L-methionine levels are not sufficient around parturition[53]. It has been established that choline deficiency induces the generation of ROS[54] and oxidative damage in rats[55], ruminants[56] and fishes[57]. Moreover, dietary supplementation with choline enhances the antioxidative capacity in IUGR pigs[58]. New evidence has shown that choline deficiency-induced oxidative damage is associated with changes in the transcription of antioxidant enzymes and Nrf2 signaling in the liver and intestine[57]. Furthermore, in mammals, L-glutamate is an abundant amino acid in milk that is required for the synthesis of glutathione and alleviates oxidative stress by increasing antioxidant enzyme activities[59, 60]. Glutathione is decomposed into L-gamma-glutamyl amino acid, and L-gamma-glutamyl amino acid is further converted to pyroglutamic acid. A high level of pyroglutamic acid also contributes to glutathione deficiency and could be an indicator of the oxidative state[61]. Similarly, in our study, the concentration of pyroglutamic acid was significantly downregulated. Collectively, the alteration of these metabolites’ intensities is conducive to the accumulation of glutathione, which is consistent with our previous enzyme activity assays showing that a high dose of SMM914 markedly increased GSH-Px activity (p < 0.05) in the liver. (Fig. 5f)
On the other hand, several metabolites in lipid metabolism are involved in oxidative injury[62, 63]. Cholic acid increases both the hepatic and systemic expression of oxidative stress[63]. Simultaneously, metabolic syndrome, which is both a cause and effect of oxidative stress, has been reported to be associated with elevated deoxycholic acid levels[64]. Deoxycholic acid can combine with taurine or glycine to form taurochenodeoxycholate or glycochenodeoxycholate, respectively, which are considered as hydrophobic bile acids and induce the phosphorylation of NADPH oxidase and the formation of ROS[62]. In addition, excessive stress can cause the development of neurological disorders[60]. Classic stress hormones including corticosterone and cortisol were also found to be decreased. Interestingly, these metabolites in lipid metabolism are closely related to serum ALT, which reflects damage in the liver[65]. Under psychological and emotional stress conditions, corticosterone and cortisol induce the oxidative load in the brain, with a significant increase in pro-oxidant markers in constantly changing environments[66]. In human infants, after maternal separation at weaning, separation anxiety is an inevitable phenomenon that may raise cortisol levels and even alter the gut microbiota composition through the gut-brain axis[67, 68].
The reshaped colon microbiota in piglets by SMM914
The changeover from milk to solid feed strongly influences the development of the gut microbiota[69, 70]. Disorders in the composition of the microbiota can induce oxidative stress and chronic metabolic diseases through the liver-gut axis[71, 72]. In this study, the colonic microbiota was further investigated by using 16S rDNA gene amplicon sequencing. All samples from weaned piglets approached the saturation plateau based on Shannon-Wiener rarefaction curves (Fig. S5), suggesting that the sampling was sufficient for nearly all bacterial species. The shared and specific OTUs are shown in a Venn diagram (Fig. 6a). The bacterial community of the three groups shared 687 OTUs. There were 12 unique OTUs in the LD group and 35 in the HD group. Twenty-five OTUs were detected in the LD and HD groups but not in the control group. No differences were observed between the control and HD groups in terms of α-diversity (Fig. S6).
To intuitively measure the extent of the similarity of the overall microbiota, the results of PCA based on distance revealed a separate clustering of samples between the HD group and the control group, but the colonic specimens of the LD group were not separated from those of the control group. These results indicated that the high dose of SMM914 reshaped the microbiota structure of the colon but not remodeled the LD group (Fig. 6b).
At the family level, the relative abundance of Lactobacillaceae, Lachnospiraceae, Christensenellaceae and Ruminococcaceae in the LD group and HD group were increased by 43.54%, 11.14%, 44.71%, and 32.37% and by 219.79%, 17.17%, 418.68%, and 57.91%, respectively, compared with the control group (Fig. S7). Specifically, the results revealed that the genus Lactobacillus was significantly increased in the HD group compared with the control group (p < 0.05). SMM914 also promoted the growth of the genus Lachnospiraceae AC2044 (p < 0.001) and the genus Lachnospiraceae_uncultured (p < 0.05) in the HD group (Fig. 6c).
Previous compelling investigations have demonstrated that during the suckling period, Lactobacillus plays a protective role against oxidative damage by upregulating the expression of glutathione reductase and glutathione S-transferase[73, 74]. The Lachnospiraceae family participates in the breakdown of carbohydrates and potentially contributes to antioxidative properties[75]. For example, methionine attenuates oxidative stress in rats, which was achieved through higher abundances of Lactobacillus and Lachnospiraceae[76]. Furthermore, dietary provision of sodium butyrate in broilers is reported to significantly depress the MDA concentration in the jejunal mucosa, which is associated with the microbial community, including a striking increase in Lachnospiraceae[77]. The Ruminococcaceae family is always negatively related to disease severity[78]. The Christensenellaceae_R_7 group plays a positive role in intestinal immunomodulation[79]. In the present work, the genus Christensenellaceae_R_7 (p < 0.01), the genus Ruminococcaceae UCG-005 (p < 0.01) and the genus Ruminococcaceae UCG-014 (p < 0.05) showed enrichment in the HD group (Fig. 6c, d).
Conversely, at the family level, the relative abundances of Bacteroidaceae and Prevotellaceae in the HD group were decreased by 81.57% and 65.56%, respectively (Fig. S7). Specifically, the genus Bacteroides was observed to decrease in the HD group compared with the control group (p < 0.05). SMM914 also inhibited the relative abundance of Prevotella (p < 0.05) and Prevotella 2 (p < 0.05) at the genus level. These decreased bacteria have been reported to be associated with oxidative stress. For example, chitosan oligosaccharides increase the antioxidant capacity by inhibiting the abundance of harmful bacteria, including Bacteroides and Prevotella[74]. Improving the cellular antioxidant potential is a promising approach for inflammatory bowel disease (IBD) prevention. The dysbiotic microbiota of IBD is mostly characterized by an increase in Prevotellaceae and a decrease in Ruminococcaceae and Lachnospiraceae[80]. Hence, it appeared that SMM914 administration selectively promoted the transition of a microbial community to adapt to oxidative stress by assisting those more favorable genera but simultaneously inhibiting undesirable ones.