Background

Bacterial infections frequently spread from host to host by the tainted food or water contact. They rapidly spread among the host causing various diseases [1]. The use of antibiotics is now a common treatment for pathogen infections, but long-term use of these drugs can result in bacterial resistance as well as other health risks. In recent years, probiotics have gained widespread interest due to their excellent performance in disease prevention and alleviation. Probiotics exert their benefits through four main mechanisms: improvement of the barrier function, immunomodulation, metabolic function, and inhibition of potential pathogens [2, 3]. In addition to these, the presence of probiotics can also induce strong ecological and evolutionary forces to reshape the gut native microbial communities [4, 5]. Moreover, these changes in microbiome facilitate the treatment of various diseases and the maintenance of host health [6, 7]. To better utilize probiotics, the potential mechanisms of probiotics in regulating intestinal microbiota are therefore needed for further investigation.

Gut microbiota is a complex system that plays an important role in regulating physiology, immune, and nutrition of host [8,9,10]. Recent studies have demonstrated that gut microbes form an interactive ecological network to maintain host health, while the lack of a stable microbiota structure contributes to host disease [11, 12]. For example, antibiotic treatment increases the risk of gastrointestinal infections in humans, such as infection with Clostridium difficile [13]. Germ-free animals have considerable physiologic and immunologic variations from their counterparts raised conventionally, suggesting that microbiome plays a significant role in physiology and immunology of host. Moreover, changes in humans gut microbes have been strongly associated with diabetes, nonalcoholic fatty liver disease, obesity, and cancer [14,15,16,17]. Therefore, a thorough understanding of the factors that lead to microbiome variance in hosts is necessary to comprehend how microbiota affect host physiology and how to regulate microbiota to promote host health [18]. Diet is one of the most important external factors that might influence the composition of the gut microbiota [19]. A high-protein [20], high-fat diet [21], probiotics [22], and antibiotics [23] can all cause alterations in some bacterial groups. Recent studies have reported that antibiotic-treated and normally reared zebrafish exhibited different susceptibilities to infection by pathogenic bacteria, which were also observed in zebrafish with different gut microbiota compositions [24], suggesting that zebrafish is an animal model that can be used to study the correlation between gut microbes and host health [18, 25, 26].

Most of the gut microbiota’s contribution to host physiology is related to microbial metabolism, with bacteria contributing the most to ecosystem function in terms of relative genetic content [27, 28]. In general, microbes metabolize exogenous and endogenous substrates into nutrients for direct use by the host, and metabolites can also regulate the immune system by affecting the physiology and gene expression of host cells [29]. Additionally, the presence of different metabolic activity can help the microbiota occupy ecological niches and limit pathogen colonization at various sites through competition [30, 31]. Metabolites from specific classes of microbiota, such as short-chain fatty acids [32], bile acids [33], tryptophan [34], and B vitamins [35] not only affect host health, but also have an impact on interactions between host gut commensal bacteria. For example, large portion of vitamins produced by gut microbiota may be taken up by non-vitamin-producing gut microbes participating in symbiotic relationships, which suggested that vitamins may has the ability to reshape microbial communities [36, 37]. Vitamin B12 is a necessary water-soluble vitamin that is needed for maintaining neuronal health and hematopoiesis [38]. Some studies have shown that vitamin B12 deficiency can cause megaloblastic anemia and neurological disorders [39] and that supplementation is beneficial in several inflammatory diseases including sepsis, arthritis, Alzheimer’s disease, multiple sclerosis, and chronic fatigue syndrome [40]. Recent research suggested that vitamin B12 was crucial for maintaining human health in other areas, such as the structure and function of the gut microbiome [41]. Despite growing evidence that specific commensal microbes in the host gut can produce vitamin B12, the effect of microbially produced B12 on host resistance to bacterial infection is unclear.

Probiotics, including Bacillus and lactic acid bacteria, were used as a promising approach for decreasing infections [42]. Bacillus velezensis is an important biological control agent that is widely used in animal disease control [43]. Zebrafish, as an omnivorous freshwater fish belonging to the Cyprinidae family, shares homology with the human genome [44]. The zebrafish model is widely used in the researches of resistance against bacterial infection, gut microbiome, and disease prevention and cure [45,46,47]. Here, Bacillus velezensis and zebrafish were used as probiotic and animal models, respectively, to describe the effect of probiotics modulating the intestinal commensal bacteria of zebrafish on host resistance to pathogen infections. Specifically, we (I) investigated microbial changes in the gut caused by probiotics and analyze the key indigenous bacteria using 16S rRNA gene profiling sequencing, (II) investigated functional characteristics and key metabolites of this indigenous bacterium, using whole-genome sequencing combined with HPLC, and (III) investigated the mechanism of vitamin B12 in enhancing host resistance to pathogen infections.

Results

Protective effect of oral supplementation with Bacillus velezensis against Aeromonas hydrophila infection depends on gut microbiota

A bacterial strain 1704-Y was isolated from the zebrafish gut and identified as Bacillus velezensis based on morphological observation and phylogenetic analysis of gyrB gene (GenBank accession no. OM176702) (Fig. S1a-c) [48], and showed to be potentially safe for fish use by the hemolysis and antibiotic susceptibility tests (Fig. S1d, f; Table S2). In order to study the effects of strain 1704-Y on the prevention of pathogenic infection in fish, zebrafish were fed a basic diet supplemented with or without Bacillus velezensis 1704-Y for 28 days (Group Y/CK), and then bath infected with Aeromonas hydrophila (Group TY/TCK) (Fig. 1a). The results showed that dietary administration of BV1704-Y for 28 days significantly improved the survival rate of zebrafish bath infected with A. hydrophila (Group TY/TCK, P < 0.05, Fig. 1b). Meanwhile, a common cocktail of antibiotics was used to deplete the intestinal bacteria according to previous studies [49, 50]. Zebrafish were given mixed antibiotics (120 mg/kg metronidazole, 120 mg/kg neomycin sulfate, and 60 mg/kg vancomycin) for 1 week before feeding trial (Fig. 1a). However, no protective effect was observed when the fish were fed a diet containing an antibiotic mix for 7 days before oral administration of BV1704-Y (Fig. 1b). Moreover, dietary supplementation of BV1704-Y significantly decreased Aeromonas load in the fish gut, liver, and kidney after the challenge (Group TY/TCK, P < 0.05), whereas after administration with the antibiotic mix, the decline in Aeromonas load was not found in the gut and liver (Fig. 1c–e). These results suggest that a diet with the antibiotic mix removed the protective effect of oral supplementation of BV1704-Y, thereby indicating gut microbiota played a key role in the protection of fish from A. hydrophila infection.

Fig. 1
figure 1

Microbiota is essential to protect fish against A. hydrophila infection after B. velezensis 1704-Y supplementation. a The experimental design. (i) The zebrafish were fed a basic diet supplemented with/without B. velezensis 1704-Y (1 × 107 CFU/g diet) for 28 days (Group Y/CK), respectively, and then bath infected with A. hydrophila AH2006-3 J at a concentration of 1 × 108 CFU/mL (Group TY/TCK); (ii) The zebrafish were fed a diet containing an antibiotic mix (120 mg/kg metronidazole, 120 mg/kg neomycin sulfate and 60 mg/kg vancomycin) for 7 days (Group AY/ACK), and then received the same treatments as (i) (Group TAY/TACK). b Kaplan–Meier graph of the zebrafish survival after bath infection with A. hydrophila. * indicates significant difference (P < 0.05) between different groups. c–e Aeromonas load (Aero gene copies/g of fish tissues) in fish tissues (gut, liver and kidney) sampled prior to bath infection or at 10 days post-infection. Significant differences (P < 0.05) between different groups are indicated with different lowercase letters above the bars

Full size image

Resistance to A. hydrophila infection is conferred by one member of fish gut microbiota, Cetobacterium somera

To determine how gut microbiota affected the protective effect induced by BV1704-Y, first, a 16S rRNA gene amplicon sequencing method was used to compare gut microbiota composition before and after oral administration. The results of α-diversity showed that there was no obvious difference in bacterial richness and diversity between the fish fed with and without BV1704-Y. Interestingly, A. hydrophila infection significantly increased the diversity of the microbiota in the fish orally administrated with BV1704-Y (Group Y/TY, P < 0.05), as supported by Shannon index and number of Observed_OTUs, but no significant difference was detected in these metrics (including Simpson index) of the control fish between pre- and post-infection (Fig. 2a, Fig. S2a, b). The principal coordinate analysis (PCoA) based on Bray–Curtis distances revealed that bacterial communities in the fish fed with BV1704-Y were markedly distinct from the control after bath infection with A. hydrophila (Group TY/TCK, ANOSIM R = 0.647, P = 0.001) (Fig. 2b). In addition, the composition of gut microbiota at phylum and genus levels also showed the differences in gut microbiome among different groups (Fig. 2c, Fig. S2c). The relative abundance of the most abundance phyla, Fusobacteria, was significantly increased in the fish supplemented with BV1704-Y compared to the control fish (Y vs CK, P = 0.0221) (Fig. S2d), while the abundance was also higher in TY group than that in TCK group (P = 0.0293) (Fig. S2e). A set of 9 genera were present in all samples and considered as core gut microbiome [51]. The top 5 most abundant core genera (the cumulative relative abundance was 90%) were distributed in that four dominant phyla. Among them, one genus (Cetobacterium) belonged to Fusobacteria, three genera (Aeromonas, Enterobacteriaceae_unclassified, Shewanella) belonged to Proteobacteria, and one genus (Flavobacterium) belonged to Bacteroidetes. Meanwhile, Cetobacterium was the genus with the highest relative abundance (67–92%) among the core microbes. Subsequently, linear discriminant analysis effect size (LEfSe) showed that two genera including Cetobacterium (P = 0.0221) and Microbacteriaceae_unclassified (P = 0.013) were enriched in Y group while another two genera including Vibrio (P = 0.0183) and Enterobacteriaceae_unclassified (P = 0.0046) were depleted compared with CK group (Fig. 2d,e). Moreover, LEfSe identified seven different bacterial genera enriched in TY group, while five other bacterial genera were depleted in TY group compared with the TCK group (Fig. S2f-r). Of note, only Cetobacterium was the bacterium that differed both between CK and Y groups and between TCK and TY groups. Furthermore, Pearson correlation between the top 20 most abundant genera and infection status or diet with BV1704-Y showed that bacterial taxa enriched in the Y group including Cetobacterium had a positive correlation with the diet containing BV1704-Y, and also had a negative correlation with infection status (Fig. 2f). Moreover, the results of the antibiotic treatment test also showed that feeding BV1704-Y did not protect zebrafish against A. hydrophila infection when the abundance of Cetobacterium in the gut was decreased (Fig. 1b, Fig. S2s). Interestingly, numerous previous studies also found that dietary supplemented with probiotics increased the abundance of Cetobacterium in the fish gut and protect the host from pathogenic infections [24, 52,53,54,55].

Fig. 2
figure 2

Resistance to A. hydrophila infection is conferred by Cetobacterium somerae. a Shannon index comparison among the different groups. The zebrafish in Group Y and CK were fed a basic diet supplemented with/without Bacillus velezensis 1704-Y (BV1704-Y), respectively, and then bath infected with A. hydrophila 2006-3 J (AH2006-3 J) at a concentration of 1 × 108 CFU/mL (Group TY/TCK). b A principal coordinate analysis (PCoA) based on Bray–Curtis distance from the different groups (Y, CK, TY, and TCK) (ANOSIM R = 0.647, P = 0.001). c Relative abundance of the top 20 genera in the fish gut from the different groups. d Discriminative biomarkers identified by linear discriminant analysis effect size (LEfSe) with logarithmic LDA score > 3.0. e Relative abundance of selected different taxa. Data are expressed as box plot. ∗ P < 0.05, ∗  ∗ P < 0.01 by Mann–Whitney U test with Bonferroni-adjusted P-values. f Heat map of Pearson’s correlation coefficients between the top 20 genera and the diets (Y:CK, left) or infection status (CK:TCK, right). Dark red indicates a stronger positive correlation, dark blue indicates a stronger negative correlation, and white indicates no correlation. Black asterisk (*) means FDR-corrected P-value < 0.05. g The experimental design. The zebrafish were fed a basic diet supplemented with/without Cetobacterium somerae CS2105-BJ (1 × 107 CFU/g diet) for 28 days (Group Ceto/CK), and then bath infected with AH2006-3 J at a concentration of 1 × 108 CFU/mL (Group TCeto/TCK). h Kaplan–Meier graph of the zebrafish survival after bath infection with AH2006-3 J. i Cetobacterium load (gene copies/g of fish gut) in the gut of fish sampled prior to bath infection. j Aeromonas load (Aero gene copies/g of fish tissues) in fish tissues (gut, liver and kidney) sampled prior to bath infection or at 10 days post-infection. Significant differences (P < 0.05) between different groups are indicated with different lowercase letters

Full size image

We next sought to determine if Cetobacterium had the ability to enhance host defenses against the pathogen infection. First, a strain of Cetobacterium somerae CS2105-BJ was isolated from the gut contents of healthy zebrafish, and phylogenetically characterized by 16S rRNA sequence analysis (GenBank accession no. ON248483). The characteristics of C. somerae CS2105-BJ was rod-shaped cells, central swelling, gram-stain-negative, and micro-aerotolerant. Subsequently, we performed an infection experiment in which the zebrafish were fed basic diets with or without C. somerae CS2105-BJ (CS2105-BJ) for 28 days (Group Ceto/CK), and then bath infected with A. hydrophila (Group TCeto/TCK) (experimental design illustrated in Fig. 2g). The results showed that the survival of the fish administrated with CS2105-BJ was significantly higher than that of the control fish after the challenge infection (P < 0.05, Fig. 2h). In addition, we found that dietary administration of CS2105-BJ significantly increased Cetobacterium load in fish gut (P < 0.05, Fig. 2i), and also resulted in a remarkable reduction in Aeromonas load in gut, liver, and kidney of the fish at 10 days post the infection with A. hydrophila (P < 0.05, Fig. 2j). Collectively, the data indicated that C. somerae from the zebrafish gut made a substantial contribution to protection of the fish against the bacterial infection.

Functional annotation of complete genome unravels probiotic characteristics of C. somerae CS2105-BJ

We next performed genomic data analysis in order to get a comprehensive view of the protective potency of C. somerae CS2105-BJ for fish. The strain genome, de novo assembled using PacBio sequencing reads, is the first complete genome sequence available for this species and contains a single circular chromosome of 1,904,440 bp with six plasmids (plasmid 1: 704,842 bp; plasmid 2: 181,744 bp; plasmid 3: 128,183 bp; plasmid 4: 40,685 bp; plasmid 5: 9413 bp; plasmid 6: 4104 bp) (Fig. 3a). The average GC content is 29.15%. Meanwhile, there were no virulence factors found in the genome of C. somerae CS2105-BJ. The assembled results were also corrected with high-quality Illumina sequencing reads (Table S3). Furthermore, a total of 2732 protein-coding genes were predicted, and among them, 2371 (86.7%) genes were annotated as functional genes, and 361 (13.3%) genes were hypothetical genes. Approximately 57% of the protein-coding genes (1565 genes) were classified to 39 KEGG functional categories and 213 functional pathways (Fig. S4). Of particular concern is that the functional category of “metabolism of cofactors and vitamins” (112 genes) was the fourth most abundant categories, followed by “carbohydrate metabolism” (289 genes), “amino acid metabolism” (140 genes), and “membrane transport” (139 genes). In addition, the results of Clusters of Orthologous Groups (COG) annotations showed that the genome of CS2105-BJ has a complete set of genes required for the de novo synthesis of vitamin B12 (called B12 hereafter), also known as cyanocobalamin (Fig. 3b). It is well known that B12 can provide various beneficial functions to human and animal health, such as the protection of gut against gastric ulcerations, the modulation of gut microbial ecology, and the stimulation of the immune responses [56,57,58,59,60]. Additionally, some recent data have demonstrated that B12 was able to contribute to the host defense against pathogenic infection [61]. We therefore speculated that the protective effect against A. hydrophila infection was associated with B12 biosynthesis capability of C. somerae CS2105-BJ. We next evaluated the ability of this C. somerae CS2105-BJ strain to produce B12 in vitro and in vivo. The B12 production was increased with proliferation of the bacteria in vitro and reached a maximum value of 140 ng/mL (Fig. S5a). Furthermore, our results showed that oral supplementation of C. somerae CS2105-BJ significantly enhanced B12 level in the fish gut (Fig. S5b), and importantly, a strong positive correlation between B12 level and Cetobacterium load was observed (r = 0.9575; P < 0.0001) (Fig. S5c), suggesting the concentration of B12 in the fish gut mainly depends on C. somerae.

Fig. 3
figure 3

Genome analysis reveals the ability of C. somerae CS2105-BJ to synthesize vitamin B12 de novo. a Circular genomic map of CS2105-BJ chromosome and six plasmids. From the innermost to outermost circle, Circle 1 represents genome size; Circle 2 (dark purple and bottle green) represents GC skew; Circle 3 (black) shows GC plot; Circles 4 and 7 are color-coded according to the COG classification of the genes located on the forward and reverse strands, respectively. Circles 5 and 6 show the CDSs (dark blue), tRNA genes (dull red), and rRNA regions (purple). b Genomic organization of vitamin B12 biosynthetic genes. The pink arrows represent the genes for Uroporphyrinogen-III synthesis; the purple ones represent genes involved in the corrin ring synthesis; the orange one represents cobalt chelatase gene for insertion of cobalt ions into the corrin ring; the green ones represent genes for the attachment of the aminopropanol arm and assembly of the nucleotide loop in vitamin B12; the blue ones represent the genes encoding ABC transport systems for vitamin B12

Full size image

In order to further affirm the protective effect of B12 produced by C. somerae, we used excess zinc to replace cobalt ions, and thereby reduced the amount of vitamin B12 in the fish gut according to previous studies [62,63,64]. Zebrafish were randomized into four groups: control group (CK), Cetobacterium supplemented group (Ceto), excess zinc supplemented group (Zn), combined treatment with Cetobacterium and excess zinc group (Ceto + Zn). The results showed that vitamin B12 was significantly decreased after zinc treatment (Fig. S6a). Moreover, the concentration of B12 in Ceto group was significantly higher than that in CK group, and Ceto + Zn group (P = 0.024; P = 0.025, respectively) (Fig. S6a). Interestingly, after infection with A. hydrophila, the survival rate of zebrafish in Ceto group was also significantly higher than that in CK and Ceto + Zn groups, respectively (P = 0.038; P = 0.021) (Fig. S6b). These results suggested that vitamin B12 produced by gut microbes plays a key role in protecting the host from A. hydrophila infection. Although we did not examine other metabolites and also could not rule out the possible role of the other metabolites (may be also important), our data sufficiently support the conclusion that vitamin B12 is indispensable for host resistance to pathogenic infections.

The protection effect of B12 against A. hydrophila infection is reliant on the gut microbiota

Then, we wonder whether B12 could also play a protective role against A. hydrophila infection. The experimental design is illustrated in Fig. 4a. Notably, dietary administration of exogenous B12 significantly improved the survival rate of zebrafish following bath infection with A. hydrophila (P < 0.05, Fig. 4b), and also reduced the Aeromonas load in liver, and kidney of the fish in TB group compared to the controls in TCK group (Fig. 4c, d). Furthermore, we observed a strong negative relationship between B12 level and Aeromonas load in liver and kidney (r =  − 0.9567, P < 0.0001; r =  − 0.9130, P = 0.0002; Fig. 4e, f). These results suggested a protective function for B12 in protection against Aeromonas infection.

Fig. 4
figure 4

Gut microbiota are the basis of B12 protection against A. hydrophila infection in zebrafish. a The experimental design. (i) The zebrafish were fed a basic diet supplemented with/without vitamin B12 (200 μg/kg diet per day) for 28 days (Group B/CK), respectively, and then bath infected with A. hydrophila strain at a concentration of 1 × 108 CFU/mL (Group TB/TCK). (ii) The zebrafish were fed a diet containing an antibiotic mix (120 mg/kg metronidazole, 120 mg/kg neomycin sulfate and 60 mg/kg vancomycin) for 7 days (Group AB/ACK), and then received the same treatments as (i) (Group TAB/TACK). b Kaplan–Meier graph of the zebrafish survival after bath infection with A. hydrophila. * indicates significant difference (P < 0.05) between different groups. c, d Aeromonas load (Aero gene copies/g of fish tissues) in fish tissues (liver and kidney) sampled prior to bath infection or at 10 days post-infection. Significant differences (P < 0.05) between different groups are indicated with different lowercase letters above the bars. e, f Linear correlation between the B12 content and pathogen load in liver and kidney, respectively. Linear correlation was performed with Pearson’s linear correlation

Full size image

More than 80% of gut microbes require exogenous B12, while less than 25% synthesize it themselves [65]. Moreover, B12 as a key metabolite in shaping the composition of the human gut microbiota has also recently gained attention [57, 66]. We postulated that the protective effect of B12 may be related to the gut microbiota. Previous studies in zebrafish have shown that the use of antibiotics can disrupt and clear gut microbes [67]. Before the feeding trial, we fed zebrafish with a diet containing mixture antibiotics (120 mg/kg metronidazole, 120 mg/kg neomycin sulfate and 60 mg/kg vancomycin) to deplete the commensal microbes, according to a published work (Fig. 4a) [50]. Impressively, the decline of Aeromonas load in the liver and kidney was not observed after the zebrafish bath infection with A. hydrophila, which was accompanied by an abolition of the protective effect (Fig. 4b–d). In addition, we also found that the protective effect of B12 on zebrafish was concentration-dependent (Fig. S7). We speculated that the small portion of high exogenous B12 were unabsorbed and reached the distal gut where they were available to interact with the microbiota [57, 65]. Although we cannot rule out the involvement of other mechanisms in the protection effects of B12, the data from these experiments support the hypothesis that B12 protected the host against A. hydrophila infection is dependent on the gut microbiota.

Vitamin B12 induces alterations in gut microbiome structure and function

We therefore analyzed the impact of B12 on the bacterial communities in the zebrafish gut. Compared to the control group (CK), the Shannon index (P = 0.0064) was higher in B12-supplemented group (B) while the Observed_OTU (P = 0.3343) and Chao1 index (P = 0.3966) was similar between the two groups (Fig. S8). To estimate the overall structure of the gut microbiota, the principal coordinate analysis (PCoA) of the variation between microbiome based on Bray–Curtis was constructed. The results showed that there was a clear separation between the two groups (ANOSIM R = 0.7817, P = 0.003) (Fig. 5a), suggesting that dietary administration of B12 significantly changed the gut microbiome community structure. Moreover, this analysis showed that the samples from TB were clustered together, suggesting that zebrafish fed with vitamin B12 had a more stable composition of gut microbial communities after infection with A. hydrophila (Fig. S9a). Similar to the results of PCoA analysis, there were significant differences in the relative abundances of several bacterial taxa between the CK and B group. The analysis at the phylum level indicated that the most abundant phyla were Proteobacteria (46.6–58.0%), Firmicutes (21.3–11.1%), and Fusobacteria (2.28–12.7%), followed by Actinobacteria (4.91–7.30%), Bacteroidetes (9.48–1.46%), and Planctomycetes (4.88–2.69%) (Fig. 5b). Additionally, a higher ratio of Bacteroidetes/Firmicutes (mean 0.446 versus 0.131) was found in group B compared to CK (also in group TB compared to TCK) (Fig. 5b, Fig. S9b), indicating that dietary administration of B12 had a huge contribution to the gut health [68, 69]. Concurrently, the heatmap of the top 50 genera also revealed a significant impact of B12 on gut microbiota (Fig. S10). The changes in bacterial composition between the two groups were further assessed using linear discriminant analysis effect size (LEfSe), which was used to identify the specific bacterial genera that were typical of the different treatments (Fig. 5c). The results revealed that Shigella and Escherichia, which were connected with gut microbiome dysbiosis [70, 71], were decreased in B12-supplemented group, while many potentially beneficial microbes, such as Bacteroides, Lachnospiraceae_uncalssified, Cellvibrio, and Clostridium, were increased.

Fig. 5
figure 5

Vitamin B12 induces alterations in gut microbiota structure and function. a Principal coordinate analysis (PCoA) of Bray–Curtis distance was analyzed based on OTU level for microbiota beta diversity (ANOSIM R = 0.7817, P = 0.003). b Phylum-level taxonomic distributions of the microbial communities in gut of zebrafish fed with different diets. c Liner discriminant analysis effect size (LEfSe) was used to analyze the difference in microbial abundance between control and B12 supplemented group. The LDA value threshold was set at 4.0. d Bacterial community phenotypes of the gut microbiome were predicted using BugBase. Statistical significance was identified by the Wilcoxon test with false discovery rate (FDR)-corrected pairwise P-values. *, P < 0.05. e Functional alterations of the gut microbiome in zebrafish fed with control (CK) and B12-supplemented diet (B). Statistical significance was determined by using LEfSe, with a P value of < 0.05 (Wilcoxon test) and a linear discriminant analysis (LDA) score (log10) of > 2.5 being considered significant

Full size image

The results of phenotype prediction according to the gut microbiome communities matched our experimental findings well (Fig. 5d). Gram-negative bacteria, which carry the most common co-pathogens [72], showed a significant decrease in abundance in group B, while gram-positive bacteria increased (P < 0.05), suggesting that dietary administration of B12 decreased potential pathogenicity. Of interest, the prediction results also displayed a significant decrease of potential pathogenicity (P < 0.05) in group B (Fig. 5d). In addition, the gram-negative bacteria and potential pathogenicity was significantly lower in TB group comparted to TCK group (Fig. S9d). Anaerobic bacteria were key regulators in maintaining the intestinal homeostasis [73]. The relative abundance of anaerobic bacteria drastically increased after B12 treatment, indicating that B12 might facilitate the enrichment of anaerobic microbes in gut. Moreover, the relative abundance of anaerobic bacteria in TB group was higher than that in TCK group (Fig. S9d). Moreover, the Pearson correlation coefficient revealed a strong positive correlation between the B12 content and the anaerobic microbiota abundance (r = 0.7845, P = 0.0015) (Fig. S11). As is widely recognized, anaerobic microbiota are greatly influenced by gut redox potential [74]. Our data indicated that the gut redox potential, a key indicator reflecting the intestinal oxygen status [73], was decreased in group B compared to control (Fig. S12a). Meanwhile, the B12 content correlated negatively with the redox potential in gut according to the results of Pearson correlation analysis (r =  − 0.8878, P = 0.0001) (Fig. S12b). These results suggested that dietary administration of B12 might reduce the redox state in the gut. Interestingly, the stress tolerance in TB group was stronger than that in TCK group (Fig. S9d), suggesting that the microbial community in the TB group was more stable. To get insights into the functional profiles of gut microbiota affected by B12, we performed PICRUSt analyses to predict the function of gut microbiota, and analyzed KEGG level 3 pathways with LEfSe (Fig. 5e). Compared to CK group, the sequences in group B showed enrichment of bacterial proliferation and colonization pathways involving bacterial motility proteins and bacterial chemotaxis, which might promote interactions between key constituents of the microbiota and the host [75, 76], suggesting that B12 mediated microbe-microbe and microbe-host interactions. After infection with A. hydrophila, the sequences in TB group enrichment of quorum sensing and metabolism (Fig. S9e) suggest stronger microbial interactions in the TB group [77]. Collectively, these results indicated that B12 supplementation significantly enhanced the diversity and community structure of the gut microbiome, and also had an impact on the redox status of the gut, improving the interaction between intestinal microbiota.

B12 enhanced the complexity and stability of the gut ecological network

In the gut, various microbes interacted with each other to form a complicated ecological network to reduce the risk of disease occurrence and maintain the health of the host [12]. To identify potential interactions of the gut microbiota changed by B12 supplementation, we constructed co-occurrence networks using a molecular ecological network analysis pipeline (MENs) based on a random matrix theory (RMT) [78]. The samples in CK and B group each had a 0.88 threshold value applied to them [79, 80]. The networks created at the OTU level showed that all gut microbiota networks’ topologies fit the power law distribution well (R2 > 0.65), indicating that they possessed scale-free properties. The network total nodes, total links, average degree, average path distance, and average clustering coefficient were calculated for gut bacterial community in CK and B groups. To check the statistical significance of the created network indices, random networks were created (Table 1). Multiple network topological characteristics, including average path distance, average clustering coefficient, and connectedness, demonstrated that the gut microbial networks in B and CK were considerably different (Table 1). Compared to CK, the network in B had more nodes and links, increasing the density of connections and resulting in more complex network patterns (Fig. S13, Table 1). Meanwhile, the higher average degree, average clustering coefficient, and connectedness also reflected the increased complexity of the network in group B [81]. Collectively, these results suggested that B12 supplementation increased complexity of gut microbiome ecological network in zebrafish.

Table 1 Major topological properties of the empirical MENs of microbial communities in group CK and B and their associated random MENs
Full size table

We concentrated on large modules (> 5 nodes) in both the CK and B networks by emphasizing the relevance of OTUs and the proportion of major phyla to find potential alterations in gut microbial interactions after B12 consumption. Of interest, negative correlations account for 79.3% of observed links in B network, showing that taxa generally tended to co-exclude (negative correlations, green links) rather than co-occur (positive correlations, red links), yet the opposite was true for the network in CK (Fig. 6a; Table 1), suggesting that the gut microbial ecological network in B was more stable than that in CK [82]. The network in group B had the larger modules (M1, M2, M4, and M5) that comprised many more nodes (32, 18, 24, and 22) than that in CK as a result of the network modules in group B becoming more connected (Fig. 6a). In addition, the network analyses for TCK and TB also showed that there were more negative correlations in the TB group (Fig. S14). Moreover, the network in TB group had larger modules than that in TCK group (Fig. S14), suggesting that the gut microbial ecological network in TB group was more stable than that in TCK group. Furthermore, the composition of OTUs in the modules altered significantly after B12 administration. Specifically, Proteobacteria and Firmicutes dominated the large modules (> 5 nodes) in CK network, while only Proteobacteria dominated the large modules (> 5 nodes) in B network (Fig. 6a). Then, the values of among-module connectivity (Pi) and within-module connectivity (Zi) of each OTU were used to identify possible keystone taxa. We classified these nodes into four groups: module hubs (Zi > 2.5 and Pi ≤ 0.62), network hubs (Zi > 2.5 and Pi > 0.62), connectors (Zi ≤ 2.5 and Pi > 0.62), and peripherals (Zi ≤ 2.5 and Pi ≤ 0.62) [81]. Due to their importance in the network topology, network hubs, module hubs, and connectors have been proposed as keystone taxa [81]. In this study, only one network hub and one module hub were detected in the B network. Meanwhile, compared with the CK network, the B network had more connectors (Fig. 6b), indicating that B12 supplementation significantly enhanced the interaction relationship between intestinal microbiota. Interestingly, the detected network hub (OTU0155) and module hub (OTU0026), as well as most of the connectors, were rare taxa, suggesting that less abundant bacteria play regulator roles in the microbial ecological network [83]. Taken together, B12 supplementation increased the complexity of the gut microbial ecological network and improved the interactions between gut microbes.

Fig. 6
figure 6

Vitamin B12 influence the modules and the keystone taxa in the gut ecological network. a Network modules in different groups. Large modules (> 5 nodes) are shown in circular layout. Major phyla are indicated by the node colors. Positive and negative correlations are indicated by red and green connections, respectively. The matching pie charts for each network in the right panel indicate the distribution of the major phyla. The module ID of each large module is indicated by M1 to M6. b Classification of nodes in CK and B networks to find possible keystone OTUs. Each symbol represents an OTU. Pale green symbols represent the nodes in group B. Pink symbols represent the nodes in group CK. Zi > 2.5 and Pi > 0.62 indicates network hubs; Zi > 2.5 and Pi ≤ 0.62 indicate module hubs; Zi ≤ 2.5 and Pi > 0.62 indicate connectors; and Zi ≤ 2.5 and Pi ≤ 0.62 indicate peripherals. Detailed taxonomic information for node is listed in Table S4

Full size image

To quantify the effects of B12 supplementation, microbial diversity, potential keystone taxa, and network complexity on the pathogen resistance of host, a partial least squares path model (PLS-PM) was constructed (Fig. 7a). The results indicated that B12 supplementation was positively correlated with bacterial diversity (0.51), network complexity (0.72), and the potential keystone taxa (0.43) but not significantly correlated with the infection level (Fig. 7). Unexpectedly, the bacterial diversity had no significant effects on the potential keystone taxa, network complexity, and infection level. Meanwhile, network complexity was also significantly impacted by the potential keystone taxa (0.48). In addition, we also found that network complexity (− 0.40) showed significant effects on the infection level, suggesting that a complex and stable network had positive effects on the pathogen resistance (Fig. 7). Overall, these results indicated that B12 supplementation improved host pathogen resistance mainly through influencing potential keystone species in the gut and enhancing the complexity of the gut microbial ecological network.

Fig. 7
figure 7

Effects of the major factors on the pathogen resistance as determined by the PLS-PM analysis. a PLS-PM showing the cascading relationships of different factors. An observable variable or a latent variable is represented by a box. The loading for bacterial diversity, the potential keystone taxa, the network complexity, and infection level that create the latent variables are shown in the dashed rectangles. After 1000 bootstraps, path coefficients are calculated and represented by the width of the arrow (red stands for positive relationship, green stands for negative relationship). The dashed arrow indicates a coefficient that did not differ significantly from 0 (P > 0.05). The GoF statistic was used to evaluate the model, and the GoF value was 0.74. b Standardized effects of each factor on zebrafish pathogen resistance profiles calculated from the results of partial least squares path modeling. The direct and indirect impacts are added together to form the total effects

Full size image

B12 maintains normal expression of gut tight junction proteins after the pathogen infection

Permeability of the gut barrier is the basis for the formation of infection outcomes in the gut by pathogens [84, 85]. Therefore, we assessed the expression of main tight junction proteins (Claudin15, Occludin, and Zo-1) of the gut by Western blot (WB). Our results showed that dietary administration of B12 significantly upregulated Claudin15, Occludin, and Zo-1 proteins (P < 0.01) (Fig. 8). In contrast, no difference in the expression of Claudin15, Occludin, and Zo-1 was observed between ACK and AB (Fig. 8), indicating that the effect of vitamin B12 on gut tight junction proteins disappears after antibiotics disrupt the gut microbiome. These results suggest that B12 influenced gut tight junction protein expression levels through the regulation of intestinal microbes. Moreover, protein levels of Claudin15 and Zo-1 after the pathogen infection were significantly decreased. Interestingly, B12 effectively maintained the normal expression of gut tight junction proteins in zebrafish after the infection. However, the protein levels of Claudin15, Occludin, and Zo-1 in TAB group were similar to those in TACK group, suggesting that oral administration of antibiotics significantly reduced the effect of B12 on the maintenance of gut barrier integrity. To further confirm these results, mRNA (Zo-1, Occludin and Claudin15) levels were determined by RT-PCR as well. As shown in Fig. S15, vitamin B12 increased Zo-1, Occludin, and Claudin15 expression in the mRNA level (P < 0.05) while the infection decreased the expression of these genes. Similar to the results of western blot, RT-PCR results demonstrated that no difference in the expression of Claudin15, Occludin, and Zo-1 was observed between ACK and AB (TACK and TAB), indicating that the effect of vitamin B12 on gut tight junction disappeared after the antibiotic treatment. In addition, we also analyzed the expression of intestinal tight junction proteins-related genes in zebrafish fed with B. velezensis and Cetobacterium, and found that dietary supplemented with B. velezensis or Cetobacterium significantly improved the expression of Zo-1, Occludin, and Claudin15, which had a similar trend with B12 (Fig. S16).Overall, the results of the present study suggested that B12-influenced gut microbiota network maintained the normal expression of gut tight junction proteins in pathogen-infected zebrafish.

Fig. 8
figure 8

B12 enhances the tight junctions in the gut of zebrafish. a Western blots showing the expression of Zo-1, Occludin, and Claudin15 in the gut of zebrafish. b–d Densitometric analysis of Western blots from protein samples of the gut. Data were normalized for β-tubulin expression and expressed as fold change. Values represent means ± SD. Significant differences (P < 0.05) between different groups are indicated with different lowercase letters above the bars. CK: The zebrafish were fed a basic diet; B: The zebrafish were fed a basic diet supplemented with vitamin B12; ACK: The zebrafish treated with antibiotics for 7 days prior to administration of basic diet; AB: The zebrafish treated with antibiotics for 7 days prior to administration of B12; TCK: The zebrafish were fed a basic diet and then bath infected with A. hydrophila; TB: The zebrafish were fed a basic diet supplemented with vitamin B12 and then bath infected with A. hydrophila. TACK: The zebrafish treated with antibiotics for 7 days prior to administration of basic diet, and then fed a basic diet and then bath infected with A. hydrophila; TAB: The zebrafish treated with antibiotics for 7 days prior to administration of basic diet, and then fed a basic diet supplemented with vitamin B12 and then bath infected with A. hydrophila

Full size image

Discussion

Gut microbes are key factors in host defense against pathogen infection [86], and are incredibly important to host health [66]. In the present study, we identified a distinct mechanism by which probiotics drive intestinal commensal bacteria to produce vitamin B12, a microbial-derived metabolite, protecting the host from pathogenic infections. These findings highlight the relevance of the gut microbiota and associated metabolites in protecting host against the pathogen infection. Our data also support the idea of the using of Cetobacterium or even vitamin B12, as the key interventions for prevention of pathogen infection.

The characterization of gut microbiota composition is a hot topic for researchers to study the intersection of host microbiome and health [3]. It is well known that probiotics could modulate the composition and function of gut microbiota [87, 88]. We found that Bacillus velezensis 1704-Y supplementation could protect zebrafish against Aeromonas hydrophila infection by modulating the gut microbiota. Similar results were found in analogous studies in mammals, where pre-addition of Lactobacillus could significantly change the structure of the gut microbiome to improve the resistance to Escherichia coli [89]. Wang et al. found that probiotics attenuate obesity comorbidities through specific impacts on the gut microbiota in mice [90]. Moreover, studies in some animal models had demonstrated that probiotics protect host against pathogen infection through the modulation of gut microbiota [91,92,93]. Although the mechanisms by which probiotics regulate gut microbes still need further study, the current results all suggest that protection of the host from pathogen infections by modulating gut microbes is one of the beneficial pathways of probiotics.

Meanwhile, this study also found that Cetobacterium, the gut indigenous microbiota, was the key microbe to protect zebrafish against Aeromonas hydrophila infection after dietary supplementation of probiotic Bacillus velezensis 1704-Y. A growing number of studies have demonstrated that probiotic treatment can increase the relative abundance of Cetobacterium in the gut of freshwater fish [42, 94]. Cetobacterium is an anaerobic indigenous bacterium present in the gut of most freshwater fish [95]. In recent years, anaerobic indigenous gut bacteria, which played a crucial role in human health and disease, have received increasing attention from researchers. Anaerobic indigenous gut bacteria could consume dietary fiber and produce short-chain fatty acids to benefit the host. For example, Faecalibacterium prausnitzii, a major commensal anaerobic gut bacterium, exhibited anti-inflammatory effects on Crohn disease patients and could alleviate intestinal inflammation [96]. In addition to this, Akkermansia muciniphila, a mucin-degrading anaerobic bacterium, had been proved that could produce butyric acid to provide energy for gut epithelial cells and maintain the gut barrier and health [97, 98]. These suggested that gut indigenous anaerobic bacteria and their metabolites played a non-negligible role in maintaining the health of the host. We also found that Cetobacterium was a dominant member of gut microbiota of healthy fish, while its levels were significantly reduced in the gut of infected fish, suggesting that Cetobacterium was a sensor of health, especially for fish infected with pathogenic bacteria. A recent research demonstrated that Aeromonas veronii infection induced a significant decrease in the relative abundance of Cetobacterium in the gut of Yangtze finless porpoise [53]. Similarly, Ofek et al. also proved that diseased tilapia had a lower relative abundance of Cetobacterium in the gut compared to healthy tilapia [99]. In this study, we also found that Aeromonas infection significantly decreased the abundance of Cetobacterium in the gut. Meanwhile, both correlation analysis and antibiotic treatment tests also showed that reducing the level of Cetobacterium in the gut increased the susceptibility of zebrafish to pathogenic bacteria. Based on the results of previous studies and the results of this experiment, we made a reasonable hypothesis that Cetobacterium is a key factor maintaining the fish health and providing the protection against some pathogenic bacteria. Moreover, genomic analysis showed that Cetobacterium has all the genes needed to synthesize vitamin B12 de novo, and HPLC results from in vitro metabolism tests showed that Cetobacterium could produce vitamin B12. Approximately 80% of gut microbes appear to require vitamin B12, while less than 25% of gut microbes can synthesize vitamin B12 [65], indicating that vitamin B12 was an essential factor in maintaining normal life activities of bacteria. Currently, only Fusobacteria, Veillonella, Klebsiella, Pseudomonas, Lactobacilli, and Bifidobacteria had the ability to produce vitamin B12 in the gut [100]. Moreover, recent studies had proved that Akkermansia muciniphila might also had the ability to synthesize vitamin B12 [101]. Furthermore, Cetobacterium has the ability to produce vitamin B12 allowing it to impact these bacteria to interact with their hosts and other members of the gut microbiota, which further deepened our understanding of how this important anaerobic indigenous gut bacterium affected fish health. The data herein of microbiota and metabolism analysis suggested that Cetobacterium was a key anaerobic indigenous gut bacterium for maintaining host health in freshwater fish.

Here, we discovered that vitamin B12 has a novel function as a regulator of host gut microbial interactions, which helps to strengthen interactions within the gut microbiota and improves host resistance to pathogen infections. We found that the protective effect of B12 on zebrafish was concentration-dependent. We speculated that oral high-dose B12 supplements were largely unabsorbed and reached the distal gut where they were available to interact with the microbiota. Unlike other water-soluble vitamins, which are largely absorbed and enter the circulation, vitamin B12 absorption in the ileum becomes saturated around 2 μg/meal [102]. There are many studies focused their attention on studying the relationship between B12 receptors and host health [103,104,105]. Hansen et al. proved that dietary vitamin B12 did not affect transcription of tcn1 and tcn2 in the proximal intestine and in the distal intestine, and the cubn and amn were expressed in the distal intestine but were not affected by diet [106]. Kelly et al. had proved that oral high-dose B12 supplements is delivered to the distal gut, where it was available to interact with the microbiota, and they also detected higher concentrations of B12 in the fecal contents of the mice supplemented with excess vitamin B12 [107]. In the present study, we found that higher concentrations of B12 were detected in the hindgut of zebrafish fed with high concentrations of B12, which is similar to the previous studies. Also, we did not detect the expression of vitamin B12 receptors, but it is important to understand the vitamin B12 receptor distribution in the maintenance of zebrafish health. Although this study did not prove the upper limit of saturable uptake of vitamin B12 in the zebrafish gut, these results showed that higher concentrations of B12 were detected in the hindgut of zebrafish supplemented with high concentrations of B12, which is similar to the previous studies [107, 108]. Moreover, the dietary administration of vitamin B12 significantly increased gut microbial diversity and altered microbial composition in this study. Additionally, it increased the ratio of Bacteroidetes/Firmicutes, which was closely associated with health of gut [68]. Most of the opinions are that the ratio of Bacteroidetes/Firmicutes was associated with obesity [109, 110]. Bin et al. also found that enterotoxigenic Escherichia coli infection caused a significant decrease in the ratio of Bacteroidetes/Firmicutes in the gut of piglets and caused diarrhea [111]. These results indicated that Bacteroidetes/Firmicutes played an important role in host health, which was similar with our data. Meanwhile, vitamin B12 supplementation significantly reduce the redox potential in the gut, suggesting that vitamin B12 could modulate redox homeostasis. High oxygen environment could provide a more suitable environment for pathogenic bacteria to expand their virulence [112]. Disturbances in the redox balance in the gut could potentiate inflammation, impair barrier function, prevent colonization with a healthy microbiome, and threaten host health. Our findings proved that vitamin B12 could prevent disturbances of the redox balance, suggesting that vitamin B12 reduced the oxygen environment in the gut and provided a suitable living environment for probiotic bacteria, which in turn may promote microbial interactions. The study from Busti et al. proved that the reduction of the oxygen level in the gut environment contributed to the modulation of gut microbiota favoring the presence of obligate anaerobic [113]. Indeed, the supplementation of B12 significantly increased the complexity and stability of the gut ecological network. For example, the number of nodes and links increased due to the supplementation of B12, indicating enhanced microbial interactions. The gut microbiome contains thousands of microbes that interact with each other to form complicated networks, and this stability of the gut network is considered important for health of host [82]. Correspondingly, major shifts in microbial community composition are often associated with ill health [82, 114]. In the ecological network, the species were used as nodes and their relationships as links [115], which is essential for characterizing species interactions and dynamics of gut network [116]. We also observed much higher modularity in B12 addition group. Modularity is a very important concept in ecological network. It could be originated from specificity of interactions, convergent evolution, and natural selection, and it could be important for system stability [117]. Yang et al. demonstrated that probiotic supplementation led to more module to maintain gut microbiota homeostasis [118]. The higher modularity indicated that the supplementation of B12 enhance the stability of gut ecology system. Previous studies have suggested that competition promotes stability in the gut ecosystem and that cooperation creates dependencies that foster instability in microbial communities [82]. Indeed, we observed a higher number of negative interactions in B12 addition group, suggesting that the supplementation of B12 could improve the stability of the gut ecological network. Overall, the data of the present study indicated that vitamin B12 is an important regulator in maintaining the interactive relationships between members of the gut microbiota.

Subsequently, we used the PLS-PM model to explore the relationship among the gut microbiota network, the B12 supplementation, and pathogen resistance of zebrafish. Before this, a systematic review of the effects of vitamin B12 on gut microbiome was conducted by Guetterman et al., which demonstrated that B12 was associated with gut microbiome outcomes, including beta diversity, alpha diversity, relative abundance of bacteria, and functional capacity [41]. However, few studies have been conducted to evaluate the impact of vitamin B12 on the pathogen resistance of host. The results of PLS-PM showed that B12 had no directed correlation with the pathogen resistance, but had directed correlation with the gut microbiota network. Meanwhile, the gut microbiota network had directed correlation with the pathogen resistance of zebrafish. These suggested that vitamin B12 could not directly interact with the host to protect the zebrafish against the pathogen infection but affected pathogen resistance by increasing the stability of the gut microbiota network, which further responded to the importance of vitamin B12 as a regulator among microbiota for the host. Moreover, the results of west blot analysis showed that B12 did not directly affect gut tight junction proteins, but upregulated them through regulating gut microbes. Tight junction protein expression has been demonstrated to control gut barrier functions, and an increase in tight junction protein levels could prevent or reverse pathogen impacts [119]. Until now, it remains unclear what the underlying molecular mechanisms are on the effects of vitamin B12 on the gut barrier. However, it is known that commensal bacteria and probiotics have been shown to promote gut barrier integrity both in vitro and in vivo [120,121,122,123]. Previous studies demonstrated that Claudin15, Occludin, and Zo-1 were the key proteins in maintaining the integrity of the gut epithelial barrier [124,125,126]. In this study, the expression of these proteins (Claudin15, Occludin, and Zo-1) was upregulated in the zebrafish receiving B12 supplementation. Meanwhile, we found that the effect of vitamin B12 on the expression level of gut tight junction proteins disappeared after the antibiotic treatment. We speculated that the effect of vitamin B12 on gut tight junction proteins depended on gut microbes. These data suggested that B12-influenced gut microbiota network might alleviate the increased gut permeability caused by pathogen infections. However, at this stage, we did not investigate the mechanisms by which the supplementation of B12 enhance the stability of gut microbiota network. Deeper verification tests based on meta transcriptome and metabonomic are necessary, and further research on this issue will contribute to better utilization of vitamin B12.

Conclusions

The use of probiotics and their metabolites is a practical alternative to promote animal health and prevent disease. We proved that the microbiome alterations influenced by Bacillus velezensis BV1704-Y could control the pathogenic load in tissues and improve the survival of zebrafish. Further analysis found that Cetobacterium, anaerobic indigenous gut microbe, might be a sensor of health, especially for fish infected with pathogenic bacteria. Genomic analysis and metabolic assays suggested that Cetobacterium had the ability to produce vitamin B12. Supplementation of vitamin B12 reduces the redox potential in the gut, induces alterations in gut microbiome structure and functions, and improves microbial interactions and enhances the stability of the gut microbiota network. Moreover, B12 supplementation did not directly influence the pathogen resistance of zebrafish, but by impacting the gut microbiota network. In addition, the stable gut microbiota network upregulated the tight junction proteins of the intestine and protected host against pathogen infections (Fig. 9). Our findings provided a new mechanism of action of probiotics in enhancing host resistance to pathogen infections from the perspective of probiotic action on the gut microbiome. And also demonstrates the novel function of B12 as a regulator to enhance microbial interactions. This discovery may deepen the understanding of the impact of probiotics on host health and help to elucidate the health benefits of vitamin B12 against pathogen infections.

Fig. 9
figure 9

Mechanisms of probiotic protection of the host against pathogen infections. Dietary supplementation with Bacillus velezensis BV1704-Y induces an increase in the abundance of the indigenous gut microbiota (Cetobacterium) and thus metabolizes sufficient amounts of vitamin B12. Vitamin B12 is used by the surrounding microbiota to form a more stable and complex gut ecological network while reducing the redox potential in the gut and maintaining the anaerobic state of the intestinal lumen, which further promotes the expression of intestinal tight junction proteins (Claudin15 and Zo-1) and prevents the infestation of Aeromonas

Full size image

Materials and methods

Zebrafish and experimental design

All experiments were done using 4-month-old AB wild type zebrafish (0.45 ± 0.05 g; 3.5 ± 0.2 cm), which were obtained from the China Zebrafish Resource Center (Wuhan, China). All zebrafish maintained under laboratory conditions with a 14-h light and 10-h dark cycle and adapted to the laboratory conditions for 2 weeks. Each tank containing dechlorinated and aerated water (pH 7.5 ± 0.5) at 28 ± 1 °C. More than 6.5 mg/L of dissolved oxygen was present in the tank.

Experiment 1: B. velezensis BV1704-Y supplementation experiment

To study the effects of B. velezensis BV1704-Y on resistance of zebrafish to Aeromonas hydrophila infection, a total of 120 zebrafish were randomly divided into two groups with three 20-L tanks (20 fish/tank) each. Zebrafish were fed with control diet (CK) or B. velezensis BV1704-Y-supplemented diet (1 × 107 CFU/g diet) (Y) for 28 days [127]. In order to investigate the role of gut microbiota in the process of probiotic protection of the host against pathogen infection, a common cocktail of antibiotics was used to deplete the intestinal bacteria according to previous studies [49, 50]. Zebrafish received an antibiotic mix consisting of vancomycin (2 g/kg diet), metronidazole (4 g/kg diet), and neomycin sulfate (4 g/kg diet) in the diet for 7 days, and then antibiotics-containing diets were replaced with the control and B. velezensis BV1704-Y-supplemented diet (AY), which were fed for 28 days [127], respectively. All diets were 3-mm-diameter pellets. Fish were hand-fed diets at 3% of body weight twice a day (9:00 and 16:00) and consumed all feed offered within 30 s after feeding [106], and the amount of feed intake was about 0.8 g of feed per feeding. The gut, kidney, and liver were collected under aseptic conditions from each zebrafish immediately after the feeding trial, and stored at − 80 °C until further analysis. Then the rest of zebrafish in each group (15 fish in each replicate) were bath infected with A. hydrophila, while the control and B. velezensis 1704-Y-containing diets were kept fed during the infection period. The experimental design is shown in Fig. 1a.

Experiment 2: C. somerae CS2105-BJ supplementation experiment

A total of 120 zebrafish were randomly divided into two groups with three 20-L tanks (20 fish/tank) each. Zebrafish were fed with control diet (CK) or C. somerae CS2105-BJ-supplemented diet (1 × 107 CFU/g diet) (Ceto) for 28 days, and then received the same treatments as Experiment 1. The experimental design is shown in Fig. 2g.

Experiment 3: vitamin B 12 supplementation experiment

A total of 120 zebrafish were randomly divided into two groups with three 20-L tanks (20 fish/tank) each. Zebrafish were fed with control diet (CK) or vitamin B12-supplemented diet (B) (200 μg/kg diet) for 28 days. Then received the same antibiotic treatment as in Experiment 1 and marked as ACK and AB, respectively. Subsequent operations were the same as in Experiment 1, except that the B. velezensis 1704-Y-supplemented diet was replaced with a B12-supplemented diet. The experimental design is shown in Fig. 4a.

Isolation of Cetobacterium

Zebrafish from B. velezensis BV1704-Y-treated were collected. Gut was removed and grinded into homogenate with 200 μl sterile saline solution. Serial dilutions of 100 μl (10−3, 10−4, and 10−5) containing homogenate were inoculated on BHI agar and incubated at 28 °C for 48 h in an anaerobic environment. The single colony cultured in the medium was chosen and incubated again in the anaerobic bottle with 5 mL BHI broth medium. PCR was utilized with 16S rRNA gene universal primers (27F and 1492R) to verify the bacterial solution. The right samples were kept, and whole-genome sequencing was done for added verification.

Infection of Aeromonas hydrophila AH2006-3 J

Aeromonas hydrophila AH2006-3 J (GenBank accession no. OP778940) was isolated in disease outbreak fish pond and screened as the most invasive strain from dozens of different A. hydrophila strains. For Aeromonas hydrophila infection, zebrafish were immersed in water containing bacteria with a final concentration of 1 × 108 CFU/mL for 10 days [67]. The water and bacteria were changed every 2 days, and mortality was recorded every day. At the end of experiment, the gut, kidney, and liver of zebrafish were sampled under aseptic conditions and stored at − 80 °C until further analysis.

Bacterial load quantification

Total gut, liver, and kidney DNA was extracted using a QIAamp-DNA Stool Kit (Qiagen, Hilden, Germany). Primers for amplification of genes are listed in Table S1. Amplified sequences were cloned into pMD19-T plasmids (Takara, Dalian, China), to perform a 10-fold dilution and generate a standard curve for calculation of the bacterial load [128, 129]. The values obtained from bacterial copies were calculated relative to the weight of the tissues.

Microbiome analysis

The gut total DNA was extracted by the QIAamp PowerFecal DNA Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The V3–V4 regions of the bacterial 16S rRNA gene were amplified using the 341F-806R primers (341F: 5′—CCTAYGGGRBGCASCAG—3′; 806R: 5′- GGACTACNNGGGTATCTAAT—3′) [130]. High-throughput sequencing was performed on the Illumina MiSeq platform at LC-Bio Co., Ltd (Hangzhou, China). In order to create feature tables and feature sequences, the sequences that were obtained from sequencing were subjected to quality filtering and modification. The sequences were aligned using the SILVA Database (v.138) [131] as a reference database, and the taxonomic information was categorized by RDP classifier (v2.2) with 80% confidence [132]. R software (v3.5.2) was used to calculate and display alpha diversity, including the Shannon, Simpson, Richness, and Chao1 indices. Bray–Curtis distance-based PCoA analysis was performed using the vegan R package. Using the LEfSe method [133], we further selected the significant microbiome characteristics in different groups at genus taxonomic rank. PICRUSt2 [134] was used to infer the expected metagenomes and the function of the gut microbiota, and the LEfSe method was used to determine the differentially abundant KEGG pathways between groups.

Genome sequencing and de novo assembly

The genomic DNA of Cetobacterium was extracted for whole-genome sequencing. Following that, Personal Biotechnology Company (Shanghai, China) used the Pacific Biosciences platforms and Illumina MiSeq platforms to sequence the genome. Quality control was performed by using AdapterRemoval [135] and SOAPec [136]. By using SPAdes [137] and A5-miseq [138], the filtered reads were assembled to create scaffolds and contigs. The data received by Pacbio platform sequencing were assembled using Canu software [46]. Subsequently, all assembled results were combined to produce a comprehensive sequence. Finally, using pilon software [139], the genome sequence was obtained after the rectification. The complete genome sequence was deposited at GenBank under the BioProject accession number CP092519-CP092525.

Determination of vitamin B12 and redox potential

To extract vitamin B12 from the hindgut, the contents were disrupted by boiling for 15 min in 0.1 M phosphate buffer containing 0.01% potassium cyanide at pH 6.0. After centrifugation at 4500 × g for 15 min, the supernatants were collected and passed through 0.22-μm filters [59]. Vitamin B12 was detected using a modified HPLC method that was previously reported [59, 140]. At room temperature, all chromatographic separations were carried out. The mobile phases were a mixture of methanol with 0.1% formic acid (A) (Aladdin, Jinan, China) and ultra-purified water with 0.1% formic acid (B), which was degassed by an ultrasonic water bath; the flow rate was 0.5 mL per min. The gradient elution was programmed as follows: 0–2 min, 20% A; 2–3 min, 20–25% A; 3–11 min, 25–35% A; 11–19 min, 35–20% A; 20–22 min, 100–100% A; 22–26 min, 100–20% A; and 26–36 min, 20% A. A Diode Array Detector (1260 Infinity II, US) was used to measure the column eluate at 361 nm, and the injection volume was 100 μl.

For the redox assessment, 20 mg of gut contents was diluted in 2 mL of distilled water and centrifugated at 8000 rpm for 10 min. The redox potential was measured in accordance with the manufacturer’s instructions using a pH/redox meter (REX, Shanghai, China) [73, 74].

Network analysis

The Molecular Ecological Network Analysis (MENA) pipeline (http://ieg2.ou.edu/MENA/), as previously published [81], was used to conduct network analysis to examine microbial relationships and network complexity of gut microbiota in various groups. Based on the OTU abundances that had been log-transformed, the Pearson coefficient was determined. Prior to network formation, the relevant similarity threshold (St) was automatically determined using random matrix theory (RMT) [78]. Following that, all gut microbial networks were created using the same St (i.e., 0.88). The network graphs were displayed using the Cytoscape software (3.8.2).

Various indexes, including average degree, average path distance, average clustering coefficient, connectedness, and modularity were used to describe the characterization of individual nodes in the network and the overall topologies or structures of different networks. To test the significance of the constructed empirical MENs, 100 random networks were generated for each empirical network. The means and standard deviations computed from the 100 randomizations for each characteristic were compared to the corresponding empirical MENs [81].

Each node (i.e., OTU) in networks was evaluated for its connectivity using the metrics of within-module connectivity (Zi) and among-module connectivity (Pi) to find potential keystone taxa that may have an impact on the patterns of gut microbiota assemblage [81]. This can divide all nodes into four categories: Zi > 2.5 and Pi > 0.62 indicated network hubs; Zi > 2.5 and Pi ≤ 0.62 indicated module hubs; Zi ≤ 2.5 and Pi > 0.62 indicated connectors; and Zi ≤ 2.5 and Pi ≤ 0.62 indicated peripherals. Connectors, module hubs, and network hubs can all be seen of as potential keystone taxa [141].

Partial least squares path modeling analysis

The PLS-PM was performed using R package of “plsmp” to quantify the effects of different factors (i.e., bacterial diversity reflected by Shannon and Richness; keystone taxa box was reflected by connectors and hubs; network complexity was reflected by degree and connectedness) on the pathogen resistance of zebrafish, which indicated by pathogen load and mortality.

Western blot analysis

Gut samples were homogenized in RIPA buffer supplemented with 1% of protease and phosphatase inhibitors (Beyotime Biotechnology, Shanghai, China), and protein content was measured with a BCA Protein Assay kit (CWBIO, Suzhou, China). Total proteins (20–50 μg) were resolved using 10% SDS-PAGE gel electrophoresis and transferred to PVDF membrane (0.22 μm pore; Millipore, Billerica, USA). After blocking with 5% (w/v) skim milk at room temperature for 1 h, membranes were then incubated with the primary antibodies at 4 °C overnight (dilutions of respective antibodies are given in Table S5). Next day, membranes were incubated with the HRP-labeled secondary antibody for 40 min at room temperature and the chemiluminescent substrate was used to detect the protein bands. Densitometry analysis of bands was done using ImageJ software.