A strain affiliated to Bacillus amyloliquefaciens alleviates 1 high-carbohydrate diet-induced metabolic syndrome by 2 restoration of acetate-producing bacteria in fish intestines

23 Background: Increasing the utilization efficiency of high-carbohydrate diet has the 24 potential to promote “protein sparing effects” in farmed fish; however, many fish 25 utilize carbohydrates poorly. The intestinal microbiota plays an important role in 26 carbohydrate degradation. Whether the addition of functional bacteria could increase 27 the carbohydrate utilization efficiency and alleviate high-carbohydrate diet-induced 28 adverse effects is unknown. 29 Results: A bacterial strain that could degrade starch in vitro was isolated from the 30 intestines of Nile tilapia ( Oreochromis niloticus ). The bacterium was affiliated to 31 Bacillus amyloliquefaciens (designated as B. amy SS1 ) based on 16S rRNA gene 32 sequencing. Three diets, including control diet (CON), high-carbohydrate diet (HCD), 33 and high-carbohydrate diet supplemented with B. amy SS1 (HCB), were used to feed 34 Nile tilapia for 10 weeks. The beneficial effects of B. amy SS1 on weight gain and 35 protein accumulation were observed. The HCB decreased blood glucose levels and 36 reduced lipid deposition compared with the HCD group. To detect the possible 37 mechanism, the intestinal microbiota composition was characterized using 38 high-throughput sequencing. The HCB increased the abundance of short-chain fatty 39 acid-producing bacteria. Gas chromatographic analysis indicated that the 40 concentration of acetate increased dramatically in the HCB group compared with that 41 in the HCD group. Glucagon-like peptide-1 (GLP-1) levels increased in the intestine 42 and serum of the HCB group. Different concentrations of sodium acetate (low (HLA), 43 900 mg/kg; medium (HMA), 1800 mg/kg, and high (HHA), 3600 mg/kg) were added to the HCD to feed the fish for eight weeks. The HMA and HHA groups mirrored the 45 effects of the HCD supplemented with B. amy SS1 by increasing serum GLP-1 levels. 46 Increased acetate concentrations stimulated GLP-1 production, which might account 47 for the effects caused by the addition of B. amy SS1 to the HCD. 48 Conclusions: This study systematically analyzed the influence of B. amy SS1 on fish 49 metabolism, suggesting that B. amy SS1 treatment alleviates the metabolic syndrome 50 caused by HCD by enriching acetate-producing bacteria in fish intestines. Regulating 51 the intestinal microbiota and their metabolites might represent a powerful strategy for 52 fish nutrition modulation and health maintenance in future. 53

superior degradative ability with respect to resistant starch, and the released products 77 from resistant starch can be utilized by other gut bacteria to produce short chain fatty 78 acids (SCFAs), which have wide-ranging impacts on host physiology, including 79 serving as an energy source for host cells or stimulating the production of gut 80 hormones [12,13]. A high-carbohydrate diet altered the fecal microbiome by 81 increasing the carbohydrate degradation members and SCFAs excretion in humans 82 [14]. However, a study in the grass carp (Ctenopharyngodon idellus) showed that 83 SCFA levels were lower in the hindgut when they were fed with a 84 high-fiber/low-protein diet compared with that under a high-protein/low-fiber diet 85 [15], indicating that the gut microbiota in the grass carp, which is a predominantly 86 herbivorous fish, does not tend to ferment fiber to SCFAs. These results suggested 87 that the limited utilization efficiency of carbohydrates by the intestinal microbiota 88 5 might account for the glucose intolerance of fish. 89 The intestinal microbiota shows a great potential for maintaining glucose 90 homeostasis; however, the response to regulation by the intestinal microbiota in the 91 context of glucose homeostasis is strongly linked with the baseline microbiota 92 composition. Research on humans with prediabetes showed that exercise-induced 93 changes in the gut microbiota correlated with improved glucose metabolism and 94 insulin sensitivity [16]. The microbiome of responders exhibited an enhanced capacity 95 to produce SCFAs and catabolize branched-chain amino acids, suggesting that the gut 96 microbiota is a key determinant for the variability of glycemic control [16]. A similar 97 observation was made in a cohort of healthy individuals exposed to barley 98 kernel-based bread (BKB), which suggested that humans harboring a higher 99 Prevotella/Bacteroides ratio exhibited improved glucose metabolism following 3-day 100 consumption of BKB [17]. Fish harbor a Proteobacteria-dominated microbiota, which 101 is different from the dominant microbiota in human or mice [18,19]. Whether 102 regulation of the intestinal microbiota could increase the carbohydrate utilization 103 efficiency and alleviate the adverse effects caused by high-carbohydrate diets in fish 104 remains unknown. 105 Nile tilapia (Oreochromis niloticus) is an economically important fish species 106 and is an ideal fish model for nutritional and metabolic studies because of its fast 107 growth, high resistance to disease, and available genomic information [20]. In the 108 present study, we isolated a strain that could degrade starch in vitro from the intestine 109 of Nile tilapia. 16S rRNA gene sequencing showed that the strain was affiliated to Bacillus amyloliquefaciens (designated as B. amy SS1). Three diet treatments, 111 including control diet (CON), high-carbohydrate diet (HCD), and high-carbohydrate 112 diet supplemented with B. amy SS1 (HCB) were used to feed Nile tilapia for ten 113 weeks. The host physiology and metabolic characteristics were identified in these 114 three groups and the possible mechanism by which B. amy SS1 regulates carbohydrate 115 utilization was investigated.

118
A strain isolated from the intestine of Nile tilapia improved the growth 119 performance of fish. 120 To isolate bacteria that could degrade starch in fish gut, starch was used as the main 121 carbon source in the culture medium. About two hundred colonies were screened and 122 one colony, which had a larger transparent zone on starch agar medium after the 123 addition of iodine solution, was selected for the further research. 16S rRNA gene 124 sequencing showed that the strain was affiliated to Bacillus amyloliquefaciens ATCC 125 39320 (Fig. 1a). The selected strain was named as B. amy SS1 in the present study.

126
DNS analysis confirmed the amylase activity of B. amy SS1 in vitro (Fig. 1b) and gas 127 chromatography showed that B. amy SS1 could ferment corn starch to mainly produce 128 acetate and butyrate (Fig. 1c).

129
To detect whether the addition of B. amy SS1 could influence the growth 130 performance of fish under an HCD, three treatments, including CON, HCD, and HCD 131 with B. amy SS1 (HCB) were used to feed fish for 10 weeks. Weight gain was 132 7 detected every two weeks. The results showed that the average weight was 133 significantly higher in the HCD group than in the CON group; moreover, the average 134 weight was further increased by B. amy SS1 treatment in the HCB group (Fig. 1d).

135
The addition of B. amy SS1 to the HCD resulted in the highest weight gain among the 136 three groups (Fig. 1e) and the feed efficiency was higher in the HCB group compared 137 with that in the HCD group (Fig. 1f). showed that the addition of B. amy SS1 reduced the high fasting glucose level caused 145 by the HCD (Fig. 2a). The IGTT test showed that the addition of B. amy SS1 146 markedly reduced the persistently higher blood glucose level caused by the HCD (Fig.   147 2b, c), i.e., B. amy SS1 improved glucose tolerance. Considering the important role of 148 insulin in glucose homeostasis, the fasting insulin level was detected; however, no 149 significant difference was found among the groups (Fig. 2d), suggesting that the 150 addition of B. amy SS1 to the HCD might elevate insulin sensitivity rather than its 151 amount. Glucose homeostasis induced by B. amy SS1 was further supported by 152 significantly decreased liver glycogen levels (Fig. 2e). To investigate whether the 153 insulin signaling pathway was activated by the addition of B. amy SS1, the expression 154 8 levels of crucial proteins, including phosphatidylinositol 3-kinase (PI3K) and protein 155 kinase B (AKT), were detected using western blotting. The total levels of PI3K and 156 AKT were similar among the groups; however, the levels of phosphorylated PI3K and 157 AKT were significantly increased by B. amy SS1 administration (Fig. 2f, g), 158 suggesting that the addition of B. amy SS1 to the HCD improved glucose tolerance via 159 activating the PI3K/AKT insulin signaling pathway.

160
Enhanced glycolysis might improve glucose homeostasis; therefore, three key 161 enzymes of glycolysis, hexokinase (HK), phosphofructokinase (PFK) and pyruvate 162 kinase (PK) were analyzed. The glycolytic enzyme activities in the liver were all 163 increased in B. amy SS1-treated fish ( Fig. 2h-j). The mRNA expression of glycolysis 164 targeted genes, including gck, pfk, pk, and ir in the liver were downregulated in the 165 HCD group, but upregulated by the addition of B. amy SS1 (Fig. 2k). These data 166 strongly suggested that the addition of B. amy SS1 to the HCD enhanced glycolysis by 167 activating the pivotal enzymes related to glycolysis in the liver.

169
The addition of B. amy SS1 to the HCD reduced lipid deposition by activating the 170 AMPK/ACC signaling pathway to increase energy expenditure in Nile tilapia. 171 An HCD causes excess lipid accumulation in fish, which further aggravates the 172 metabolic imbalance [22]. The hepatic somatic index (HSI) was mostly increased in 173 the HCD group compared with that in the CON group, and the HCB group showed a 174 decreased trend in HSI, although no significant difference was detected (Fig. 3a). The 175 hepatic lipid content was significantly increased in the HCD group compared with 176 9 that in the CON group, but it was decreased by the addition of B. amy SS1 (Fig. 3b).

177
The addition of B. amy SS1 to the HCD also exhibited protective effects against 178 HCD-induced liver damage, including lower content of triglyceride (TG), 179 non-esterified fatty acid (NEFA), and total cholesterol (T-CHO) in the liver (Fig. 3c-e). 180 Furthermore, hematoxylin eosin staining (H&E) and oil red O staining also indicated 181 that the addition of B. amy SS1 to the HCD markedly reduced lipid accumulation ( Fig.   182 3f-i). The mRNA levels of genes related to lipid synthesis, including fas, accα, dgat2, 183 and pparγ, showed no significant difference in the liver among the groups (Fig. 3j).

184
However, compared with that in the HCD group, the HCB group showed substantial 185 up-regulation of genes targeted to lipolysis, including atgl, cpt1, hsl, and pparα in the 186 liver (Fig. 3k). These findings suggested that the addition of B. amy SS1 to the HCD 187 activated lipolysis to decrease lipid deposition in the liver.

188
To address whether activated lipolysis was associated with energy homeostasis, 189 the levels of key proteins involved in this process were detected using western  Besides lipid accumulation in the liver, we also detected the content of total lipid 199 in the body. Notably, a decrease in the total lipid content was observed in the B. amy 200 SS1-treated fish (Fig. 3n). Moreover, mesenteric fat index (MFI) was lower in HCB 201 group compared with that in the HCD group (Fig. 3o). B. amy SS1 administration also 202 reduced the cell size of adipocytes (Fig. 3p, q). Meanwhile, the serum TG, NEFA,  The addition of B. amy SS1 to the HCD increased protein accumulation by 210 activating the mTOR/S6 signaling pathway in Nile tilapia. 211 We further assessed the impact of B. amy SS1 on body protein accumulation. The 212 results showed that the addition of B. amy SS1 to the HCD increased the carcass index 213 and carcass protein content significantly (Fig. 4a, b). The mRNA expression of mtor 214 and s6, which are related to protein synthesis, were detected. The results indicated that 215 mtor was significantly up-regulated by B. amy SS1 administration, but no significant 216 difference was observed for s6 among the groups (Fig. 4c). Western blotting analysis 217 demonstrated that the levels of phosphorylated mTOR and S6 increased significantly 218 after the addition of B. amy SS1 to the HCD; however, no significant difference was 219 found in the total levels of these proteins (Fig. 4d, e). Overall, these data implied that

225
The gut microbiota has critical roles in host nutrition and metabolic processes.

257
The results of the previous section showed that the abundance of SCFA-producing  IGTT showed that blood glucose levels were obviously reduced in the HMA and 283 HHA groups compared with those in the HCD group (Fig. 6a-c), suggesting that the 284 addition of certain concentrations of sodium acetate to the HCD could improve 285 glucose homeostasis of fish. Additionally, the HSI was noticeably decreased in the 286 14 HMA and HHA groups (Fig. 6d). In parallel, liver TG levels were reduced in the 287 HMA and HHA groups (Fig. 6e, f). Accordingly, liver histology via H&E staining 288 showed that the percentage of the lipid area exhibited was reduced substantially in the 289 HMA and HHA groups compared with that in the HCD group (Fig. 6g, h). 290 Importantly, we also found that GLP-1 levels were elevated in the serum from the 291 HMA and HHA groups (Fig. 6i). Western blotting analysis revealed that levels of        Table S1. The total weight of fish in each tank was recorded every two weeks, and the 450 feeding amount was adjusted accordingly.  Table S2.

518
Quantification and statistical analysis was conducted using Image J. The total RNA was isolated from tissues by using the TRIzol Reagent (Magen, China).

551
The total RNA concentration was measured using a NanoDrop 2000C 552 spectrophotometer. RNA with an absorbance ratio OD 260/280 between 1.9 to 2.2 and 553 an OD 260/230 greater than 2.0 was used for subsequent analysis. As the template, 554 800 ng of total RNA was used to synthesize cDNA using a PrimeScript RT Reagent  All authors contributed experimental assistance and intellectual input to this study.  Quantitation of the levels of p-mTOR and p-S6 were normalized to that of GAPDH.

802
Data are expressed as mean ± SEM (n = 6). One-way ANOVA with Tukey's 803 adjustment was used for data analysis.  and p-mTOR were normalized to that of GAPDH. Data are expressed as mean ± SEM 822 (n = 6). One-way ANOVA with Tukey's adjustment was used for data analysis.