Carbohydrates are a relatively economical source of energy in aquafeeds, which could reduce the consumption of protein and lipid for energy(Reference Stone1). Accordingly, carbohydrates are widely used in aquafeeds, although fish has relatively no specific nutritional requirements(2), especially the carbohydrates of starch source, which has good binding and expansion properties and is the pivotal material for the production of puffed feed(Reference Sørensen, Nguyen and Storebakken3). However, numerous studies have found that carbohydrates could not be efficiently used by carnivorous fish(Reference Kamalam, Medale and Panserat4,Reference Polakof, Panserat and Soengas5) and long-term intake of high-carbohydrate diet (HCD) leads to varying degrees of damage in fish, including growth inhibition, excessive hepatic glycogen accumulation, intestinal microbial disruption and impaired antioxidant capacity(Reference Amoah, Coyle and Webster6–Reference Chen, Jiang and Liu9). Therefore, mitigating the negative effects of HCD in carnivorous fish could be helpful for the extensive application of carbohydrates in feed.
At present, with the ever more advanced research related to probiotics, specific functional probiotics have been demonstrated to partly mitigate the negative effects of HCD in fish(Reference Xu, Ding and Zhou10). Clostridium butyricum, as an anaerobic endospore-forming gram-positive bacterium, has the capacity to actively ferment carbohydrates, and its main ultimate metabolites are SCFA, including butyric, propionic and acetic acids(Reference Tran, Li and Ma11,Reference Cassir, Benamar and La Scola12) . In the past aquaculture-related studies, C. butyricum has been demonstrated to be effective in improving the immune response of tilapia (Oreochromis niloticus)(Reference Li, Zhou and Ling13), common carp (Cyprinus carpio L.)(Reference Meng, Wu and Hu14) and freshwater prawn (Macrobrachium rosenbergii)(Reference Sumon, Ahmmed and Khushi15). However, a recent review on mammals declared that SCFA besides their ability to inhibit inflammatory responses also have regulatory effect on glucose metabolism(Reference He, Zhang and Shen16). In past studies, it has been discovered that SCFA have a potential role in mediating the utilisation of glucose in the muscle and liver of mammals through the free fatty acid receptor-2 (FFAR2/GPR43) and free fatty acid receptor-3 (FFAR3/GPR41) pathways(Reference Kondo, Kishi and Fushimi17–Reference Zadeh-Tahmasebi, Duca and Rasmussen19). However, research on the effect of C. butyricum cultures (CBC) on glucose metabolism in fish is lacking.
A previous study showed that butyric acid (the predominant SCFA product of C. butyricum) improves insulin sensitivity and increases energy expenditure in mice(Reference Gao, Yin and Zhang20). Glucose utilisation cannot be achieved without the secretion and regulation of insulin, which are regulated through insulin signalling pathway. In general, through insulin binding to the insulin receptors (IR), phosphatidylinositol 3-kinase (PI3K) is attracted to the insulin receptor substrates (IRS), and thereby AKT is phosphorylated to regulate various metabolic processes such as glycolysis and gluconeogenesis(Reference Titchenell, Lazar and Birnbaum21). In addition, the insulin signalling pathway-related genes are also mediated by AMP-activated protein kinase (AMPK)(Reference Zheng, Yang and Wu22), and past studies have reported that SCFA can directly activate the AMPK signalling pathway by increasing the AMP/ATP ratio(Reference Kondo, Kishi and Fushimi17,Reference Gao, Yin and Zhang20) . Therefore, we speculate that investigating the changes in insulin signalling pathways and AMPK may provide a theoretical basis for the improvement of carbohydrate utilisation in largemouth bass by CBC in this study.
In mammals, diabetes triggers cellular damage, lipid peroxidation and insulin resistance as a result of increased free radical production or impaired antioxidant defence(Reference Maritim, Sanders and Watkins Iii23). In addition, ecological dysbiosis of gut microbiota is also considered as a symptom of diabetes(Reference Singer-Englar, Barlow and Mathur24). Also, carnivorous fish with diabetic features have been found to have impaired antioxidant capacity and gut microbiota at high carbohydrate levels(Reference Zhou, He and Jin25). Previous studies have revealed that the dietary inclusion of C. butyricum increased the antioxidant enzyme activities and facilitated the enrichment of beneficial intestinal microbiome in aquatic animals(Reference Duan, Wang and Dong26,Reference Duan, Zhang and Dong27) . Therefore, CBC as one of the options to solve the negative impacts of HCD has potential research value.
As a typical carnivorous fish, largemouth bass (Micropterus salmoides) has been widely cultured in China(28), and it is also a suitable model for glucose metabolism research due to its limited glucose utilisation(Reference Chen, Jiang and Liu9). Meanwhile, probiotic cultures are characterised by abundant metabolites and active bacteria, which are great choices to be used as functional additives(Reference Tao, Wang and Qiu29). Therefore, the aim of the present study was to investigate the potential effects of supplementation of CBC in HCD on carbohydrate utilisation, antioxidant capacity and intestinal microbiota of largemouth bass.
Materials and methods
Experimental diets
Previous study in our laboratory has demonstrated that, compared with 5 % inclusion level, 10 % α-starch produced abnormal hepatic glycogen accumulation and further impaired hepatic function of largemouth bass(Reference Liu, Liu and Wang30). Therefore, two isonitrogenous (crude protein content, 51 %) and isolipidic (crude lipid content, 11 %) diets formulated with 5 and 10 % α-starch were used as low-carbohydrate diet (LCD) and HCD (Table 1). Another three diets were supplemented with 0·25 % (HCP (0·25)), 0·50 % (HCP (0·50)) and 1·00 % (HCP (1·00)) CBC based on HCD (Table 1). The CBC used in present study was produced with liquid fermentation at 37°C for 20 h under anaerobic condition. The fermentation broth was centrifuged at 5000 × g for 10 min at 4°C, and the slurry was mixed with 1 % sterile water, then freeze-dried to obtain the bacterial powder. The cell quantity of bacterial powder was detected as 5·0 × 108 colony-forming unit (CFU)/g using iron sulfite agar method at 37°C for 24 h under anaerobic condition(Reference Harmon, Kautter and Peeler31). Meanwhile, the quantity of C. butyricum cells in the experimental diets was examined with the same method as above, and the results were 0·0 CFU/g in LCD and HCD groups, 6·2 × 105 CFU/g in HCP (0·25) group, 1·3 × 106 CFU/g in HCP (0·50) group, and 2·5 × 106 CFU/g in HCP (1·00) group. The experimental diets were prepared with reference to the procedure described previously(Reference Liu, Liu and Wang30) and then stored at -20°C after manufacture.
Experimental procedure
The present experiment strictly followed the requirement of Animal Care and Use Committee of Shanghai Ocean University and conducted in the constant temperature circulating water system of the joint laboratory of Shanghai Ocean University and Guangdong Evergreen Feed Co., Ltd. Juvenile largemouth bass sourced from our laboratory staff cultured from the larval stage (larval largemouth bass obtained from a commercial hatchery in Guangdong, China) to the specifications required for the present experiment. After that, juvenile largemouth bass (initial weight 35·03 ± 0·04 g) were randomly divided into fifteen reinforced plastic tanks (water volume of 1000 litres) with twenty-eight fish per tank, following anaesthesia with eugenol (1:10 000) (Shanghai Chemical Reagents Co., Ltd) and fed twice a day until apparent satiation using the experimental diet for 60 d, with each experimental diet being randomly assigned to three tanks. About 10 % of the water was replenished daily. All experimental tanks were under natural photoperiod, and water quality was monitored during the feeding trial: water temperature, 27 ± 1°C; pH, 7·2 ± 0·2; dissolved oxygen, 6–7 mg/l; ammonia nitrogen, < 0·05 mg/l.
Sample collection
Before sample collection, all experimental fish were fasted for 24 h, anesthetised with eugenol (1:10 000). The two individuals from each tank were randomly selected and stored at −20°C for body composition analysis. Then, twelve fish of each tank were randomly selected, after body weight and length measurement, blood sample was collected from tail vein, which was maintained at 4°C for 1 h, and then serum was separated by centrifugation at 3000 × g for 10 min at 4°C. Then, six of the above twelve fish were dissected foe viscerosomatic index (VSI) and hepatosomatic index (HSI) calculation, and muscle from lateral surface of fish above lateral line and whole liver were separated for biochemical composition analysis. Livers of the rest six fish were collected for gene expression analysis. The intestines of twelve fish were dissected into hindgut and anterior midgut, with the hindgut used for gene expression analysis and the anterior midgut contents scraped for intestinal microbiota analysis.
Biochemical analysis
The proximate composition, including moisture, crude protein and crude lipid content of whole fish, muscle and liver, as well as ash content of whole fish were analysed. Moisture content was measured by drying the samples in an oven (BPG-9140 A, Shanghai Yiheng Scientific Instrument Co., Ltd) at 105°C to a constant weight, and ash content was determined by burning the samples in a muffle furnace (SX2, Shanghai Laboratory Instrument Works Co., Ltd) at 550°C to a constant weight(32). Crude protein content was obtained with the Kjeldahl method (n × 6·25) by a Kjeldahl apparatus (Kjeltec 2200, Foss Analytical A/S)(32). Crude lipid content was determined by chloroform-methanol extraction method(Reference Folch, Lees and Sloane Stanley33). Glycogen content was detected with the commercial assay kit provided by Nanjing Jiancheng Bioengineering Institute (A043-1-1) by the potassium hydroxide/anthrone method(Reference Seifter, Dayton and Novic34).
The liver samples were homogenised by a tissue homogeniser (Tiss-Basic2, Shanghai Jingxin Industrial Development Co., Ltd) with ice-cold phosphate buffer (1:9, w/v), which were then centrifuged (3500 × g , 10 min, 4°C) to separate the supernatant for antioxidant capacity analysis with commercial assay kits provided by Nanjing Jiancheng Bioengineering Institute (A045-2-2, A007-1-1, A001-3-2, A003-1-1 and A015-2-1). The soluble protein content was assayed with the Coomassie brilliant blue method(Reference Bradford35). The malondialdehyde was measured with thiobarbituric acid-reactive substances method(Reference Buege36). The activity of total antioxidative capacity was determined by 2,2'-azinobis(3-ethylbenzothiazoline-6-sulfonate) (ABTS+)-based methods(Reference Erel37). The activity of superoxide dismutase (SOD) was detected using xanthine oxidase and the water-soluble tetrazolium (WST-1)(Reference Peskin and Winterbourn38). The activity of catalase (CAT) was determined by measuring the variation in the concentration of hydrogen peroxide(Reference Aebi39).
RNA extraction and real-time quantitative PCR
Total RNA of liver and hindgut was extracted, respectively, with RNAiso Plus. After RNA quality being tested, it was reverse-transcribed to first-strand cDNA, used for real-time quantification PCR, with the Prime Script™ RT reagent Kit. Real-time quantification PCR was conducted with quantitative thermal cycler (Mastercycler EP Realplex, Eppendorf) with the following procedures: 95°C for 2 min, followed by 40 cycles of 95°C for 10 s, 57°C for 10 s, 72°C for 20 s and a melting curve analysis. The specific primers used for real-time quantification PCR are shown in Table 2. The relative gene expression was calculated with the 2-ΔΔCt method(Reference Livak and Schmittgen40), and β-actin was used as the reference gene.
Table 1. Formulation and proximate composition of experimental diets (% DM)
LCD, low-carbohydrate diet; HCD, high-carbohydrate diet.
* Supplied by Zhejiang Xinxin Tian’en Aquatic Feed Corporation (Jiaxing, China): fishmeal, crude protein, 71·84 %, crude lipid, 7·66 %; shrimp meal, crude protein, 61·72 %, crude lipid, 11·20 %; fermented soyabean meal, crude protein, 51·67 %, crude lipid, 2·62 %; maize gluten meal, crude protein, 60·41 %, crude lipid, 5·12 %; blood meal, crude protein, 87·06 %, crude lipid, 1·51 %; wheat gluten meal, crude protein, 80·05 %; crude lipid, 1·13 %; brewer’s yeast, crude protein, 51·05 %, crude lipid, 0·52 %.
† Vitamin Premix (mg/kg diet): vitamin B1, 17·80; vitamin A, 16 000 mg; vitamin B2, 48; vitamin B6, 29·52; vitamin B12, 0·24; vitamin C, 800; vitamin D3, 8000 mg; vitamin E, 160; vitamin K3, 14·72; choline chloride, 1500; folic acid, 6·40; niacinamide, 79·20; inositol, 320; calcium pantothenate, 73·60; biotin, 0·64.
‡ Mineral Premix (mg/kg diet): I (Ca (IO3)2), 1·63; Mn (MnSO4), 6·20; Cu (CuSO4), 2·00; Fe (FeSO4), 21·10; Se (Na2SeO3), 0·18; Co (CoCl2), 0·24; Zn (ZnSO4), 34·40.
Table 2. Primers used in the present study
nrf2, nuclear factor erythroid 2-related factor 2; keap1, Kelch-like ECH-associated protein 1; sod1, superoxide dismutase 1; sod2, superoxide dismutase 2; cat, catalase; ampkα1, AMP-activated protein kinase α 1; ira, insulin receptor a; irb, insulin receptor b; irs, insulin receptor substrate; pi3kr1, phosphatidylinositol-3-kinase p85 α; akt1, serine/threonine kinase 1; gk, glucokinase; pfkl, phosphofructokinase liver type; pk, pyruvate kinase; pepck, phosphoenolpyruvate carboxykinase; fbp1, fructose-1,6-bisphosphatase-1; g6pc, glucose-6-phosphatase catalytic subunit; ffar3, free fatty acid receptor 3; mct1, monocarboxylate transporter 1.
DNA extraction, amplification and sequencing of gut 16S rRNA genes
The present results showed that the dietary inclusion of 1·0 % CBC had a significant positive effect on carbohydrate utilisation and antioxidant capacity of largemouth bass. Therefore, considering the logistic and cost limitations, microbiological analysis was performed only in HCD and HCP (1·00) (defined as HCP group in the intestinal microbiota analysis) groups. Bacterial DNA from HCD and HCP groups was extracted with the E.Z.N.A.® Soil DNA kit (Omega). DNA integrity and concentration were determined using 1% agarose gel electrophoresis and NanoDrop 2000 UV-Vis spectrophotometer (Thermo Scientific), respectively. Then, amplification of the hypervariable region V3–V4 of the bacterial 16S rRNA gene was performed in a PCR thermal cycler, using the 338F/806R primer pair, with the following procedures: 95°C for 3 min, followed by 27 cycles of 95°C for 30 s, 55°C for 30 s and 72°C for 45 s, and 72°C for 10 min. After the PCR products were extracted and purified, the paired-end sequencing of the purified products was conducted using the Illumina MiSeq PE300 platform (Shanghai Bio-Pharm Technology Co. Ltd). The raw reads were stored in the NCBI Serial Read Archive (SRA) database with the accession number PRJNA962001.
Bioinformatics analysis
The raw 16S rRNA sequences were demultiplexed and quality-filtered using the Quantitative Insights into Microbial Ecology (QIIME) quality filters(Reference Caporaso, Kuczynski and Stombaugh41) and then were processed with FLASH software to merge them(Reference Magoč and Salzberg42). After that, operational taxonomic units according to the 97 % similarity threshold were clustered using UPARSE pipeline and chimeric sequences were removed using UCHIME(Reference Edgar43). The phylogenetic affiliation of each operational taxonomic unit representative 16S rRNA gene sequence was analysed by RDP Classifier based on the 16S rRNA database (Silva 136) according to a confidence threshold of 70 %(Reference Wang, Garrity and Tiedje44). The α-diversity analysis and β-diversity analysis were calculated with QIIME and displayed with R software, for evaluating changes in anterior midgut micro-organisms of largemouth bass with dietary supplementation of CBC. Largemouth bass anterior midgut bacterial diversity and operational taxonomic unit abundance in HCD and HCP groups were statistically analysed using Welch’s t test, with P < 0·05 as significant. Linear discriminant analysis effect size analysis was used to identify the different abundant taxa between the HCD and HCP groups(Reference Segata, Izard and Waldron45).
Calculation and statistical analysis
The relevant variables were obtained using the following formula:
All statistical assessments were performed after data normality and homoscedasticity tests with SPSS 19.0 for Windows. The data of the HCD treatments and the LCD group were analysed by Dunnett’s test. All data from HCD treatments were statistically analysed by one-way ANOVA with Duncan’s multiple range test on SPSS 19.0 for Windows. In addition, depending on the experimental design and to evaluate possible trends of experimental results, polynomial comparisons (linear and quadratic) were used to detect all the variables and parameters tested in HCD treatments. When linear and quadratic significance regressions were detected simultaneously, linearity, as the first priority, was chosen to describe the observed trend. In the present study, experimental results were presented as mean ± sem, and the significance was set at P < 0·05.
Results
Growth performance and feed utilisation
No mortality was observed in all experimental fish in this study. Compared with the LCD group, final body weight and SGR were significantly reduced in the HCD group; however, final body weight and SGR were linearly increased with inclusion of CPC in the HCD (P < 0·05) (Table 3). In addition, HSI and VSI were significantly increased with increasing levels of carbohydrates (P < 0·05) (Table 3). With the inclusion of CPC in the HCD, both HSI and VSI were linearly reduced, but VSI was consistently higher than the LCD group (P < 0·05) (Table 3). However, no statistical difference was produced in the CF (P > 0·05) (Table 3). The relevant parameters of feed utilisation, FCR and FI, in HCD treatments were observed to be significantly lower than in the LCD group, although FI showed a linear increase with the addition of CBC in the HCD (P < 0·05) (Table 3). However, PER from all the HCD treatments was significantly higher than that of the LCD group (P < 0·05).
Table 3. Effects of dietary CBC on the growth performance and feed utilisation of largemouth bass
CBC, Clostridium butyricum cultures; LCD, low-carbohydrate diet; HCD, high-carbohydrate diet; NA, not applicable; IBW, initial body weight; FBW, final body weight; SGR, specific growth rate; CF, condition factor; HSI, hepatosomatic index; VSI, viscerosomatic index; FCR, feed conversion ratio; PER, protein efficiency ratio; FI, feed intake.
Dunnett’s test was conducted for comparing LCD and other dietary groups, and the significantly different values (n 3) within a row were marked ‘*’ at the superscript (P < 0·05). Duncan’s multiple range test was performed in all high-carbohydrate diet groups, and values within a row with a common superscript letter are not significantly different from the other dietary groups (P > 0·05) but with a different superscript letter are significantly different from the other dietary groups (P < 0·05).
Body composition analysis
The crude protein and ash contents of whole fish did not show statistical differences in all treatments (P > 0·05) (Table 4). Compared with the LCD group, the moisture content of whole fish did not produce significant differences in HCD group (P > 0·05), and the crude lipid content of whole fish in the HCD was significantly increased (P < 0·05) (Table 4). The moisture content of whole fish increased in a quadratic manner with the addition of dietary CBC, while crude lipid content of whole fish linearly decreased (P < 0·05) (Table 4). In addition, the moisture and crude protein content of liver and muscle tissues presented no significant differences among all treatments (P > 0·05) (Table 4). The crude lipid content of liver was significantly decreased in HCD group compared with LCD treatment and increased quadratically with the dietary inclusion of CBC (P < 0·05) (Table 4). Glycogen content of liver and crude lipid content of muscle had similar trends, and both were enhanced in HCD group compared with LCD group and showed the linear decrease with a dose-dependent manner in HCD treatments (P < 0·05) (Table 4).
Table 4. Effect of dietary CBC on the body composition (% wet weight) of largemouth bass
CBC, Clostridium butyricum cultures; LCD, low-carbohydrate diet; HCD, high-carbohydrate diet.
Dunnett’s test was conducted for comparing LCD and other dietary groups, and the significantly different values (n 3) within a row were marked ‘*’ at the superscript (P < 0·05). Duncan’s multiple range test was performed in all high-carbohydrate diet groups, and values within a row with a common superscript letter are not significantly different from the other dietary groups (P > 0·05) but with a different superscript letter are significantly different from the other dietary groups (P < 0·05).
Expression of hepatic ampkα1 and insulin pathway-related genes
The relative expression of AMP-activated protein kinase α 1 (ampkα1) increased linearly in a dose-dependent manner, but the expression was dramatically decreased in the HCD and HCP (0·25) groups compared with the LCD group (P < 0·05) (Fig. 1(a)). The results of insulin receptor a (ira) and insulin receptor b (irb) expression were observed similar to ampkα1, both showing a linear increase in the HCD treatments, and the HCD and HCP (0·25) groups remained statistically different with the LCD group (P < 0·05) (Fig. 1(b–c)). Meanwhile, CBC inclusion led to a linear increase of insulin receptor substrate (irs) and phosphatidylinositol-3-kinase p85 α (pi3kr1) expression in HCD treatments (P < 0·05) (Fig. 1(d–e)). The results of comparison with the LCD group indicated that the expression of irs in the HCD group and pi3kr1 in the HCD and HCP (0·25) groups was significantly reduced (P < 0·05) (Fig. 1(d–e)). In addition, the variation of serine/threonine kinase 1 (akt1) expression was also consistent with a linear increase; however, the expression in all HCD treatments was still significantly lower than in the LCD group (P < 0·05) (Fig. 1(f)).
Fig. 1. Relative expression of AMP-activated protein kinase α 1 (ampkα1) (a) and insulin signalling pathway-related genes, insulin receptor a (ira) (b), insulin receptor b (irb) (c), insulin receptor substrate (irs) (d), phosphatidylinositol-3-kinase p85 α (pi3kr1) (e) and serine/threonine kinase 1 (akt1) (f), in liver of largemouth bass fed the experimental diets for 60 d. Dunnett’s test was conducted for comparing LCD and other dietary groups, and the significantly different values (mean ± sem, n 3) were marked ‘*’ at the superscript (P < 0·05). Duncan’s multiple range test was performed in all high-carbohydrate diet groups, and values with a different superscript letter are significantly different from the other dietary groups (P < 0·05). Polynomial comparison results: ampkα1, P linear = 0·005, R2 linear = 0·284; ira, P linear = 0·007, R2 linear = 0·534; irb, P linear = 0·008, R2 linear = 0·677; irs, P linear = 0·001, R2 linear = 0·656; pi3kr1, P linear = 0·001, R2 linear = 0·675; akt1, P linear = 0·010, R2 linear = 0·497. LCD, low-carbohydrate diet; HCD, high-carbohydrate diet.
Expression of hepatic glucose metabolism-related genes
The inclusion of CBC in HCD did not significantly influence the relative expression of glucokinase (gk), and no significant difference was found between the LCD group and HCP (1·00) group (P > 0·05) (Fig. 2(a)). Meanwhile, the relative expression of phosphofructokinase liver type (pfkl) reached its peak in a linear manner with dietary inclusion of CBC, while the expression in the HCD group became significantly lower as compared with the LCD group (P < 0·05) (Fig. 2(b)). In addition, the relative expression of pyruvate kinase (pk) showed the similar linear increase in the HCD treatments (P < 0·05) and did not produce a statistical difference with the LCD group (P > 0·05) (Fig. 2(c)).
Fig. 2. Relative expression of glycolysis, glucokinase (gk) (a), phosphofructokinase liver type (pfkl) (b) and pyruvate kinase (pk) (c), and gluconeogenesis, fructose-1,6-bisphosphatase-1 (fbp1) (d), glucose-6-phosphatase catalytic subunit (g6pc) (e) and phosphoenolpyruvate carboxykinase (pepck) (f), in liver of largemouth bass fed the experimental diets for 60 d. Dunnett’s test was conducted for comparing LCD and other dietary groups, and the significantly different values (mean ± sem, n 3) were marked ‘*’ at the superscript (P < 0·05). Duncan’s multiple range test was performed in all high carbohydrate diet groups, and values with a different superscript letter are significantly different from the other dietary groups (P < 0·05). Polynomial comparison results: pfkl, P linear = 0·002, R2 linear = 0·623; pk, P linear = 0·019, R2 linear = 0·438; fbp1, P linear = 0·033, R2 linear = 0·379. LCD, low-carbohydrate diet; HCD, high-carbohydrate diet.
The relative expression of fructose-1,6-bisphosphatase-1 (fbp1), glucose-6-phosphatase catalytic subunit (g6pc) and phosphoenolpyruvate carboxykinase (pepck) did not produce significant differences between the HCD and LCD groups (P > 0·05) (Fig. 2(d–f)). With the inclusion of CBC in HCD, the expression of fbp1 presented a linear increase (P < 0·05), while the expression of g6pc and pepck did not produce statistical differences (P > 0·05) (Fig. 2(d–f)).
Expression of intestinal SCFA transport-related genes
The relative expression of intestinal free fatty acid receptor 3 (ffar3) in largemouth bass was linearly elevated in the HCD treatments, and the expression was significantly reduced in the HCD group compared with the LCD group (P < 0·05) (Fig. 3(a)). Meanwhile, with the dietary addition of CBC, a quadratic rise in the relative expression of monocarboxylate transporter 1 (mct1) appeared in the largemouth bass gut (P < 0·05), but all HCD treatments did not produce significant differences with the LCD group (P > 0·05) (Fig. 3(b)).
Fig. 3. Relative expression of the SCFA transport-related genes, free fatty acid receptor 3 (ffar3) (a) and monocarboxylate transporter 1 (mct1) (b), in gut of largemouth bass fed the experimental diets for 60 d. Dunnett’s test was conducted for comparing LCD and other dietary groups, and the significantly different values (mean ± sem, n 3) were marked ‘*’ at the superscript (P < 0·05). Duncan’s multiple range test was performed in all high-carbohydrate diet groups, and values with a different superscript letter are significantly different from the other dietary groups (P < 0·05). Polynomial comparison results: ffar3, P linear = 0·024, R2 linear = 0·412; mct1, P quadratic = 0·029, R2 quadratic = 0·546. LCD, low-carbohydrate diet; HCD, high-carbohydrate diet.
Hepatic antioxidant capacity
Largemouth bass hepatic malondialdehyde concentration was significantly higher in the HCD group than LCD group and declined in a linear manner with the addition of CBC (P < 0·05) (Table 5). Statistical analysis with the LCD group indicated that antioxidant enzyme activities, including CAT and total superoxide dismutase, were significantly lower in the HCD group (P < 0·05) (Table 5). Unlike change in malondialdehyde, both CAT and total superoxide dismutase activities showed a linear increase in all HCD treatments (P < 0·05) (Table 5). The variation of total antioxidant capacity was similar to that of antioxidant enzyme activities, presenting lower in HCD group than LCD group and a linear increase with inclusion of CBC in the HCD (P < 0·05) (Table 5).
Table 5. Effect of dietary CBC on the hepatic antioxidant enzyme activities of largemouth bass
CBC, Clostridium butyricum cultures; LCD, low-carbohydrate diet; HCD, high-carbohydrate diet; MDA, malondialdehyde; T-AOC, total antioxidant capacity; CAT, catalase; T-SOD, total superoxide dismutase.
Dunnett’s test was conducted for comparing LCD and other dietary groups, and the significantly different values (n 3) within a row were marked ‘*’ at the superscript (P < 0·05). Duncan’s multiple range test was performed in all high-carbohydrate diet groups, and values within a row with a common superscript letter are not significantly different from the other dietary groups (P > 0·05) but with a different superscript letter are significantly different from the other dietary groups (P < 0·05).
At the transcriptional level, variations of hepatic antioxidant enzyme-related gene expression corresponded to antioxidant enzyme activities, with the relative expression of catalase (cat), superoxide dismutase 1 (sod1) and superoxide dismutase 2 (sod2) being increased linearly among the HCD treatments (P < 0·05) (Fig. 4(c–e)). Compared with the LCD group, the relative expressions of cat and sod1 were significantly lower in the HCD and HCP (0·25) groups, while the relative expression of SOD2 still remained remarkably lower than the LCD group with increasing dose of CBC (P < 0·05) (Fig. 4(c–e)). Meanwhile, the relative expression of nuclear factor erythroid 2-related factor 2 (nrf2) displayed a linear increase in all HCD treatments (P < 0·05), but there was no statistical difference between all HCD treatments and the LCD group (P > 0·05) (Fig. 4(a)). However, the relative expression of keap1 decreased linearly with the inclusion of CBC and the expression in the HCD group was significantly higher than that in the LCD group (P < 0·05) (Fig. 4(b)).
Fig. 4. Relative expression of Nrf2/Keap1 signalling pathway-related genes, nuclear factor erythroid 2-related factor 2 (nrf2) (a), Kelch-like ECH-associated protein 1 (keap1) (b), catalase (cat) (c), superoxide dismutase 1(sod1) (d) and superoxide dismutase 2 (sod2) (e), in liver of largemouth bass fed the experimental diets for 60 d. Dunnett’s test was conducted for comparing LCD and other dietary groups, and the significantly different values (mean ± sem, n 3) were marked ‘*’ at the superscript (P < 0·05). Duncan’s multiple range test was performed in all high-carbohydrate diet groups, and values with a different superscript letter are significantly different from the other dietary groups (P < 0·05). Polynomial comparison results: nrf2, P linear = 0·008, R2 linear = 0·526; keap1, P linear = 0·011, R2 linear = 0·491; cat, P linear < 0·001, R2 linear = 0·801; sod1, P linear < 0·001, R2 linear = 0·838; sod2, P linear < 0·001, R2 linear = 0·717. LCD, low-carbohydrate diet; HCD, high-carbohydrate diet.
Microbiota composition and diversity analysis
Alpha diversity analysis revealed that the Chao index, Ace index and Simpson index of intestinal microbiota in juvenile largemouth bass from the HCP group were slightly lower compared with the HCD group, while the Shannon index was slightly higher (P > 0·05) (Fig. 5). The results of β-diversity comparisons analysis, including principal coordinate analysis, non-metric multidimensional scaling and hierarchical clustering tree, demonstrated that the intestinal microbiota of largemouth bass in HCP group differed from the HCD group (Fig. 6).
Fig. 5. The α-diversity comparisons analysis, including Chao species richness index (a), Ace species richness index (b), Shannon diversity index (c) and Simpson diversity index (d) of microbial communities in the gut of largemouth bass between the HCD and HCP groups. Values (mean ± sem) in bars that have no asterisks are not significantly different (P > 0·05; Welch’s t test) between HCD and HCP groups (n 3). HCD, high-carbohydrate diet.
Fig. 6. The β-diversity comparisons analysis, including principal component analysis (PCoA) (a), non-metric multidimensional scaling (NMDS) (b) and hierarchical clustering tree (c) of microbial communities at genus level in the gut of largemouth bass between the HCD and HCP groups. HCD, high-carbohydrate diet. OUT, operational taxonomic units.
The bar charts showed the dominant intestinal bacterial communities at the phylum and genus levels for the two groups, respectively (Fig. 7). The Firmicute, Fusobacteriota and Proteobacteria were the dominant phylum in HCD and HCP groups (Fig. 7). Meanwhile, the results of community analysis at the genus level revealed that the Mycoplasma, Cetobacterium and Achromobacter were the predominant microbes (Fig. 7). Welch’s t test showed that the dietary supplementation of CBC significantly increased the relative abundance of Fusobacteriota and Cetobacterium in the gut of largemouth bass fed the HCD and significantly decreased the relative abundance of Firmicute and Mycoplasma (Fig. 7).
Fig. 7. Relative abundances (%) of intestinal bacteria and comparison of bacterial abundances in the gut of largemouth bass from HCD and HCP groups at the phylum (a, b) and genus (c, d) level, and the phyla and genera with relative abundances lower than 1 % were assigned as ‘others’. *0·01 < P ≤ 0·05, **0·001 < P ≤ 0·01 (Welch’s t test, n 3). HCD, high-carbohydrate diet.
Venn diagram analysis showed that a total of ten common phyla were detected in the HCD and HCP groups, while five and two unique bacterial phyla were identified in the HCD and HCP groups, respectively (Fig. 8). At the genus level, a total of fifty-two shared genera were identified in the HCD and HCP groups, and 100 and twenty-one unique genera were found in the HCD and HCP groups, respectively (Fig. 8). The relatively abundant genera (>5 %) among the shared genera were Mycoplasma (52·17 %), Cetobacterium (33·06 %) and Achromobacter (12·76 %) (Fig. 8). The unique intestinal genera (>5 %) of largemouth bass in the HCD group were Exiguobacterium (5·79 %); meanwhile, the unique intestinal genera (>5 %) in the HCP group were Geodermatophilus (24·26 %), Parviterribacter (22·06 %), Blautia (6·62 %), Microvirga (6·62 %), Eubacterium (5·88 %) and Deinococcus (5·15 %) (Fig. 8).
Fig. 8. Venn diagram analysis of microbial communities in the gut of largemouth bass between the HCD and HCP groups. The number of microbial communities (a) and overlapping (b) and unique (c, d) bacterial species at phylum level in the gut of largemouth bass were identified (Fig. 8). The number of microbial communities (a) and overlapping (b) and unique (c, d) bacterial species at genus level in the gut of largemouth bass were identified (Fig. 8). HCD, high-carbohydrate diet.
The results of largemouth bass gut microbial community data analysed by linear discriminant analysis effect size from domain to genus level revealed significant difference in the taxonomic distribution of gut microbial communities between HCD and HCP groups (Fig. 9). The gut of largemouth bass fed the diet with CBC exhibited significant enrichment for genus Cetobacterium and Fusobacteriaceae (from phylum to family) (Fig. 9). In addition, the HCD group had a higher relative bacterial abundance in gut of largemouth bass, including phylum Firmicutes, class Bacilli, Mycoplasma (from order to genus) (Fig. 9).
Fig. 9. Cladogram showing the phylogenetic distribution of the bacterial lineages associated with inclusion of CBC in HCD. Taxonomic representation of statistically and biologically consistent differences among intestinal microbiota of largemouth bass between the HCD and HCP groups (a). Differences were represented by the colour for the most abundant class (red indicates HCD group and blue indicates HCP group). Histogram of linear discriminant analysis (LDA) scores for differentially abundant taxon (b). For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article. CBC, Clostridium butyricum cultures; HCD, high-carbohydrate diet.
Discussion
C. butyricum has the properties of high temperature resistance and fermentation of carbohydrates to produce SCFA, and it is a selectable functional additive for aquafeeds at present(Reference Tran, Li and Ma11). In mammals, C. butyricum has been demonstrated to prevent and treat hyperglycemia and associated metabolic dysfunction(Reference Jia, Li and Feng46). Based on the beneficial role of SCFA in regulating energy metabolism(Reference He, Zhang and Shen16) and the limited glucose utilisation capacity of largemouth bass(Reference Gou, Chen and Xu47), the present experiment was conducted to investigate the effect of CBC in largemouth bass fed HCD. Past studies have shown that the effect of C. butyricum on growth performance may be related to the level of dietary supplementation in teleosts(Reference Zhang, Dong and Lai48–Reference Meng, Cai and Li50). In the present study, largemouth bass final body weight and SGR produced a linear increase with the inclusion of CBC in HCD, but dietary CBC inclusion did not product statistical effect compared with the HCD group, which may be related to the dose of supplementation. In addition, the elevated HSI and VSI caused by the HCD were alleviated with the addition of CBC. The change in HSI and VSI might indicate that liver problems due to excessive accumulation of hepatic glycogen, such as impaired liver function(Reference Xu, Chen and Liu51), were partly repaired. A former study has been shown that higher carbohydrate levels lead to alterations in hybrid grouper (Epinephelus fuscoguttatus ♀ × E. lanceolatus ♂) feed utilisation, as evidenced by lower FI and FCR and higher PER(Reference Li, Li and Zhang8). Meanwhile, the reduction of FI in high carbohydrate levels was also found in largemouth bass(Reference Harmon, Kautter and Peeler31). Feed utilisation variations in carnivorous fish associated with HCD were suggested to be caused by excessive total energy intake and therefore improved protein-sparing effect of dietary carbohydrate(Reference Stone, Allan and Anderson52,Reference Li, Sang and Turchini53) . The results of the present study showed that HCD with the inclusion of CBC did not have a substantial effect on feed utilisation of largemouth bass, and the influence of different carbohydrate content on feed utilisation was similar to those described above.
In general, excessive accumulation of hepatic glycogen is one of the characteristics of carnivorous fish with diabetes-like features fed HCD(Reference Hemre, Mommsen and Krogdahl7,Reference Li, Sang and Zhang54,Reference Wang, Li and Han55) . In the present study, hepatic glycogen content of largemouth bass was significantly elevated in the HCD group, while overaccumulation of glycogen in largemouth bass liver was alleviated with dietary supplementation of CBC. The modulatory effect of C. butyricum on fish hepatic glycogen at high dietary carbohydrate levels has not been reported yet, but its metabolite butyric acid has been shown to reduce blood glucose level in type 2 diabetic mice(Reference Jia, Li and Feng46). Therefore, the insulin signalling pathway and AMPK were analysed at the transcription level in order to further validate the change in carbohydrate utilisation of largemouth bass after feeding CBC.
The insulin signalling pathway is a key pathway for maintaining glucose homoeostasis through regulating glucose metabolism(Reference Titchenell, Lazar and Birnbaum21,Reference Beale56) . However, the expression of insulin signalling pathway-related genes are generally inhibited in carnivorous fish fed HCD(Reference Capilla, Médale and Navarro57–Reference Polakof, Moon and Aguirre61). In the present study, the HCD also inhibited insulin signalling; however, the addition of CBC improved this situation as evidenced by increased expression of ira, irb and irs. This is similar to the results found in butyrate-related studies of chickens(Reference Mátis, Kulcsár and Turowski62) and rats(Reference Aguilar, da Silva and Navia-Pelaez63). Activation of ir and irs usually triggers the downstream stimulation of PI3K/AKT1 pathway and thus promotes the role of insulin for regulating metabolism(Reference Titchenell, Lazar and Birnbaum21). In the present study, the expression of pi3kr1 and akt1 increased linearly with the inclusion of CBC in HCD, which is consistent with the variation tendency of insulin pathway-related genes(Reference Zhou, Xu and Zhang64). In addition, the previous studies have shown that AMPK is always involved in the regulation of glucose metabolism(Reference O’Neill, Maarbjerg and Crane65), and the phosphorylation and activation of insulin receptor are also associated with AMPK(Reference Chopra, Li and Wang66). In the present study, the expression of ampkα1 of largemouth bass liver was significantly elevated with the inclusion of CBC in HCD, and this result was similar to that in studies of SCFA(Reference Yoshida, Ishii and Akagawa67). The above results demonstrate the activation of insulin signalling pathway and AMPK by CBC in largemouth bass fed HCD.
In general, activation of the insulin signalling pathway and AMPK improves glucose metabolism by promoting glycolysis and inhibiting gluconeogenesis(Reference Chopra, Li and Wang66,Reference Shi, Liu and Xu68) . In present study, the inclusion of CBC in HCD significantly enhanced the expression of glycolysis-related genes, such as pfkl and pk. Similar results were reported from the study correlated with sodium butyrate in type 2 diabetic mellitus(Reference Zhang, Zhao and Gui69). However, in the present study, largemouth bass gluconeogenesis was not significantly inhibited with inclusion of CBC, which has been reported in other largemouth bass studies(Reference Chen, Jiang and Liu9,Reference Harmon, Kautter and Peeler31) . The potential mechanism for the regulation of gluconeogenesis by CBC in largemouth bass required further investigated. In summary, the activation of glycolysis partially explains dietary CBC inclusion alleviates the hepatic glycogen accumulation of largemouth bass induced by HCD.
A previous study showed that butyric could promote the phosphorylation of AMPK and improve insulin sensitivity in mammals(Reference Gao, Yin and Zhang20,Reference Hong, Jia and Pan70) . It may imply that the metabolites of C. butyricum are the main reason for the above changes. At present, many researches associated with C. butyricum on aquaculture have clearly demonstrated that dietary C. butyricum inclusion would increase the content of intestinal SCFA in various fish and crustacean(Reference Meng, Wu and Hu14,Reference Li, Hou and Zhao49,Reference Meng, Cai and Li50,Reference Duan, Zhang and Dong71) . However, there are fewer studies targeting SCFA transport-related genes in C. butyricum research on aquatic animals, which is probably a key factor for improving carbohydrate utilisation of largemouth bass in the present study(Reference He, Zhang and Shen16). The ffar3 and mct1 are the main vehicles for propionic and butyric acids to enter cells, and ffar3 is also an effective target for the treatment of type 2 diabetes(Reference Milligan, Stoddart and Smith72–Reference Yao, Cai and Fei74). In the present study, largemouth bass intestinal ffar3 and mct1 expression increased linearly in the HCD treatments, which probably indicates that SCFA, especially propionic acid and butyric acid, entered more into the largemouth bass cells with dietary inclusion of CBC. Meanwhile, this is also the potential reason for the improved carbohydrate utilisation of largemouth bass in the present study.
HCD usually cause damage to the antioxidant capacity in largemouth bass(Reference Harmon, Kautter and Peeler31,Reference Ma, Mou and Pu75) . It has been reported the beneficial effect of C. butyricum on improving antioxidant capacity in mammals(Reference Li, Shang and Liu76), fish(Reference Zhang, Dong and Lai48) and shrimps(Reference Wangari, Gao and Sun77). In the present study, lipid peroxidation products malondialdehyde and total antioxidative capacity showed a linear decrease and increase with inclusion of CBC, respectively, which partly indicates greater antioxidant enzyme activities and reduced reactive oxygen species. Meanwhile, antioxidant enzyme activities, including CAT and total superoxide dismutase, also increased linearly with the inclusion of CBC in the HCD. The above changes were similar to the results in tilapia(Reference Zhang, Dong and Lai48). In addition, the expression of antioxidant enzyme-related genes, including cat, sod1 and sod2, followed similar trends as enzyme activity levels, showing increased linearly in the HCD treatments. The results of enzyme activity and transcript levels partly demonstrate that the inclusion of CBC has a promotive impact on the antioxidant capacity of largemouth bass with HCD treatments. In mammals, the expression of antioxidant enzyme-related genes was identified to be regulated by the Nrf2/Keap1 signalling pathway(Reference Lee, Calkins and Chan78,Reference Reisman, Yeager and Yamamoto79) . In the present study, the expression of nrf2 and keap1 showed the linear increase and decrease with the dietary inclusion of CBC, respectively. Activation of antioxidant enzyme-related genes through similar variations has also been observed in zebrafish (Danio rerio)(Reference Timme-Laragy, Van Tiem and Linney80), blunt snout bream (Megalobrama amblycephala)(Reference Liang, Mokrani and Ji81) and common carp(Reference Shang, Yu and Yin82). Therefore, the beneficial role of CBC in the antioxidant capacity of juvenile largemouth bass fed HCD is suggested to be mediated by Nrf2/Keap1signalling pathway.
The formation and establishment of fish intestinal microbiota is a complex process, influenced by feed and water environment(Reference Huang, Zhou and Chen83,Reference Nayak84) . In the present study, Firmicutes, Fusobacteria and Proteobacteria were the dominant phyla in the intestine of largemouth bass. It was partly consistent with the results of another high-carbohydrate-related study in largemouth bass, with the dominant phylum Fusobacteria, Tenericutes, Firmicutes and Proteobacteria (Reference Chen, Jiang and Liu9). Past studies in teleosts have identified HCD commonly causing dysbiosis of intestinal microbiota(Reference Zhou, He and Jin25,Reference Huang, Zhong and Kang85) . In general, the inclusion of probiotics in diets would change the diversity or abundance of intestinal microbiota in fish species(Reference Li, Zhou and Ling13,Reference Liu, Li and Guo86,Reference Guimarães, da Silva Guimarães and Natori87) . In the present study, the inclusion of CBC significantly reduced the relative abundance of the Firmicutes in the intestine of largemouth bass following HCD. Past studies observed the potential relationship between the Firmicutes and type 2 diabetes with higher abundance in type 2 diabetics(Reference Remely, Hippe and Zanner88,Reference Hung, Hung and Tsai89) . Meanwhile, the abundance of Firmicutes was found to have a positive correlation with serum glucose content in grass carp(Reference Pan, Li and Xie90). Therefore, the decreased abundance of Firmicutes may have potentially positive consequences for largemouth bass with HCD. Mycoplasma is a common bacterium found in the water environment and aquatic animals, and it has been discovered in study of tilapia that infection with Mycoplasma led to extensive and marked inflammatory lesions in the gills, liver and spleen(Reference Ahmed and El-Hady91). The decreased abundance of Mycoplasma (from phylum to genus) probably indicates the reduction of potentially pathogenic bacteria from the largemouth bass intestine. In past studies, Fusobacteria has been characterised by the consumption of carbohydrates and peptones to produce butyric acid(Reference Arellano, Jomard and El Kaddouri92). Meanwhile, Cetobacterium is considered as beneficial intestinal bacteria that favours anaerobic metabolism in fish(Reference Zhang, Fan and Yi93), and acetate-producing Cetobacterium somerae has been reported to play a causative role for improving glucose homoeostasis in zebrafish(Reference Wang, Zhang and Ding94). In the present study, the dietary inclusion of CBC significantly increased the relative abundance of Fusobacteria and Cetobacterium in the intestine of largemouth bass. This might be potentially linked to the above which mentioned changes in SCFA transport and insulin pathway-related genes.
In conclusion, dietary inclusion of CBC in HCD significantly alleviated glycogen accumulation and improved carbohydrate utilisation in largemouth bass, possibly through activating of AMPK and insulin signalling pathways. Also, dietary inclusion of CBC significantly improved the antioxidant capacity and intestinal microbiota of largemouth bass fed HCD. However, the underlying regulatory mechanisms of SCFA in glucose metabolism needed to be further explored.
Acknowledgements
The authors thank Zheng Chen and Zishuo Fang for their help in experimental diets preparation and sampling.
This work was financially supported by National Key Research and Development Program of China (2022YFD2400803) and ‘Chenguang program’ supported by Shanghai Education Development Foundation and Shanghai Municipal Education Commission (19CG56).
J. T.: Investigation, Data curation, Formal analysis, and Writing – Original Draft; Y. G.: Methodology, Writing – Review and Editing; S. C.: Investigation and Methodology; W. L.: Investigationand Methodology; R. X.: Investigation; H. Z.: Methodology; N. C.: Conceptualisation and Project administration; X. H.: Conceptualisation and Funding acquisition; S. L.: Conceptualisation, Supervision, Writing – Review & Editing, and Funding acquisition.
The authors declare that there are no conflicts of interest.
