Vitamin D3 (VD3) belongs to the family of steroid hormones and plays a crucial role in maintaining the metabolic homoeostasis of Ca and P and promoting bone development in animals(Reference Christakos, Dhawan and Verstuyf1). In mammals, VD3 is mainly metabolised and degraded by three cytochrome P450 (CYP) enzymes, 25-hydroxylase, 1 α-hydroxylase and 24-hydroxylase, which are encoded by the genes of cyp2r1, cyp27b1, cyp24a1, respectively(Reference Jones, Prosser and Kaufmann2). Moreover, VD3 exerts its physiological functions through its active form 1α,25(OH)2D3 by binding to vitamin D receptor (VDR)(Reference Christakos, Dhawan and Verstuyf1,Reference Pike, Meyer and Lee3) .
The metabolic process of VD3 is evolutionarily conserved(Reference Jones, Prosser and Kaufmann2). Previous studies demonstrated that a complete endocrine VD3 system, including three metabolic enzymes and VDR, first appeared in fish(Reference Hanel and Carlberg4). In most terrestrial higher animals, the photolysis of 7-dehydrocholesterol and subsequent conversion to VD3 is induced in skin by thermal isomerisation exposure to UV light at 290–320 nm(Reference Hanel and Carlberg4). In contrast, many fish live in deep water where the sunlight cannot penetrate; therefore, it is believed that VD3 found in the fish liver and adipose tissues mainly comes from the food chain(Reference Sunita Rao and Raghuramulu5). In addition, unlike mammals having only one VDR, most fish have two subtypes of VDR (vdra and vdrb)(Reference Howarth, Law and Barnes6,Reference Craig, Sommer and Sussman7) .
It has been reported that VD3 deficiency in fish is manifested as epidermis thinning, the necrosis of underlying muscle tissues and a significant decrease in blood Ca concentrations(Reference Lock, Waagbø and Bonga8). Moreover, VD3 contributes to increase the activity and number of osteoblasts in fish(Reference Cerezuela, Cuesta and Meseguer9,Reference Fenwick10) . According to the previous reports, the minimum VD3 supplemental levels of tilapia (Oreochromis niloticus O. aureus), channel catfish (Ictalurus punctatus) and rainbow trout (Oncorhynchus mykiss) were 360, 500 and 1600 μg/kg diet, respectively(11). The requirement of dietary VD3 for optimum growth of Wuchang bream (Megalobrama amblycephala) was around 5000 μg/kg(Reference Miao, Ge and Xie12), while 750 and 431 μg/kg of dietary VD3 were required for the optimum growth of marine fish orange-spotted grouper (Epinephelus coioides) and Japanese sea bass (Lateolabrax japonicus), respectively(Reference He, Ding and Watson Ray13,Reference Zhang, Li and Mai14) . In addition, it has been known that fish oil as an important lipid source in aquatic animal feeds is rich in VD(Reference Hodar, Vasava and Mahavadiya15). Therefore, it is important to supplement adequate VD in fish feeds when fish oil is replaced with other lipid sources that contain much lower VD3. On the other hand, it seems that most fish possess a high tolerance to excessive VD3 (Reference Lock, Waagbø and Bonga8,Reference Rychen and Aquilina16) . However, the European panel on additives and products or substances used in animal feed (FEEDAP) has authorised the maximum content of dietary VD3 which is 3000 μg/kg feed in fish(Reference Rychen and Aquilina16).
In addition to the regulation of the homoeostasis of Ca and P, VD3 also has a wide range of immunomodulatory functions(Reference Sassi, Tamone and D’Amelio17), and VDR is extensively expressed in immune cells, such as monocytes, dendritic cells, B cells and T cells(Reference Kongsbak, von Essen and Boding18,Reference Ooi, McDaniel and Weaver19) . It is noteworthy that 1α,25(OH)2D3 can promote the expression of antimicrobial peptides in macrophages(Reference Liu, Stenger and Li20,Reference Gombart, Borregaard and Koeffler21) , thus improving the host resistance to pathogenic infection(Reference Liu, Stenger and Li20,Reference Martineau, Jolliffe and Hooper22,Reference Wan, Tang and Rekha23) . Furthermore, the studies have also shown that VD3 affects the immunity and antibacterial activity in fish(Reference Cheng, Tang and Guo24–Reference Cheng, Ma and Guo27). Furthermore, accumulating evidence from mammalian studies also shows that VD3 regulates the metabolism of lipids and carbohydrates, and VD3 deficiency leads to obesity, hyperglycaemia and related metabolic syndromes(Reference Sassi, Tamone and D’Amelio17). Unlike mammals, fish are less able to utilise dietary carbohydrates and have a low level of insulin secretion, which are considered to be congenital ‘diabetics’(Reference Polakof and Panserat28). So far, the effects of VD3 on the glucose metabolism of fish have not been reported. However, the study on zebrafish has demonstrated that 1α,25(OH)2D3 promotes fatty acid oxidation in fish adipose tissues(Reference Peng, Shang and Wang29).
Interestingly, VD3 has a significant impact on the gut microbiota. The study on VDR knockout animals has suggested that the lack of VD3 signalling pathways leads to an imbalance of gut microbiota and induces various metabolic diseases(Reference Sittipo, Lobionda and Lee30,Reference Wang, Thingholm and Skieceviciene31) . Interestingly, the synthesis of VD3 is significantly reduced in germ-free mice(Reference Bora, Kennett and Smith32). However, there are no reports about the interactions between VD3 and gut microbiota in fish.
Turbot (Scophthalmus maximus L.) is an important commercial carnivorous fish(Reference Pereiro, Figueras and Novoa33). In this study, we determined the precise requirement of turbots for VD3 in the feed by the addition of different doses of dietary VD3 and explored the potential physiological functions of VD3 in fish.
Materials and methods
Animal ethics
The Experimental Animal Ethics Committee of Ocean University of China has approved all animal care and handling procedures in the present study.
Reagents
All ingredients in the experimental diets (except casein and VD3) were supplied by Great Seven Biotechnology Co., Ltd.; casein and VD3 were purchased from Sigma; VD3 was added in the form of cholecalciferol; methanol used for HPLC was purchased from Merck KGaA; all reagents used for the measurement of body and diet composition were purchased from Sinopharm; Trizol was purchased from Takara; HiScript® III RT SuperMix for qPCR was purchased from Vazyme; SYBR green qPCR kit was purchased from Accurate Biology; 4 % paraformaldehyde solution was purchased from Biosharp; the reagents for the dehydration and staining of histological sections were purchased from Thermo Scientific; QIAamp DNA Stool Mini Kit was purchased from Qiagen and PBS was purchased from Solarbio.
Fish maintenance
Turbots (Scophthalmus maximus L.) weighing around 13 ± 0·08 g were purchased from a commercial fish farm in Shandong Province, China. The feeding trial was carried out in a flow-through system located in Longhui Aquatic Product CO. Ltd. The fish were acclimatised in the system for 2 weeks. After the acclimation period, the fish were weighed and fasted for 24 h before they were randomly allocated to fifteen tanks (500 l) with a lid, and sixty fish per tank. Each diet was randomly assigned to triplicate tanks. The fish were fed twice daily at 07.00 and 19.00 hours, and the feeding experiment was performed for 8 weeks, since significant differences in growth were attained already at 8 weeks. The water quality during the feeding experiment was monitored as follows: the temperature ranged from 16 to 18°C; salinity was from 27 to 29‰; the concentrations of ammonia-nitrogen and nitrite were less than 0·1 mg/l and the dissolved oxygen was approximately 7 mg/l. The husbandry and handling of the fish in the present study were performed strictly according to the Management Rule of Laboratory Animals (Chinese order no. 676 of the State Council, revised 1 March, 2017).
Diet formulation
The previous reports showed that the requirement of dietary VD3 for the optimum growth of marine fish, such as orange-spotted grouper and Japanese sea bass, was 750 and 431 μg/kg, respectively(Reference He, Ding and Watson Ray13,Reference Zhang, Li and Mai14) . Accordingly, five formulations of experimental diets with different VD3 contents (0, 200, 400, 800 and 1600 μg/kg) were designed in our study. The composition of the experimental diets is shown in Table 1. Briefly, casein and gelatin were utilised as the dietary protein sources, and crystal amino acids were added to meet the basic nutritional requirements of turbots(Reference Kaushik34). Fish oil, soyabean oil and soya lecithin were added as the primary lipid source. All the ingredients were ground into a fine powder through a 120-mesh sieve. After VD3 was completely blended with other ingredients, fish oil and soyabean oil were further kneaded with the premixed ingredients. After the oil was fully mixed with all ingredients, some water (100 g/kg diet) was added to increase the diet viscosity. Finally, a manual granulator was used to pelletise the diets. The diets were dried in the shade and stored at –20°C until further use.
Ingredients and proximate compositions of the basal diet (g/kg)
* Amino acid premix (g/100 g diet): arginine, 1·69; histidine, 0·55; isoleucine, 0·22, leucine, 0·14; lysine, 0·73; phenylalanine, 0·50; threonine, 0·61; valine, 0·13; alanine, 1·32; aspartic acid, 1·63; glycine, 1·62; serine, 0·42; cystine, 0·40; and tyrosine, 0·10.
† Vitamin premix (mg/kg diet): retinyl acetate, 32; a-tocopheryl acetate, 240; menadione sodium bisulphite, 10; ascorbic acid, 120; cyanocobalamin, 10; biotin, 60; choline dihydrogen citrate, 7000; folic acid, 20; inositol, 800; niacin, 200; D-calcium pantothenate, 60; pyridoxine HCl, 20; riboflavin, 45; thiamine HCl, 25; microcrystalline cellulose, 16 473.
‡ Mineral premix (mg/kg diet): MgSO4·7H2O, 1200; CuSO4·5H2O, 10; FeSO4·7H2O, 80; ZnSO4·H2O, 50; MnSO4·H2O, 45; CoCl2, 5; Na2SeO3, 20; calcium iodine, 60; zeolite powder, 8485.
§ Attractant (g/kg diet): betaine, 4; DMPT, 2; threonine, 2; glycine, 1; inosine-5′-diphosphate trisodium salt, 1.
Sampling
At the end of the feeding experiment, all the fish fasted for 24 h. Turbot from each group were anaesthetised with 20 mg/l tricaine and weighed. The liver from each fish was collected and weighed for the calculation of the hepatosomatic index. Serum, liver and gut samples from three fish per tank were collected and frozen in liquid N2immediately, followed by storage at −80°C for further analysis. Another three fish from each tank were euthanised with 20 mg/l tricaine and stored at −80°C for the measurement of body composition.
The measurement of VD3 in the diets
The contents of VD3 in the diets were confirmed by HPLC as described previously(Reference Horvli and Lie35). Briefly, 0·2 g of the diet was mixed with 1 ml of ascorbate ethanol solution (5 g/l) and 200 μl of potassium hydroxide solution (500 g/l), and the mixture was saponified at 50°C for 4 h. After that, 2 ml of ethyl ether was added into the saponification solution and followed by the centrifugation for 10 min at 7000 g for three times. The ethyl ether phase after the centrifugation was collected and redissolved with 300 μl of methanol after nitrogen blowing. The HPLC conditions were as follows: mobile phase methanol, flow rate 1 ml/min, column temperature 25°C and detection wavelength 265 nm. As shown in Table 1, the actual contents of VD3 in five diets were 3·53, 190·35, 380·46, 789·23 and 1549·22 μg/kg, respectively.
The measurement of 1α,25(OH)2D3 in serum
The concentrations of 1α,25(OH)2D3 in fish serum were measured with the 1,25-Dihydroxyvitamin D3 (Calcitriol) ELISA Kit (Abbexa) according to the manufacturer’s instructions.
The measurement of Ca and phosphate in serum
The concentrations of Ca and phosphate in fish serum were measured with the Calcium Assay Kit (Jiancheng Biotech Co.) and Phosphate Assay Kit (Jiancheng Biotech Co.) according to the manufacturer’s instructions.
The composition analysis of fish body and diets
The body composition of turbots was confirmed by previously described methods(Reference Cunniff36). Briefly, the samples were dried at 105°C to determine the moisture contents. Besides, the contents of crude protein and lipids were measured by Kjeldahl method of nitrogen determination (FOSS, Sweden) and soxhlet ether extraction (Buchi, Switzerland), respectively. The ash content of the samples was assessed by burning in a muffle furnace for 10 h.
Quantitative real-time PCR analysis
The liver and hindgut samples from three turbots per tank (nine fish per group) were collected as described above. Total mRNA was isolated using Trizol, and cDNA was synthesised from total RNA by PrimeScript RT reagent kit according to the manufacturer’s instructions. The real-time PCR was performed in a thermo-cycler CFX96 instrument (BioRad), and the expression of target genes was normalised to β-actin. The sequences of all primers used in the present study are provided in Table 2.
Primer sequences used for qRT-PCR
Transcriptomic analysis
To further identify the potential effects of VD deficiency on turbot, two groups, that is, VD deficiency (0 μg/kg) and VD optimum (400 μg/kg), were compared in this experiment. Briefly, 1 µg RNA extracted from the livers of the turbots in 0 μg and 400 μg groups (three turbots per tank and total nine fish per group) was used as input material for the RNA sample preparations. Sequencing libraries were generated using NEBNext® UltraTM RNA Library Prep Kit for Illumina® (NEB), and index codes were added to attribute sequences to each sample. The clustering of the index-coded samples was performed on a cBot Cluster Generation System using TruSeq PE Cluster Kit v3-cBot-HS (Illumina). After cluster generation, the library preparations were sequenced on an Illumina Novaseq platform and 150 bp paired-end reads were generated.
Feature Counts v1.5.0-p3 was used to count the read numbers mapped to each gene. Furthermore, fragments per kilobase per million mapped reads (FPKM) of each gene was calculated based on the length of the gene and read count mapped to this gene to calculate the gene expression level.
Differential expression analysis of two groups was performed using the DESeq2 R package (1.16.1). DESeq2 provides statistical routines for determining differential expression in digital gene expression data using a model based on the negative binomial distribution. Genes with an adjusted P-value < 0·05 found by DESeq2 were assigned as differentially expressed. Kyoto Encyclopedia of Genes and Genomes is a database resource for understanding high-level functions and utilities of the biological system. The cluster Profiler R package was used to test the statistical enrichment of differential expression genes in Kyoto Encyclopedia of Genes and Genomes pathways.
The assessment of gut micromorphology
The hindguts of three turbots per tank (nine fish per group) were fixed with paraformaldehyde (BL539A, Biosharp) for 24 h and transferred to 75 % alcohol for the preservation. The fixed tissue around 1 cm long was cut, dehydrated routinely with a series of alcohols and embedded in paraffin. Tissue sections of approximately 7 microns were then cut, placed on slides and stained with haematoxylin and eosin. The slides were examined under a light microscope (Olympus, DP72) equipped with a camera (Nikon E600) and Cell Sens Standard Software (Olympus).
Analysis of gut microbiota
Total genome DNA of the whole gut from the turbots in 0, 400 and 800 μg groups (three turbots per tank and total nine fish per group) was extracted using the QIAamp DNA Stool Mini Kit (Qiagen) according to the manufacturer’s protocols. 16S rDNA of distinct regions (16S V3-V4) was amplified using specific primers. All PCR were carried out with Phusion® High-Fidelity PCR Master Mix (New England Biolabs).
Sequencing libraries were generated using TruSeq® DNA PCR-Free Sample Preparation Kit (Illumina) following the manufacturer’s recommendations and index codes were added. The library quality was assessed on the Qubit@ 2.0 Fluorometer (Thermo Scientific) and Agilent Bioanalyzer 2100 system. The library was sequenced by an Illumina NovaSeq platform.
Sequences analysis was performed by Uparse software (Uparse v7.0.1001, http://drive5.com/uparse/)(Reference Edgar37). Sequences with ≥ 97 % similarity were assigned to the same operational taxonomic units (OTU). The representative sequence for each OTU was screened for further annotation. For each representative sequence, the Silva Database (http://www.arb-silva.de/)(Reference Quast, Pruesse and Yilmaz38) was used based on the Mother algorithm to annotate taxonomic information. OTU abundance information was normalised using a standard of sequence number corresponding to the sample with the least sequences. Beta diversity analysis was used to evaluate differences of samples in species complexity.
Bacterial challenge
In order to further verify the difference of anti-infectious ability of turbot under different VD conditions (deficiency, optimum and excess), ten fish from 0, 400 and 800 μg groups were randomly selected for a bacterial challenge at the end of the feeding trial. Edwardsiella tarda was isolated from diseased turbot and determined by sequencing. For challenge experiments, bacterial cultures were centrifuged at 8000 g for 1 min, and the pellet was subsequently suspended in PBS to a final 2 × 107 CFU/ml before use.
After fasted for 24 h, the turbots were anaesthetised with tricaine (20 mg/l). The weight of each fish was recorded before intraperitoneal injection with 0·5 ml PBS containing 2 × 107 CFU/ml E. tarda using medical syringes. Fish mortality was monitored four times per day until 7th day, and the final weight of each fish was recorded. The average daily weight loss was calculated by (final weight – initial weight)/days.
Calculations and statistical analysis
The growth parameters were calculated as follows: weight growth rate (%) = 100 × (final individual weight – initial individual weight)/initial individual weight; specific growth rate (% per d) = 100 × (Ln final individual weight – Ln initial individual weight)/number of days; viscerosomatic index (%) = 100 × viscera weight of final individual fish/final individual weight (g); hepatosomatic index (%) = 100 × liver weight of final individual fish/final individual weight; survival rate (%) = 100 × (final number of fish)/(initial number of fish). The broken-line regression model and the quadratic regression model were used to estimate the optimum VD3 requirement of turbot based on the weight growth rate.
Results were presented as means with their standard error of the mean unless otherwise stated. Raw data were analysed by the one-way ANOVA after normality and homogeneity of variance was confirmed. Multiple comparisons were conducted with Tukey’s post-hoc test. P < 0·05 was considered as statistical significance. Statistical analysis was performed using the GraphPad Prism 9 (GraphPad Software Inc.).
Result
The growth and body composition of the turbot
To identify the optimum dose of VD3 addition in the diet for turbot growth, the juvenile turbots with average initial weight 13 ± 0·08 g were fed with the diets containing five different levels of VD3: 0, 200, 400, 800 and 1600 μg/kg, that is, 0, 5, 10, 20 and 40 μg/kg for 8 weeks. The feed efficiency calculated as [(final body weight − initial body weight)/feed intake] for each group was 0·74, 0·77, 0·78, 0·75 and 0·80, respectively, which was quite similar among all groups. As the results showed in Table 3, the fish fed with 400 μg/kg VD3 in the diet displayed the highest values of final body weight, specific growth rate and weight growth rate. The broken-line model showed the optimal dietary requirement of VD3 was 397·01 μg/kg under the conditions of this study (Fig. 1(a)). Further analysis based on a regression model also exhibited a significant correction between weight growth rate and dietary VD contents, and the optimal requirement of turbot for dietary VD3 was estimated as 846·25 μg/kg according to the quadratic regression model (Fig. 1(b)). Besides, the viscerosomatic index of the fish was lowest at 400 μg diet, while the hepatosomatic index did not differ significantly among the groups. Furthermore, there was no significant difference in crude protein, water content and ash content between the treatment groups, except that the contents of crude fat in the fish fed 400 μg/kg dietary VD3 were significantly lower than those in the 0 μg group (Table 4).
The effects of different dietary vitamin D3 contents on growth indices* (Mean values with their standard errors of the mean)
* n 9 fish. Values in the same row sharing the same superscript letter are not significantly different determined by the one-way ANOVA.
† WGR (%): weight growth rate (%) = 100 × (final individual weight (g) – initial individual weight (g))/initial individual weight.
‡ SGR (% per d): specific growth rate (% per d) = 100 × (Ln final individual weight (g) – Ln initial individual weight (g))/number of days.
§ VSI (%): viscerosomatic index (%) = 100 × viscera weight (g) of final individual fish/final individual weight (g).
|| HIS (%): hepatosomatic index (%) = 100 × liver weight (g) of final individual fish/final individual weight (g).
¶ Survival rate (%) = 100 × (final number of fish)/(initial number of fish).
Effects of dietary VD3 levels on WGR of turbot. (a) Based on the broken-line model of WGR corresponding to dietary VD3 contents (y = 0·0869x + 167·36, R 2 = 0·9926; y = − 0·0065x + 204·44, R 2 = 0·9379), the optimum level of dietary VD3 for juvenile turbot was estimated to be 397·01 μg/kg. (b) Based on the quadratic regression model of WGR corresponding to dietary VD3 contents (y = –0·00004x2 + 0·0757x + 170·47, R 2 = 0·8791), the optimum level of dietary VD3 for juvenile turbot was estimated to be 846·25 μg/kg. Error bars were indicated as means and standard deviations (n = 24 fish).
The effects of different dietary vitamin D3 contents on the body composition of turbots (dry weight)* (Mean values with their standard errors of the mean)
* n 3 of replicates of nine fish. Values in the same row sharing the same superscript letter are not significantly different determined by the one-way ANOVA.
The histological structure of turbot hindgut
Moreover, we analysed the effects of VD3 on the morphology of fish guts. According to our results, the hindgut anatomy of turbots fed with different doses of VD3 displayed no difference (Fig. 2(a)). The villus height (Fig. 2(b)), enterocytes height (Fig. 2(c)) and lumen diameter (Fig. 2(d)) of the turbot hindguts in all groups exhibited no significant difference. Besides, the gene expression of occludin and zo-1, which are the markers of intestinal barrier, did not change significantly in all groups (Fig. 2(e) and (f)).
Effects of dietary VD3 contents on the hindgut histology of turbot. (a) The hindguts of turbot in different VD3 groups were collected and sectioned. After the fixation by haematoxylin and eosin (H&E), the hindgut morphology in different groups was observed. The images were representative of at least three independent experiments. (b)–(d) The micromorphology, including villus height (b), enterocyte height (c) and lumen diameters (d) of the turbot guts in three groups was evaluated. (e), (f) The gene expression of occludin and zo-1 in the hindgut of the turbots fed with different VD3 doses was analysed by RT-PCR (n 9 fish). Error bars indicate means with their standard error of the mean. zo-1, zona occludens-1.
The composition and diversity of gut microbiota in the turbot
On the other hand, the compositions of gut microbiota in 0, 400 and 800 μg groups were analysed. A total number of 665 735 clean reads were generated, covering 5829 OTU (97 % similarity level) after the sequence quality control. Venn diagram showed that three groups shared 2448 OTU, and the number of unique OTU in 0, 400 and 800 μg diets was 416, 1166 and 393, respectively (Fig. 3(a)). Interestingly, the fish fed 400 μg/kg VD3 in the diets exhibited the highest diversity of gut microbiota (Fig. 3(b)). In addition, Firmicutes, Proteobacteria and Bacteroidetes were identified as the predominant bacterial phyla in the guts of turbots from all groups. The relative abundance of Cyanobacteria, Acidobacteriota and Deinococcota was significantly improved in the gut of the turbots in the 400 μg group compared with that in the other two groups. Moreover, the lower relative abundances of Firmicutes in the 400 μg group were found compared with that in the 0 μg group, although no statistical difference was observed (Fig. 3(c)). At the genus level, the 800 μg diet increased the relative abundance of Bacteroides, Acinetobacter and Lactobacillus. Meanwhile, the 400 μg diet reduced relative abundances of the Roseburia and Faecalibacterium but improved relative abundances of Gardnerella than the 0 μg diet (Fig. 3(d)).
Effects of dietary VD3 contents on the composition and diversity of gut microbiota of turbot. (a) Every circle in Venn diagram represents one group. The value from the overlapping part of different circles represents the shared OTU between groups, and the value from the non-overlapping part of one circle represents the unique OTU of that group. (b) The beta diversity index of intestinal microbiota from three groups (0, 400, 800 μg) was calculated. Error bars indicate means with their standard error of the mean; *P < 0·05, **P < 0·01. (c), (d) The taxonomy classification of reads at the phylum (c) or genus (d) taxonomic level. Only top ten most abundant (based on the relative abundance) bacterial phyla or genera were shown. Other phyla or genera were all assigned as ‘Others’, n 9 fish.
Intestinal inflammation and anti-infectious ability of the turbot
In the following experiments, we measured the gene expression of several inflammatory cytokines, including il-1β, il-8 il-6 and tnf-α in the liver and hindgut of turbots in the five groups. As shown in Fig. 4(a), the gene expression of all pro-inflammatory cytokines was down-regulated in the liver of turbot in the 400 μg group compared with other groups. The similar results were obtained in the hindgut of the turbots, except that the gene expression of il-6 in the 800 μg group was lower than that in the 400 μg group (Fig. 4(b)).
Effects of dietary VD3 contents on immunity and anti-infection ability of turbot. (a), (b) The gene expression of inflammatory cytokines in the liver and hindgut of turbots from different VD3 groups. The gene expression of inflammatory cytokines in the liver (a) and hindgut (b) of turbot in different VD3 groups was analysed by RT-PCR (n 9 fish). Error bars indicate means with their standard error of the mean. The different letters above the bars indicate significant differences. il-1β, interleukin-1beta; il-8, interleukin-8; il-6, interleukin-6; tnf-α, tumour necrosis factor-alpha. (c), (d) The mortality and weight loss in the infected turbots fed with different VD3 doses. Ten turbots were randomly selected in 0, 400 and 800 μg VD3 groups at the end of feeding experiment, and each fish was intraperitoneally injected E. tarda (1 × 107 bacteria per fish). (C) The survival rates of the turbots were recorded every 24 h in 6 d (n 10). (D) The daily mass change of each turbot before death was calculated. Error bars indicate means with their standard error of the mean ± sem. *P < 0·05.
According to our results in bacterial challenge, the survival rate of the fish in the 400 μg group was around 80 % at 6th day after infection, compared with 30 % in the 0 μg group and 20 % in the 800 μg group (Fig. 4(c)). Corresponding to the mortality, the daily weight loss of turbots in the 0 μg and 800 μg groups was also significantly higher than that in the 400 μg group (Fig. 4(d)).
Vitamin D metabolism in the turbot
First, the gene expression of VD metabolic enzymes, including cyp2r1, cyp27b1 and cyp24a1 in different tissues of juvenile turbots, was analysed. According to our results, the genes of cyp2r1, cyp27b1 and cyp24a1 were mainly expressed in the liver, and a lower abundance of cyp27b1 transcript in other tissues was also detected (Fig. 5(a)). We further measured the contents of 1α,25(OH)2D3 in the turbot sera of all groups, and the results demonstrated that the sera from the turbot in the 400 μg group contained the maximum level of 1α,25(OH)2D3 (Fig. 5(b)). Consistently, the gene expression of cyp2r1 and cyp27b1 reached the highest level in the livers of the turbots in the 400 μg group (Fig. 5(c) and (d)). In contrast, the gene expression of cyp24a1 was lower in the 200 μg and 400 μg groups, compared with that in 0, 800 and 1600 μg groups (Fig. 5(e)). In addition, the gene expression of fibroblast growth factor 23 (fgf23), a factor for negative feedback during VD3 metabolism, was elevated with the increase of VD3 concentration in the diets (Fig. 5(f)).
VD metabolism of turbot fed with different dietary VD3. (a) The gene expression of cyp2r1, cyp27b1, cyp24a1 in different tissues of juvenile turbots was measured by PCR. The image was representative of at least three independent experiments. (b) 1α,25(OH)2D3 concentrations in the serum of turbots fed with different VD3 were determined by ELISA (n 9 fish). (c)–(f) The gene expression of cyp2r1 (C), cyp27b1 (D), cyp24a1 (E), and fgf23 (F) in the livers of the turbots fed with different VD3 diets was analysed by RT-PCR (n 9 fish). Error bars indicate means with their standard error of the mean. The different letters above the bars indicate significant differences. cyp2r1, cytochrome P450, family 2, subfamily R, polypeptide 1; cyp27b1, cytochrome P450, family 27, subfamily B, polypeptide 1; cyp24a1, cytochrome P450, family 24, subfamily A, polypeptide 1; fgf23, fibroblast growth factor 23.
The concentrations of Ca and phosphate in the turbot
Considering the well-known effects of VD3 on the regulation of Ca and P homoeostasis in animals, the concentrations of Ca and phosphate in the turbot sera were analysed. It appeared that the concentrations of Ca (Fig. 6(a)) and phosphate (Fig. 6(b)) were at a similar level in turbot sera from all groups, regardless of the VD3 dose in the diets. However, the gene expression of Ca transporter trpv6 and P transporter slc20a2 in the hindgut of turbots from different groups showed an opposite trend to 1α,25(OH)2D3 content in turbot sera (Fig. 6(c) and (d)).
Ca and phosphate metabolism of turbot in different VD3 groups. (a), (b) The concentrations of Ca (a) and phosphate (b) in serum of the turbots with different dietary VD3 levels were analysed (n 9 fish). (c), (d) The gene expression of trpv6 and slc20a2 in the hindgut of the turbots fed with different VD3 doses was analysed by RT-PCR (n 9 fish). Error bars indicate means with their standard error of the mean. The different letters above the bars indicate significant differences. trpv6, transient receptor potential cation channel, subfamily V, member 6; slc20a2, solute carrier family 20 member 2.
The metabolism of lipids and carbohydrates in the turbot
To further identify the potential effects of VD deficiency on turbot, the transcriptomic sequencing of turbot livers from two groups, that is, VD deficiency (0 μg/kg) and VD optimum (400 μg/kg) groups, was performed. As the results showed, a total of 1175 genes were differentially expressed (adjusted by P-value < 0·05) between two groups. Among these genes, the transcripts of 454 genes were up-regulated, while the transcripts of 721 genes were down-regulated in the 400 μg group compared with those in the 0 μg group (Fig. 7(a)). The enrichment results of Kyoto Encyclopedia of Genes and Genomes metabolic pathways showed that a series of pathways in nutritional metabolism were significantly influenced, such as fatty acid biosynthesis, PPAR signalling pathway, protein export and amino sugar metabolism (Fig. 7(b)). The heatmap of differentially expressed genes also revealed that many key genes in lipid metabolism (Fig. 7(c)) and glucose metabolism (Fig. 7(d)) changed significantly in VD-deficient group.
Effects of dietary VD3 deficiency on the turbot liver. Transcriptomic analysis of the liver from turbot fed with 0 μg and 400 μg VD3 diet was conducted. (a) Cluster analysis of different gene changes between 0 and 400 μg vitamin D3 treatment. (b) The top twenty statistics of KEGG pathway enrichment for differentially expressed genes (DEG). GeneRatio is the ratio of number of differentially expressed genes enriched in a certain pathway to total number of DEG. (c), (d) The heatmap of differentially expressed genes in the metabolism of lipids (C) and glucose (D). n 9 fish. The list genes include farnesyl-diphosphate farnesyltransferase 1 (fdft1), sterol O-acyltransferase 2 (soat2), fatty acid synthase (fasn), peroxisome proliferator-activated receptor gamma (pparγ), fatty acid desaturase 2 (fads2), stearoyl-CoA desaturase (scd), ATP citrate lyase a (aclya), ATP citrate lyase b (aclyb), phosphoglycerate dehydrogenase (phgdh), phospholipase D family member 3 (pld3), sterol regulatory element binding transcription factor 1 (srebf1), protein phosphatase 1(ppp1r3ca), EBP cholestenol delta-isomerase (ebp), forkhead box O1 a (foxo1a), glucokinase (gck), insulin receptor substrate 4a (irs4a), insulin receptor substrate 2a (irs2a); 6-phosphogluconolactonase (pgls), galactose-1-phosphate uridylyltransferase (galt), glutamine–fructose-6-phosphate transaminase 1 (gfpt1), phosphomannomutase 1 (pmm1), phosphoglucomutase 3 (pgm3), galactokinase 1 (galk1), GDP-mannose pyrophosphorylase B (gmppb), isocitrate dehydrogenase (NADP(+)) 1 (idh1), acetoacetyl-CoA synthetase (aacs), phosphogluconate dehydro (pgd).
Discussion
In this study, we confirmed that VD3 has extensive effects on different physiological processes in fish. We identified that dietary VD3 influenced the growth, intestinal health and pathogen resistance in turbot, although the homoeostasis of Ca and phosphate appeared not affected. Moreover, we found that the nutritional metabolism was disturbed in the turbots with VD3 deficiency in the diets. We have demonstrated for the first time that VD3 influences the composition and diversity of gut flora in fish. To our knowledge, this is also the first study to investigate the effects of dietary VD3 on the metabolism of VD3 itself in fish.
The significance of VD3 in fish has been reviewed by Lock and co-workers(Reference Lock, Waagbø and Bonga8). Actually, Barnett et al. proved for the first time the importance of VD3 addition in fish feeds(Reference Barnett, Cho and Slinger39). In this study we attempted to investigate the effects of VD on turbot from three levels of dietary VD3: deficiency, optimum and excess. According to our results, the turbot fed with 400 IU/kg dietary VD3 displayed the optimal growth performance (Fig. 1(a)), inflammatory status (Fig. 4(a) and (b)) and the highest concentration of the active VD3 metabolite in serum (Fig. 5(b)) in all five groups, hence the group fed with 400 IU/kg VD3 was considered as the optimum one. In the analysis of intestinal microbiota and anti-infectious ability of turbot, the fish fed with 0, 400 and 800 IU/kg VD3 were selected to be representative of VD3 deficiency, optimum and excess, respectively. To further investigate the effects of VD deficiency on turbot, two groups, i.e., VD deficiency (0 IU/kg) and VD optimum (400 IU/kg) were compared in transcriptomic analysis.
Based on the prediction by the broken-line model, the optimal requirement of dietary VD3 in the feed for the growth of juvenile turbot is around 400 μg/kg (Fig. 1(a)), which is close to marine fish orange-spotted grouper(Reference He, Ding and Watson Ray13) and Japanese sea bass(Reference Zhang, Li and Mai14). It has been known that the predicted requirement of a nutrient for the maximal growth of animals could be greatly different depending on the selected mathematical model. Usually, the broken-line model predicts the nutrient requirement that is lower than that predicted by curvilinear curve-fitting procedures(Reference Baker40). As our results showed, the optimal dietary VD3 requirement predicted by quadratic regression model for the growth of juvenile turbot is 846·25 μg/kg (Fig. 1(b)). However, it seems the broken-line model fits better the data in our study considering the R 2 value is less than 0·90 in the quadratic regression model.
In fact, dietary VD3 requirement of cultured fish greatly varies in different experiments. For example, the dietary VD3 requirement of freshwater species Wuchang bream (initial weight was 17·71 ± 0·22 g) was estimated to be around 5000 μg/kg using the second-order polynomial regression model(Reference Miao, Ge and Xie12). As predicted by the broken-line model based on the weight gain, the dietary VD3 requirement of juvenile black carp (Mylopharyngodon piceus, initial weight 4·73 ± 0·13 g) and tilapia (initial weight around 0·80 g) was 534·2 IU/kg(Reference Wu, Lu and Wang41) and 374·8 IU/kg(Reference Shiau and Hwang42), respectively. Surprisingly, no significant difference in the growth and body composition was detected when the 21-d-old fry of freshwater species Rora (Labeo rohita, initial weight around 0·1 g), were fed with VD3 deficient or 1650 IU/kg VD3 supplemented feeds Reference Ashok, Rao and Raghuramulu(43) . In addition, the study on the marine fish orange-spotted grouper (initial weight 81·5 ± 0·1 g) showed that the optimum addition level of dietary VD3 was 750·19 IU/kg estimated by the broken-line model based on WGR Reference He, Ding and Watson Ray(13) . The vast disparities among the different experiments may be caused by the various species and developmental stage of cultured fish. Considering the previous report that the fish fed with diets low in fish meal and high in plant protein seemed to require more dietary VD3 to reach the optimal growth(Reference Prabhu, Lock and Hemre44,Reference Dominguez, Montero and Zamorano45) , the different composition of the feeds should be taken into account when the optimal requirement of dietary VD3 for cultured fish was evaluated.
On the other hand, it seemed that most fish exhibited high tolerance to excessive VD3 in the diet, since many fish did not display significant impairment on their growth(Reference Rychen and Aquilina16,Reference Graff, Hoie and Totland46) . Similarly, our result showed that the growth performance and the intestinal anatomy of turbot were not significantly influenced when the higher doses of VD3 up to 1600 μg/kg were added. However, when the fish were fed with the diets containing the higher doses of VD3 than 400 μg/kg, the inflammation was induced in the gut and liver (Fig. 4). More importantly, the pathogen resistance of the fish in the 800 μg group was significantly lower than that in the 400 μg group, which confirmed that the immune status of the fish in the 800 μg group was impaired. A recent report has also demonstrated that the lower or higher than adequate dose of dietary VD3 exhibits adverse effects on antioxidant capacities and innates immunity in black carp Reference Wu, Lu and Wang(41) . According to the perspective of ‘precise nutrition’, we claim that although most fish seem tolerant to much higher doses of VD3 in the diets based on their growth performance, the immune status and other physiological functions might be impaired when the higher doses of VD3 were fed to the fish. It is noteworthy that FEEDAP Panel has authorized the maximum content of dietary VD3 is 3000 IU/kg feed in fish Reference Rychen and Aquilina(16) . Nonetheless, some previous studies have demonstrated high doses of VD3 are beneficial to the innate immunity of the fish(Reference Cerezuela, Cuesta and Meseguer9,Reference Cheng, Ma and Guo27,Reference Jiang, Shi and Zhou47) . The discrepancy could be caused by the difference in the species and the methods how to prepare VD3-containing diets.
Previous studies have proved the functions of VD3 on anti-inflammation and host immune regulation. For example, the VD/VDR pathway was involved in protecting the intestinal barrier during colon inflammation and relieved the symptoms of dextran sulphate sodium-induced colitis in mice(Reference Zhao, Zhang and Wu48). The study on Atlantic salmon demonstrated that the co-incubation of VD3 with macrophages reduced the adhesion of Aeromonas salmonicida subsp. salmonicida to macrophages(Reference Soto-Davila, Valderrama and Inkpen26). Meanwhile, dietary VD3 increased the activity of lysozyme in serum and the expression of hepcidin in the liver of juvenile black carp(Reference Wu, Lu and Wang41,Reference Graff, Hoie and Totland46) . The study on European perch (Dicentrarchus labrax L.) also showed that VD3 increased the phagocytosis of leucocytes in the head kidney, while inhibited the expression of il-1β in the head kidney and intestinal tract(Reference Dioguardi, Guardiola and Vazzana49). Furthermore, the addition of dietary VD3 inhibited the up-regulation of pro-inflammatory cytokines induced by bacterial infection in yellow catfish(Reference Cheng, Tang and Huang25) and Jian carp (Cyprinus carpio var. jian)(Reference Jiang, Shi and Zhou47). Consistently, our results demonstrated that the gene expression of the pro-inflammatory cytokines, including il-1β, il-8, il-6 and tnf-α, was significantly lower in the liver and gut of the turbots in the 400 μg group (Fig. 4(a) and (b)), indicating that VD3 deficiency or overdose could induce inflammation in liver/gut axis in fish. In addition, our results also demonstrated that VD3 significantly improved the anti-infection ability of turbots (Fig. 4(c)), and our recent published report has depicted the molecular mechanisms how VD3 enhances the pathogen resistance in turbot(Reference Liu, Shao and Lan50). It is well known that VD has a large impact on innate immunity(Reference Cantorna, Rogers and Arora51). For example, VD3 significantly enhanced the expression of antimicrobial peptides in human macrophages(Reference Gombart, Borregaard and Koeffler21) and in the fish cells(Reference Dioguardi, Guardiola and Vazzana49,Reference Estévez, Mostazo and Rodriguez52) . Interestingly, the VD3/VDR-type I interferon axis seems involved in the immunomodulatory functions of VD3 in yellow catfish(Reference Cheng, Ma and Guo27).
Our result from transcriptomic analysis in the turbot livers clearly demonstrated the metabolism of fatty acids and carbohydrates were interfered in the VD3-deficient group (Fig. 7). In accord with this result, the crude fat contents in turbots fed VD3-deficient diet significantly increased compared with those in the fish fed 400 μg/kg dietary VD3 (Table 4), and the decrement on body weight was lowest in the 400 μg group during the bacterial infection ( Fig. 4(d)). In fact, the studies in higher animals have shown that VD3 deficiency is closely related to obesity, hyperglycaemia and related metabolic syndromes(Reference Zuk, Fitzpatrick and Rosella53). Consistently, the transcriptomic analysis showed that 1α,25(OH)2D3 significantly affected the lipid metabolism pathway in the early embryo of zebrafish(Reference Craig, Zhang and McNulty54). Furthermore, significant fat accumulation was also observed in cyp2r1 knockout zebrafish, and the promotion of fatty acid oxidation by VD3 in fish was also confirmed(Reference Peng, Shang and Wang29). Importantly, it has been well known that gut microbiota plays a vital role in maintaining the homoeostasis of the host intestinal environment, and its imbalance often leads to various metabolic diseases(Reference Round and Mazmanian55). The studies in mammals have provided the evidence that VDR is a key genetic factor for shaping the host microbiome(Reference Wang, Thingholm and Skieceviciene31,Reference Jin, Wu and Zhang56,Reference Garcia, Moore and Kahan57) . Our results showed that the fish fed VD-deficient diet exhibited the lowest diversity of gut microbiome, and the addition of VD3 in the diets shifted the composition of gut microbiome in turbots, increasing the abundance of the beneficial bacteria, including lactobacillus (Fig. 3). A recent report claimed that VDR affected the metabolism of carbohydrates, proteins/amino acids, lipids and exogenous organisms by regulating microbial metabolites(Reference Chatterjee, Lu and Zhang58). Our experiments on zebrafish also demonstrated that the regulation of lipid and glucose metabolism by dietary VD3 was dependent on intestinal flora (unpublished results). It would be intriguing to further clarify how VD/VDR signalling pathway impacts the nutritional metabolism via the regulation of microbial metabolites in fish.
Regarding to VD metabolism in fish, several points are different from that in mammals. Firstly, 1α,25(OH)2D3 is mainly synthesised in mammalian kidney, while 1α,25(OH)2D3 synthase (encoded by cyp27b1) is also expressed in fish liver, suggesting that the liver could be the primary source of 1α,25(OH)2D3 in fish(Reference Takeuchi, Okano and Kobayashi59,Reference Sunita Rao and Raghuramulu60) . Moreover, 1α,25(OH)2D3 is the primary circulating form of VD3 metabolite in fish, instead of 25(OH)D3 in mammals(Reference Fraser61). Consistent with the previous reports, the gene expression of cyp27b1 in turbots was mainly detected in liver, and a much lower abundance of cyp27b1 transcript was also identified in kidney and gut (Fig. 5(a)). In addition, we discovered that the content of 1α,25(OH)2D3 in turbot serum reached the highest level when 400 μg/kg VD3 was added in the diets, and the absence or overdoses of VD3 in the diets lowered the serum contents of 1α,25(OH)2D3 in turbots (Fig. 5(b)).
In fact, VD metabolism is strictly regulated in higher animals, and fgf23 has been identified as a negative regulator of VD metabolism in higher animals(Reference Henry62). When the content of 1α,25(OH)2D3 is too high in vivo, fgf23 inhibits the expression of cyp27b1 and reduces the synthesis of 1α,25(OH)2D3 to prevent poisoning(Reference Bikle, Murphy and Rasmussen63). Based on our results that the expression of cyp2r1 and cyp27b1 decreased and the expression of cyp24a1 increased in liver when the fish were fed with the higher doses of VD3 than 400 μg/kg in the diets, we inferred that it might be caused by the increase in fgf23 expression when the fish were fed with dietary overdoses of VD3, leading to the reduced 1α,25(OH)2D3 productions. In addition to fgf23, the evidence has showed that gut microbiota also influences VD3 metabolism, since germ-free mice have lower serum levels of 25(OH)D3 than those in conventional mice(Reference Bora, Kennett and Smith32). Hence, further studies are worth to be conducted to clarify how the gut microbiota affect VD3 metabolism in fish.
As early as the 1920s, VD3 was identified to prevent rickets(Reference McCollum, Simmonds and Becker64). So far, the regulation of Ca and P homoeostasis by VD3 in land animals and in fish has been extensively studied. Our results demonstrated that the concentrations of Ca and phosphate in turbot sera were stable, regardless of VD3 doses in the diets (Fig. 6(a) and (b)). It is noteworthy that 1α,25(OH)2D3 concentration was still around 640 pg/ml in the serum of the turbot fed with 0 μg/kg dietary VD3 for 2 months (Fig. 5b)). Different from mammals, it is believed that fish cannot synthesise VD3 in vivo; they acquire VD3 via food chain(Reference Sunita Rao and Raghuramulu65). According to our analysis, there still was a very low amount of VD3 (3·53 μg/kg) in the VD3 absent diet, which could come from the fish oil in the diet and contribute to the source of 1α,25(OH)2D3 in the turbot serum from 0 μg group. A previous study has also shown that the bone development of European sea bass (Dicentrarchus labrax) seems not to be affected by the low dietary VD3 (Reference Darias, Mazurais and Koumoundouros66). Thus, the basal content of 1α,25(OH)2D3 in the serum of fish in the 0 μg group could involve in the regulation of Ca and phosphate homoeostasis in fish. Moreover, the expression of Ca and P transporters was increased when the contents of 1α,25(OH)2D3 were lowered in turbot sera, which could also be beneficial to the maintenance of Ca and phosphate homoeostasis in fish.
Conclusion
Our study assessed the optimal VD3 requirement in the feed for turbot and demonstrated the effects of dietary VD3 on intestinal health, anti-infection ability and metabolism in fish. The results in the present study deepened our understanding on the physiological functions and metabolism of VD3 in fish and provided a reference to the evaluation of precise requirement for dietary VD3 in aquatic animals.
Acknowledgements
This study was supported by the National Natural Science Foundation of China (grant no. 31972802); Natural Science Foundation of Shandong Province (grant no. ZR2019MC041); Youth Talent Program Supported by Laboratory for Marine Fisheries Science and Food Production Processes, Pilot National Laboratory for Marine Science Technology (Qingdao) (grant no. 2018-MFS-T11).
R. S. designed and performed the experiments, analysed the data and wrote the manuscript; J. L., Y. L., X. L., J. Z. and W. X. performed the experiments; K. M. supervised the project; Q. A. supervised the project and wrote the manuscript; M. W. supervised the project, designed the experiments, analysed the data and wrote the manuscript.
The authors declare no competing financial interests.
