Dietary Trivalent Chromium Exposure Up-Regulates Lipid Metabolism in Coral Trout: The Evidence From Transcriptome Analysis

Diet quality greatly affects an animal’s performance and metabolism. Despite the fact that trivalent chromium [Cr(III)] is considered an essential element and is widely used in nutritional supplements for animals and humans, the potential toxicity of Cr(III) is unclear. Here, liver transcriptome sequencing was performed on coral trout (Plectropomus leopardus) exposed to 200 mg kg–1 of dietary organic Cr(III) [as chromium picolinate (CrPic)] for 8 weeks. One-hundred-and thirteen differentially expressed genes (DEGs) were identified in response to Cr(III) stress, in comparison to the control, including 31 up-regulated and 82 down-regulated DEGs. Clusters of Orthologous Groups of proteins (COG) classifies DEGs into 15 functional categories, with the predominant category being related to lipid transport and metabolism (9.73%). The Kyoto Encyclopedia of Genes and Genomes (KEGG) assigned DEGs to six major categories with robust DEGs as part of the lipid metabolism pathway (18.58%). Moreover, KEGG functional enrichment analysis showed that these DEGs are primarily related to steroid biosynthesis, terpenoid backbone biosynthesis, and steroid hormone biosynthesis pathways, of which steroid biosynthesis was the most significant pathway, and 12 key up-regulated DEGs (dhcr7, dhcr24, ebp, lss, msmo1, sqle, cyp51, tm7sf2, sc5dl, fdft1, nsdhl, and hsd17b7) were found for steroid biosynthesis pathways. To validate the RNA sequencing data using quantitative real-time PCR (qRT-PCR), qRT-PCR results indicate that the expression of genes encoding HMGCR, TM7SF2, TRYP2, CTRL, EBP, LSS, and CYP51 were induced, while those encoding THRSP, LCE, and MCM5 were reduced, consistent with RNA-seq results. This findings provides the first evidence that a long-term high dose of Cr(III) intake causes lipid metabolism disorder and potential toxicity in fish. Cautious health risk assessment of dietary Cr(III) intake is therefore highly recommended for the commercial and/or natural diets of aquatic animals, which has previously largely been ignored.


INTRODUCTION
An animal's growth performance is largely influenced by food quality, and the safety of food affects its quality. In August 2005, the Food and Drug Administration (FDA) in the United States officially approved the production of organic chromium (CrPic) (FDA, 2005). CrPic a trivalent chromium [Cr(III)] and its safety is controversial. Despite this, Cr(III) is considered an essential element and is widely used in nutritional supplements for animals and humans (Tian et al., 2013;Hamilton et al., 2018;Mayorga et al., 2018;Krol et al., 2020).
Several studies have shown that CrPic had cardioprotective effects in hyperlipidemic rats (Shafik et al., 2017), hepatoprotective activity in alloxan dosed mice (Fan et al., 2012), and antioxidative effects in rats (Marmett et al., 2018). Specifically, Cr(III) is considered to be a glucose tolerance factor (GTF) and has an increasing insulin sensitivity effect. Cr is considered to be an essential micronutrient and is linked to many processes regulated in the human body, including activation of insulin receptors by the oligopeptide chromudulin to maintain glucose homeostasis, thereby increasing insulin sensitivity and signal transduction (NIH, 2013). Cr deficiency may cause glucose intolerance, fasting hyperglycemia, increased circulating insulin, and can even impair growth (Vincent, 2000). Wang et al. demonstrated that AMP-activated protein kinase activates the increase of dietary CrPic inhibited secretion of insulin-resistant 3T3-L1 adipocytes . CrPic could also improve protein deposition via the regulation of mRNA levels of ubiquitin (Ub), insulin-like growth factor 1 receptor (IGF-1R), and insulin-like growth factor 1 (IGF-1) in rat skeletal muscle cells (Peng et al., 2010), and glucose uptake and metabolism via up-regulating the mRNA levels of glycogen synthase (GS), glucose transporter 4 (GLUT4), uncoupling protein-3 (UCP3), and insulin receptor (IR) in skeletal muscle cells (Qiao et al., 2009). Dietary CrPic supplementation has been found to stimulate weight loss in humans (Tian et al., 2013).
Conversely, isolated studies have reported that CrPic caused harmful effects, such as mutagenicity in a Drosophila model (Hepburn et al., 2003) and in mammalian cells (Andersson et al., 2007). Additionally, mitochondrial damage and apoptosis was reported in the CHO AA8 cell line (Jiang et al., 2018) and in a bacterial model (Jiang et al., 2018). Furthermore, the deleterious effects of dietary CrPic supplementation have been confirmed in rats, including oxidative DNA damage (Andersson et al., 2007), increased lipid peroxidation (Refaie et al., 2009), and decreased antioxidative enzyme activity (Shafik et al., 2017). CrPic intake was also found to lead to oxidative damage in pigs (Tan et al., 2008), and apoptotic effects in chick embryo fibroblasts (Bai et al., 2014). In humans, in some cases, dietary Cr(III) may lead to Cr(III) accumulation in the kidneys (Stoecker, 1999). These findings raise concerns about the potential toxicity and safety of CrPic as a dietary supplement.
In aquatic animals, a few studies have demonstrated that dietary CrPic supplementation could significantly improve growth performance and carbohydrate metabolism in fish, such as mirror carp (Cyprinus idellus; Ahmed et al., 2012), and Nile tilapia (Oreochromis niloticus; Li et al., 2018). However, other studies reported that dietary CrPic supplementation did not affect the health status and the growth of rainbow trout (Selcuk et al., 2010) and Nile tilapia (Pires et al., 2015). These inconsistent results might be due to differences in species, exposure time and/or the dietary CrPic concentrations used in those studies. Nevertheless, to date, the toxicity of CrPic has been rarely investigated in aquatic animals, especially in a long-term dietary CrPic exposure scenario, which has probably been occurring in animals feeding on Cr(III) enriched commercial feeds or contaminated natural diets.
In the present study, therefore, we investigated the chronic toxicity of dietary organic Cr(III) [as chromium picolinate (CrPic)] in juvenile coral trout (Plectropomus leopardus), a commercially and ecologically important coral fish listed in the Red List of Threatened Species as being Near Threatened (IUCN, 2004). The aim of this study was to use transcriptomics to cast light on the mechanisms of effects elicited by dietary CrPic exposure.
After 2 weeks of acclimatization, the coral trout were randomly separated into two groups-the control group and dietary CrPic treatment group (CrPic). Each group had three triplicates containing 18 fish in total, with six fish in each treatment held in one tank. In the control group, the coral trout were fed a commercial diet [the total Cr concentration in the diet was 2.21 ± 0.14 mg kg −1 , this is a baseline concentration of Cr (Supplementary Table 1)]. The CrPic group were fed the same commercial diet, but with a nominal Cr concentration of 200 mg kg −1 added as chromium picolinate (Sigma-Aldrich). The measured value for the Cr content of the CrPic diet was 195.67 ± 9.97 mg kg −1 (Supplementary Table 1). The nominal Cr concentration of 200 mg kg −1 was based on our preliminary experiment and previously published studies (e.g., Kim and Kang, 2016;Staniek and Wojcik, 2018;Ren et al., 2018). The commercial diet and chromium picolinate were ground and mixed thoroughly, then pellets were reformed with a diameter of 3.0 mm using a laboratory press. Pellets were then oven dried at 60 • C and stored at −20 • C. During the feeding experiment, the seawater in the tank was replaced by fresh filtered seawater six times per day. The concentration of total Cr was <0.045 mg L −1 in the water. On the 56th day of the feeding experiment, the fish were euthanized and the livers were collected and then frozen in liquid nitrogen for later preparation for transcriptome sequencing.

Transcriptome Sequencing
The livers from the control and CrPic group (36 fish in total; n = 18 per group) were used for RNA sequencing, which was completed by Majorbio Bioinformatics Technology Co., Ltd. (Shanghai, China). In brief, using TRIzol reagent (Life Technologies, California, United States) total RNA was extracted from each liver sample according to the manufacturer's instructions. We then tested RNA degradation and pollution on 1% agarose gels, and we tested the RNA extract's purity and concentration using the 260/280 nm ratio determined by a Nanotrop 2000 (IMPLEN, CA, United States). We used 2 µg of RNA per liver sample as the input material for the transcriptome sequencing. To generate sequencing libraries, we used NEBNext Ultra RNA Library Prep Kit according to the manufacturer's recommendations. We then sequenced libraries on an Illumina Hiseq 2500 platform and generated 150 bp pairedend reads.

Quantitative Real-Time PCR Analysis
We randomly selected 10 differentially expressed genes for qRT-PCR verification. β-actin was used as an internal reference gene. The primers were designed by Primer Premier 6 (Supplementary Table 2). The amplification efficiency of the tested targeted genes is shown in Supplementary Table 3. Complementary DNA (cDNA) were synthesized. The qRT-PCR reaction was run on a QuantStudio 6 Flex Real-Time PCR System (Thermo Fisher Scientific, United States), using 2× SYBR Green qPCR Mix (Vazyme, Biotech, China) as recommended by the manufacturer's protocol. A melting curve was generated for each qRT-PCR product. The relative mRNA expression levels of the selected genes were calculated using the 2 − CT method (Pfaffl, 2001).

Statistical Analysis of mRNA Data
Statistical analyses (t-test and one-way ANOVA) were performed to test the difference in the mean values between different groups using SPSS 19.0 software (SPSS Inc., Chicago, IL, United States).
The Shapiro-Wilk test and Levene's test were used to verify normality and homoscedasticity of the data, respectively. All the data was presented as mean ± standard deviation (SD, n = 6).

RNA-Seq
We obtained 62,923,240 and 56,630,392 mean raw reads in the control and CrPic group, respectively ( Table 1). By removing reads from the undetermined base, containing adapter reads and low-quality raw reads, we obtained clean reads. Following this, there were 62,062,992 and 55,821,019 mean clean reads in the control and CrPic group, respectively. The obtained Q20 and Q30 were more than 94%, and the error rate was less than 0.05%. The 7 http://www.geneontology.org/ The control groups were Con_1, Con_2 and Con_3. The treatment groups were CrPic_1, CrPic_2 and CrPic_3.
Frontiers in Physiology | www.frontiersin.org GC content was ca. 50% in the control and CrPic-treated groups, indicating high quality transcriptome sequencing in our study.

The Differential Expression Genes (DEGs)
There were 113 differential expression genes [31 down-regulated and 82 up-regulated DEGs (Supplementary Figure 1)] obtained based on p < 0.05 and the gene expression values of >twofold change (Figure 1 and Supplementary Table 4).

GO, COG, KEGG Functional Classification Analysis for DEGs
In this study, the 113 DEGs annotated in GO were assigned to three categories, which were biological process (62.83%), cellular component (63.72%), and molecular function (38.94%) (Figure 2). These DEGs were further divided into 20 subcategories (Supplementary Table 5).
Using GO in the DAVID bioinformatics software (version 6.8) to analyze the down-regulated and upregulated DEGs, we found that 12 DEGs were involved in cellular process (10.62%), whilst 14 DEGs each were found for single-organism process (12.39%), catalytic activity (12.39%), and binding (12.39%). Interestingly, we further found that these DEGs were all up-regulated and involved in the main abundant subcategories (Supplementary Figure 2).
In the COG database, the 113 DEGs were classified into 15 functional categories (Figure 3 and Supplementary Table 6). The predominant category was the cluster for lipid transport and metabolism (9.73%), followed by energy production and conversion (4.42%) and post-translational modification, protein turnover, and chaperones (3.54%) (Supplementary Table 6).
and Supplementary Table 4). We also observed that for lipid metabolism, 18 out of 21 DEGs were up-regulated (Supplementary Figure 3).

KEGG and GO Enrichment Analysis of DEGs
The analysis of KEGG enrichment revealed that top of 18 pathways and three of these pathways were significantly enriched (p < 0.05, Fisher's exact test followed by Bonferroni), namely those for steroid biosynthesis, terpenoid backbone biosynthesis, and steroid hormone biosynthesis (Figure 5 and Supplementary Figures 5, 6). To confirm these significant enrichment pathways, we conducted a GO enrichment analysis and observed six significantly enriched pathways (p < 0.05, Supplementary Figure 4), including those for lipid biosynthesis, isoprenoid biosynthesis, lipid metabolic processes, isoprenoid metabolic processes, single-organism biosynthetic processes, and cellular lipid metabolic processes. These results showed that lipid metabolism is involved in the significantly enriched pathways, which are strongly correlated with the KEGG functional classification analysis.

DISCUSSION
Our study provided the first evidence that a long-term, high dose of dietary CrPic exposure significantly affects lipid metabolism in aquatic animals at the gene expression level. The three most enriched pathways were steroid biosynthesis, terpenoid backbone biosynthesis, and steroid hormone biosynthesis.

Steroid Biosynthesis Was Up-Regulated by Dietary CrPic Exposure
Steroid biosynthesis is a well-known component of lipid metabolism in animals (Si et al., 2018), and it plays a vital role in responding to environmental pressure of aquatic animals (e.g., salinity change: Xu et al., 2015). We found evidence that the steroid biosynthesis was significantly affected by dietary CrPic supplementation in marine fish by identifying several steroid biosynthesis-related genes that were up-regulated by CrPic exposure (e.g., ebp, lss, dhcr7, dhcr24, msmo1, sqle, cyp51). As far as we know, this is the first study showing that dietary CrPic has an effect on the expression of steroid biosynthesisrelated genes in fish. The ebp gene, specifically, codes for the enzyme sterol isomerase, which plays an important function in regulating the pathway of sterol biosynthesis in animals (Sanchez-Pulido and Ponting, 2014). Sterols act as an important compounds, and it works in lots of biological processes, such as possessing transport capability and acting as cell membrane components (Schaller, 2003). In this study, we observed that the expression of the ebp gene was up-regulated in response to dietary CrPic exposure, suggesting an increased capacity to stimulate biosynthesis of sterols in the livers of P. leopardus. In addition, lss gene transcription was also increased under dietary CrPic exposure. The lss gene encodes the enzyme lanosterol synthase, which is a key enzyme in the cholesterol biosynthetic pathway. Lanosterol is one of the upstream precursors in the pathway of sterol biosynthesis for many animals (Chen et al., 2007). Increased lss expression thus indicated enhanced synthesis of cholesterol in the livers of P. leopardus.
The cyp51 gene specifically encodes for the lanosterol 14-alpha demethylase protein, which is a key enzyme for the conversion of lanosterol to cholesterol, and thus acts as an important component in lipid metabolism and the steroid biosynthesis pathway (Lewinska et al., 2014). Long-term exposure of dietary CrPic led to the up regulation of cyp51, indicating active cholesterol synthesis in P. leopardus. Conversely, other studies have reported that long-term exposure of difenoconazole downregulates the expression of sterol-genesis genes (e.g., cyp51) in zebrafish (Mu et al., 2015). While another study reported that the lss and dhcr24 (24-dehydrocholesterol reductase) gene expressions were up-regulated in the steroid biosynthesis pathway in Nile tilapia (O. niloticus) under salinity stress (Xu et al., 2015). Long-term exposure to dietary CrPic up regulated both of these genes in P. leopardus, lending further evidence that liver cells were required to increase cholesterol synthesis in response to long-term Cr exposure and that this was likely a stress response. An interesting and recent study found that eight genes (FDFT1, SQLEA, LSS, MSMO1, NSDHL, EBP, SC5D, DHCR7) were remarkably up-regulated in the steroid biosynthesis pathway under low temperature (12 and 14 • C) (Shi et al., 2020). These results strongly indicate that up-regulation of the steroid biosynthesis related genes may play an important role in Cr defense and different tolerance responses (e.g., cold stress, salinity stress).
Frontiers in Physiology | www.frontiersin.org 9 February 2021 | Volume 12 | Article 640898 FIGURE 7 | qRT-PCR validation of up-regulated and down-regulated genes analyzed by RNA-Seq. qRT-PCR was performed for 10 selected genes that were identified as differentially expressed between control and CrPic-treated group. The Y -axis shows the relative mRNA expression levels.
diphosphate (DMAPP) and isopentenyl pyrophosphate (IPP) transformed to farnesyl-PP (requiring the product of fppfarnesyl pyrophosphate synthase), and farnesyl diphosphate synthease helps produce terpenoids (the product of fdps; Wasko et al., 2011). Previous studies have shown that the biosynthesis pathway of the terpenoid backbone and steroid were activated in the brain of grass carp (Ctenopharyngodon idellus) under low temperatures conditions. They found that six genes (HMGCR,HMGCS1, ACAT2, MVD, IDI1, MVK) were up-regulated at 12 and 14 • C in the terpenoid backbone biosynthesis pathway (Shi et al., 2020). Consistent with previous studies, we found that five genes (HMGCR, PMVK, MVD, IDI, FDPS) were significantly up-regulated in the terpenoid backbone biosynthesis pathway under organic Cr (III) [as chromium picolinate (CrPic)] exposure; some of the selected genes (HMGCR, EBP, LSS) were verified by qRT-PCR. Another study also showed that the expression of terpenoid backbone and steroid biosynthesis were affected in the liver of Atlantic salmon (Salmo salar) by heat stress. They found that hmgcr was down-regulated after 24 h of heat stress in terpenoid backbone biosynthesis, and they also found that the expression of genes (fdft, lss sc5d, sqle) associated with the steroid biosynthesis pathway were reduced (Shi et al., 2019). In the present study, we found the related gene expressions of those key enzymes involved in the terpenoid backbone biosynthesis pathway and steroid biosynthesis pathway, were all up-regulated by CrPic exposure. Therefore, the up-regulation of these enzymes, related to terpenoids biosynthesis and steroid biosynthesis, showed a requirement for increased lipid metabolism in the fish exposed to dietary CrPic.

Steroid Hormone Biosynthesis Was Up-Regulated by Dietary CrPic Exposure
Steroid hormones play a critical role in regulating stress responses, water and salt balance, and nutrient metabolism (Schiffer et al., 2019). Hydroxysteroid dehydrogenase (hsd) and cytochrome P450 (cyp) enzymes are two major enzymes involved in steroid hormone biosynthesis (Schiffer et al., 2019). Hydroxysteroid 17-beta dehydrogenase 7 (hsd17b7) and cytochrome P450 family 7 subfamily A member 1 (cyp7a1), are important enzymes required for steroid hormone biosynthesis. In our study, we found that the expression of cyp7a1 and hsd17b1 genes were up-regulated. These enzymes are vital to the biosynthesis of steroid hormones, indicating that the steroid hormone biosynthesis of the fish was significantly stimulated by dietary CrPic supplementation. This is the first report to indicate that steroid hormone biosynthesis is influenced under dietary CrPic exposure in teleost fish. Both cyp17a1 and hsd3b2 are key components of cortisol production, a hormone produced in response to various pressure backgrounds, it can be used as a biomarker when aquatic animals respond to stress (Tokarz et al., 2015). A previous study has confirmed that fish homeostasis under cutaneous stress increased cortisol levels (Kulczykowska et al., 2018), suggesting that long-term Cr exposure causes stress to P. leopardus.

Signal Transduction Was Indispensable in Lipid Metabolism Disorder
In the present study, in the CrPic group, genes from several cellular regulatory pathways related to signal transduction were expressed differentially. The affected pathways included calcium signaling (chrm5, mcub, p2rx7), MAPK signaling (dusp1, hsp70, IL-1R1, tgfb3), FoxO signaling (pepck, tgfb3), AMPK signaling (acod, hmgcr, pepck), ErbB signaling (hbegf ), and Wnt signaling (chd8). Mitochondria performs critical roles in many cell signaling pathways. Several mechanisms are involved, including altered Ca 2+ levels and reactive oxygen species (ROS) generation (Chandel, 2014). In our results, the expression of a gene central to the regulation of Ca 2+ homeostasis in mitochondria and the cytosol, the mcub gene, was up regulated. The product of this gene negatively regulates the activity of the mitochondrial Ca 2+ uniporter (MCU), decreasing Ca 2+ transport into the mitochondria (Mammucari et al., 2018). This can increase cytosolic Ca 2+ , which activates signaling pathways that regulate many cellular processes (Bohovych and Khalimonchuk, 2016). At the same time, Ca 2+ in the mitochondrial matrix is decreased, limiting the production of ROS (Jhun et al., 2016) due to the decreased activity of key Ca dependent enzymes that are required for oxidative metabolism in the mitochondria (Mammucari et al., 2018). The main production of ROS in the cell is mitochondria, and reduced ROS production by this organelle will affect the many cellular processes regulated through ROS (Chandel, 2014). Moreover, in mammal cardiomyocytes, MCUb was up-regulated after injury, apparently as a stress response (Lambert et al., 2019), which was consistent with the transcription level response observed in the fish liver cells in response to stress caused by dietary CrPic intake in this study.
Overall, the changes in transcription level expression of the genes observed in our study suggests that the cells of the fish liver adapted to dietary CrPic exposure by altering several signaling pathways, which could in turn potentially affect many aspects of the liver's cell functions and biochemistry.

CONCLUSION
In conclusion, the present study successfully constructed a database of transcriptome changes in juvenile coral trout under long-term, high dose exposure to dietary CrPic. We identified 113 differentially expressed genes (DEGs) significantly enriched in the CrPic treatment. Our study demonstrates that the RNA-Seq analysis is highly effective in determining molecular/cellular level responses to potentially toxic contaminants in aquatic animals. The findings here demonstrate that dietary CrPic exposure significantly up-regulates lipid metabolism in teleost fish at the transcription level, mainly by affecting steroid and terpenoid backbone biosynthesis pathways, and altered signal transduction pathways. This suggests that a high dose of organic dietary Cr(III) exposure results in stress and sub-lethal toxic effects in aquatic animals. Cautious health risk assessment of dietary Cr(III) intake is therefore highly recommended in both commercial and natural diets for aquatic animals, which has previously been largely ignored.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm. nih.gov/, PRJNA684012.

ETHICS STATEMENT
The animal study was reviewed and approved by the Hainan University Institutional Animal Use and Care Committee (Hainan, China).

AUTHOR CONTRIBUTIONS
LW analyzed all the data, wrote, and revised the manuscript. YL and HY performed the qRT-PCR experiment and collected the materials. JX, CH, and IG revised the manuscript. ZG and DH designed the study and revised the manuscript. All authors have read and approved the final manuscript.