Dietary Cholesterol Supplementation Inhibits the Steroid Biosynthesis but Does Not Affect the Cholesterol Transport in Two Marine Teleosts: A Hepatic Transcriptome Study

Cholesterol has been used as additive in fish feeds due to the reduced use of fish meal and fish oil. In order to evaluate the effects of dietary cholesterol supplementation (D-CHO-S) on fish physiology, a liver transcriptome analysis was performed following a feeding experiment on turbot and tiger puffer with different levels of dietary cholesterol. The control diet contained 30% fish meal (0% fish oil) without cholesterol supplementation, while the treatment diet was supplemented with 1.0% cholesterol (CHO-1.0). A total of 722 and 581 differentially expressed genes (DEG) between the dietary groups were observed in turbot and tiger puffer, respectively. These DEG were primarily enriched in signaling pathways related to steroid synthesis and lipid metabolism. In general, D-CHO-S downregulated the steroid synthesis in both turbot and tiger puffer. Msmo1, lss, dhcr24, and nsdhl might play key roles in the steroid synthesis in these two fish species. Gene expressions related to cholesterol transport (npc1l1, abca1, abcg1, abcg2, abcg5, abcg8, abcb11a, and abcb11b) in the liver and intestine were also extensively investigated by qRT-PCR. However, the results suggest that D-CHO-S rarely affected the cholesterol transport in both species. The protein-protein interaction (PPI) network constructed on steroid biosynthesis-related DEG showed that in turbot, Msmo1, Lss, Nsdhl, Ebp, Hsd17b7, Fdft1, and Dhcr7 had high intermediary centrality in the dietary regulation of steroid synthesis. In conclusion, in both turbot and tiger puffer, the supplementation of dietary cholesterol inhibits the steroid metabolism but does not affect the cholesterol transport.


Introduction
Cholesterol, as an essential substance not only for the formation of cell membranes but also for the synthesis of bile acids, steroid hormones, and vitamin D, is the most abundant steroid compound in fish [1,2]. As vertebrate, fish are able to synthesize cholesterol themselves. Therefore, cholesterol is generally considered as a nonessential nutrient for fish, resulting in only little research interest on the role and mechanism of dietary cholesterol in fish. Limited previous studies in this research area have focused on the effects of dietary cholesterol supplementation (D-CHO-S) on the growth and feed utilization of fish [3][4][5], and less research has been done regarding the effects on steroid metabolism or other relevant physiological processes. In previous experiments, we have evaluated the effects of D-CHO-S on the growth, tissue biochemical parameters, and expression of lipid metabolism-related genes in turbot and tiger puffer [6], but not the effects on cholesterol homeostasis.
Cholesterol homeostasis is a balance of catabolism, synthesis, intestinal absorption, and biliary secretion. Previous studies have shown that progesterone receptor membrane component 1 (Pgrmc1), cytochrome P450 (CYP), lanosterol 14-alpha demethylase (Cyp51), hydroxysteroid dehydrogenase/ketosteroid reductase, methylsterol monooxygenase 1 (Msmo1), and lanosterol synthase (2,3-oxidosqualene-lanosterol cyclase) (Lss) play key roles in steroid synthesis [7][8][9][10][11]. The cholesterol absorption can be regulated by ATP-binding cassette (ABC) transporters [12]. For example, ABC subfamily A, member 1 (Abca1) serves the efficient cholesterol transport from enterocytes to high-density lipoproteins (HDL) in the serum [13]. Abca1 is also expressed in the liver where it mediates excretion of cholesterol into bile [14]. Other members of ABC family, subfamily G member 5 (Abcg5) and member 8 (Abcg8), are implicated in cholesterol absorption [15]. Member 1 of this subfamily (Abcg1) can also mediate the efflux of free cholesterol to mature HDL [16]. Reverse cholesterol transport (RCT) is known as HDL-mediated transport of cholesterol from peripheral tissues to the liver, where cholesterol can then be removed from via biliary secretion [17][18][19]. The ABC family also plays important roles in RCT. To more comprehensively investigate the roles of dietary cholesterol in fish physiological processes, in this study, a transcriptomic assay was used to screen the metabolic processes in turbot and tiger puffer, which were responsive to D-CHO-S. In previous studies, it has been observed that D-CHO-S inhibits the synthesis of cholesterol and promotes the synthesis of bile acid [3-5, 20, 21], mainly due to reduced hydroxymethylglutaryl CoA reductase (HMG-CoAr) activity and upregulated cytochrome P450 7A1 (cyp7a1) gene expression. However, the available information is still not comprehensive. Nowadays, the transcriptomic analysis technology has become a useful tool in metabolic studies. The transcriptomic analysis may help to elucidate the metabolic progresses and signaling pathways underlying the changes in phenotypic responses.
Turbot (Scophthalmus maximus) and tiger puffer (Takifugu rubripes) are important aquaculture species [22]. On turbot and tiger puffer, as a follow-up study of Meng et al. [6], the current study is aimed at investigating the physiological responses, in particular the cholesterol metabolismrelated ones, of turbot and tiger puffer to D-CHO-S, with transcriptome sequencing. Liver samples from two experimental groups with or without extra cholesterol supplementation were used for the transcriptomic assay. The results obtained from this study will provide basic data for future research in this research area.

Experimental Diets and Feeding
Trial. Liver samples were collected from a previous study, and therefore, the detailed procedures for diet preparation and feeding trial have been described previously [6]. Two of the five groups with graded levels of D-CHO-S in the previous study were used in this study. Briefly, the control diet contained 30% fish meal level without fish oil (Table 1) [23][24][25][26]. A commercial cholesterol reagent (AR, purity > 95%, Macklin) was supplemented into the control diet at the level of 1.0% to obtain the treatment diet, which was named CHO-1.0. A 10-week feeding experiment with turbot (21 g) and tiger puffer (12 g) was conducted. Each diet was assigned to triplicate polyethylene tanks (200 L, 30 fish per tank). Six fish were randomly collected from each tank for the collection of liver and intestine samples. A whole intestine sample was divided into three parts: anterior intestine (the part near the pyloric caeca), mid intestine (the part near the cecum), and hind intestine (the cecum). All sampling and fish rearing protocols in this study were approved by the Animal Care and Use Committee of Yellow Sea Fisheries Research Institute.

Transcriptome Sequencing and Bioinformatic Analysis.
The detailed procedures for RNA isolation, construction of cDNA library, and sequencing have been described in previous publications [27]. Six samples individually collected from the six fish per tank were pooled. Therefore, at last, six pooled samples in total were used for the preparation of six individual cDNA libraries.

Aquaculture Nutrition
Raw reads in fastq format were processed with in-house Perl scripts, during which clean reads can be obtained after low-quality reads, ploy-N-containing reads, and adaptorcontaining reads were removed. Meanwhile, other features such as GC content, Q30, and Q20 can be calculated. All the downstream analyses used the clean data.
Prior to the differential expression analysis, the counts of read were adjusted by edge R through one scaling normalized factor. The analysis of differential expression between the two experiment groups was conducted with DESeq2 R (1.20.0). DESeq2 determines the differential gene expression (P < 0:05) with a model, which involves the negative binomial distribution.
Gene Ontology (GO) enrichment analysis of the differentially expressed genes (DEG) was conducted with cluster Profiler R, during which the bias of gene length can be corrected. Adjusted P < 0:05 indicates significant enrichment by DEG. The Profiler R was also used to test the KEGG enrichment by DEG.

Quantitative Real-Time Polymerase Chain Reaction
(qRT-PCR). Due to the fact that the transcriptomic analysis revealed very few DEG related to cholesterol transport, which was unexpected, qRT-PCR was conducted on selected cholesterol transport-related genes, in order to verify this result, as well as to validate the accuracy of the transcriptomic analysis. The qRT-PCR experiment was conducted for both fish species.
The qRT-PCR methods (see Table 2 for primers), regents (Accurate Biotechnology and Tsingke Biological Technology), and equipment (Roche LightCycler 96, Basel, Switzerland) were the same to our previous publications [6]. The relative mRNA expression was evaluated with the 2 −ΔΔCT method [28]. All data were subjected to T test for independent samples in SPSS 16.0 for Windows. P < 0:05 indicates significant difference. The results are expressed as mean ± standard error.

Sequence Assembly.
In turbot, a total of 133,323,428 and 130,010,624 clean reads were generated for the control group and the CHO-1.0 group, respectively, corresponding to a total clean base of 19.99 and 19.51 G, respectively. For the two groups, the average Q20 and Q30 (the percentage of base with Phred value > 20 and 30, respectively) were 97.68% and 93.76%, respectively. This suggested that the sequencing had high accuracy.

Aquaculture Nutrition
In tiger puffer, a total of 123,437,502 and 127,602,710 clean reads were generated for the control and CHO-1.0, respectively, corresponding to total clean bases of 18.51 and 19.14 G, respectively. For the two groups, the average Q20 and Q30 were 97.83% and 94.03%, respectively. Raw reads were deposited at NCBI's Sequence Read Archive. The accession nos. were PRJNA933270 (turbot) and PRJNA933719 (tiger puffer) (D0 and D1 in the archived files match the control and CHO-1.0, respectively).

Differentially
Expressed Genes (DEG) between the Two Experiment Groups. In turbot, a total of 722 genes had significantly different expression (P < 0:05) between the control and CHO-1.0 (Figure 1(a)). Diet CHO-1.0 upregulated the mRNA expression of 382 genes and downregulated that of 340 genes. In tiger puffer, a total of 581 genes had signifi-cantly different expression (P < 0:05) between the two treatments (Figure 1(b)). Compared to the control, diet CHO-1.0 upregulated the mRNA expression of 335 genes and downregulated that of 246 genes.

qRT-PCR Validation of Gene Expression Related to
Cholesterol Transport. Since few of these DEG are related to cholesterol transport, which is an important component of cholesterol homeostasis, to get more comprehensive information about the regulation of cholesterol transportrelated transcription by D-CHO-S, the gene expression of 10 cholesterol transport-related genes, which showed no significant difference between the control and CHO-1.0 from the transcriptomic analysis, was investigated with the qRT-PCR method (Figures 5 and 6). The results showed that very few changes were observed in response to D-CHO-S in both fish species. Significant changes were observed only in turbot liver and tiger puffer hind intestine. In turbot liver, D-CHO-S significantly downregulated the mRNA expression of abca1 but significantly upregulated that of abcg5 and abcb11a (Figure 5(a)). In tiger puffer liver, D-CHO-S significantly (P < 0:05) downregulated the mRNA expression of npc1l1 and abca1 (Figure 6(a)). In tiger puffer hind intestine, D-CHO-S significantly downregulated the transcription of npc1l1 (Figure 6(d)).

The Protein-Protein Interaction (PPI) Network
Construction Based on the Steroid Biosynthesis-Related DEG. In turbot, msmo1, lss, nsdhl, ebp, hsd17b7, fdft1, and dhcr7 had higher intermediary centrality than other DEG, indicated by larger node degrees in the PPI network, whereas in tiger puffer, the intermediary centrality of all DEG was the same (Figure 7). In turbot, the clustering coefficient of dhcr24, sc5d, and sqlea was higher, while in tiger puffer, the clustering coefficient of all DEG was the same.
Among the shared DEG between the two fish species, msmo1, lss, dhcr24, and nsdhl might play key roles in the steroid synthesis, because they were more sensitive to D-CHO-S. In turbot, this was confirmed by the PPI network analysis, Table 3: Differentially (P < 0:05) expressed genes (DEG) between the two dietary groups related to steroid biosynthesis. "−" before log 2 FC value represents downregulation in the CHO-1.0 group compared to the control group. FC: fold change. which showed that msmo1, lss, and nsdhl had high intermediary centrality and dhcr24 had high clustering coefficient. Nevertheless, in tiger puffer, all DEG had the same intermediary centrality. This could be related to the fact that in tiger puffer, the PPI network was constructed based on less DEG, and thus, there was not enough information to support a high-quality PPI network. Different responsive genes were also observed between turbot and tiger puffer. Four more cholesterol biosynthesis-related DEG, sqlea, cyp51, tm7sf2, and hsd17b7, were observed in turbot compared to tiger puffer. Sqlea, which stereospecifically oxidized squalene, is a crucial enzyme in the biosynthesis of steroid [44,45]. Cyp51 is a key enzyme in the conversion of lanosterol to cholesterol [46]. Tm7sf2 reduces the C14-unsaturated bond of lanosterol [47], and Hsd17b7 is a bifunctional protein in the metabolism of steroid hormone [48]. The difference mentioned above between the two species may be due to the different lipid storage patterns of the two fish species. Cholesterol is synthesized in essentially all tissues of the body, but the liver is presumed to be the primary site. However, for tiger puffer, besides lipid metabolism, the liver also functions as lipid storage organ, which may weaken the function of cholesterol metabolism. Our previous studies have suggested that in tiger puffer, the intestine is probably a lipid metabolism center, but the liver may function as a pure lipid storage organ [49]. Another explanation of this difference could be the fact that tiger puffer body composition has a higher buffer capacity than turbot in response to dietary regulation, which has been indicated by our previous studies [6].
In the present study, the expression of all steroid biosynthesis-related DEG described above was significantly downregulated by D-CHO-S. As shown in Figure 4 [29], all these genes associate with the cholesterol synthesis process and are key enzymes in the different reaction stages of cholesterol synthesis. In both turbot and tiger puffer, downregulation of the steroid biosynthesis genes clearly indicates that the cholesterol synthesis is inhibited when cholesterol was supplemented at 1%. This result was similar to the findings in Atlantic salmon (Salmo salar) [50]. The endogenous cholesterol synthesis was probably spared by the exogenous cholesterol supply in the diet. On the other hand, exogenous cholesterol supply may help to guarantee enough substrate for conversion to bile acids. This has been confirmed by our previous research [6], which showed that in both fish species, D-CHO-S downregulated the gene expression of hmg-coar, a key enzyme for the biosynthesis of cholesterol [51], but upregulated the cyp7a1 expression, which has a limiting role in the biosynthesis of bile acid [52].

Aquaculture Nutrition
Cholesterol transport is another important process of cholesterol homeostasis. The intestinal absorption and reverse cholesterol transport (RCT), which is the transport of cholesterol from extrahepatic tissue towards the liver for eventual excretion [17][18][19], all rely on ATP-binding cassette proteins. Abcg1 catalyzes the efflux of cholesterol from cell to HDL [53], and Abcg2 mediates the transport of intracellular substrate outside the cells [54,55]. Abcg5 and Abcg8 10 Aquaculture Nutrition collaboratively facilitate the transmembrane transport of sterol. They play important roles in the selective excretion of sterol by the liver into bile [56,57]. Abcb11, which is found in the canalicular membrane, is a primary transporter for the continuous secretion of hepatic bile acids to bile duct [58]. However, the present results showed that D-CHO-S had almost no significant effect on the mRNA expression of these cholesterol transport-related genes. When a certain amount of cholesterol is added to the feed, it is presumed that the cholesterol synthesis decreases and cholesterol excretion capacity increases [51]. This feedback regulation has been observed in another turbot study [20]. Nevertheless, when the cholesterol content is too high, it may be difficult for the fish to fully metabolize them, and the feedback regulation may thereby be impaired. Another explanation of the present result was that the increased cholesterol absorption may offset the increased excretion.
The very few changes in the relative mRNA expression of cholesterol transport-related genes in response to D-CHO-S include upregulation of abcg5 and abcb11a and downregulation of abca1 and npc1l1, mostly in the liver.  Figure 7: Protein-protein interaction (PPI) network based on the steroid biosynthesis-related differentially expressed genes (DEG) in turbot (a) and tiger puffer (b). Larger node size indicates larger node degree, which is the number of proteins interacting with this node. Brighter node color indicates higher clustering coefficient. Larger edge size indicates higher interacting intensity.
11 Aquaculture Nutrition has also been observed in Atlantic salmon [50]. Because dietary cholesterol is absorbed into small intestinal epithelial cells through Npc1l1 protein, this downregulation could be a typical feedback regulation of fish to reduce the absorption of cholesterol, probably to prevent potential negative influence of excess cholesterol. Abca1 plays a key role in the efflux of intracellular cholesterol to apolipoproteins and the formation of nascent HDL, which facilitates the RCT. The downregulation of abca1 could indicate an impaired RCT capacity, which was consistent with the reduced HDL-C/ LDL-C ratio in the serum [6]. It is well known that LDL helps to carry cholesterol from the liver to the peripheral tissues, while HDL helps to transport cholesterol in an opposite direction [59,60]. Therefore, the HDL-C/LDL-C ratio can be used as an indicator of RCT [61,62]. In other turbot studies, when the D-CHO-S level was equal to or greater than 1% (total dietary cholesterol level equal to or greater than 1.25%), the cholesterol easily accumulated in the peripheral tissues [3,20,21]. The upregulation of abcg5 expression by D-CHO-S could be a direct response of Abcg5 to increased cholesterol uptake, considering the roles of Abcg5 in the transmembrane transport of cholesterol across the enterocyte membranes. Anyway, compared to the significant influence on cholesterol biosynthesis, the influence of D-CHO-S on cholesterol transport was minor. Based on the current information, it is difficult to speculate the reasons, and future studies are needed regarding this topic.

Conclusions
In conclusion, in both turbot and tiger puffer, two important mariculture species, the supplementation of dietary cholesterol inhibits the steroid biosynthesis but seldom affects the cholesterol transport.

Data Availability
Raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.