1. Introduction
The use of fishmeal in aquaculture animal diets has been a hot topic in aquatic animal nutrition and feed research. Studies have shown that fishmeal, a source of marine animal protein, is irreplaceable or exceptional in aquatic animal diets [
1]. However, declining fishmeal supplies and increasing prices have prompted research into alternatives to fishmeal protein in aquafeeds [
2]. Plant proteins are more widely accessible, less expensive, and simpler to produce, making them a good option to replace fishmeal. Currently, many plant proteins are used in aquaculture diets, such as soybean protein concentrate [
3], cottonseed meal [
4], peanut meal [
5], fermented soybean meal [
6], and rapeseed meal [
7]. Studies have also explored the effect of plant proteins on the development and physiological function of aquatic animals [
8], and the results show that optimum levels of plant protein improve growth performance and immunity, while excessive consumption may cause a loss of growth performance and immunity due to anti-nutritional components including phytic acid, protease inhibitors, free cotton phenols, and lectins [
9,
10,
11].
Cottonseed protein concentrate (CPC), a high-protein product derived from cottonseed flakes, has received attention due to its large-scale availability [
12]. China and India are the two major cotton-producing countries, producing twice as much as the United States [
13]. Although CPC is less cost-effective per unit of protein compared to fishmeal and soybean meal [
14], it has the potential to reduce aquaculture production costs significantly if used correctly. To produce CPC, cottonseed flakes are subjected to aqueous alcohol extraction to diminish the content of soluble carbohydrates and anti-nutritional factors [
12]. Compared to other alternative protein sources such as canola, rapeseed, sunflower, linseed, lupine, soybean, and corn gluten, less research has been conducted on the substitution of CPC for fishmeal in aquaculture [
15]. One of the main limiting factors in using CPC is the presence of gossypol (C
30H
30O
8), a yellow polyphenolic molecule composed of reactive aldehydes and hydroxyl groups, which is the principal anti-nutritional component in CPC that can harm fish [
16]. Therefore, to ensure normal fish development, the ratio of CPC to fishmeal in the diet should not be too high.
Currently, research has focused on identifying optimal levels of substitution of fishmeal with other ingredients for different species of fish, including catfish (
Pseudobagrus ussuriensis) [
16], black sea bass (
Centropristis striata) [
14], black sea bream (
Acanthopagrus schlegelii) [
17], Nile tilapia (
Oreochromis niloticus) [
18], and turbot (
Scophthalmus maximus) [
19]. Previous studies have discovered that CPC can substitute 20% to 50% of fishmeal protein without negatively impacting aquatic animal development. However, substituting 36% of the fishmeal in the diet of golden pompano (
Trachinotus ovatus) with CPC resulted in considerable losses in whole-body crude protein [
12], while replacing 50% of the fishmeal resulted in a significant weight reduction in red drum (
Sciaenops ocellatus) [
20]. The mechanism behind such adverse consequences is not completely understood [
21,
22]. It was revealed that substituting 45% of fishmeal in diets with CPC increased gossypol remnants [
17,
22] and caused intestinal inflammation [
23], but their link to the negative effects of CPC replacement is unknown.
High-throughput mRNA sequencing (RNA-Seq) is a commonly used technique to uncover gene expression differences in non-model animals without reference genome data [
24]. Moreover, several studies have been carried out in order to investigate the differential gene expression in metabolism pathways through transcriptomic analysis in various fish species, for instance, rainbow trout (
Oncorhynchus mykiss) [
25], Atlantic salmon (
Salmo salar) [
26,
27], and grass carp (
Ctenopharyngodon idella) [
28]. The liver is very important in metabolism because it performs a variety of functions, including gluconeogenesis, storage, and regulation of nutrients, and detoxification of harmful substances. Additionally, it is involved in the regulation of bile secretion. However, the effect of a fishmeal-free diet on the metabolism process in fish livers remains unclear. Thus, the liver transcriptomic analysis may help in better understanding the metabolic processes induced by dietary feed stuff augmentation in aquaculture.
Amur sturgeon (
Acipenser schrenckii), a native freshwater fish indigenous to the Amur River basin of China and Russia [
29], has remarkable evolutionary, economic, and conservation value and is one of China’s most widely farmed sturgeons [
30]. Despite its popularity, there is limited research on its dietary requirements and the utilization of plant protein in its diet. This study aims to use RNA-Seq to uncover the molecular mechanism behind the transcriptome profile of the liver of
A. schrenckii when fed a CPC-based fishmeal-free diet. Additionally, its effects on growth, body composition, and physiological and metabolic responses will be analyzed to evaluate the efficiency of nutrient delivery and utilization of CPC in the
A. schrenckii diet. This study will give useful information about the physiological and metabolic effects of feeding sturgeons a CPC-based fishmeal-free diet and contribute to the development of improved nutrient supply and CPC utilization techniques in aquatic animal diets.
2. Materials and Methods
2.1. Experimental Diets
The control group used 50% fishmeal as the main protein component, while the treatment group used 50% cottonseed protein concentrate (CPC). Balancing the nutritional content of the diet by substituting fish oil and crystalline amino acids for gross energy and essential amino acids (EAAs). To adjust the overall phosphorus level of the test diet, calcium dihydrogen phosphate, Ca(H
2PO
4)
2, was added. The trace amount of 0.1% yttrium oxide (Y
2O
3) was supplemented into diets as an inert marker to measure the digestibility.
Table 1 displays the diet formulation and composition.
In the preparation of the feed, the ingredients were ground to a particle size of 250 μm using a laboratory grinder. Afterward, these ground ingredients were blended with additional components, including micronutrients such as vitamins and minerals. The dry mix was stirred for 15 min to ensure homogeneity. Next, the lipid source and water were added and blended for an additional 10 min using a mixer. The final feed mixture was pelletized into 1.5-mm pellets using the HX-200G pelletizer (Minan Instruments, Jining, China). Diets were then oven dried at 55 °C to reach a moisture content of 8–12%, followed by storage in sealed plastic bags at −20 °C before feeding. The chemical composition of the feed was analyzed using standard techniques, including determinations of moisture, ash, gross energy, crude protein, and crude lipid content [
31]. The amino acids in the diets were determined using the techniques described by Fountoulakis and Lahm (1998) [
32] and Yust et al. (2004) [
33]. High-performance liquid chromatography (HPLC) was employed to measure dietary gossypol concentrations [
34].
2.2. Farming Management
The Heilongjiang River Fisheries Research Institute Committee for the Welfare and Ethics of Laboratory Animals approved all techniques used in animal experiments (Ethics approval number: 20200615). A total of 180 healthy sturgeons (initial weight: 21.32 ± 0.18 g) were purchased from the Engineering and Technology Center of Sturgeon Breeding and Cultivation (Beijing, China). Sturgeons were randomly allocated to tanks in three replicates, with a density of 30 fish per tank (280 L). After a 14-day acclimatization period in the laboratory setting, sturgeons were given the control diet before the start of the experiment. For 56 days, the sturgeons were raised in an aquatic recirculation system with a constant water flow of around 2.0 L/s. Twice-daily manual feeding was conducted at 9:00 a.m. and 4:00 p.m. until the fish showed signs of satiation. One hour after feeding, any remaining feed was drained from the tank and then oven-dried at 65 °C for 24 h to determine the quantity of feed consumed and its efficiency. According to Stone et al. (2008), feces samples were obtained by hand stripping, which began in line at the front of the pelvic fins and ended at the anus [
35]. The feces samples were collected over two weeks to allow for sufficient material for chemical analysis. The trial was conducted under a 12 h dark/12 h light photoperiod. Throughout the trial, daily monitoring of water quality was performed utilizing a YSI 6600 V2-2 instrument (YSI Co., Yellow Springs, OH, USA), and the results indicated dissolved oxygen levels of 7.1–8.3 mg/L, ammoniacal nitrogen of <0.2 mg/L, pH of 7.3–7.8, nitrate of <0.4 mg/L, and nitrite of <0.2 mg/L.
After the trial, sturgeons were starved for 24 h. The weight and total length of each fish were recorded and used to calculate its condition factor (CF). The total weight of sturgeons per tank was aggregated, and the weight growth rate (WGR) and feed conversion rate (FCR) were calculated accordingly.
2.3. Sample Collection
Randomly selected sturgeons from each tank were sedated using tricaine methanesulfonate (MS-222, 60 mg/L). Subsequently, blood, whole body, liver, and mid-intestine samples were collected. Four fish were selected from each tank for body composition analysis, amino acid profile analysis, and blood chemical analysis. Blood was obtained from their caudal veins using a syringe containing sodium ethylenediaminetetraacetic acid (EDTA) as an anticoagulant. The collected blood was immediately subjected to centrifugation for 10 min at 7000× g to separate the supernatant. Afterward, the serum samples were stored at −20 °C until analyzed. Mid-intestinal samples were obtained from four individuals per tank and preserved at −80 °C for subsequent enzymatic analysis of digestive function. Additionally, livers were selected from six individuals per group for histological examination. Lastly, three individuals in each tank were randomly sampled for transcriptomic analysis of the livers. A rapid extraction procedure was employed to collect liver tissue, which was immediately placed in liquid nitrogen.
2.4. Analysis of Body Composition, Apparent Digestibility and Serum Indices
The proximate compositions of diets, feces, and the whole body were analyzed following the protocols provided by Al-Mentafji (2016), which included moisture, crude protein, crude lipid, ash, and gross energy [
31]. Moisture content was determined by drying in an oven at 105 °C for 3 h, and crude protein content was analyzed using a Kjeltec machine (2300, Foss Ltd., Höganäs, Sweden). The ether extraction procedure was performed on the Soxtec System (ST-6A, Naai Ltd., Shanghai, China) to determine crude lipid. Ash was determined by muffle furnace incineration at 600 °C for 2 h. The gross energy was measured using a bomb calorimeter (LRY-600A, Chuangxin Ltd., Hebi, China). The yttrium content was quantified using mass spectrometric analysis on the LCMS-8030 system (Shimadzu Ltd., Kyoto, Japan).
The amino acid composition was determined following acid hydrolysis using the Fountoulakis and Lahm technique (1998) [
32] on an automated amino acid analyzer (L-8900, Hitachi Ltd., Tokyo, Japan). Tryptophan was hydrolyzed in sodium hydroxide, neutralized, and measured by high-performance liquid chromatography (Dionex, Thermo Fisher Scientific, Waltham, MA, USA) [
33].
Serum samples were analyzed using the HITACHI 7170A biochemistry analyzer (Tokyo, Japan) for quantification of triglycerides (TG), total protein (TP), aspartate transaminase (AST), glucose (GLU), alanine transaminase (ALT), and cholesterol (CHOL). Serum ammonia levels were quantified using a commercially available assay kit (no. A086-1-1, Nanjing Jiancheng Biology Engineering Institute, China) following the protocol provided by the manufacturer.
2.5. Digestive Physiology
The mid-intestine was homogenized with a solution of 0.86% physiological saline in a volume ratio of 9:1, and the homogenate was centrifuged at 1000×
g for 10 min at 4 °C. The supernatant was further centrifuged and stored at −80 °C for future analysis. Amylase (AMS) and lipase (LPS) activities were determined using commercially available kits from Nanjing Jiancheng Bioengineering Institute (Nanjing, China), following the manufacturer’s instructions. AMS activity was assessed by measuring the amount of hydrolyzed starch (no. C016-1-1). LPS activity was measured at 580 nm using kit no. A054-2-1. Protease (PRT) activity was determined using the Folin-phenol Ciocalteu’s reagent technique [
36]. The mid-intestinal protein content was determined through the application of the Coomassie brilliant blue protein assay method. A microplate reader (Synergy 2, BoiTek, Minneapolis, MN, USA) was used for all absorbance measurements.
Liver segments were processed, embedded in standard paraffin, and sliced into 6-µm-thick sections using a microtome (HistoCore MULTICUT, Leica, Wetzlar, Germany). The sections were subjected to deparaffinization, hydration, staining with hematoxylin and eosin, and then mounting with neutral resin. A microscope (IX51, Olympus, Tokyo, Japan) was used to examine a total of 18 liver sections from each group.
2.6. RNA Isolation and Sequencing
Livers from three sturgeons per tank were homogenized in liquid nitrogen and mixed as one transcriptome sample for analysis. Total RNA was extracted according to the protocol of the Trizol reagent kit (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). The qualified RNA sample was treated with DNase I, and magnetic beads with Oligo (dT) attached were used to enrich mRNA. The first strand of cDNA was then synthesized with random primers. Subsequently, RNaseH, deoxyribonucleoside triphosphates (dNTPs), DNA polymerase I, and buffer were added to synthesize the second strand of cDNA. After purification and screening, PCR amplification was conducted to construct a cDNA library. The sequencing of a qualified cDNA library was conducted on an Illumina HiSeq 2500 (Illumina Inc., San Diego, CA, USA) for data analysis.
2.7. Gene Function Annotation
After removing the adapter sequences and low-quality reads, high-quality clean data were acquired. The Trinity software was used to assemble and splice the sequences to obtain the transcript sequences [
37]. To compare the assembled unigenes sequences with the Gene Ontology (GO) [
38], Cluster of Orthologous Groups of Proteins (COG) [
39], and Kyoto Encyclopedia of Genes and Genomes (KEGG) [
40] databases by the BLAST software. The predicted amino acid sequences of unigenes were evaluated using the HMMER 3.1 software, and the results were compared with the entries of the Protein Family (Pfam) database to obtain annotation information on the corresponding gene functions [
41].
2.8. Differentially Expressed Genes (DEGs) Analysis
After sequence alignment, the annotation file of the reference genome was used to calculate the fragments per kilobase of transcript per million mapped read values (FPKM) value of each transcript in the samples, which was used as the expression level of the transcript [
42]. The higher the transcript abundance, the higher the gene expression level. By performing sample-to-sample differential significance analysis of the expression level of each transcript, the DEGs were obtained. The significance was judged by the false discovery rate (FDR) ≤ 0.01 and Fold Change ≥ 2.
2.9. Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR)
To validate the transcriptome library screening, qRT-PCR was used to assess the expression of eight genes associated with metabolism. Primers (
Table 2) were designed using the Premier 5.0 software according to the transcriptome sequences and synthesized by Shanghai Genechem Co., Ltd (Shanghai, China). The qRT-PCR assays were performed using SYBR Green PCR Master Mix (Takara, Dalian, China) on a Bio-Rad CFX Connect Real-Time System (Bio-Rad Inc., Hercules, CA, USA). The melting curve analysis was performed with conditions ranging from 65 °C to 95 °C for 30 s, and the fluorescence was recorded. All qRT-PCR was repeated three times. The relative expression of the target was calculated using the 2
−ΔΔCt method, with β-actin serving as an endogenous reference [
43].
2.10. Calculations and Statistical Analysis
WGR, FCR, and CF were calculated via the following formulas:
The apparent digestibility (ADC) of the dry matter, gross energy, crude protein, and crude lipid in the diet was calculated using the following equation: ADC of nutrient or energy (%) = 100 × (1 − (dietary Y2O3/fecal Y2O3) × (fecal nutrient or energy/dietary nutrient or energy).
The Shapiro-Wilk and Levene’s tests were used to evaluate the normality and homogeneity assumptions of these data, respectively. A paired Student’s t-test and a Wilcoxon signed rank test were applied to compare these data with normal and non-normal distributions, respectively. Statistical significance was regarded at a p-value < 0.05, and statistical analysis was conducted using SPSS 23.0 (SPSS Inc., Chicago, IL, USA). Column graphs were generated by GraphPad Prism 9.0 (GraphPad Software, San Diego, CA, USA).
3. Results
3.1. Growth
No mortality was observed throughout the 56-day trial period. The substitution of fishmeal with CPC significantly reduced WGR (
p < 0.05), but it was accompanied by a significant increase in FCR (
p < 0.05) (
Table 3).
3.2. Body Composition
Replacing fishmeal with CPC decreased the contents of crude lipids and protein in the whole body (
p < 0.05) while increasing ash content (
p < 0.05) (
Table 4). Dietary CPC caused a significant decrease in the levels of EAAs such as phenylalanine, lysine, valine, isoleucine, arginine, and total amino acid contents (
p < 0.05), but histidine showed an opposite trend (
p < 0.05) (
Table 5). Non-essential amino acids (NEAAs) such as glycine, alanine, tyrosine, and proline decreased significantly (
p < 0.05), while serine increased (
p < 0.05).
3.3. ADC of Nutrients
There were significant reductions in ADC of crude lipid, crude protein, dry matter, and gross energy compared to the control group (
p < 0.05) (
Table 6).
3.4. Digestive Physiology
Table 7 presents the results of the mid-intestinal enzyme activity of
A. schrenckii. PRT, LPS, and AMS activities were significantly lower in the CPC group (
p < 0.05).
Liver histology showed normal results in individuals fed the fishmeal diet, with abundant granular and eosinophilic cytoplasm, central nuclei, and large nucleoli. Little lipid deposits were present (
Figure 1A). Sturgeons fed CPC-containing diets, on the other hand, showed fatty infiltration of hepatocytes and nuclei peripheral displacement in hepatocytes (
Figure 1B).
3.5. Hematological Parameters
Table 8 showed that dietary CPC had a significant increment in the activity of AST and ALT, as well as the concentration of ammonia in serum (
p < 0.05). By contrast, significantly decreased levels of TG were observed in the CPC group (
p < 0.05). No significant differences in the concentrations of TP, GLU, and CHOL were detected between the two groups (
p > 0.05).
3.6. Transcriptomic Sequencing
To obtain quality metrics of the raw reads, we performed quality control checks on the sequencing data using FastQC (v.0.11.4) software. The quality of the sequencing data was assessed, with an average of 4.2 GB of clean data obtained per sample and a Q30 base percentage of over 85%. Raw data were deposited in the NCBI Sequence Read Archive (SRA) database under BioProject number PRJNA736603.
3.7. Gene Functional Annotation and Categorization
A volcano plot was constructed (
Figure 2A) as well as an M-versus-A (MA) plot (
Figure 2B) to illustrate the general trend of gene expression levels and comparative multiples among samples. A total of 2758 DEGs were identified, with 846 up-regulated and 1912 down-regulated in the treatment group.
GO functional analysis classified 902 DEGs into three categories (
Figure 3). The top three subcategories of the biological process were the “cellular process”, “metabolic process”, and “single-organism process”. For the cellular component category, the major subcategories were “cell portion”, “organelle”, and “membrane part”. At the same time, “binding activity”, “catalytic activity”, and “transporter” were the most ones in the molecular function category.
3.8. Analysis of DEGs
According to the COG function classification of the unigenes, the most enriched group was “general function prediction only” (19.32%), followed by “amino acid transport and metabolism” (9.66%), “carbohydrate transport and metabolism” (7.69%), “signal transduction mechanisms” (7.51%), “replication, recombination, and repair” (7.33%), “energy production and conversion” (7.33%), “lipid transport and metabolism” (7.51%), “transcription” (5.19%), and “inorganic ion transport and metabolism” (4.47%) (
Figure 4).
These TOP 50 pathways were observed, including metabolism (25 pathways), organismal systems (8 pathways), diseases (2 pathways), environmental information processing (9 pathways), and cellular processes (6 pathways) based on the KEGG analysis (
Figure 5).
Subsequently, pathway enrichment analysis was performed, and the top 20 significantly enriched pathways were identified, including steroid biosynthesis, fatty acid metabolism, pyruvate metabolism, propanoate metabolism, cytokine-cytokine receptor interaction, biosynthesis of amino acids, fatty acid biosynthesis, glycolysis/gluconeogenesis, biosynthesis of unsaturated fatty acids, PPAR signaling pathway, and other pathways (
Figure 6).
3.9. Gene Expression Profiling
The top 100 DEGs in the liver transcriptome of sturgeons fed CPC diets were identified through gene expression profiling analysis (
Table S1). The genes discovered were linked to a range of metabolic pathways, including secondary metabolites biosynthesis, transport, and catabolism, carbohydrate transport and metabolism, amino acid transport and metabolism, nucleotide transport, lipid transport and metabolism, inorganic ion transport and metabolism, coenzyme transport and metabolism. Furthermore, they were connected with various biological processes, including chromatin structure and dynamics, transcription, chromosome partitioning, cell division, cell cycle control, extracellular structures, defense mechanisms, vesicular transport, secretion, intracellular trafficking, signal transduction mechanisms, energy production and conversion, chaperones, protein turnover, and post-translational modification.
This study focused on the DEGs related to metabolic pathways. Genes related to amino acid transport and metabolism, such as elastase-1 (EL1) and excitatory amino acid transporter 1 (EAAT1), were generally upregulated by diets containing the CPC ingredient. Conversely, transcripts for argininosuccinate synthase (ASS1), cytosolic carboxypeptidase 2 (CCP2), y+ L amino acid transporter 2 (y+ LAT2), and mast cell protease 1A-like (MCP-1) showed downregulation. Similarly, fructose-1,6-bisphosphatase (FBPASE) was downregulated, while alpha-enolase (ENO1) was upregulated. Concerning lipid transport and metabolism, liver polyprenol reductase (SRD5A3) was downregulated, while most other genes in this category were upregulated, including fatty acid desaturase 1 (FADS1), fatty acid desaturase 2 (FADS2), fatty acid-binding protein 7 (FABP7), fatty acid-binding protein 2 (FABP2), fatty acid-binding protein 1 (FABP1), fatty acid-binding protein 10-A (FABP10A), elongation of very long chain fatty acids protein 6 (ELOVL6), 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR), acyl-CoA desaturase-like (ADS), elongation of very long chain fatty acids protein 5 (ELOVL5), diphosphomevalonate decarboxylase (MVD), and fatty acid-binding protein 3 (FABP3). In secondary metabolite biosynthesis, transport and catabolism, and cytochrome P450 3A27 (CYP3A27) were downregulated, while ATP-binding cassette sub-family B member 8 (ABCB8), cholesterol side-chain cleavage enzyme, mitochondrial (Precursor) (CYP11A1), and bile salt export pump-like (BSEP) were upregulated. In nucleotide transport and metabolism, inorganic ion transport metabolism, and coenzyme transport and metabolism, adenylosuccinate synthetase isozyme 1 B (ADSSL1) and farnesyl pyrophosphate synthase (FDPS) were upregulated, while solute carrier family 26 member 9 (SLC26A9) was downregulated when sturgeons were fed the CPC diet.
3.10. Quantification of qRT-PCR
The current research verified the DEGs, which were identified with the aid of RNA-Seq, using qRT-PCR of eight crucial metabolic enzymes (
Figure 7). The results of the RNA-Seq and qRT-PCR assays for the eight DEGs revealed an up-regulation pattern of alanine aminotransferase 2 (ALT2), phosphoenolpyruvate carboxykinase mitochondrial isoform (mPEPCK), glucose 6-phosphatase (G6Pase), fructose 1,6-bisphosphatase (FBPase), fatty acid synthase (FAS), and carnitine palmitoyltransferase 2 (CPT2), and a down-regulation pattern of hexokinase (HK2) and pyruvate kinase (PK).