Effects of dietary protein level on reproductive performance, transcriptomics and metabolomics of ovaries of female northern snakehead (Channa argus) broodstock

Abstract

The study investigated the effects of dietary protein levels on reproductive performance, egg quality, and larval health of the northern snakehead (Channa argus) broodstocks. A total of 450 female northern snakeheads (16 months old), averaging 731.49 ± 3.51 g, were randomly distributed into 15 cement pools, with each pool containing 30 fish. Five isolipidic diets comprising 34%, 38%, 42%, 46%, and 50% crude protein (labeled as CAP34, CAP38, CAP42, CAP46, and CAP50) were formulated and fed to experimental fish for 4 months. The results showed that the fertilization rates of northern snakehead broodstock in the CAP42 and CAP46 groups were significantly higher than those in other groups (P = 0.049). The optimal dietary protein levels for northern snakehead broodstock were 41% and 42.2% based on fertilization rate and malformation rate. The plasma vitellogenin (VTG) and estradiol (E2) contents were reduced significantly with the increasing dietary protein levels from 38% to 46% (P < 0.05). Ovarian amino acid (AA) content increased significantly with the increasing dietary protein levels from 34% to 42%, followed by a plateau, while highly unsaturated fatty acid content was significantly higher in the CAP42 group than in the other groups (P < 0.05). Metabolomic and transcriptomic results indirectly indicated that 42% of dietary protein benefited steroid hormone biosynthesis, while 50% of dietary protein facilitated AA metabolism, and 34% of dietary protein enhanced arachidonic acid metabolism. In conclusion, 42% dietary protein is the optimal level for improving gonad development and egg quality in female C. argus by promoting nutrient accumulation and steroid hormone synthesis.

1. Introduction

In modern aquaculture, the development of high-quality broodstock is a crucial step in ensuring the sustainable and stable production of seedlings. High-quality fish eggs are sourced from well-managed and high-quality broodstock. Egg quality is directly influenced by various factors, including the endocrinological state of females during oogenesis, the quantity and quality of consumed food, physicochemical parameters of water, as well as stress experienced by broodstock (Sheikhzadeh et al., 2012). Among these factors, broodstock nutrition plays an important role as it provides all the essential nutrients required prior to the initiation of exogenous feeding, which are maternally supplied and incorporated into the oocyte during vitellogenesis (Dupont et al., 2014). During gonadal maturation, fish require proteins, lipids, and glycogen synthesized from both exogenous and endogenous sources to be absorbed and stored in oocytes to provide the necessary energy for embryonic development (Singh et al., 2021). However, the regulation of broodfish nutrition remains one of the least investigated areas in aquaculture research, largely owing to the complexity of the underlying biological mechanisms (Xiao et al., 2024).

Protein serves as a critical source of amino acids (AAs) and function as a reservoir of essential substrates for biosynthetic processes during early embryogenesis (Lanes et al., 2012). In fish, proteins serve as the main nitrogen source for protein synthesis during sexual maturation, and dietary protein levels significantly influence the spawning performance and offspring quality of broodstock. Studies on yellow catfish (Pelteobagrus fulvidraco) (Chen et al., 2022; Xiong et al., 2021), greater amberjack (Seriola dumerili) (Sarih et al., 2019), bagrid catfish (Mystus nemurus Cuv. & Val.) (Abidin et al., 2006), rohu (Labeo rohita) (Afzal Khan et al., 2005), and Nile tilapia (Oreochromis niloticus L.) (El-Sayed et al., 2003), have confirmed that optimal dietary protein levels can enhance egg quality and increase the number of hatching larvae, whereas inadequate protein intake may impair female fertility. The protein requirements of fish have been the subject of extensive research in recent decades, yielding significant advancements (NRC, 2011). However, the precise nutritional requirements of broodstock have been relatively neglected. In practice, the use of high-protein or nutrient-excess diets is commonly employed to ensure broodstock development. Traditional nutrition research examining the effects of dietary protein on broodstock has focused only on fertility, gamete quality, and biochemical composition. Therefore, it is essential to investigate the precise protein requirements of broodstock feed and further elucidate the molecular mechanisms through which protein regulates gamete quality and fertility.

Northern snakeheads (Channa argus) are highly valued freshwater aquatic product in China due to their delicious taste and high nutritional value, with an annual production exceeding 0.5 million tons over the past decade (FBMA, 2023). In addition to being an important commercial fish, the northern snakehead also serves as a vital parental source for the breeding of hybrid snakehead. In recent years, the decline in reproductive performance and egg quality among female northern snakeheads has been attributed to parental malnutrition. Previous studies have found that a dietary protein intake of 48% is required for optimal growth performance in juvenile northern snakehead (Sagada et al., 2017). However, the dietary nutrient requirements of broodstock will differ from those of other growth stages (Izquierdo et al., 2001). Consequently, this study seeks to examine the ideal dietary protein levels for northern snakehead broodstock and further explore the regulatory mechanisms by which dietary protein levels affect the reproductive performance and egg quality of the northern snakehead through transcriptomics and non-targeted metabolomics analysis.

2. Materials and methods

2.1. Animal ethics statement

All animal care and experimental protocols received approval from the Institutional Animal Care and Use Ethics Committee of the Chinese Academy of Fishery Sciences (approval No. 2024SJRC).

2.2. Experimental diets

Ingredients and chemical compositions of experimental diets are shown in Table 1. Fish meal, corn gluten meal, and soybean meal were used as protein sources. Fish oil and soybean oil (2:1) served as lipid sources, and cassava starch was used as a carbohydrate source. Five isoenergetic practical diets were designed, with protein levels increasing by 4% increments, ranging from 34% to 50%. All ingredients were weighed, pulverized, thoroughly blended for 20 min, and subsequently formed into pellets. The pellets were dried at 80 °C for 30 min to a moisture content of about 6% and stored at 4 °C until use. The experimental diets were analyzed for crude protein, crude lipid, moisture, and crude fiber content using the procedures outlined by the Association of Official Analytical Chemists (AOAC, 2006). Crude protein was determined by multiplying the nitrogen content by a factor of 6.25. Nitrogen levels were determined following the method 984.13 using an automatic Kjeldahl nitrogen analyzer (FOSS Tecator 8400, Pestop Biotechnology Group Co., Haganas, Sweden). The Soxhlet extraction technique, utilizing petroleum ether as the solvent (Soxtec System HT Tecator, Pestop Biotechnology Group Co., Haganas, Sweden), was applied to assess the crude lipid content (method 2000.03). Moisture content was measured by drying the samples to a constant weight at 105 °C, in accordance with the method 934.01. Crude fiber content (method 978.10) was analyzed using an automatic fiber analyzer (Ankom200, Ankom Technology Corp., Macedon, NY, USA). The determination of crude ash was performed using incineration at 550 °C for 8 h in a muffle furnace (Thermolyn, Thermo Fisher Scientific Inc., Waltham, MA, USA) according to method 942.05.

Table 1.

Ingredients and chemical compositions of the experimental diets (dry matter basis, %).

Items	Groups1
	CAP34	CAP38	CAP42	CAP46	CAP50
Ingredients
Fishmeal	26.00	32.00	38.00	44.00	55.00
Corn gluten meal	16.00	16.00	16.00	16.00	15.00
Soybean meal	15.00	15.00	15.00	15.00	11.00
Wheat flour	5.00	5.00	4.00	3.00	2.50
Cassava starch	20.00	16.00	12.00	9.00	5.00
Fish oil	3.50	3.20	3.00	2.75	2.35
Soybean oil	3.50	3.20	3.00	2.75	2.35
Vitamin premix2	0.50	0.50	0.50	0.50	0.50
Mineral premix3	4.00	4.00	4.00	4.00	4.00
Vitamin C	0.50	0.50	0.50	0.50	0.50
Squid paste	0.50	0.50	0.50	0.50	0.50
Ca(H2PO4)2	1.00	1.00	1.00	1.00	1.00
Choline chloride	0.09	0.09	0.09	0.09	0.09
Ethoxyquinoline	0.11	0.11	0.11	0.11	0.11
Cellulose	4.30	2.90	2.30	0.80	0.10
Total	100.00	100.00	100.00	100.00	100.00
Proximate compositions
Crude protein	34.44	38.77	42.85	46.57	50.76
Crude lipid	7.84	7.88	8.00	8.49	8.80
Crude fiber	6.32	4.87	4.12	3.22	2.01
Moisture	5.64	6.06	5.28	5.90	5.48
Organic matter	88.90	87.70	86.50	85.30	83.3
Gross energy, kJ/g	19.32	19.39	19.46	19.60	19.59

¹CAP34, CAP38, CAP42, CAP46, and CAP50 represent 34%, 38%, 42%, 46%, and 50% dietary protein levels, respectively.

²Vitamin premix (mg/kg diet): vitamin B₁ 20, vitamin B₂ 20, vitamin B₆ 20, vitamin B₁₂ 0.02, folic acid 5, calcium pantothenate 50, inositol 100, niacin 100, biotin 0.1, cellulose 5561.88, ascorbic acid 4000, vitamin A 11, vitamin D 2, vitamin E 100, vitamin K 10.

³Mineral premix (mg/kg diet): NaCl 500.0, MgSO₄·7H₂O 8155.6, NaH₂PO₄·2H₂O 12,500.0, KH₂PO₄ 16,000.0, Ca(H₂PO₄)·2H₂O 7650.6, FeSO₄·7H₂O 2286.2, C₆H₁₀CaO₆·5H₂O 1750.0, ZnSO₄·7H₂O 178.0, MnSO₄·H₂O 61.4, CuSO₄·5H₂O 15.5, CoSO₄·7H₂O 0.91, KI 1.5, Na₂SeO₃ 0.60, corn starch 899.7.

Organic matter = Dry matter - Crude ash.

The gross energy was analyzed using an isothermal automatic calorimeter (1281, Parr Instrument Co., Moline, IL, USA).

2.3. Experimental conditions

The broodstock (16 months old) were purchased from Foshan Bairong Aquaculture Co., Ltd. (Foshan, Guangdong, China), and all were bred in the same batch and raised in a pond. A total of 450 female snakeheads (average body weight 731.49 ± 3.51 g) were randomly allocated to 15 cement pools at a density of 30 fish per pool, with each diet randomly assigned to three replicates. Experimental diets were hand-fed twice a day to apparent satiation for 4 months (from December 2023 to April 2024). Fish were raised under a natural light cycle. Dissolved oxygen and water temperature were recorded daily, averaging 5.57 ± 0.33 mg/L and 27.1 ± 3.6 °C, respectively.

2.4. Spawning and egg collection

Eight male northern snakeheads with the same genetic background were used in the artificial insemination experiment. Three females were randomly selected from each tank to assess their reproductive performance. Two types of hormones, human chorionic gonadotropin (HCG) and luteinizing hormone releasing hormone A2 (LHRH-A2) (Ningbo Second Hormone Factory, Ningbo, Zhejiang, China), were used for artificial fertilization. The female fish were injected twice, with an interval of 10 h between the two injections. The initial injection dose for females was set at 6 μg/kg body weight of LHRH-A2. For the second injection, it was a combination of 900 IU of HCG and 12 μg/kg of LHRH-A2 body weight. These components were dissolved in 0.68% normal saline to achieve a total injection volume of 2 mL/kg body weight. Once ovulation occurred, 8 male fish were anesthetized with tricaine methanesulfonate (MS-222; 100 mg/L, Sigma–Aldrich, St. Louis, MI, USA). After that, they were dissected, and their testes were removed and then transferred to a sperm preservation solution for semen extraction. Gentle pressure was applied to the female’s abdomen to release the eggs. Then, 2 mL of semen was added, followed by a small amount of water, and the mixture was gently stirred for 15 s to complete the artificial insemination process.

The fertilized eggs from each female were placed and incubated in a square incubator (80 cm × 40 cm × 40 cm). Approximately 300 fertilized eggs were subsequently selected to calculate the fertilization rate. The fertilized eggs were then allowed to hatch into larvae, and the hatching rate was calculated based on the number of hatched larvae. A total of 150 larvae were randomly chosen and transferred to Petri dishes for a 12-d starvation experiment, and the number of dead larvae was recorded daily without any operational intervention. Two hundred larvae were randomly selected from each group and placed into glass tanks in triplicate. The larvae were fed enough fairy shrimp (Chirocephalus diaphanous) every day. The body length and weight of larvae were recorded at 7, 14, and 21 d after hatching, with 30 larvae assessed during each measurement, respectively.

2.5. Sample collection

At the end of the feeding trial, all females were subjected to a 24-h fasting period prior to being weighed and counted. Six females were randomly chosen from each tank and anesthetized with 100 mg/L MS-222. Out of these, three females from each tank were separately sampled for accurate measurement of body length and weight to determine the condition factor (CF). Blood was obtained from the caudal vein of three female fish in each tank using a heparinized syringe. The collected samples were centrifuged at 4 °C and 2500 × g for 5 min to separate plasma, which was stored at −80 °C until further analysis of biochemical indexes and hormone levels. The fish were subsequently dissected, and the viscera, liver and ovary were weighed to calculate the gonadosomatic index (GSI), hepatosomatic index (HSI), and viscerosomatic index (VSI). Ovarian samples from two female individuals in each tank were preserved in liquid nitrogen for subsequent omics studies.

2.6. Assessment of serum steroid hormones

The levels of 11-ketotestosterone (11-KT; Cat. H088), 17-β estradiol (E2; Cat. H102), and vitellogenin (VTG; Cat. H362) in the serum were quantified using enzyme-linked immunosorbent assay (ELISA) kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China), following the manufacturer’s protocols.

2.7. Ovarian immunohistochemical

The immunohistochemical analysis of the ovary was carried out to assess the expression levels of estrogen receptor (ER) and aromatase cytochrome P450 (CYP19A). The antibodies against ER and CYP19A were purchased from Proteintech Group Inc. (Wuhan, Hubei, China). Formalin-fixed, paraffin-embedded tissues (6 μm) were deparaffinized, rehydrated, and subjected to antigen retrieval in citrate buffer (pH 6.0) via microwave heating (95 °C, 15 min). Endogenous peroxidase was blocked with 3% H₂O₂ (10 min), followed by 5% bovine serum albumin (BSA; Shanghai Yuanye Bio-Technology Co., Ltd., Shanghai, China) blocking (30 min, room temperature). The tissue sections were incubated with primary antibodies (diluted at 1:200) overnight at 4 °C, followed by incubation with horseradish peroxidase (HRP; Shanghai Yuanye Bio-Technology Co., Ltd., Shanghai, China)-conjugated secondary antibodies (diluted at 1:500) for 1 h at room temperature. 3′, 3′-diaminobenzidine (DAB) staining (2–5 min) and hematoxylin counterstaining were performed. Slides were dehydrated, cleared in xylene, and mounted. Negative controls used the antibody diluent instead of the primary antibodies.

2.8. Analysis of AAs and fatty acids in the ovary

Approximately 0.5 g of freeze-dried ovarian tissue was homogenized with 5 mL of a chloroform/methanol (2:1, vol/vol) mixture in a 5-mL centrifuge tube and extracted for 12 h. For the methyl esterification process, 1 mL of potassium hydroxide-methanol solution (0.4 mol/L) was added to a centrifuge tube. The mixture was then mixed vigorously for 3 min and allowed to stand for 30 min. Subsequently, ultra-pure water was added to the mixture, and the upper layer of the solution was extracted for further analysis. The resulting upper layer solution (200 μL) was mixed with hexane (800 μL), filtered using a 0.22-μm membrane, and analyzed for fatty acid content using a gas chromatograph (GC-2010, Serial No. A09056, Shimadzu Corp., Kyoto, Japan).

Ovary samples (100 mg) were homogenized and hydrolyzed in vacuo with 6 mol/L HCl at 110 °C for 22 to 24 h. The hydrolysate was evaporated, reconstituted in 0.02 mol/L HCl citrate buffer, and filtered (0.22-μm). Quantification employed ion-exchange chromatography coupled with post-column ninhydrin derivatization using an amino acid analyzer (L-8900, Hitachi Ltd., Tokyo, Japan). Data are expressed as g/kg dry weight.

2.9. Transcriptomic analysis

Three ovaries were randomly selected from each of the CAP34, CAP42, and CAP50 groups for transcriptomic analysis. The total RNA of the ovary was extracted using TRIzol Reagent (Life Technologies Inc., Carlsbad, CA, USA) according to the instruction manual. The high-quality RNA was used to construct the RNA-Seq libraries using the NEBNext Ultra RNA Library Prep kit (NEB 7530, New England Biolabs Inc., Ipswich, MA, USA), which were sequenced using Illumina Novaseq 6000 by Gene Denovo Biotechnology Co. (Guangzhou, Guangdong, China). The clean reads were generated by eliminating reads containing adapter sequences, ploy-N sequences, and low-quality reads from the raw data. The resulting clean reads were subsequently mapped onto a C. argus genome using the aligner HISAT2. The data analysis was performed based on previous method (Fei et al., 2024). Differential expression analysis of RNAs was conducted using DESeq2 software between two distinct groups (Love et al., 2014). The transcripts with a false discovery rate (FDR) < 0.05 and an absolute fold change (FC) ≥ 2 were identified as differentially expressed transcripts. The cluster Profiler R package was used to perform Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of differentially expressed genes (DEGs) (Xie et al., 2011). The DEGs identified in the ovary of northern snakehead are shown Table S1.

2.10. Metabolomic analysis

Six ovarian samples from each of the CAP34, CAP42, and CAP50 groups were immediately frozen in liquid nitrogen and subsequently suspended in 1 mL of extraction solvent (a mixture of acetonitrile-methanol–water containing an internal standard). The mixture was subsequently homogenized and sonicated in an ice-water bath. The resulting suspension was incubated at −20 °C and then centrifuged at 7500 × g for 10 min at 4 °C. The supernatants were collected and subjected to liquid chromatography-tandem mass mpectrometry‌/mass mpectrometry (LC-MS/MS) analysis using a UHPLC system (1290, Agilent Technologies Inc., Santa Clara, CA, USA). Raw MS data were converted to mzXML format using ProteoWizard MSConvert and processed in XCMS for peak alignment. Metabolite identification was achieved by matching precise m/z values (<10 parts per millionerror) and MS/MS fragmentation patterns against an in-house database of authenticated standards. Preprocessing included missing value imputation via the k-nearest neighbors (KNN) algorithm, removal of extreme outliers, and total peak area normalization. Differentially expressed metabolites (DEMs) were screened through orthogonal partial least squares-discriminant analysis (OPLS-DA) using the ropls R package (variable important in projection [VIP] ≥1), complemented by Student’s t-test (P < 0.05) and fold-change thresholds (>1.5). Subsequent functional annotation incorporated pathway mapping (KEGG/HMDB/LipidMaps), hierarchical clustering (Cluster 3.0), and metabolic pathway enrichment analysis via MetaboAnalyst 4.0.

2.11. Quantitative real-time PCR (qRT-PCR) analysis

The StepOnePlu Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) was utilized for qRT-PCR with SYBR Green Master Mix (QPK-201, Toyobo Co., Ltd., Osaka, Japan), following the manufacturer’s protocols. The thermal cycling protocol was as follows: initial denaturation at 95 °C for 2 min; followed by 40 cycles of denaturation at 95 °C for 5 s and combined annealing/extension at 60 °C for 30 s (single-step). A melt curve analysis was subsequently conducted to verify amplification specificity by heating from 65 °C to 95 °C with a continuous fluorescence measurement (increment of 0.5 °C/s). Ten DEGs identified from ovarian transcriptomes were selected for validation. Gene-specific primers were designed using genomic and transcriptomic references (Table S2). Reactions were performed in triplicate technical replicates, and relative expression levels were quantified via the 2^−ΔΔCt method.

2.12. Statistical analysis

The data are expressed as mean and standard error of the mean (SEM). The Shapiro–Wilk test and Levene’s test were employed to evaluate the normality and homogeneity of variance in the data. All data were analyzed using one-way analysis of variance (ANOVA) with the model:

Y_ij = μ + T_i + ϵ_ij,

where Y_ij denotes the observed value of the dependent variable; μ represents the overall mean; T_i is the fixed effect of treatment; and ϵ_ij is the residual error. Multiple comparisons between treatment groups were performed using Duncan’s test. Statistical significance was defined as P < 0.05. Statistical analysis was performed using IBM SPSS Statistics 23.0 software (SPSS, International Business Machines Corpo., Chicago, IL, USA).

3. Results

3.1. Morphological indexes and reproductive performance

In this study, all females had attained sexual maturity. At the end of the feeding trial, there was no statistically significant difference in the mean body weight of the experimental fish among the groups (P > 0.05). As shown in Table 2, broodstock fed a 42% protein diet exhibited significantly higher VSI but lower HSI relative to other dietary groups (P < 0.05). However, no significant differences were found in the CF and GSI of the broodstock among the groups (P > 0.05). Fertilization rate peaked at 42% dietary protein before declining, whereas the malformation rate demonstrated an inverse trend, with the CAP42 and CAP46 groups showing significantly reduced malformations (P < 0.05). Absolute and relative fecundity remained unaffected by dietary protein levels (P > 0.05). Egg diameter increased with higher dietary protein levels, and the CAP50 group exhibited a significantly larger egg diameter compared to the CAP34 group (P < 0.05). The body length and body weight of the 7 d post-hatching (DPH) larvae initially increased and then decreased with increasing dietary protein levels, while those of the 14DPH larvae showed a positive correlation with dietary protein levels, and the same trend was observed in the 21DPH larvae.

Table 2.

Spawning performance and post-hatching larvae size of northern snakehead broodstock fed diets with varying protein levels.

Items	Groups1					SEM	P-value
	CAP34	CAP38	CAP42	CAP46	CAP50		ANOVA	Linear	Quadratic
IBW, g	727.41	732.22	734.44	727.48	735.92	3.511	0.875	0.343	0.783
FBW, g	717.21	743.17	739.65	742.3	768.48	16.712	0.786	0.232	0.873
FI2, g/(fish·d)	0.87	0.88	0.91	0.88	0.78	0.050	0.679	0.673	0.153
CF3, g/cm3	0.92	0.93	0.91	0.98	0.94	0.014	0.355	0.288	0.574
GSI4, %	7.37	8.72	7.52	8.02	7.44	0.257	0.401	0.756	0.567
HSI5, %	1.16ab	1.13ab	0.92b	1.15ab	1.38a	0.068	0.037	0.256	0.055
VSI6, %	13.64b	14.72ab	16.64a	15.45a	14.34ab	0.315	0.012	0.332	0.021
Absolute fecundity7, %	32,057.02	32,162.43	29,760.24	31,395.86	30,355.45	1314.424	0.415	0.662	0.900
Relative fecundity8, %	46.35	45.64	43.07	46.13	40.56	1.374	0.335	0.258	0.501
Fertilization rate	76.98bc	80.64ab	86.57a	81.53ab	68.67c	2.335	0.049	0.263	0.013
Hatching rate9, %	57.65	58.56	63.55	60.25	68.55	1.854	0.150	0.263	0.012
Malformation rate, %	10.57a	8.66b	6.54c	6.02c	11.26a	0.362	0.045	0.555	0.025
Egg diameter, mm	1.42b	1.48ab	1.47ab	1.48ab	1.52a	0.013	<0.001	0.001	0.165
Larvae body length 7DPH, mm	6.99ab	6.93ab	7.05a	6.81b	6.85b	0.054	0.038	0.142	0.024
Larvae body weight 7DPH, mg	2.29ab	2.48a	2.37a	1.96b	2.16b	0.052	<0.001	0.005	0.001
Larvae body length 14DPH, mm	7.79c	8.77bc	9.05ab	9.59a	9.64a	0.134	<0.001	0.001	0.001
Larvae body weight 14DPH, mg	5.13c	6.96b	7.93b	10.06a	10.19a	0.284	<0.001	0.001	0.001
Larvae body length 21DPH, mm	12.01b	11.75b	12.91a	12.86a	12.23ab	0.163	0.002	0.012	0.001
Larvae body weight 21DPH, mg	18.57b	18.82b	21.83a	22.59a	22.21a	0.880	0.042	0.005	0.001

IBW = initial body weight; FBW = final body weight; FI = feed intake; CF = condition factor; GSI = gonadosomatic index; HSI = hepatosomatic index; VSI = viscerosomatic index; DPH = day post-hatching; SEM = standard error of the mean.

Within a row, means without a common superscript letter differ at P < 0.05, n = 6.

¹CAP34, CAP38, CAP42, CAP46, and CAP50 represent 34%, 38%, 42%, 46%, and 50% dietary protein levels, respectively.

²FI = 100 × Total feed intake/Days/[(IBW + FBW)/2].

³CF = 100 × (Body weight/Body length³).

⁴GSI = 100 × Gonad weight/FBW.

⁵VSI = 100 × Viscera weight/FBW.

⁶HSI = 100 × Liver weight/FBW.

⁷Absolute fecundity = Total eggs produced per female.

⁸Relative fecundity = Total egg production per female/mean weight of female.

⁹Hatching rate = 100 × Total number of larvae/Total number of fertilized eggs.

Regression analysis between fertilization rate and malformation rate with dietary protein levels showed that the optimal dietary protein levels were 41.0% and 42.2%, respectively (Fig. 1). As shown in Fig. 2, the results of starvation tolerance experiments of larvae post-hatching showed that dietary protein levels ranging from 42% to 50% significantly increased larval survival rate compared to the CAP34 and CAP38 groups, with the survival rates ranging between 25% and 36% at 11 d post-hatching (P < 0.05).

Fig. 1.

Quadratic regression of fertilization rate and malformation rate of broodstock northern snakehead.

Fig. 2.

Survival rate of post-hatching larvae of northern snakehead with different dietary protein levels. CAP34, CAP38, CAP42, CAP46, and CAP50 represent 34%, 38%, 42%, 46%, and 50% dietary protein levels, respectively. ∗ Indicates a statistically significant difference between groups at P < 0.05, n = 3.

3.2. Amino acid and fatty acid composition of the ovary

The composition of AAs and fatty acids in the ovary of northern snakehead are shown in Table 3, Table 4, respectively. The crude protein and total AA were significantly influenced by the dietary protein levels and were lower in the CAP34 and CAP38 groups than in the other groups (P < 0.05). A total of 16 AAs were identified in the eggs, and all of them were significantly affected by dietary protein levels. The contents of all AAs in the ovary showed an increasing trend with the increase in dietary protein level. Both linear regression and quadratic regression analyses revealed significant positive correlations between the content of various AAs in the ovary and the dietary protein level. Some of them, such as Lys, Val, Arg, Asp, Ser, Glu, and Ala, reached a plateau when the dietary protein level exceeded 42% (Table 3).

Table 3.

Effects of dietary protein levels on the ovarian amino acid (AA) profile (g/kg, wet weight) of northern snakehead broodstock.

Items	Groups1					SEM	P-value
	CAP34	CAP38	CAP42	CAP46	CAP50		ANOVA	Linear	Quadratic
Thr	0.78bc	0.77c	0.85ab	0.84a	0.90a	0.015	0.007	0.001	0.003
Phe	0.79c	0.80bc	0.87ab	0.86ab	0.93a	0.016	0.006	<0.001	<0.001
Lys	1.29b	1.30b	1.45a	1.44a	1.51a	0.028	0.005	<0.001	<0.001
His	0.39b	0.39b	0.43a	0.42ab	0.45a	0.009	0.014	0.002	0.010
Val	0.95b	0.96b	1.05a	1.04a	1.11a	0.018	0.007	<0.001	0.002
Met	0.48c	0.50bc	0.53ab	0.53ab	0.57a	0.009	0.006	<0.001	0.021
Ile	1.13c	1.16bc	1.26a	1.24ab	1.33a	0.023	0.004	<0.001	0.001
Leu	1.24c	1.25bc	1.37a	1.35ab	1.44a	0.024	0.007	<0.001	0.002
Arg	0.93b	0.92b	1.01a	1.01ab	1.07a	0.013	0.003	0.001	0.003
EAA	7.98b	8.04b	8.82a	8.73a	9.31a	0.161	0.005	<0.001	0.002
Tyr	0.66c	0.68bc	0.72b	0.73b	0.79a	0.022	0.010	<0.001	<0.001
Asp	1.26b	1.24b	1.38a	1.37a	1.45a	0.019	0.006	<0.001	0.002
Ser	1.04b	1.03b	1.18a	1.15a	1.23a	0.037	0.003	<0.001	0.003
Glu	1.97b	1.94b	2.20a	2.17a	2.28a	0.021	0.004	<0.001	0.003
Pro	0.77c	0.78bc	0.85ab	0.83ab	0.89a	0.012	0.016	0.001	0.005
Gly	0.56b	0.56b	0.60ab	0.60ab	0.63a	0.011	0.043	0.002	0.008
Ala	1.39b	1.40b	1.55a	1.53a	1.64a	0.028	0.003	<0.001	0.001
NEEA	7.65b	7.63b	8.46a	8.37a	8.91a	0.154	0.004	<0.001	0.002
Total AA	15.63b	15.68b	17.30a	17.10a	18.21a	0.301	0.005	<0.001	0.002
Crude protein	16.42b	16.45b	17.66a	17.58a	18.67a	0.264	0.007	<0.001	0.002

EAA = essential amino acids; NEAA = non-essential amino acids; SEM = standard error of the mean.

Within a row, means without a common superscript letter differ at P < 0.05, n = 6.

¹CAP34, CAP38, CAP42, CAP46, and CAP50 represent 34%, 38%, 42%, 46%, and 50% dietary protein levels, respectively.

Table 4.

Effects of dietary protein levels on the ovarian fatty acid profile (g/kg, wet weight) of northern snakehead broodstock.

Items	Groups1					SEM	P-value
	CAP34	CAP38	CAP42	CAP46	CAP50		ANOVA	Linear	Quadratic
C14:0	2.59a	2.45a	2.61a	2.68a	1.95b	0.179	0.001	0.427	0.537
C16:0	12.72a	13.62a	12.8a	13.93a	9.85b	0.778	0.024	0.343	0.308
C18:0	2.80	2.79	2.76	3.07	2.19	0.157	0.120	0.420	0.423
∑SFA	18.11a	18.85a	18.17a	19.6a	13.98b	1.107	0.021	0.363	0.353
C16:1n-9	14.88	15.54	15.48	15.23	15.62	0.164	0.157	0.328	0.553
C18:1n-9	68.26	79.19	70.97	74.87	66.22	1.610	0.056	0.481	0.092
C20:1n-9	4.12	4.34	4.36	4.13	4.04	0.082	0.201	0.546	0.404
∑MUFA	87.25	99.07	90.81	94.22	85.89	1.744	0.084	0.559	0.116
C18:2n-6	22.96	29.12	17.41	33.31	22.88	2.355	0.417	0.818	0.942
C20:4n6	1.97b	2.10ab	2.25a	1.95b	2.07b	0.031	0.001	0.871	0.296
∑n-6 PUFA	25.74	32.15	20.53	36.27	25.89	2.355	0.422	0.801	0.929
C18:3n-6	0.57	0.61	0.49	0.58	0.57	0.019	0.221	0.786	0.754
C18:3n-3	3.41	3.84	3.26	4.18	3.58	0.168	0.503	0.592	0.814
C20:5n-3	7.49b	7.08b	8.70a	8.05ab	8.67a	0.234	0.002	0.040	0.131
C22:6n-3	23.92b	24.51b	27.90a	23.47b	24.9b	0.634	0.021	0.844	0.537
∑n-3 PUFA	35.40b	36.03ab	40.35a	36.28b	37.72b	0.701	0.047	0.345	0.342
∑HUFA	33.38b	33.69b	38.84a	33.47b	35.64ab	0.806	0.007	0.468	0.480

SAF = saturated fatty acids; MUFA = monounsaturated fatty acids; PUFA = polyunsaturated fatty acids; HUFA = highly unsaturated fatty acids; SEM = standard error of the mean.

Within a row, means without a common superscript letter differ at P < 0.05, n = 6.

¹CAP34, CAP38, CAP42, CAP46, and CAP50 represent 34%, 38%, 42%, 46%, and 50% dietary protein levels, respectively.

As shown in Table 4, the content of saturated fatty acids (SFA) such as C14:0 and C16:0 in the ovary were significantly lower in the CAP50 group than in the other groups (P < 0.05). The content of highly unsaturated fatty acids (HUFAs), especially arachidonic acid (ARA; C20:4n6), eicosapentaenoic acid (EPA; C20:5n3), and docosahexaenoic acid (DHA; C22:6n3), initially increased and then decreased with increasing dietary protein level, with the maximum content observed at a dietary protein level of 42% (P < 0.05).

3.3. Plasma steroid hormones and vitellogenin

Dietary protein levels significantly modulated plasma E2 and VTG concentrations in northern snakehead broodstock (P < 0.05, Table 5). Although plasma 11-KT levels remained stable across the dietary protein gradients, both E2 and VTG concentrations showed progressive declines within the 38%-46% protein range.

Table 5.

Effects of dietary protein levels on plasma sex hormone level of northern snakehead broodstock.

Items	Groups1					SEM	P-value
	CAP34	CAP38	CAP42	CAP46	CAP50		ANOVA	Linear	Quadratic
11-KT, ng/L	117.99	117.55	169.31	139.96	133.90	8.224	0.917	0.946	0.681
E2, ng/L	98.03a	100.43a	82.86ab	67.49b	70.99b	4.342	0.033	0.003	0.013
VTG, μg/mL	14.97a	14.32a	13.36ab	11.88b	13.28ab	0.344	0.039	0.013	0.020

11-KT = 11-ketotestosterone; E2 = 17-β estradiol; VTG = vitellogenin; SEM = standard error of the mean.

Within a row, means without a common superscript letter differ at P < 0.05, n = 6.

¹CAP34, CAP38, CAP42, CAP46, and CAP50 represent 34%, 38%, 42%, 46%, and 50% dietary protein levels, respectively.

3.4. Immunohistochemical analysis

Estrogen receptor and CYP19A1 immunoreactivity were marked in follicle cells and oocyte cell membranes (Fig. 3). No positive signals were detected in the negative controls. As illustrated in Fig. 3A, ER and CYP19A1 antigens were localized to the oocyte cell membranes and follicle cells, with an intense positive reaction observed specifically in the CAP38 group. The optical density of CYP19A1 and ER in ovaries was significantly increased at a 38% dietary protein level (P < 0.05), but showed a decreasing trend when the level exceeded 38% (Fig. 3B).

Fig. 3.

Immunohistochemical analysis of CYP19A1 and ER in the ovaries of broodstock northern snakehead fed diets with different protein levels. Immunohistochemical sections of ovaries are shown at 34% (A), 38% (B), 42% (C), 46% (D), and 50% (E) dietary protein levels at 100 × magnification. The positive antigen was dyed brown with 3′, 3′-diaminobenzidine (DAB). (F) The optical density of CYP19A1 and ER in the ovary. Different lowercase letters above columns represent significant differences among treatments at P < 0.05 (n = 6). CYP19A1 = cytochrome P450 family 19 subfamily a member 1; ER = estrogen receptor.

3.5. Transcriptomic profiles

Based on the transcriptomic analysis of nine ovarian samples from CAP34, CAP42, and CAP50 groups, high-quality sequencing data were obtained following stringent quality control procedures. After filtering, the total number of clean reads was 44.59 million for CAP50, 39.47 million for CAP42, and 45.13 million for CAP34. The GC content ranged consistently between 45.36% and 46.78%, and Q30 values exceeded 94.96% across all samples. Alignment to the C. argus reference genome confirmed the high integrity of the data, with 92.62%-93.81% of reads mapping to annotated regions (Table S3). Differential expression analysis revealed that the comparison between CAP50 and CAP42 was the most divergent, with 764 upregulated and 105 downregulated DEGs. In contrast, the comparison between CAP50 and CAP34 resulted in the fewest number of DEGs (Fig. 4A). The Venn diagram visualization (Fig. 4B) illustrated both overlapping and unique DEGs among the CAP34, CAP42, and CAP50 groups. Notably, the 130 (124 ± 6) DEGs produced in CAP50 vs. CAP42 and CAP50 vs. CAP34 were identical, possibly suggesting that the ovary undergoes similar regulation during its development in broodstock fed low-protein or high-protein diets. This study investigated the differences between broodstock fed with high- and low-protein diets, so enrichment analysis was performed on the DEGs obtained from the comparisons of CAP50 vs. CAP42 and CAP50 vs. CAP34 to explore their functional categories. The functional enrichment analysis of the shared DEGs revealed significant associations with various biological processes (Fig. 4C), including cellular process, biological regulation, and single-organism process. In terms of molecular functions, the predominant categories were binding and catalytic activity. Regarding cellular components, key annotations included cell, cell part, and organelle. The GO and KEGG analyses of the comparisons of CAP42 vs. CAP34 are presented in Fig. S1.

Fig. 4.

Transcriptome analysis among the comparison groups. (A) Number of differentially expressed genes (DEGs) in the ovary. (B) Venn diagram of DEGs in the ovary. (C) Gene Ontology (GO) enrichment analysis of DEGs in the ovary. CAP34, CAP42, and CAP50 represent 34%, 42%, and 50% dietary protein levels, respectively.

Kyoto Encyclopedia of Genes and Genomes pathway analysis of ovarian DEGs revealed significant enrichment in 16 (CAP50 vs. CAP34) and 23 (CAP50 vs. CAP42) biological pathways (Fig. 5A and B). During ovarian maturation, DEGs were primarily annotated to the endocrine and immune systems, as well as canonical signaling cascades including PI3K-Akt, JAK-STAT, Rap1, and Toll-like receptor pathways. In addition, extracellular matrix (ECM)-receptor interaction was enriched in both comparison groups, indicating pleiotropic regulation by dietary protein levels during ovarian development. To identify key regulatory factors, we prioritized ovarian DEGs associated with the endocrine system, immune system, steroid hormones, organismal system, signal transduction, and metabolic pathways. The expression patterns of DEGs at various dietary protein levels were visualized using heatmaps constructed with Z-score normalized values. These key genes exhibited predominantly high expression in the CAP42 group, while displaying low expression in both the CAP50 and CAP34 groups (Fig. 5C).

Fig. 5.

Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis and heat map of the differentially expressed genes (DEGs). (A and B) Enriched KEGG pathways analysis of the DEGs in the ovary from comparisons between CAP50 vs. CAP34 and CAP50 vs. CAP42, and the figure was the pathway list of the smallest P-value top 20. The sizes of the data points represent the number of DEGs enriched in the corresponding pathway. The larger the dot, the more genes it contains. (C) Heat map of DEGs involved in key pathways in the ovary. CAP34, CAP42, and CAP50 represent 34%, 42%, and 50% dietary protein levels, respectively. ECM = extracellular matrix.

3.6. Metabolic profile

To facilitate inter-group comparison, ovarian metabolites identified via both positive and negative ionization modes were systematically reorganized and integrated for subsequent analysis. The OPLS-DA score plots demonstrated clear distinctions among all comparison groups (Fig. S2). The classification parameters, R²Y and Q², exhibited stable values, reflecting sufficient model fit and predictive capacity, thus confirming the feasibility of analyzing subsequent DEMs.

Venn analysis identified a total of 315 unique metabolites across all comparative groups (Fig. S3). Specifically, the comparison between CAP50 and CAP34 revealed 181 DEMs, including 37 upregulated and 144 downregulated metabolites (Fig. 6A). There were 34 metabolites upregulated and 119 metabolites downregulated in the ovary between CAP50 vs. CAP42 groups (Fig. 6B). To investigate the metabolic pathways involved in ovarian maturation of northern snakehead fed different protein diets, metabolites and their correlation with maternal ovarian metabolic pathways were examined. The findings suggested that the primary metabolic pathways associated with differential metabolites in the ovaries were butanoate metabolism, synthesis and degradation of ketone bodies, and lysine degradation metabolism between CAP50 and CAP34 groups (Fig. 6C). However, histidine metabolism, retinol metabolism, starch and sucrose metabolism and citrate cycle were significantly enriched in the CAP50 vs. CAP42 groups (Fig. 6D).

Fig. 6.

Statistical comparison of metabolites and analysis of differential metabolites and key metabolic pathways. (A and B) Volcano map of differentially expressed metabolites (DEMs) in the ovary from comparisons between CAP50 vs. CAP34 and CAP50 vs. CAP42. The vertical and horizontal axes display the expression levels of the metabolites in the two groups using -log₁₀ (P-value) and log₂ (FC), respectively. Each point in the graph represents a specific metabolite whose size corresponds to the VIP value. Red dots, significantly up-regulated genes (VIP > 1 and P < 0.05); Yellow dots, significantly down-regulated genes (VIP > 1 and P < 0.05); Blue dots, no significant difference (VIP < 1 and P > 0.05). (C and D) Kyoto Encyclopedia of Genes and Genomes (KEGG) topology analysis bubble diagram of the key metabolites in the ovary from comparisons between CAP50 vs. CAP34 and CAP50 vs. CAP42. The horizontal axis indicates the metabolite’s relative importance within the pathway, while the vertical axis illustrates the enriched significance [-log₁₀ (P-value)] of the metabolite involved in the pathway. (E) A heatmap of the top 30 DEMs in the ovaries. The colors indicate the expression of the metabolites in the group of samples. CAP34, CAP42, and CAP50 represent 34%, 42%, and 50% dietary protein levels, respectively. FC = fold change; VIP = variable important in projection.

Hierarchical cluster analysis was performed on the top 30 significantly enriched differential metabolites in each ovary (Fig. 6E). Multiple DEMs involved in lipid metabolism and AA metabolism, including 2-hydroxyestradiol, prostaglandin G2, adrenosterone, phosphocholine, phosphorylcholine, allopregnanolone, arachidonic acid, eicosapentaenoic acid, progesterone, 3-methyl-l-histidine, and alpha-ketoglutarate, were significantly up-regulated in the CAP50 group. DEMs associated with steroid hormone biosynthesis, such as 5α-androsterone, androsterone glucuronide, phosphocreatine, epinephrine, L-saccharopine, 5α-androstan-17β-ol-3-one, testosterone, and 5-androstene- 3β,17β-diol, were significantly up-regulated in the CAP42 group. Furthermore, metabolites such as leukotriene D4, prostaglandin J2, phosphoserine, eicosanoid acid, pyruvate, prostaglandin E2, and prostaglandin H2, which are involved in arachidonic acid metabolism, were significantly enriched in the CAP34 group.

3.7. Transcription of genes involved in ovarian development

As shown in Fig. 7, ten significantly differentially expressed genes were selected from the transcriptome for further validation of their expression patterns during ovarian development and sex hormone production. The RNA-Seq data were validated by selecting 10 DEGs from the ovary (pik3, ar, er, amh, tgfb1, foxo3, ifr2, thbs1, ghr, and ilr2) for qRT-PCR analysis. The results demonstrated consistent patterns between the RNA-Seq and qRT-PCR data. Compared with the CAP38 and CAP42 groups, the expression levels of most critical genes in the ovaries were significantly decreased in the CAP50 group.

Fig. 7.

Relative mRNA levels of key differentially expressed genes involved in the regulation of ovarian development in the ovary of northern snakehead. CAP34, CAP38, CAP42, CAP46, and CAP50 represent 34%, 38%, 42%, 46%, and 50% dietary protein levels, respectively. Different lowercase letters above columns represent significant differences among treatments at P < 0.05 (n = 6).

4. Discussion

Before exogenous feeding, in the period of peri-conception and embryogenesis, vertebrates exhibit significant developmental plasticity and are highly sensitive to nutritional alterations, enabling the organisms to adapt to adverse post-natal circumstances (Izquierdo et al., 2015). In this study, the results showed that the nutritional status of female fish has a profound effect on the later performance of offspring. According to the regression analysis, the reproductive performance was the best when the dietary protein level was between 41% and 42.2%, although the dietary protein level did not significantly affect the egg mass and hatching rate. Similar results were observed in our previous study on the maternal protein requirements of yellow catfish, which demonstrated that higher dietary protein levels are not advantageous for female fish during the reproductive period (Chen et al., 2022). In this study, a 42% dietary protein level was sufficient for the gonadal development of the female northern snakeheads, because all nutrients were mainly used for gonadal development rather than growth during sexual maturation. The results are consistent with previous studies on the female rohu (L. rohita) (Afzal Khan et al., 2005), Nile tilapia (O. niloticus) (El-Sayed and Kawanna, 2008), and swordtails (Xiphophorus helleri) (Chong et al., 2004), all of which observed that maintaining dietary protein content within an appropriate range can promote gonadal development and enhance spawning performance, whereas excessive or insufficient protein levels may delay gonadal maturation or impair fecundity.

Egg quality is a well-established predictor of subsequent larval viability, quality, and stress resistance (Stuart et al., 2020). Egg diameter has been demonstrated to be an indicator of spawning success in fish, based on the premise that larger eggs possess a higher metabolic reserve and thus are of superior quality (Gruenthal et al., 2014; Nazari et al., 2009). This study demonstrated a significant positive correlation between broodstock dietary protein levels and egg diameter. Notably, 7DPH larvae exhibited superior biometric parameters (body length and weight) under moderate-protein diets (≤42%) relative to high-protein regimens (46%-50%). However, following the complete shift from endogenous to exogenous nutrition, the growth performance of 14DPH larvae in the high-protein group (42%-50%) markedly surpassed that of the low-protein group (34%-38%). It has been shown that egg size is positively correlated with viability, larval yolk sac volume, and days to starvation (Stuart et al., 2020). In this study, the superior starvation tolerance observed in larvae from broodstock fed a diet containing 42%-50% protein suggests that maternally derived resources fundamentally enhance larval metabolic efficiency and resilience during the nutritionally critical post-yolk absorption stage. The present results suggest a transgenerational nutritional legacy, wherein maternal dietary strategies influence not only gamete quality but also the developmental potential and stress adaptability of offspring.

In addition to fecundity, comprehensive egg quality evaluation involves morphological parameters, nutrient composition, and hatching success. Vitellogenesis requires substantial energetic and material investment to establish sufficient embryonic nutrient reserves. AAs and fatty acids are recognized as essential energy sources for embryonic and larval development (Thiruvasagam et al., 2024). In this study, when the dietary protein level was 34% and 38%, the concentrations of crude protein and AAs in the ovaries of northern snakehead significantly decreased. This may be attributed to the insufficient dietary protein content, which fails to provide adequate protein nutrition for ovarian development. Once the dietary protein level reaches saturation at 42%, further increases in dietary protein will not influence the AA composition of the ovaries. This aligns with observations in Nile tilapia (O. niloticus), where similar protein thresholds regulated ovarian AA profiles (Gunasekera et al., 1996). In this study, dietary protein significantly affected the content of ovarian HUFA, particularly when the maternal diet contained 42% protein, which resulted in a significant increase in the content of ARA, EPA, and DHA. In embryonic development and early larval stages, these fatty acids play a crucial structural role as components of the phospholipids in the fish bio-membranes (Izquierdo, 1996). Additionally, they serve as a source of metabolic energy for the embryo, precursors for the synthesis of eicosanoids that regulate cell proliferation and differentiation, and regulators of gene transcription and expression (Li et al., 2024). The results of this study indicate that the maternal pathway can provide more available energy and nutrients for embryonic development by increasing the accumulation of AAs and essential fatty acids under the optimal dietary protein level, which also confirms that the fertilization rate of northern snakehead is the highest when the maternal diet contains 42% protein. This relationship enables effective early nutritional manipulation of the embryo via the broodstock diet in fish (Calder, 2012; Nhan et al., 2020).

In the present study, dietary CAP34 and CAP38 were found to enhance the secretion of E2 and VTG in plasma, whereas dietary protein levels exceeding 42% significantly suppressed the synthesis of E2 and VTG. This reduction may be attributed to hepatic AA metabolic stress, competitive inhibition of steroidogenic substrates, and energy allocation (Hardie et al., 2012; Laplante and Sabatini, 2012), as evidenced by the reduced expression of the transcription factor foxo3 and estrogen receptor er genes in the high-protein group. Reduced positive signals of CYP19A1 and ER were observed in oocytes of broodstock fed a dietary protein level of 34%, suggesting potential inhibitory effects on ovarian development. Furthermore, aromatase functions to transform androgen into estrogen, a process that stimulates the liver to synthesize VTG (Ye et al., 2022). During the endogenous nutrition phase of larvae, VTG functions as a precursor for egg yolk formation, providing essential nutrients and energy required for fry development, which underscores the critical importance of both yolk quantity and quality in ensuring the early survival and development of fry (Hiramatsu et al., 2015). Therefore, it was hypothesized that insufficient supply of feed protein may inhibit ovarian development, and the reduced secretion of sex hormones resulting from excessive feed protein might be attributed to the increased liver metabolic burden on the broodstock.

The development of oocytes and follicles is regulated by pathways associated with cell proliferation and survival, such as the PI3K-Akt signaling pathway (Wang et al., 2016). In this study, the DEGs in the ovaries of northern snakehead under varying protein levels were significantly enriched in the PI3K-AKT signaling pathway and ECM-receptor interaction pathway, indicating that these pathways are crucial for maintaining ovarian development and oocyte maturation. In addition, the relative mRNA expression levels of genes associated with various cellular processes of ovarian, such as pi3k, amh, tgfb1, foxo3, ifr2, ghr, and ilr2 were analyzed. The results revealed differences in the transcript levels of these genes among groups fed different dietary protein levels. As one of the core mechanisms that regulate growth and development, the PI3K-AKT signaling pathway plays a pivotal role in mediating ovarian follicle and oocyte maturation induced by growth factors and maturation-inducing hormone (MIH) (Andrade et al., 2017; Weber et al., 2025). In this study, low protein intake was observed to inhibit pi3k transcription in the ovaries of the northern snakehead. This observation led to the hypothesis that insufficient dietary protein may suppress pi3k activation, thereby promoting apoptosis. Similar findings have been documented in abalones (Haliotis discus hannai), where both excessively low and high dietary protein levels were found to downregulate pi3k expression, leading to impaired growth performance (Ma et al., 2021).

The transcription factor FOXO3a plays a key role in ovarian activity. When activated, it has anti-proliferative and pro-apoptotic effects, and protects quiescent cells from oxidative stress (Kops et al., 2002). In the present study, the foxo3 gene was significantly downregulated in the ovaries of the high protein diet group, while the fertilization rate was the lowest in the highest protein group, implying that high protein feeding may lead to oxidative stress in oocytes. It has been reported that activation of FOXO3a protects mouse oocytes from oxidative stress (Di Emidio et al., 2014). In the present study, the expression levels of amh in the ovaries of the CAP38 group were significantly higher than those in other groups. Additionally, the high expression of the antibody proteins CYP19A1 and ER in the ovaries of the CAP38 group indirectly confirms that AMH may play a role in regulating vitellogenesis and oocyte maturation. Some studies have reported a favorable association between AMH and both oocyte quality and the achievement of successful fertilization (Fong et al., 2008; Zhang et al., 2020). In this study, the reduced fertilization rate observed in the high-protein group was consistent with the low expression of amh. Existing research have shown that the expression level of amh in the ovaries of fish increases during the vitellogenesis and oocyte maturation stages, which is considered a marker of oocyte maturation (Liu et al., 2021; Rocha et al., 2016). These findings suggest that amh may serve as a reference indicator for assessing ovarian developmental status and egg quality. Interestingly, tgfβ1, which belongs to the same TGF-β superfamily as amh, showed the same expression pattern. Our previous studies in maternal blotched snakehead (Channa maculata) have also found that tgfβ1 is highly expressed in high-quality eggs (Fei et al., 2024).

Metabolomics analysis in this study further revealed that northern snakehead broodstocks precisely modulated their metabolic pathways in response to ovarian development. In this study, the differential metabolites in the ovaries of female snakehead fed diets with different protein levels were primarily associated with AA metabolism, lipid metabolism, steroid hormone biosynthesis, and arachidonic acid metabolism. This indicates that these pathways may function as essential metabolic routes, playing a critical role in the gonad development of the northern snakehead. These pathways involved in the regulation of ovarian development have also been validated in the rare minnow (Gobiocypris rarus) (Gao et al., 2017) and Chinese sturgeons (Acipenser sinensis) (Leng et al., 2019). In this study, the DEMs of AA metabolism were mainly enriched in lysine and histidine metabolism pathways, indicating a critical role of these pathways in ovarian development. However, the regulatory network of AA metabolism during fish ovarian development remains unclear. Maternal low protein intake can lead to ovarian AA deficiency, potentially impacting ovarian function.

Several prostaglandins derived from arachidonic acid (ARA) metabolism stimulate pituitary secretion of luteinizing hormone and follicle-stimulating hormone, which are crucial for gonadal development (Xu et al., 2016). Leukotriene D4 and prostaglandin E2 were significantly enriched in the ovaries of the CAP34 group. These substances are pro-inflammatory mediators derived from arachidonic acid metabolism, synthesized through the cyclooxygenase (COX) and lipoxygenase (LOX) pathways (Takahashi et al., 2018). This suggests that maternal low protein intake enhances local inflammatory responses, potentially leading to ovarian tissue damage, oxidative stress, and impaired oocyte development. The precursors of steroid hormones, such as androsterone glucuronide, epinephrine, testosterone, and 5-androstene-3 β metabolites, were significantly up-regulated in the CAP42 group, suggesting substantial enrichment of steroid hormone biosynthesis pathways and further elucidating molecular characteristics related to ovarian function development. This is consistent with the high expression of genes involved in steroid hormone synthesis in the CAP42 group, suggesting that a 42% dietary protein level can promote steroid hormone synthesis to maintain oocyte development in female northern snakehead.

5. Conclusion

The ovarian development status, nutrient composition, sex hormone concentrations, metabolomics profiling, and transcriptomic analysis all demonstrated that dietary protein levels can influence the ovarian development and reproductive performance of broodstock, as well as the quality of their larvae. A dietary protein level of 42% was found to be most conducive to oocyte growth, ovarian AA and lipid deposition, and it was hypothesized that this outcome was attributable to enhanced regulatory effects on AA metabolism, lipid metabolism, and steroid hormone synthesis. Furthermore, enrichment of ARA metabolites in the ovaries of the low-protein group suggests that a low-protein diet may induce inflammatory or responses or aberrant steroid hormone secretion in northern snakehead. These findings deepen advance the understanding of the nutritional regulation of ovarian development in fish, and provide feasible metabolic targets for improving broodstock reproductive efficiency.

Credit Author Statement

Shuzhan Fei: Writing – original draft, Methodology, Investigation, Data curation. Junhao Zhang: Software, Methodology, Investigation, Data curation. Haiyang Liu: Software, Investigation, Formal analysis. Qing Luo: Visualization, Validation, Conceptualization. Jian Zhao: Resources, Project administration, Funding acquisition, Conceptualization. Mi Ou: Writing – review & editing, Supervision, Funding acquisition.

Declaration of competing interest

We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, and there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the content of this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: