Effect of Quinoa Germ Meal as a Replacement for Fish Meal in Diets on Intestinal Health of Juvenile Turbot (Scophthalmus maximus)

Simple Summary

This study investigates the potential of quinoa germ meal as a dietary alternative protein source for juvenile turbot. Five experimental diets were formulated, comprising a control group and four experimental groups with replacement levels of 10%, 20%, 30%, and 40%. With increasing replacement levels, most growth parameters, notably feed intake, declined significantly. All experimental groups displayed decreased muscle crude lipid content and altered intestinal morphology. Transcriptome analysis revealed that quinoa germ meal modulated gene expression in ribosome-related and protein digestion and absorption pathways, and notably, high replacement levels activated the cellular senescence pathway in the intestine. These findings provide a theoretical basis for optimizing quinoa germ meal as a sustainable protein source in aquaculture.

Abstract

Quinoa germ meal (QGM) is a protein-rich by-product with potential as an alternative protein source; however, its effects on growth performance and intestinal health in marine carnivorous fish remain unclear. Juvenile turbot (Scophthalmus maximus) were fed five isonitrogenous (45.6% crude protein) and isolipidic (9.8% crude lipid) diets for 8 weeks: a fishmeal-based control diet (C) and four experimental diets in which fishmeal was replaced with QGM at 10% (Q10), 20% (Q20), 30% (Q30), and 40% (Q40). Growth performance, muscle proximate composition, intestinal histomorphology, and intestinal transcriptomic profiles were analyzed. Growth performance parameters, including final body weight, weight gain rate, specific growth rate, daily feed intake, and condition factor, decreased significantly with increasing QGM inclusion levels compared with the control (p < 0.05). Feed conversion ratio increased significantly only when replacement exceeded 30% (p < 0.05), while survival rate was unaffected (p > 0.05). Muscle crude lipid content was significantly reduced in all QGM-fed groups (p < 0.05), whereas crude protein, moisture, and ash contents were unchanged. Intestinal mucosal fold height increased in the Q30 and Q40 groups, while submucosal width decreased in the Q40 group (p < 0.05). Transcriptomic analysis revealed a dose-dependent increase in differentially expressed genes, mainly enriched in ribosome-related pathways, linoleic acid metabolism, and protein digestion and absorption. High dietary inclusion of QGM (>30%) impaired growth performance in juvenile turbot, whereas low inclusion levels (≤20%) exerted minimal adverse effects. Quinoa germ meal represents a potential alternative protein source, yet its effective utilization requires further optimization to maintain growth performance.

1. Introduction

Aquaculture currently represents the fastest-growing sector in animal production, boasting an average annual growth rate of 7.5% since 1970 [1]. This rapid expansion has inevitably precipitated a parallel surge in the demand for commercial aquafeeds. Fishmeal (FM) has long served as a traditional protein source in aquaculture feed due to its high protein content and balanced amino acid profile [2], but the FM supply has been threatened by factors such as overfishing and climate change [3,4]. Therefore, extensive research has been conducted in recent years to identify novel protein sources for partial replacement of FM in aquaculture feed, with plant-based alternatives receiving particular attention. These sources include soybean meal (SBM) [5], cottonseed meals [6], corn gluten meal [7], and pea protein [8]. However, the application of these conventional plant protein sources is often limited by various drawbacks. Nutritional constraints, such as unbalanced essential amino acid profiles, low palatability, and the presence of anti-nutritional factors, can lead to reduced growth performance and intestinal inflammation in fish [9], and conventional crops such as corn and soybeans require arable land and are susceptible to environmental fluctuations [10]. For example, China’s soybean supply relies heavily on international trade, which is characterized by high vulnerability and uncertainty [11]. In contrast, quinoa (Chenopodium quinoa) is an emerging crop that can thrive in marginal environments [12] without competing for arable land; thus, it has significant potential for sustainable development.

Quinoa has been cultivated for 5000–7000 years in the Andean region of South America [13]. Due to its rich genetic diversity, quinoa can adapt to and thrive in the harshest environmental conditions [12]. Quinoa cultivation in China spans 24 provinces, and planting areas are distributed across an elevational range from 154 m below sea level to over 5000 m above sea level [14]. Quinoa has remarkable nutritional value, as it contains 12.9–16.5% protein, 1.8–9.5% fat, and 46–77% complex carbohydrates (including starch and dietary fiber) [15,16]. It is also a significant source of antioxidants, and its rich free and bound phenolic compounds play a key role in mitigating oxidative stress and cellular damage [17].

Research on quinoa as an alternative protein source in aquafeed is still in its early stages, and limited scientific evidence is currently available. Kumar et al. [18] replaced FM with quinoa husk in the diet of Pangasianodon hypophthalmus and found that replacement at 25% effectively mitigated the fish’s stress gene responses to multiple stressors while also enhancing growth performance and modulating immune responses. Ahmed et al. [19] reported that the use of quinoa seeds as dietary supplements for Nile tilapia (Oreochromis niloticus) could improve the fish’s survival rate and hematological, digestive, antioxidant and immunological parameters; enhance its immune response against Aeromonas sobria infection; and restore the histological morphology of its intestine, liver, and spleen tissues. In a study of the components of quinoa, Mufari et al. [20] found that the germ contains double the protein content and triple the lipid content of whole quinoa. To date, however, no systematic studies of the use of quinoa germ meal (QGM) as an alternative protein source to replace FM in aquafeed have been reported.

The intestinal tract is an important organ that allows animals to digest and absorb nutrients from food [21]. It is a key research target for evaluating the efficacy of plant protein substitution for FM because its health status directly affects the growth performance and immune response of fish [22]. Blachier et al. [23] found that plant protein substitution for FM could induce intestinal inflammation, disrupt intestinal barrier function, compromise intestinal integrity, and ultimately impair the growth performance of Atlantic salmon (Salmo salar). Additionally, SBM-induced enteritis is a prevalent issue in aquaculture. Therefore, analyzing intestinal histomorphological changes and the expression of inflammatory markers is a useful approach for evaluating the impact of plant protein replacement on digestive and absorptive functions in fish [24,25].

In this study, we formulated an isonitrogenous and isolipidic basal diet using FM as the primary protein source, and we then incorporated graded levels of QGM in place of FM to produce the test diets. Our objective was to investigate the effects of QGM substitution levels on growth performance, muscle composition, and intestinal health of juvenile turbot (Scophthalmus maximus) and to elucidate the molecular mechanisms involved in intestinal-related pathways under different dietary QGM inclusion levels.

2. Materials and Methods

2.1. Experimental Diets

The diets were formulated using FM as the primary protein source, with fish oil and soybean lecithin serving as lipid sources and wheat flour and gluten as carbohydrate sources. A basal diet containing 45.6% crude protein and 9.8% crude lipid was formulated using FM as the main protein source (Diet C). Four additional FM-substituted diets were prepared by replacing 10%, 20%, 30%, and 40% of FM in the basal diet with QGM; these diets were designated as Q10, Q20, Q30, and Q40. All experimental diets were formulated to be isonitrogenous and isolipidic. To maintain these nutritional standards during the substitution of fish meal with quinoa germ meal, the inclusion levels of wheat gluten and wheat flour were adjusted in the basal formula. Additionally, crystalline amino acids (lysine, methionine, and threonine) were supplemented to match the essential amino acid profile across groups and meet the nutritional requirements of juvenile turbot [26]. Table 1 and Supplementary Table S1 present the experimental formulations of the diets.

Table 1.

Diet recipes and chemical composition as analyzed in this study.

	C	Q10	Q20	Q30	Q40
Fish meal a	60	54	48	42	36
Quinoa germ meal b	\	9.4	18.8	28.2	37.6
Vital wheat gluten	3	3.88	4.74	5.64	6.5
Fish oil	4	4	4	4	4
Wheat flour	27.9	23.25	18.7	13.95	9.4
Soy lecithin	2	2	2	2	2
Vitamin premix c	0.5	0.5	0.5	0.5	0.5
Mineral premix d	1	1	1	1	1
Choline chloride	0.5	0.5	0.5	0.5	0.5
Monocalcium phosphate	1	1	1	1	1
Calcium propionate	0.05	0.05	0.05	0.05	0.05
Butylated hydroxyanisole	0.25	0.25	0.25	0.25	0.25
Butylated hydroxytoluene	0.25	0.25	0.25	0.25	0.25
Lysine	\	0.19	0.29	0.52	0.62
Methionine	\	0.04	0.1	0.14	0.19
Threonine	\	0.14	0.27	0.45	0.59
Total	100	100	100	100	100
Proximate composition of the diets (dry matter, %)
Crude protein	49.7	49.98	50.11	50.32	50.34
Crude lipid	10.6	10.42	10.89	10.83	10.64
Moisture	11.3	11.45	11.65	11.88	11.75
Ash content	11.7	11.42	11.22	10.92	10.54
Lysine	3.38	3.39	3.39	3.38	3.38
Methionine	1.29	1.29	1.3	1.3	1.29
Threonine	2.03	2.05	2.03	2.03	2.03

^a Fish meal (crude protein content was 65.56%, crude lipid content was 5.3%, lysine content was 3.6%; methionine content was 0.05%; threonine content was 0.29%) was supplied by Blumar SA Co., Las Condes, Chile. ^b Quinoa germ meal (crude protein content was 41.75%, crude lipid content was 3.9%, lysine content was 1.74%; methionine content was 0.55%; threonine content was 1.84%) was supplied by Center for Excellence in Molecular Plant Sciences, Chinese Academy of Sciences, Shanghai, China. ^c Vitamin premix supplied the diet with (mg kg⁻¹ diet) the following compounds: Vitamin A (Vitamin A acetate), 310 mg/kg; Vitamin D₃, 5 mg/kg; Vitamin E (dl-α-tocopheryl acetate), 1000 mg/kg; Vitamin K₃ (menadione sodium bisulfite), 10 mg/kg; Vitamin B₁ (thiamine nitrate), 1800 mg/kg; Vitamin B₂ (riboflavin), 5 mg/kg; Vitamin B₆ (pyridoxine hydrochloride), 5 mg/kg; Vitamin B₁₂ (cobalamin), 10 mg/kg; Vitamin C (ascorbic acid), 16,000 mg/kg; niacinamide, 9080 mg/kg; D-calcium pantothenate (Vitamin B₅), 10 mg/kg; folic acid, 20 mg/kg; biotin, 60 mg/kg; cellulose, 11,685 mg/kg. ^d Mineral premix supplied the diet with (mg kg⁻¹ diet) the following compounds: ferrous sulfate, 1000 mg/kg; copper sulfate, 500 mg/kg; zinc sulfate, 2800 mg/kg; manganese sulfate, 6800 mg/kg; magnesium sulfate, 1000 mg/kg; sodium selenite, 10 mg/kg; cobalt chloride, 5 mg/kg; potassium chloride, 5000 mg/kg; calcium iodate, 50 mg/kg; sodium chloride, 10,000 mg/kg; L-lysine hydrochloride, 5000 mg/kg; zeolite powder, 47,835 mg/kg.

2.2. Trial Duration and Feeding Regimes

At the Langya Research Base (Yellow Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Qingdao, China), the current experimental trial was carried out. Juvenile turbot used in the study (n = 300) were from the same batch of seedlings purchased from Shengli Ocean Technology Co., Ltd. (Rushan City, China). Before the commencement of the trial, the fish underwent a 28-day stabilization phase involving ad libitum feeding with Diet C. After a 24 h fasting period, healthy turbot with similar body weights (17.62 ± 0.24 g) were selected, and individuals with abnormal appearance or behavior were excluded. Initial body weights were then measured, after which all fish were individually numbered. Sample size determination was based on previously reported [5] variability in key growth-related parameters at the tank-mean level, as the tank was considered the experimental unit, in combination with practical considerations, including facility availability, rearing system capacity, and sampling demands. Therefore, three replicate tanks were used per dietary treatment. A stocking density of 30 fish per tank was applied to reduce the impact of individual variability on tank-level responses, to meet the requirements for growth assessment and terminal tissue and biochemical sampling, and to account for potential mortality during the feeding trial. Fish were assigned to 15 recirculating tanks (1000 L; 30 fish/tank) in triplicate using a computer-aided randomization protocol. Throughout the experimental period, the system operated with disinfected natural seawater characterized by specific physicochemical properties: 14–16 °C temperature, 28–30‰ salinity, and a pH of 7.6–7.8 [27]. Fish were fed twice daily at 06:30 and 18:30 using a multiple small meals feeding strategy to reduce feed waste.

2.3. Sample Collection

After 8 weeks of culture, fish underwent a 24 h starvation to empty their digestive tracts. At the trial’s conclusion, every individual was assessed for growth metrics, and three representative fish were randomly retrieved from each system. Following anesthesia with 50 mg/L MS-222 [28], the dorsal muscle and intestines were dissected, and intestinal weight was recorded. The samples were immediately placed into cryopreservation tubes and stored in liquid nitrogen for subsequent analysis. Additionally, three more fish per tank were anesthetized with MS-222, and the mid-intestine segments were dissected and fixed in Bouin’s solution for histological examination.

2.4. Measurement of Growth Performance Indices

Fish growth indices were calculated as follows:

Survival Rate (SR, %) = (Final fish count/Initial fish count) × 100,

Weight Gain Rate (WGR, %) = [(Final weight − Initial weight)/Initial weight] × 100,

Feeding Rate (FR, %) = [Total feed intake/((Final weight + Initial weight)/2)/Days] × 100,

Specific Growth Rate (SGR, %) = (ln (Final weight) − ln (Initial weight))/Days × 100,

Feed Conversion Ratio (FCR) = Total feed intake/(Final weight − Initial weight),

Condition Factor (CF, g/cm³) = ((Final weight/(Body length³)) × 100.

2.5. Biochemical Analysis

To assess ingredients, diets, and muscle amino acid composition, the crude protein, crude lipid, moisture, and ash contents of the samples were measured following the methods specified in the Chinese National Food Safety Standards. Amino acid composition was determined according to GB 5009.124-2016 [29] utilizing an L-8900 amino acid analyzer (Hitachi Ltd. Co., Tokyo, Japan); protein content was measured using the Kjeldahl method [30]; crude lipid was determined by Soxhlet extraction [31]; moisture content was measured using the direct drying method [32]; and ash content was quantified through the high-temperature incineration method [33]. Saponin content in QGM and diets was determined by a High-Performance Liquid Chromatography system (Agilent 1100 Series (Wilmington, DE, USA)) [34].

2.6. Observation of Intestinal Tissue Structure

The observation of intestinal tissue structure primarily involved three steps: section preparation, imaging, and measurement [35]. Following a 24 h fixation in Bouin’s solution, intestinal samples were preserved in 75% ethanol. Standard histological processing was performed, encompassing dehydration, xylene clearing, and paraffin embedding. Serial sections (5 μm thick) were prepared using a Leica RM2125 RTS rotary microtome (Nussloch, Germany) and stained with hematoxylin and eosin (H&E). Digital imaging was conducted via the Pannoramic MIDI II system (3DHistech, Budapest, Hungary), and structural analysis was executed using the accompanying Slide Converter software (v2.3.2.53488).

2.7. Intestinal Transcriptome Analysis

Sequencing procedures, including RNA extraction from the intestine and library generation, were conducted at Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China). The Fastp tool (v0.23.4) was used to process raw reads, followed by alignment to the reference genome with HISAT2 (v2.2.1). Expression levels were calculated using the Expectation-Maximization method (RSEM). To evaluate differential expression, we utilized both DEGseq (v 1.38.0) and DESeq2(v 1.20), with the threshold for differentially expressed genes (DEGs) set as |log2FC| ≥ 1 and Padjust < 0.05. DEGs were annotated against the GO and KEGG repositories. We applied Goatools (v1.0.6) and Python Scipy (v4.1.33) to detect significantly enriched GO terms and metabolic pathways, applying a strict threshold of Bonferroni-corrected p < 0.05 compared to the entire transcriptome. Furthermore, eggNOG-mapper v2.1.11 (default mode) was employed to complement the functional annotation.

2.8. Experimental Validation of DEGs via RT-qPCR

Transcriptome reliability was assessed by RT-qPCR analysis of randomly chosen DEGs. RNA extraction was performed on pooled samples using TRIzol (Invitrogen), followed by cDNA synthesis utilizing the Tiangen First-Strand SuperMix (Tiangen, Beijing, China). Specific primers were constructed using Primer-BLAST (NCBI, https://www.ncbi.nlm.nih.gov/tools/primer-blast/, accessed on 12 January 2026), manufactured by Sangon Biotech (Shanghai, China), and validated through sequencing and efficiency analysis. Gene expression results were normalized against both β-actin and RPS4 as internal references, with all reactions run in technical triplicate. See Table 2 for detailed primer information. Gene relative expression levels were calculated using the $2^{-\Delta \Delta Ct}$ method [36], with fold changes (FCs) determined relative to the control group’s target gene mRNA levels, followed by log2 fold change (log2FC) calculation.

Table 2.

Primer sequences.

Primer	Sequence (5′–3′)	Description
mep1a.1-F	AGTTCCCATCAAGCCAAGGG	meprin A, alpha (PABA peptide hydrolase), tandem duplicate 1
mep1a.1-R	GCTGCAAATGTGGCATCTCC
slc12a1-F	GGCATCTTACGCCCTCATCA	solute carrier family 12 member 1
slc12a1-R	CCCATCAGGGAAAGCCACAT
plb1-F	TGTGTGCCTGTTCACCACAT	phospholipase B1
plb1-R	GAACCAAAGCGTGATGCGAA
calr3b-F	GCCCGTTTTGAGCCTTTCAG	calreticulin 3b
calr3b-R	GATTCTCCGTGCATGTTGGC
alas1-F	TTTCCCCCGACCAACTGAAG	Aminolevulinate, delta-, synthase 1
alas1-R	TAGATCCATCGGCTCCCCTC
nop10-F	CGGAGAGCGAATCTACACCC	NOP10 ribonucleoprotein homolog (yeast)
nop10-R	CTTGTCGTCTGGGGAGAACC
meltf-F	GACAGGGCTACATGGACTCG	Melanotransferrin
meltf-R	GCCCGTTTCATTTTGTGCCT
acta2-F	GAGCGAGGCTACAGTTTCGT	Actin alpha 2, smooth muscle
acta2-R	AGTCCAGAGCGACGTAACAC
tm4sf4-F	ACCCAAAGAGTCAAGCCACC	Transmembrane 4 L six family member 4
tm4sf4-R	AAGCTGCCCGAACACATTCT
tpm1-F	CAAGGCCGAGGCAGATGTC	Tropomyosin 1 (alpha), transcript variant X10
tpm1-R	TGCCTCTCTCGCTCTCATCT
cpb1-F	CAGAGTGACATGGAGCACGA	Carboxypeptidase B1 (tissue)
cpb1-R	GAGGAAACGGAAGCGATCCA
rpl26-F	CGGCCATAGAGGAGCTAAGG	Ribosomal protein L26
rpl26-R	TTGAAGTGCCTTTTGCGGTT
rpl18-F	CAGTGTTGCTGTCAGGACCC	Ribosomal protein L18
rpl18-R	GACCACGAGCTCTCTCGAAC
rps27a-F	CGGAAAGCCCTCTTCTCGTG	Ribosomal protein S27a
rps27a-R	TCTGCATCCTGGCGTTTCAA
endou-F	AAAGGGTGTCTATGCGTCCG	Endonuclease, polyU-specific
endou-R	CCCTTGACCTCTCCTGCAAA
rpl39-F	GTTCGCTGGACGCTTATTGG	Ribosomal protein L39
rpl39-R	CAGTCCGTCCCCAACTATGG
rpl37a-F	AGGACCAGTAAACCAGTGCG	Ribosomal protein L37a
rpl37a-R	CGTCATGTCCAGGTCGGTTA
rpl37-F	GGAAGCCGCACGGATATGAT	Ribosomal protein L37
rpl37-R	TTCCGGCGTTTACCGAAAGA
COX3-F	AACGGAAACAGGCCATCCAA	Cytochrome c oxidase subunit III
COX3-R	ATGGTAAAAGGGGCTTCGCA
lgals2b-F	CCGACTTCGTTACTCTCGGC	Lectin, galactoside-binding, soluble, 2b
lgals2b-R	CGCGGATCTTGAACTCCTGA
gpx4a-F	CTTCCTAAGGGCAGCAGTCAT	Glutathione peroxidase 4a, transcript variant X1
gpx4a-R	GTCTGTCGCCGAGAAGTCATA
RPS4-F	CAACATCTTCGTCATCGGCAAGG	Ribosomal protein S4
RPS4-R	ATTGAACCAGCCTCAGTGTTTAGC
β-actin-F	CATGTACGTTGCCATCCAAG
β-actin-R	ACCAGAGGCATACAGGGACA

2.9. Statistical Analysis

Experimental data were analyzed using SPSS Statistics 24 (IBM, Armonk, NY, USA) and visualized with GraphPad Prism 9 (Boston, MA, USA). Statistical results are presented as mean + standard deviation. For comparisons among three or more groups with a single variable, one-way analysis of variance (ANOVA) followed by Duncan’s multiple comparisons test was applied to determine statistical significance. p < 0.05 was considered to be statistically significant.

3. Results

3.1. The Nutritional Profile of QGM

The amino acid composition of QGM, SBM [37] and FM was determined (Table 3). Compared with SBM and FM, QGM showed distinct differences in individual amino acid contents. Among essential amino acids, arginine content in QGM (4.84%) was higher than that in SBM (2.80%) and FM (3.85%), while methionine content (0.55%) was higher than that in SBM (0.35%) but lower than that in FM (0.84%). The levels of histidine, isoleucine, leucine, lysine, threonine and valine in QGM were lower than those in FM and generally lower than or comparable to those in SBM. Phenylalanine content in QGM was lower than in SBM but close to that in FM, and tryptophan content was lower than in SBM and similar to FM. For non-essential amino acids, QGM showed the highest glutamate content (9.35%) compared with SBM (8.04%) and FM (7.30%). Alanine, glycine and serine contents in QGM were lower than those in FM but comparable to SBM, while aspartate, cystine, tyrosine and proline contents were lower than those in both SBM and FM. Overall, QGM differed from SBM and FM mainly by its higher arginine and glutamate contents and lower levels of several other essential and non-essential amino acids. The saponin content was 0.19% in QGM (Supplementary Table S2), undetected in the control, and ranged from 0.02% (Q10, Q20) to 0.03% (Q30 and Q40) in the treatment groups (Supplementary Table S3).

Table 3.

Amino acid composition (% dry matter basis) of quinoa germ meal and fish meal.

	Quinoa Germ Meal	Soybean Meal	Fish Meal
Ingredient
Crude protein	41.75	50.84	65.56
Crude lipid	3.9	2.11	5.3
Essential amino acids
Arginine	4.84	2.8	3.85
Histidine	1.34	1.64	1.13
Isoleucine	1.37	1.94	2.18
Leucine	2.13	3.34	3.82
Methionine	0.55	0.35	0.84
Lysine	1.74	2.74	3.6
Phenylalanine	1.81	2.18	2.01
Threonine	1.84	1.78	2.84
Valine	1.5	2.01	2.92
Tryptophan	0.4	0.63	0.45
Non-essential amino acids
Alanine	1.78	1.85	3.64
Aspartate	3.59	5.07	4.48
Cystine	0.58	0.69	1.16
Glutamate	9.35	8.04	7.3
Glycine	1.98	1.87	4.54
Serine	1.76	2.34	1.87
Tyrosine	1.27	1.38	1.6
Proline	1.51	2.21	7.54

3.2. Growth Performance

As shown in Figure 1 (Supplementary Table S3), dietary treatments did not compromise survival rates (p > 0.05). Conversely, key growth indicators, including final BW, WGR, and SGR, were significantly suppressed by higher QGM levels, with statistically significant differences detected sequentially from the control down to the Q30 group (p < 0.05). The Q30 and Q40 cohorts shared similar growth profiles. Furthermore, a significant alteration in FCR was only detected when the replacement level surpassed 30%, whereas low-substitution groups (Q10 and Q20) maintained FCR values comparable to the control.

Figure 1.

Growth performance. (A) Survival rate (%); (B) weight gain (%); (C) feed intake (%/day); (D) specific growth rate (%); (E) feed conversion ratio; (F) condition factor (g/cm³). Data are presented as mean ± standard deviation (SD) of three replicates (n = 3). Distinct letters above the bars denote statistically significant differences among treatment groups (one-way ANOVA followed by Duncan’s multiple range test, p < 0.05).

3.3. Muscle Composition

Significant differences in the crude protein, moisture, or ash content of turbot muscle tissue were not detected across any replacement levels (p > 0.05) (Figure 2 and Supplementary Table S4). Replacement of FM by QGM at all four levels significantly reduced the crude lipid level in turbot muscle relative to that of the control (p < 0.05), but the crude lipid level did not differ significantly among the four treatment groups (p > 0.05).

Figure 2.

Biochemical composition of muscle. (A) Crude protein; (B) crude lipid; (C) moisture; (D) ash content. Data are presented as mean ± standard deviation (SD) of three replicates (n = 3). Distinct letters above the bars denote statistically significant differences among treatment groups (one-way ANOVA followed by Duncan’s multiple range test, p < 0.05).

3.4. Morphology of the Intestine

The morphology of the intestinal mucosa underwent changes with increasing levels of QGM in the feed (Figure 3). In the Q30 and Q40 groups, the height of intestinal mucosal folds was significantly higher than that of the control group (p < 0.05). In the Q40 group, the width of the submucosa was significantly lower than that of the control group (p < 0.05) (Figure 4 and Supplementary Table S5).

Figure 3.

Representative histomorphological images of intestinal tissue sections stained with hematoxylin and eosin. (A) Group C; (B) group Q10; (C) group Q20; (D) group Q30; (E) group Q40. MH: mucosal folds height, SW: submucosa width.

Figure 4.

Intestinal histology statistics. (A) Mucosal fold height; (B) submucosa width. Data are presented as mean ± standard deviation (SD) of three replicates (n = 3). Distinct letters above the bars denote statistically significant differences among treatment groups (one-way ANOVA followed by Duncan’s multiple range test, p < 0.05).

3.5. Intestinal Transcriptomic Analysis

3.5.1. Differentially Expressed Genes

The repertoire of DEGs distinguishing the sample groups was determined using the DESeq2 package (Figure 5). As the replacement level of quinoa germ in the feed increased, the number of DEGs significantly rose. The Q40 vs. C comparison had the highest number of DEGs (1618 genes, including 794 upregulated and 824 downregulated genes), and the Q10 vs. C comparison contained the fewest DEGs (595 genes, including 201 upregulated and 394 downregulated genes). The Q20 vs. C comparison revealed 598 DEGs (280 upregulated and 318 downregulated genes), and 777 DEGs (363 upregulated and 414 downregulated genes) were detected in the Q30 vs. C comparison.

Figure 5.

Overview of differential expression profiles. (A) Histogram summarizing the counts of up- and downregulated DEGs. (B,C,E,F) Volcano plots illustrating the distribution of significant genes; red points denote upregulation, while green points signify downregulation. The horizontal dashed line represents the significance threshold (p < 0.05), and the vertical dashed lines indicate the fold change threshold (|log₂FoldChange| > 1). (D) Venn diagram displaying the overlap of DEGs across the four comparison groups (Q10–Q40 relative to C). Numbers within the intersections represent the count of annotated unigenes.

Venn analysis (Figure 5D) highlighted the distinctiveness of the transcriptomic profiles. The Q40 vs. C comparison exhibited the highest number of unique DEGs (949), significantly surpassing those in the Q10 (179), Q20 (251), and Q30 (167) groups. In terms of pairwise overlaps, the Q30 and Q40 groups shared the most substantial number of DEGs (344), suggesting similar regulatory mechanisms at higher substitution levels. Notably, a core set of 78 DEGs was consistently identified across all four treatment groups. For further analysis, we established the following three gene sets: (1) unique DEGs (949 DEGs specific to the Q40 vs. C group); (2) high-replacement DEGs (344 shared DEGs between the Q30 vs. C and Q40 vs. C groups); and (3) common DEGs: 78 DEGs universally present in all groups.

3.5.2. GO Enrichment, KEGG Enrichment and EggNOG Annotation Analysis of DEGs in the Unique DEG Set

GO enrichment analysis highlighted 58 significant terms (Padjust < 0.05) across the MF, CC, and BP categories (Figure 6A). The MF and CC domains were largely characterized by ribosome-associated functions, specifically the structural constituent of ribosome (GO:0003735) and the ribonucleoprotein complex (GO:1990904). Notably, the BP class contained the highest number of enriched terms. These were primarily associated with protein synthesis and metabolism, including translation (GO:0006412), peptide biosynthetic process (GO:0043043), amide biosynthetic process (GO:0043604), and broader functions such as cellular nitrogen compound metabolic process (GO:0034641) and organic substance biosynthetic process (GO:1901576).

Figure 6.

Functional annotation and enrichment analysis of DEGs identified in the Q40 group. Subplots display (A) Gene Ontology (GO) enrichment results; (B) KEGG pathway enrichment analysis; and (C) functional classification based on the eggNOG database.

KEGG analysis identified 12 pathways that were significantly enriched (Padjust < 0.05), primarily distributed across the genetic information processing (GIP), metabolism (M), and organismal systems (OS) categories (Figure 6B). In the GIP class, enrichment was focused on ribosome (ko03010) and DNA replication (ko03030). The OS category highlighted digestive functions, specifically fat digestion and absorption (ko04975) and protein digestion and absorption (ko04974). The metabolism class encompassed the widest range of pathways, including those for lipids (linoleic acid, ko00591; arachidonic acid, ko00590; alpha-linolenic acid, ko00592), amino acids (beta-alanine, ko00410; histidine, ko00340), and others, such as glutathione metabolism (ko00480) and pyrimidine metabolism (ko00240).

In the EggNOG annotation analysis of unique DEGs, the most enriched functional categories were J (translation, ribosome structure and biogenesis), O (posttranslational modification, protein turnover, and molecular chaperones), U (intracellular transport, secretion and vesicular transport), and K (transcription). These categories represent the primary functional roles of DEGs based on the EggNOG database classification (Figure 6C).

3.5.3. GO Enrichment, KEGG Enrichment and EggNOG Annotation Analysis of DEGs in the High-Replacement DEG Set

For common DEGs associated with the high-replacement ratio, 53 GO terms were significantly enriched (Padjust < 0.05; Figure 7A). In the MF category, enrichment focused on structural constituent of ribosome (GO:0003735) and RNA binding (GO:0003723). The CC category was dominated by translation-related machinery, ranging from general non-membrane-bounded organelles (e.g., nucleolus and the ribonucleoprotein complex) to specific initiation structures like the eukaryotic 43S/48S preinitiation complexes and translation initiation factor 3 complex. Consistently, the BP category highlighted core biosynthetic activities, including rRNA processing/metabolism, translation, and peptide biosynthetic processes.

Figure 7.

Functional annotation of DEGs identified in the high-replacement group. (A) Gene Ontology (GO) enrichment analysis; (B) KEGG pathway enrichment analysis; (C) EggNOG functional classification.

As shown in Figure 7B, four KEGG pathways were significantly enriched in the high-replacement dataset (Padjust < 0.05). The analysis pointed to a modulation of translational machinery (ribosome [ko03010]; ribosome biogenesis in eukaryotes [ko03008]) and tissue/digestive functions (ECM-receptor interaction [ko04512]; protein digestion and absorption [ko04974]). An overview of the top 20 enriched pathways is provided in the same figure.

EggNOG annotation analysis revealed that the most enriched functional categories were J, O, and K (Figure 7C).

3.5.4. GO Enrichment, KEGG Enrichment and EggNOG Annotation Analysis of DEGs in the Common DEG Set

For the common DEG set, GO analysis highlighted a distinct enrichment in cell cycle regulatory mechanisms. Specifically, five pathways within the BP category reached statistical significance (Padjust < 0.05; Figure 8A): regulation of cell cycle process (GO:0010564), regulation of mitotic cell cycle (GO:0007346), cell cycle regulation (GO:0051726), and spindle-associated processes such as regulation of spindle checkpoint (GO:0090224) and regulation of mitotic spindle organization (GO:0060236).

Figure 8.

Functional annotation of the common DEG set shared by all treatment groups. (A) GO enrichment analysis; (B) KEGG pathway enrichment analysis; (C) EggNOG functional classification.

Figure 8B shows the top 20 enriched pathways identified by KEGG enrichment analysis of the common DEGs. However, no pathways were significantly enriched (Padjust > 0.05). The circadian rhythm (ko04710) pathway showed the highest enrichment.

EggNOG annotation analysis revealed that the most frequently annotated functional categories were O, Z (cytoskeleton), J, and T (signal transduction mechanisms) (Figure 8C).

3.6. DEGs Involved in Various Processes

3.6.1. Ribosome-Related Processes

In this study, we detected 74 ribosomal-related DEGs. The expression levels of ifi35 (interferon-induced protein35), LOC118310109 (mucin-3A), and gba (glucosidase, beta, acid) were downregulated as quinoa replacement levels increased, whereas the expression levels of other ribosomal protein genes, such as rps20 (ribosomal protein S20), rps12, and rpl22, were upregulated under the same conditions (Figure 9A).

Figure 9.

Circular heatmaps illustrating the transcriptional profiles of specific functional categories. Gene symbols are annotated on the periphery, while sample identities are mapped to the radial cross-sections. Expression magnitude is color-coded, with red indicating high relative abundance and blue indicating low abundance. Subplots correspond to (A) ribosome; (B) structural molecule activity; (C) biosynthesis; (D) metabolism; (E) fat digestion and absorption; and (F) protein digestion and absorption.

3.6.2. DEGs Involved in Structural Molecular Activity

In total, 78 DEGs associated with structural molecular activity were identified in this study. These included major ribosomal protein genes, whose expression levels were upregulated with increasing quinoa replacement levels. In addition to ribosomal protein genes, histone-related genes such as LOC118285068 (histone H1-like), LOC118285072 (histone H2A-like), and LOC118287506 (histone H5) were identified, and their expression levels were downregulated as quinoa replacement levels increased (Figure 9B).

3.6.3. DEGs Involved in Biosynthetic Processes

We identified 105 DEGs that are associated with biosynthetic processes. They included major ribosomal protein genes, which were upregulated with increasing quinoa replacement levels. In contrast, mitochondrial acetate metabolism-related genes, such as acacb (acetyl-CoA carboxylase beta), acss1 (acyl-CoA synthetase short chain family member 1), and faxdc2 (fatty acid hydroxylase domain-containing protein 2), were downregulated as the QGM replacement levels increased (Figure 9C).

3.6.4. DEGs Involved in Metabolic Processes

Thirty-five DEGs involved in metabolic processes were detected in this study (Figure 9D). Expression levels of glutathione metabolism-related genes (e.g., such as chac2 (ChaC, cation transport regulator homolog2 (E. coli), transcript variant X1) and LOC118317139 (glutathione hydrolase7, transcript variant X1)); amino acid metabolism-related genes (e.g., cad (carbamoyl-phosphate synthetase2, aspartate transcarbamylase and dihydroorotase, transcript variant X3) and aasdhppt (aminoadipate-semialdehyde dehydrogenase-phosphopantetheinyl transferase, transcript variant X1); and a nucleotide metabolism-related gene cdab (cytidine deaminase b) were upregulated with increasing quinoa replacement levels. In contrast, other metabolism-associated DEGs, such as aldehyde dehydrogenase, cytochrome P450 enzymes, phospholipase A, and ribonucleotide reductase, were downregulated under the same conditions.

3.6.5. DEGs Involved in Fat Digestion and Absorption Processes

Sixteen DEGs involved in fat digestion and absorption processes were detected in this study (Figure 9E). Among them, expression levels of the bile salt-activated lipase-related gene LOC118284943 (bile salt-activated lipase-like) and a cholesterol reverse transport key protein-related gene LOC118312226 (phospholipid-transporting ATPase ABCA1, transcript variant X1) were upregulated with increasing QGM replacement levels. In contrast, other lipid metabolism-related genes, such as phospholipase-related gene acat2 (acyl-CoA: cholesterol acyltransferase 2) and an apolipoprotein-related gene pla2g3 (phospholipase A2 group III), showed significant downregulation in the Q40 group.

3.6.6. DEGs Involved in Protein Digestion and Absorption Processes

We identified 37 DEGs that are involved in protein digestion and absorption processes. Among them, solute carrier family-related genes simultaneously exhibited both upregulation and downregulation. Expression levels of LOC118291630 (collagen alpha-1(XIV) chain), col11a2 (collagen type XI alpha2 chain), kcnn4 (potassium calcium-activated channel subfamily N member4), and si_dkeyp-120h9.1 (Y+L amino acid transporter 2-like) were upregulated, and those of collagen family members, alpha-fetoprotein, and trypsin were downregulated as the quinoa replacement level increased. Genes in the solute carrier family exhibited bidirectional changes: slc38a2 and slc8a4b were upregulated with increasing QGM replacement levels, while slc15 expression was downregulated. (Figure 9F).

3.7. Key Pathways in Turbot Fed with QGM Replacement Diets

We identified several DEGs that play critical roles in the cellular senescence pathway (Figure 10). In the FOXO signaling pathway, downregulation of TGFβ (Transforming Growth Factor β) and HLA-G (Human Leukocyte Antigen G, MHC class I antigen) indirectly induced DNA damage, leading to upregulated p21 (Cyclin-Dependent Kinase Inhibitor1A), which subsequently suppressed CDK2 (Cyclin-Dependent Kinase2) and CycE (Cyclin E1) expression. In the mTOR signaling pathway, RHEB (Ras Homolog Enriched in Brain) showed increased expression. In the p53 signaling pathway, upregulation of C-Myc (v-Myc Avian Myelocytomatosis Viral Oncogene Homolog) and downregulation of HIPK2 (Homeodomain-Interacting Protein Kinase2) were observed. In the cell cycle pathway, upregulated p21 inhibited CDK2, CycE, and CycA (Cyclin A) expression. Phosphorylation of CDK2/CycA, combined with reduced E2F (Transcription Factor E2F2) expression, further downregulated B-MYB (Myb-Related Protein B). Concurrently, E2F downregulation suppressed FOXM1 (Forkhead Box Protein M), ultimately causing cell cycle arrest. Additionally, upstream regulators of SASP (senescence-associated secretory phenotype), 4EBP1(Eukaryotic Translation Initiation Factor4E Binding Protein 1), and NFAT(Nuclear Factor of Activated T-Cells, Cytoplasmic 1) were upregulated, while Calpain-1 expression was downregulated. The SASP-associated gene PAI-1 (Plasminogen Activator Inhibitor-1) exhibited elevated expression.

Figure 10.

Visual representation of the cellular senescence pathway mapped against the KEGG database. The boxes represent genes, with color coding indicating differential expression status: red denotes upregulation, and green denotes downregulation. The color saturation correlates with the magnitude of the expression difference (darker shades indicate larger fold changes).

The intestinal transcriptomic analysis of turbot fed with QGM replacement diets revealed that DEGs were predominantly enriched in ribosome-related genes, metabolism-related genes, biosynthesis and metabolic pathways, and genes associated with lipid and protein digestion and absorption. Notably, ribosome-related genes were upregulated, while most other functional gene categories exhibited downregulation. Critical alterations were observed in the cellular senescence pathway, where DNA damage-induced upregulation of p21 suppressed cell cycle progression via inhibition of CDK2/CycE. Concurrently, C-Myc was upregulated, while HIPK2 was downregulated. In the mTOR signaling pathway, RHEB showed elevated expression. Senescence-associated genes, including PAI-1, were upregulated, whereas TGFβ was downregulated. These molecular shifts collectively suggest the activation of cellular senescence pathways, thereby potentially facilitating cellular senescence-like processes in turbot.

3.8. Validation of DEGs by RT-qPCR Analysis

To verify the reliability of the transcriptomic profiling, six DEGs were randomly selected from each treatment group (C, Q10–Q40) for RT-qPCR validation. The expression patterns observed via RT-qPCR were highly congruent with the RNA-Seq results, thereby confirming the accuracy of the high-throughput sequencing data (Figure 11).

Figure 11.

Validation of RNA-Seq results was conducted using qRT-PCR. Data are expressed as mean + standard error with six biological replicates. (A) Q10 vs. C; (B) Q20 vs. C; (C) Q30 vs. C; (D) Q40 vs. C.

4. Discussion

4.1. The Amino Acid Composition of QGM

This investigation provides evidence that the amino acid composition of QGM differs markedly from that of the commonly used plant protein source soybean meal (SBM) and the animal protein source fish meal (FM), highlighting its distinct nutritional characteristics as a novel protein ingredient. QGM is well known for its balanced amino acid profile, particularly its relatively high content of essential amino acids [38,39]. Among essential amino acids, the methionine content of QGM was higher than that of SBM but lower than that of FM. This suggests that while QGM provides a nutritional advantage over SBM by partially compensating for methionine deficiency, supplementation is still required to match the profile of fish meal. Similarly, the contents of several essential amino acids, including lysine, threonine, and valine, were lower in QGM than in FM and were generally comparable to or slightly lower than those in SBM. These deficiencies likely act as primary limiting factors for protein synthesis, indicating that QGM has potential primarily as a partial alternative protein source that requires amino acid balancing rather than a complete replacement.

4.2. Effects of QGM Replacement Diets on Growth Performance and Muscle Composition of Juvenile Turbot

In this study, we replaced FM in the diet of turbot with graded replacement levels (10%, 20%, 30%, and 40%) of QGM. Although the inclusion of feed ingredients (e.g., wheat) varied to maintain isonitrogenous and isolipidic profiles, the balanced nutrient levels across groups minimized their potential confounding effects. Consequently, the observed physiological responses were primarily attributed to the substitution of fish meal with quinoa germ meal. Decreased FR, WGR, and SPR were consistently observed across all QGM replacement groups relative to the control. These findings align with previous reports of the effects of plant-based protein replacement diets (e.g., SBM and rapeseed meal) in turbot aquaculture [40]. The reduction in FI at ≥10% inclusion suggests that the palatability of QGM might be a limiting factor. Crucially, the presence of saponins (0.02–0.03%) in the QGM-based diets likely compromised palatability due to their bitterness, thereby serving as a primary factor for the reduced feed intake. Saponins concentrated in quinoa grains are known to generate bitter and astringent tastes, which can suppress appetite and interfere with nutrient absorption [41]. Furthermore, the utilization of quinoa is often hampered by the presence of these anti-nutritional factors and bitter substances, as confirmed in previous studies [42]. Consequently, it is reasonable to infer that the reduced feed intake serves as the primary factor explaining the observed decline in growth performance, rather than nutrient deficiency per se. The FCR of turbot increased significantly only when the QGM replacement level was 30% or higher. This indicates that at lower substitution levels, QGM primarily affected turbot growth by modulating FR rather than FCR. This phenomenon may be due to the balanced amino acid profile of quinoa protein, which has adequate essential amino acids required for growth and metabolic maintenance, as well as superior bioavailability compared to plant proteins such as SBM [43]. SBM replacement in the turbot diet has been shown to reduce the feed intake rate. However, unlike QGM, Gu et al. [5] reported that SBM impaired the FCR at all replacement levels.

Saponins are an important group of secondary metabolites in quinoa. Over the past three decades, at least 40 distinct saponin structures have been isolated from this crop. These compounds pose a significant obstacle to the utilization of quinoa as an animal feed ingredient, primarily due to their inherent bitter taste [44]. In the present trial, turbot exhibited a dose-dependent decline in CF with rising QGM levels, a trend paralleling observations from previous studies on SBM substitution [5]. CF exhibits a positive correlation with BW but a negative correlation with body length and relative intestine mass [45], and it is also influenced by multiple biotic and abiotic factors. In our experiment, the observed reduction in CF may result primarily from decreased feed intake induced by the QGM-based diets, leading to reduced BW. Additionally, QGM substitution reduced the crude lipid content in turbot muscle, likely because quinoa’s soluble dietary fiber and insoluble dietary fiber modulate lipid metabolism [46].

4.3. Effects of QGM Replacement Diets on Intestinal Histology in Juvenile Turbot

Histological analysis revealed that fish in the Q10 and Q20 groups did not exhibit alterations in intestinal morphology, indicating that low-level QGM replacement had no adverse effects on intestinal structure. In contrast, several studies have reported that SBM-based diets triggered intestinal inflammation that was characterized by reduced mucosal fold height and increased submucosal width [5,47,48], which is consistent with findings that high-starch diets induce structural anomalies such as mucosal inflammation and intestinal dilation [49]. In our study, fish in the Q30 and Q40 groups showed increased mucosal fold height. Although this feature has been previously associated with the enhanced digestibility of plant-derived components [50], given the significantly reduced growth performance in our study, it more likely reflects a compensatory hypertrophy to counteract nutritional stress rather than an actual improvement in digestive capacity. However, fish in the Q40 exhibited reduced submucosal width, which was likely attributable to the starvation response driven by significantly decreased feed intake [35].

4.4. Effects of QGM Replacement Diets on Intestinal Gene Expression Patterns in Juvenile Turbot

In turbot fed QGM at the 30% and 40% replacement levels, the intestine exhibited significant upregulation of ribosomal pathway genes and significant downregulation of histone-related genes. Ribosomes are conserved ribonucleoprotein complexes composed of ribosomal RNA and ribosomal proteins that catalyze the translation of genetic codes from mRNA into proteins in all cells [51,52]. Ribosomal proteins can convert nutritional status into changes in gene expression. Previous studies have shown that when nutrients are sufficient, the expression of ribosomal protein-coding genes is upregulated [53]. Conversely, Salem et al. [54] reported that feed restriction in rainbow trout (Oncorhynchus mykiss) downregulated ribosomal pathway genes, which contradicts the present findings. Unlike these starvation models, the present study observed a pervasive upregulation of ribosomal genes despite growth reduction. Rather than indicating sufficient nutrient supply, this discrepancy likely reflects a compensatory metabolic response, where the organism upregulates protein synthesis machinery in an attempt to maintain homeostasis against nutritional stress. Concurrently, altered histone expression may influence epigenetic modifications, thereby regulating intestinal cell proliferation, differentiation, and stress responses [55]. Thus, QGM substitution likely optimizes metabolic and defense mechanisms by suppressing histone synthesis or modification [56]. Similarly, Roostaee et al. [57] suggested that dynamic histone adjustments serve as an adaptive strategy, as nutrients remodel chromatin structure to modulate gene silencing or activation, thereby enhancing growth performance and disease resistance.

The intestine of turbot in the Q40 group exhibited downregulation of acacb, acss1, and faxdc2, indicating suppression of fatty acid synthesis and modification. Quinoa contains high dietary fiber and/or potential anti-nutritional factors that could interfere with energy metabolism, which may explain the observed downregulation. The acacb gene encodes the β subunit of acetyl-CoA carboxylase, a key regulator of fatty acid metabolism that promotes de novo fatty acid synthesis and modulates mitochondrial β-oxidation [58]. Downregulation of this gene likely reflects the organism’s prioritization of limited energy for alternative metabolic pathways by inhibiting fatty acid synthesis under nutritional stress. The gene acss1 encodes an enzyme that catalyzes acetate conversion to acetyl-CoA, thereby directly participating in mitochondrial metabolism, histone acetylation, and ATP production [59]. Reduced acss1 activity may impair acetate utilization efficiency and exacerbate energy deficits and oxidative stress during metabolic pressure. The gene faxdc2 mediates fatty acid hydroxylation and cholesterol synthesis [60], and its suppression likely disrupts lipid absorption efficiency. This scenario aligns with the observed reduction in muscle crude lipid content in the high-quinoa substitution groups. Concurrently, reduced feed intake may contribute to the downregulation of anabolic-pathway-related genes in the intestine of fish in the high-quinoa substitution group (Q40) [61].

Turbot fed the 40% QGM replacement diet exhibited differential gene expression in the intestine across metabolic pathways. Upregulation of glutathione metabolism genes (e.g., chac2, LOC118317139) and amino acid metabolism genes (e.g., cad, aasdhppt) likely reflects cellular responses to oxidative stress induced by the quinoa-based diet. Enhanced expression of glutathione-related genes is associated with antioxidant defenses; for instance, in copper-induced oxidative injury models, glutamate intervention upregulates antioxidant enzyme genes (e.g., glutathione-associated enzymes) via the nuclear factor erythroid 2-related factor 2 signaling pathway, promoting glutathione synthesis to protect intestinal cells [62]. Additionally, upregulated amino acid metabolism genes may improve intestinal integrity and function [56]. Concurrently, upregulated expression of the nucleotide metabolism gene cdab potentially synergizes with energy metabolism and immune responses to enhance cellular adaptability [63]. Downregulation of xenobiotic metabolism genes (e.g., CYP450 family genes) in turbot aligns with observations in zebrafish (Danio rerio), indicating compromised detoxification capacity in the intestine of juvenile turbot fed high-quinoa diets [64].

The intestine of fish in the Q40 group also exhibited differential expression of genes in lipid digestion and absorption pathways. Upregulation of LOC118284943 enhances lipid digestion by hydrolyzing triglycerides and cholesteryl esters, indicating adaptive improvement in lipid digestion efficiency [65]. Upregulation of LOC118312226, a key protein in cholesterol reverse transport that mobilizes cellular cholesterol to apolipoproteins, suggests that turbot maintain intestinal lipid homeostasis by augmenting cholesterol efflux [66]. Conversely, downregulation of acat2, the rate-limiting enzyme for cholesteryl ester synthesis, likely suppresses intestinal cholesteryl ester accumulation [67]. Reduced pla2g3 expression is directly associated with lipid absorption efficiency; thus, the decrease may impair phospholipid hydrolysis and inflammatory responses, resulting in diminished lipid digestion products [68].

Differential expression of genes in protein digestion and absorption pathways was also detected in the intestine of turbot in the Q40 group. Solute carriers (SLCs) are transmembrane proteins, and solute carrier families are very much conserved in human and fish species [69]. Among these, upregulation of slc38a2 typically correlates with metabolic adaptation to support cellular growth and energy demands [70], whereas downregulation of slc15 may indicate nutrient deficiency or reduced absorption efficiency [71]. Reduced carboxypeptidase expression likely impairs intestinal protein hydrolytic capacity [72], which agrees with the decrease in muscle crude protein observed in our study. Suppressed col11a2 expression may reflect altered intestinal mucosal barrier function [73], as this gene maintains extracellular matrix structure. This suppression is consistent with histologically observed gut morphological changes.

In the intestinal cellular senescence pathway of turbot fed QGM replacement diets, downregulation of TGFβ may trigger DNA damage responses, thereby activating p21 expression. p21 plays a pivotal role in senescence by inhibiting the activity of the CDK2/CycE complex and further suppressing CycA activity, collectively causing downregulation of E2F transcription factors. This blocks the G1/S phase transition, which is consistent with established cell cycle regulation and senescence mechanisms [74], and is potentially linked to intestinal adaptation or immune responses. While aberrant NFAT signaling activation in senescent intestinal epithelia typically promotes pro-inflammatory cytokine release (e.g., il-6 and il-8) [75], quinoa-based diets did not induce significant differential expression of pro-inflammatory cytokine genes (il-1β, il-8, and tnf-α in corn gluten meal [7]; il-1β, il-8, il-17a/f, il-22, and tnf-α in SBM [5]). Calpain-1 downregulation may attenuate pro-apoptotic function and indirectly extend senescent cell survival [75]. These findings provide molecular insights into the activation of senescence-related pathways, laying a foundation for future functional investigations to further substantiate these regulatory mechanisms.

Annotation of shared DEGs revealed enrichment in cytoskeleton-related genes, which could potentially increase intestinal permeability and compromise barrier integrity [76]. Both shared and unique gene annotations highlighted involvement in translation, ribosome biogenesis, and post-translational modification, indicating distinct impacts of quinoa replacement ratios on turbot intestinal health compared to conventional plant proteins. The underlying mechanisms responsible for the effect require further investigation.

5. Conclusions

Our results demonstrate that replacement of FM with QGM affects turbot growth across multiple dimensions. As the replacement ratio increased from 10% to 40%, juvenile fish exhibited progressive declines in final BW, WGR, FR, SGR, and CF. At replacement ratios ≤ 20%, growth inhibition occurred indirectly via reduced feed intake without significantly affecting feed efficiency (p > 0.05). Conversely, ratios ≥ 30% significantly increased the FCR (p < 0.05). QGM uniquely modulated muscle composition by significantly reducing crude lipid content at all replacement ratios tested. Intestinal transcriptomics revealed that differential gene expression was primarily enriched in ribosome pathways, biosynthesis, metabolism, and lipid/protein digestion–absorption pathways. Critically, although 40% replacement induced cell cycle arrest by upregulating p21 expression (affecting FOXO and mTOR signaling), no significant differential expression of pro-inflammatory cytokine genes (il-1β, il-8, and tnf-α) was detected. This finding highlights the distinct regulatory mechanisms of QGM in the intestine. The differential impacts of quinoa versus conventional plant proteins on intestinal health provide crucial insights for developing novel plant protein sources and optimizing aquafeed formulations.

Abbreviations

The following abbreviations are used in this manuscript:

QGM	Quinoa Germ Meal
FM	Fishmeal
SR	Survival Rate
WGR	Weight Gain Rate
FR	Feeding Rate
SGR	Specific Growth Rate
FCR	Feed Conversion Ratio
CF	Condition Factor

Supplementary Materials

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/biology15040304/s1: Table S1: Amino acid composition (% dry matter basis) of quinoa germ meal and fish meal. Table S2: Saponin content (mg/g dry matter basis) of QGM and diet. Table S3: Growth performance. Table S4: Biochemical composition of muscle. Table S5: Intestinal histology statistics.

Author Contributions

J.Y. participated in the methodology, writing—original draft, data curation, software, formal analysis. Z.S. participated in the data curation, software, writing—review and editing. C.M. participated in the software, formal analysis. X.W. participated in the software, formal analysis, writing—review and editing. Z.L. participated in the investigation, methodology, validation. Z.H. participated in the data curation. Y.G. participated in the data curation. Y.L. participated in the data curation, formal analysis. Y.W. participated in the data curation. A.M. participated in the funding acquisition, supervision, project administration, resources. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

The animal study protocol was approved by the Institutional Animal Care and Use Committee of the Yellow Sea Fisheries Research Institute (Qingdao, China) (protocol code 2025081 and date of approval 16 June 2025) for studies involving animals.

Informed Consent Statement

Not applicable.

Data Availability Statement

The RNA-seq data generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under the BioProject accession number PRJNA1378117.

Conflicts of Interest

The authors declare no conflicts of interest.

Funding Statement

This work was supported by the National Key R&D Program of China (grant numbers 2022YFD2400401), the Earmarked Fund for the Modern Agroindustry Technology Research System (grant numbers CARS-47-G01), the Shandong Key R&D Program (Competitive Innovation Platform) (grant numbers 2024 CXPT071-2), the Central Public-Interest Scientific Institution Basal Research Fund (grant numbers 2023TD26), and the National Key R&D Program of China (grant numbers 2022YFE0203900).