Diet-Induced Modulation of Gut Microbiota Affects Meat Quality in Grass Carp (Ctenopharyngodon idellus)

Abstract

Understanding meat quality and gut microbiota hold great potential for promoting healthy and sustainable fish production, while also contributing to fisheries management and conservation. However, differences in meat quality and the diversity, structure, and function of gut microbiota among fish across different feeding regimes remain poorly understood. This study compared meat quality and gut microbiota between grass carp (gc) and crisped grass carp (cgc) to support sustainable aquaculture and provide more tasted fish meat. Cgc exhibited higher levels of free amino acids, fatty acids, and collagen, whereas gc had greater concentrations of hydrolyzed amino acids, nucleotides, and antioxidant indices. Fatty acid composition differed significantly between the two. Gut microbiota diversity was higher in cgc, with Proteobacteria, Fusobacteriota, and Firmicutes being dominant, while gc was dominated by Proteobacteria, Bacteroidota, and Firmicutes. The microbial community structures differed significantly. Functional predictions showed 1612 COG and 2032 KEGG pathways varied between groups. Significant correlations were found between microbial abundance and meat quality traits, including fatty acids, hydrolyzed amino acids, and collagen. These findings offered valuable insights for enhancing fish nutrition, optimizing feed formulations, and improving aquaculture practices.

1. Introduction

Fish play an essential role in human nutrition, serving as a rich source of high-quality protein and bioactive fatty acids. Among aquatic species, grass carp (Ctenopharyngodon idellus) is one of the most widely cultivated freshwater fish in China due to its herbivorous diet, rapid growth, and high yield [1]. In 2023, China produced 5.94 million tons of grass carp, making it the most extensively farmed fish species globally. Its favourable traits, such as high nutritional value and palatability, have driven its popularity among consumers [2]. However, the rapid intensification of aquaculture has led to a noticeable decline in meat quality, compromising both consumer satisfaction and market value [3,4].

Meat quality in fish is a multifactorial trait influenced by amino acid composition, fatty acid profile, flavour compounds, and textural properties [5,6]. Consumer preferences are increasingly centred on nutritional content, freshness, and sensory attributes [7]. One notable strategy to enhance meat quality in grass carp is the use of a specialized diet, particularly feeding broad beans, to produce what is referred to as “crisped grass carp” (cgc), which exhibits firmer texture and improved nutritional features [3]. Despite some reports comparing meat quality traits between common grass carp (gc) and cgc, most have focused on isolated factors without examining integrative mechanisms.

Recent studies have highlighted the significant role of the gut microbiota in fish health, growth, immunity, and nutrient metabolism [8,9]. The gut microbial community composition is influenced by host species, developmental stage, environment, and notably, diet [10,11,12]. These microbial communities enzymatically transform dietary components into bioactive compounds, which in turn can influence host metabolic pathways, including glucose regulation and lipid metabolism in skeletal muscle [13,14]. Such interactions suggest a potential link between dietary-induced shifts in gut microbiota and the biochemical properties of fish meat. In order to further investigate the relationship between the gut microbiota and meat quality, this study used both grass carp and crisped grass carp groups belong to the same species and that grass carp fed a conventional diet are used as the biological control, we examined the gut microbial diversity and the meat quality of gc and cgc: comparing the meat quality traits (nucleotides, free and hydrolysed amino acids, fatty acids, antioxidant indices, and collagen content) of gc and cgc; characterizing the diversity, structure, and functional potential of gut microbiota in gc and cgc; evaluating correlations between specific microbial taxa and key meat quality indicators. By integrating microbial ecology with nutritional profiling, this study offers new insights into how dietary interventions modulate gut microbiota and ultimately affect meat quality, providing guidance for feed formulation and sustainable aquaculture practices.

2. Materials and Methods

2.1. Ethics Statement

This study followed “Laboratory animal guideline for ethical review of animal welfare (GB/T 35892-2018)” of China. The animals were handled according to the Regulations of Meizhou and Guangdong Province on the Administration of Experimental Animals and the experimental protocols approved by the Research Ethics Panel of the Jiaying University (Approval code: JYDWLL2025-07).

2.2. Sample Collection and DNA Extraction

A total of eight samples, including four healthy grass carp (gc) and four healthy crisped grass carp (cgc) samples were purchased from the same aquaculture farm in Meizhou City. These fish kept in the same water environment. The grass carp was mainly feed with mixed food and aquatic plants used as the biological control, while crisped grass carp mainly ate broad beans in the pond. The average body length of gc and cgc individuals ranged from 50.0 to 60.0 cm and 70.0 to 85.0 cm, respectively. The average body weight of gc and cgc individuals ranged from 2000 to 4000 g and 5000 to 10,000 g, respectively.

Four gc and four cgc were anesthetized with tricaine methane sulfonate (MS-222), and samples from the foregut, midgut, and reargut were collected. A total of 27 gut samples were used in this study. The lengths of foregut, midgut and reargut were measured by a calibrated scale and AxioVision software (version 4.2) [15]. We used sterile centrifuge tubes to collect the contents of various parts of the gut and then stored at −20 °C until use. An E.Z.N.A™ Mag-Bind Soil DNA Kit (Omega, M5635-02, New York, NY, USA) was used to extract the genomic DNA from gut samples. A Qubit 4.0 (Thermo, Waltham, MA, USA) was used to measure DNA concentration.

2.3. Histological Analysis

For histological examination, we used the method as previous report [16]. Briefly, the fish muscle was fixed in Bouin’s solution. The fixed samples were dehydrated and embedded in paraffin, and then serially cut into 7 μm sections on a Leica microtome. Slides were stained with hematoxylin and eosin (H&E).

2.4. PCR Amplification and 16 S rRNA Gene Sequencing

We used the 16S rRNA gene V3-V4 region to perform PCR amplification [17]. The PCR amplification reaction mixture was prepared as follows: 2 μL of microbial DNA (10 ng/μL), 1 μL of amplicon PCR forward primer (10 μM), 1 μL of amplicon PCR reverse primer (10 μM), and 2 × Hieff^® Robust PCR Master Mix (Yeasen, 10105ES03, China), bringing the total reaction volume to 30 μL. The plate was sealed and PCR amplification was performed using a thermal cycler (Applied Biosystems 9700, Waltham, MA, USA) with the following cycling program: initial denaturation, 95 °C for 3 min (1 cycle); first stage amplification (5 cycles): denaturation at 95 °C for 30 s, annealing at 45 °C for 30 s, extension at 72 °C for 30 s; second stage amplification (20 cycles): denaturation at 95 °C for 30 s, annealing at 55 °C for 30 s, extension at 72 °C for 30 s; final extension: 72 °C for 5 min (1 cycle).

The PCR products were checked using electrophoresis in 2% (w/v) agarose gels in TBE buffer (Tris, boric acid, EDTA) stained with ethidium bromide (EB) and visualized under UV light. The samples of PCR amplification product were sent to Sangon BioTech Shanghai, China, for sequencing using the Illumina MiSeq system (Illumina MiSeq, San Diego, CA, USA), according to the manufacturer’s instructions. PEAR [18] was used to assemble the two short Illumina readings after sequencing. USEARCH was used to analyze the effective tags (OTUs (operational taxonomic units) of similarity ≥ 97%) [19,20]. The RDP Database and UNITE Database were used to classify the gut microbiota OTU representative sequences.

2.5. Measurement of Collagen Content

The collagen content was measured and calculated according to a previous report [21]. Briefly, collagen content in tissue samples was measured using Diagnostic Reagent Kits (Art. No. A030-3 and A030-2; Nanjing Jiancheng Bioengineering Institute, Nanjing, China), following the manufacturer’s instructions with slight modifications. Approximately 30−100 mg of tissue was hydrolyzed in 1 mL of 6 M hydrochloric acid at 95 °C for 5 h. After hydrolysis, the pH of the solution was adjusted using reagents A and B, as per the kit instructions. The samples were then diluted, mixed, and filtered through a 0.20 μm membrane filter. Aliquots of 1 mL from each diluted sample, standard hydroxyproline solution (5 μg/mL), or ultra-pure water (blank) were mixed with 500 μL of reagent A and incubated at room temperature for 10 min. Subsequently, 500 μL of reagent B was added, and the mixture was further incubated for 5 min. Then, 500 μL of reagent C was added, and the final mixture was heated at 60 °C for 20 min. After cooling, the samples were centrifuged at 3500 rpm for 10 min at room temperature. The absorbance was measured at 550 nm using an Infinite^® Pro 200 multifunctional microplate reader (Tecan, Männedorf, Switzerland). Hydroxyproline (Hyp) concentration was determined using a standard curve generated from the known Hyp standard. To calculate collagen content, it was assumed that Hyp constitutes 12.5% of collagen in connective tissue. Therefore, the total collagen content, comprising both soluble and insoluble fractions, was calculated by multiplying the measured Hyp concentration by a factor corresponding to this percentage and expressed as mg/g protein.

2.6. Measurement of Fatty Acid

Fatty acid contents were analyzed according to a previous report [22]. Briefly, fatty acids were analyzed using a gas chromatograph (6890, Agilent, Santa Clara, CA, USA) with a flame ionization detector, following a previous method. Each 10.0 g sample was mixed with 2.0 mL undecanoic acid ester solution and hydrochloric acid, then heated at 70 °C for 40 min. After adding 10 mL of 95% ethanol, the mixture was extracted with petroleum ether. The extract was then saponified and methylated to form fatty acid methyl esters (FAMEs). FAMEs were separated using a TR-FAME GC column (100 m × 0.25 mm × 0.20 μm, Thermo Scientific, Waltham, MA, USA). Fatty acids were quantified using internal standard peak areas and identified using individual and mixed FAME standards (CRM47885, Sigma, Ottobrunn, Germany).

2.7. Measurement of Hydrolyzed and Free Amino Acids

For hydrolyzed amino acid analysis, 0.1 g of each sample was hydrolyzed in 10 mL of 6 mol/L HCl in sealed tubes at 110 °C for 24 h. After cooling, the hydrolysate was diluted to 50 mL with ultrapure water. A 1 mL aliquot was evaporated at 60 °C to dryness, reconstituted in 2 mL of 0.02 mol/L HCl, and filtered through a 0.45 μm membrane. Amino acid content was determined using a Hitachi L-8900 Automatic Amino Acid Analyzer with a sodium citrate ion exchange column and post-column ninhydrin derivatization, following the manufacturer’s protocol [23]. For free amino acid analysis, the methods were used according to a previous report [22]. Briefly, 0.4 mL of plasma or 0.2 g of freeze-dried tissue was mixed with 10% sulfosalicylic acid (1.2 mL for plasma, 1.5 mL for tissue), vortexed or homogenized for 15 min, and centrifuged at 12,000× g for 15 min at 4 °C. The supernatants were filtered through a 0.22 μm membrane and analyzed using the same amino acid analyzer. Results were expressed as nmol/mL for plasma and g/kg dry tissue for samples. Essential amino acids (EAA) included lysine, methionine, tryptophan, threonine, arginine, histidine, leucine, isoleucine, phenylalanine, and valine; branched-chain amino acids (BCAA) included leucine, isoleucine, and valine; and non-essential amino acids (NEAA) were calculated as total amino acids (TAA) minus EAA.

2.8. Measurement of Nucleotides

The nucleotides including adenosine triphosphate (ATP), adenosine diphosphate (ADP), adenosine monophosphate (AMP), inosine monophosphate (IMP), inosine (HxR), and hypoxanthine (Hx) were measured using the following methods. The analytical procedure employed an AB SCIEX 6500 Triple Quadrupole LC-MS/MS system coupled with a Shimadzu HPLC instrument (Kyoto, Japan) and a Waters C18 column. Reagents such as formic acid, acetonitrile, and methanol were sourced from Thermo (Waltham, MA, USA), while standards are obtained from Macklin. For analysis, standard solutions were prepared at various concentrations ranging from 5 to 5000 ng/mL. Tissue samples (100 mg) were homogenized in a 2 mL PE centrifuge tube with a 3 mm grinding steel bead at −20 °C, following a cycle of 60 s on and 30 s off for six iterations at 70 HZ, then centrifuged at 2000 rpm for 3 min. The resulting homogenate was extracted with acitretonile, vortexed, centrifuged, and the supernatant dried under nitrogen before being reconstituted in a 75% methanol-water solution. Chromatographic separation was performed on a Thermo C18 column (2.1 × 100 mm) maintained at 30 °C, with a flow rate of 0.3 mL/min and a gradient elution program. The mass spectrometry conditions were optimized for detection, including settings for curtain gas, collision gas, spray voltage, temperature, ion source gases, dwell time, and voltages for DP and EP. This integrated approach enables precise quantification of the specified nucleotides in tissue samples.

2.9. Data Analysis

One-way analysis of variance (ANOVA) was used to test for significant differences in meat quality (nucleotides, free amino acids, hydrolyzed amino acids, antioxidant index, collagen) between grass carp and crisped grass carp using the R vegan package [24]. Principal component analysis (PCA) was used to analyze the composition of nucleotides, free amino acids, hydrolyzed amino acids, antioxidant index and collagen performed by R vegan package [24].

Chao, Simpson and Ace diversity indices were used to analyze the alpha diversity of gut microbiota performed by Mothur [25]. One-way analysis of variance (ANOVA) was used to test for significant differences of Chao, Simpson and Ace diversity indices among each sample performed by R vegan package [24]. The heatmap analysis and non-metric multidimensional scaling (NMDS) were used to evaluate community composition of the gut microbiota among samples performed by R vegan package [24]. Adonis analysis was used to determine the significance of differences in the gut microbiota among samples. A random forest model was used to assess the variable importance on gut microbiota in grass carp and crisped grass carp performed by the random Forest package [26] in R (version 4.3.2) [27]. Welch’s t-test was used to assess differences in gut microbiota between grass carp and crisped grass carp performed by R vegan package [24]. PICRUSt (version 2.5.2) [28] was used to perform the functional prediction analysis of gut microbiota. The heatmap analysis was used to analyze the abundance of KEGG and COG performed by R vegan package [24]. Welch’s t-test analysis was used to analyze the significance of differences in KEGG and COG performed by R vegan package [24]. The heatmap analysis was also used to the correlation between abundance of gut microbiota community and content of meat quality (content of nucleotides, free amino acids, hydrolyzed amino acids, fatty acid, antioxidant index, collagen) performed by R vegan package [24].

3. Results and Discussion

3.1. Histological Analysis of Muscle: gc vs. cgc

Figure 1 showed the gc and cgc samples we collected. Aside from body size, there appears to be no difference between the two types of fish (Figure 1). However, the histological analysis revealed significant differences in muscle fiber structure between gc and cgc (Figure 2). Cgc muscle exhibited more tightly packed fibers with smaller diameters, indicative of hyperplasia, and reduced connective tissue between fibers, all contributing to greater muscle density. In contrast, gc muscle fibers were larger, more loosely arranged, and separated by wider inter-fiber spaces, suggesting higher moisture or fat content and a less compact structure. These structural distinctions implied that cgc possesses a firmer texture and potentially higher protein content, qualities that enhanced its desirability in premium markets, whereas gc’s softer texture reflected lower structural integrity. Overall, the finer and denser muscle arrangement in cgc aligned with expected improvements achieved through specific dietary or environmental conditioning.

3.2. Composition of Meat Nutrition in gc and cgc

Meat quality was a key concern for consumers, as highlighted by recent studies [2,29]. Several metabolites, such as amino acids, nucleotides, and fatty acids, were closely associated with meat flavor [25]. In order to compare the difference of the meat composition between the gc and cgc, we then examined the fatty acid, collagen, hydrolytic and free amino acid, and nucleotides in the two types of fish. As shown in Figure 3, in both cgc and gc, lysine (Lys) and histidine (His) were the most abundant among the free amino acids, with cgc exhibiting higher overall levels of free amino acids compared with gc (Figure 3A). The total free amino acids/NEAA ratio is 6.59 and 4.87 in cgc and gc, respectively. Similarly, among hydrolysed amino acids, lysine (Lys), glutamic acid (Glu), and aspartic acid (Asp) were predominant in both groups. However, gc had a higher total content of hydrolysed amino acids (23.8%) than cgc (16.75%) (Figure 3B). PCA showed that the composition of free amino acids and hydrolysed amino acids was different between cgc and gc, and explained 69.9% and 99.2% of the variation in the first two axes (Figure 4A,B). Among these, amino acids serve as critical quality indicators for evaluating the taste and flavor of meat. For example, one study reported that the total content of free amino acids was 92.3% higher in crisp grass carp compared with regular grass carp [30,31]. Our findings support this, revealing that cgc contains significantly higher levels of free amino acids, suggesting a richer and more desirable flavor profile. Conversely, gc exhibited higher levels of hydrolyzed amino acids, indicating a more balanced nutritional composition and growth.

In terms of fatty acid composition, unsaturated fatty acids were more abundant than saturated fatty acids in both cgc and gc. Within both groups, monounsaturated fatty acids (MUFA) were present in higher amounts than polyunsaturated fatty acids (PUFA; Figure 3C). Fatty acids, another group of important flavor precursors [32], were also found in greater quantities in cgc (54.7 mg/g) than in gc (0.59 mg/g); particularly, the MUFA/PUFA ratio was 1.9 and 1.07 in cgc and gc, respectively. Moreover, the fatty acid composition exhibited a similar pattern, accounting for 98.2% of the variation along the first two axes (Figure 4C). Significant differences in fatty acid profiles observed between cgc and gc may further contribute to the richer flavor and enhanced taste of cgc.

Regarding nucleotide composition, adenosine diphosphate (ADP), inosine monophosphate (IMP), and hypoxanthine riboside (HxR) were the most abundant in both groups (Figure 3D). The composition of nucleotides, showed a similar pattern, and expressed 64.75% of the variation in the first two axes (Figure 4D). Nucleotides, which influence meat quality and contribute to umami flavor, were more abundant in grass carp than in crisp grass carp [32,33], suggesting that grass carp possessed a stronger umami profile.

For antioxidant indices, superoxide dismutase (SOD) and catalase (CAT) showed the highest values in both groups, with gc exhibiting stronger antioxidant capacity overall compared to cgc (Figure 3E). The composition of antioxidant index showed a similar pattern, and expressed 81.48% of the variation in the first two axes, respectively (Figure 4E). Compounds with antioxidant properties were associated with more stable meat coloration. Our study found higher antioxidant indices in grass carp, indicating better colour stability compared to crisp grass carp.

The collagen content exceeded that of hydroxyproline in both cgc and gc. Total collagen levels were higher in cgc than in gc (Figure 3F). The composition of collagen showed a similar pattern, and expressed 100% of the variation in the first two axes (Figure 4F). Collagen content played a crucial role in maintaining meat integrity and muscle cohesion. The structure of muscle connective tissue and collagen levels largely determined the texture and hardness of fish meat [34]. Prior studies reported that collagen content in cgc was 36.7% higher than in gc [3,35]. Our results were consistent with this, indicating that the meat of cgc was firmer and harder to boil, an observation supported by previous reports [3].

3.3. Diversity and Structure of Gut Microbiota in gc and cgc

In addition to analyzing meat composition, we further examined the gut microbiota to explore potential correlations. The operational taxonomic units (OTUs), along with the Chao, ACE, and Simpson diversity indices, were all higher in cgc compared to gc, indicating a richer and more diverse microbial community in cgc (

Table 1. Number of sequences, OTUs, Coverage, and Chao, Ace, Simpson diversity indices of gut microbiota in cgc and gc.

Sample	Number	OTUs	Chao	Ace	Simpson	Coverage
cgc1f	64,896	324	362.294	372.205	0.188	0.999029
cgc1m	76,498	261	265.500	263.118	0.105	0.999882
cgc1r	79,899	360	361.250	361.024	0.111	0.999937
cgc2f	69,137	204	297.840	359.112	0.254	0.999002
cgc2m	76,259	326	387.961	400.768	0.305	0.998951
cgc2r	66,932	323	403.500	380.133	0.149	0.998954
cgc3f	64,453	326	383.000	395.766	0.076	0.998821
cgc3m	78,215	305	338.957	348.844	0.176	0.999271
cgc3r	83,685	108	108.500	108.767	0.221	0.999976
cgc4f	82,174	367	458.143	446.404	0.424	0.998929
cgc4m	83,223	170	225.039	227.930	0.364	0.999351
cgc4r	84,746	254	271.500	275.126	0.179	0.999575
gc1f	71,063	300	354.238	366.374	0.095	0.999043
gc1m	61,347	251	355.516	414.931	0.093	0.998680
gc1r	57,276	266	300.460	304.343	0.117	0.999110
gc2f	67,985	186	247.107	302.310	0.246	0.999132
gc2m	63,382	288	331.677	328.831	0.128	0.999132
gc2r	74,359	169	176.857	174.433	0.065	0.999852
gc3f	70,693	256	296.529	299.600	0.209	0.999250
gc3m	57,775	287	308.484	307.717	0.135	0.999360
gc3r	54,638	212	263.042	256.389	0.175	0.999085
gc4f	78,041	232	295.370	341.905	0.130	0.999244
gc4m	58,627	219	262.677	265.145	0.238	0.999062
gc4r	72,342	158	162.231	163.024	0.166	0.999848

). At the phylum level, Proteobacteria, Bacteroidota, and Firmicutes were the dominant taxa in gc, whereas Proteobacteria, Fusobacteriota, and Firmicutes were predominant in cgc (Figure 5). Gut microbiota played a vital role in nutrient metabolism and overall host health, making it essential to understand their diversity, structure, and function [36,37,38]. Previous studies using culture-dependent analysis or 16S rRNA gene sequencing identified Proteobacteria, Bacteroidota, and Firmicutes as the dominant gut microbiota in grass carp [17,39,40,41]. Other studies observed slight variations, reporting Cyanobacteria [42] and Actinobacteria [39] among the dominant groups. Our findings confirmed that Proteobacteria, Bacteroidota, and Firmicutes dominated in grass carp, while Proteobacteria, Fusobacteriota, and Firmicutes were predominant in crisp grass carp.

Non-metric multidimensional scaling (NMDS) analysis revealed a significant difference in the composition of gut microbiota between cgc and gc (Adonis: R² = 0.218, p = 0.006) (Figure 6A). This distinction was further supported by heatmap analysis, which mirrored the clustering patterns observed in the NMDS plot (Figure 6B). The Random Forest model identified Bacteroidota, Planctomycetota, Fusobacteriota, and Cyanobacteria as key taxa contributing to the differentiation of gut microbial composition between CGC and GC (Figure 7A). Moreover, Welch’s t-test analysis showed that Fusobacteriota and Bacteroidota had statistically significant effects on the overall microbial community structure in both groups (Figure 7B). Diet has been shown to significantly influence gut microbiota composition in fish [43,44,45]. In this study, substantial differences in microbiota composition were observed between cgc and gc, with Fusobacteriota and Bacteroidota identified as key contributing taxa. Bacteroidota are commonly found in various organisms and are involved in fermentation and the breakdown of plant-derived oligosaccharides [46,47]. These microbes were mutualistic, processing complex plant glycans into simpler molecules, and they aided in nutrient absorption, which was especially vital for herbivorous fish like gc [48]. Fusobacteriota, on the other hand, were linked to fermentative metabolism involving carbohydrates and vitamins [49,50]. Since cgc are primarily raised on a broad bean-based diet, this feeding regimen is likely to have shaped both their gut microbiota and the nutritional characteristics of the fish meat [3].

3.4. Functional and Metabolic Divergence in Gut Microbiota of gc and cgc

A total of 4178 functional COGs and 6968 functional KEGG orthologs were predicted in both cgc and gc in this study (Figure 8). Among these, K02014, K03088, K01992 and K07090 exhibited higher abundances compared to other KEGG orthologs, while COG0642, COG0745, COG1028, COG0438, COG1309, COG0583, COG2814, COG2207 and COG1538 showed higher abundances relative to other COGs. Totally, 1612 functional COGs and 2032 functional KEGG orthologs differed significantly between cgc and gc (Welch’s t-test, p < 0.05).

The functional profiles of gut microbiota serve as valuable indicators for predicting host metabolism and disease susceptibility [51,52,53,54]. In fish, gut metabolism is influenced by species, developmental stage, diet, and environmental factors [10,11,12,55]. Diet-specific effects on gut microbiota function have also been confirmed [13,56,57]. Using COG and KEGG pathway analyses, our study predicted that the gut microbiota was involved in nutrient metabolism (amino acids, carbohydrates, nucleotides, coenzymes, and inorganic ions), genetic information processing, energy production, and organismal systems.

3.5. Correlation Between Gut Microbiota Abundance and Meat Quality Composition

Significant correlations were observed between the abundance of specific microbial taxa and various meat quality components. The abundance of Fusobacteriota was positively correlated with the levels of leucine (Leu), valine (Val), tyrosine (Tyr), glycine (Gly), and glutamic acid (Glu) (Figure 9A). A significant correlation was also found between Verrucomicrobiota and proline (Pro), and between Firmicutes and glycine (Gly).

Moreover, Cyanobacteria and Bacteroidota showed significant associations with asparagine (AspNH₂), while Spirochaetota was correlated with arginine (Arg), and Actinobacteriota with aspartic acid (Asp). Proteobacteria abundance was correlated with levels of proline (Pro), tyrosine (Tyr), glycine (Gly), and glutamic acid (Glu), and Campylobacterota was significantly correlated with phenylalanine (Phe) and alanine (Ala). Notably, Bacteroidota was significantly associated with the full profile of hydrolyzed amino acids (Figure 9B). Further associations were found between Cyanobacteria, Planctomycetota, Spirochaetota, and levels of proline (Pro), arginine (Arg), glycine (Gly), and glutamic acid (Glu). In terms of fatty acid composition, Bacteroidota showed significant correlations with a wide range of fatty acids, including C12:0, C14:0, C15:0, C16:1, C20:1, C18:2, C18:3, C20:2, C20:3, C20:4, C20:5, and C22:6 (Figure 9C).

Similarly, Cyanobacteria and Planctomycetota were associated with C12:0, C14:0, C15:0, C16:1, C20:1, C18:2, C18:3, C20:2, C20:3, C16:0, C20:5, C18:0, and C18:1. Spirochaetota was significantly correlated with C20:4 and C22:6. Regarding antioxidant indices, Verrucomicrobiota was positively correlated with superoxide dismutase (SOD) levels (Figure 9D), while Cyanobacteria and Planctomycetota were associated with total antioxidant capacity (T-AOC). Spirochaetota was also significantly correlated with AMP and ATP levels (Figure 9E), and Actinobacteriota was linked to hypoxanthine (Hx). Lastly, a significant correlation was found between Actinobacteriota and total collagen content (Figure 9F).

We observed that the microbial community composition varied significantly across feeding stages, reflecting functional diversity and enhanced adaptability due to high metabolic potential. Notably, grass carp had significantly higher Bacteroidota abundance, while crisp grass carp had more Fusobacteriota. These phyla contain numerous lineage-specific genes essential for metabolism and physiology [48,50,58,59]. Our findings indicate a strong association between gut microbiota composition and the differential profiles of fatty acids, hydrolyzed amino acids, and collagen content in the two fish types.

4. Conclusions

This study revealed meat quality components, diversity, structure and function of gut microbiota and the correlation between gut microbiota abundance and meat quality components in grass carp and crisped grass carp. The results showed that the meat composition was different across the two fish types, and there was significant difference in the composition of fatty acids between the two fish types. Fusobacteriota and Bacteroidota significantly affected the composition of gut microbiota. 1612 functional COG and 2032 functional KEGG had significant differences in crisped grass carp and grass carp. The contents of fatty acids, hydrolyzed amino acids and collagen were significant correlation with the changes of gut microbiota abundance. These results provided some new insights into developing commercial fish feeds and improving existed aquaculture strategies. Additionally, we acknowledge the limitations associated with the low number of analyzed specimens and biological samples. Future studies should increase sample size to help define the scope and applicability of the conclusions.