Integrated Quality Assessment and Metabolomic Analysis of Dezhou Donkey Meat During Extended Chilled Storage

TL;DR

This study examines how Dezhou donkey meat quality and chemical composition change during chilled storage, revealing factors affecting tenderness and color loss.

Contribution

The study provides the first comprehensive metabolomic analysis of donkey meat during storage, identifying biomarkers for freshness and quality.

Findings

Extended storage improved tenderness but reduced color stability after 14 days.

Metabolomics revealed accumulation of lysophospholipids and amino acids linked to membrane breakdown and proteolysis.

Depletion of antioxidants like acetyl-L-carnitine and cysteinylglycine correlated with color deterioration.

Abstract

Donkey meat is valued for its nutritional quality, yet maintaining its freshness during chilled storage is challenging, particularly with respect to color stability, the underlying mechanisms of which remain unclear. To address this, the present study investigated changes in meat quality and metabolite profiles of Dezhou donkey meat stored at 0–4 °C for 21 days. While extended storage significantly improved tenderness due to muscle fiber degradation, it markedly reduced color stability, especially after 14 days. Using untargeted LC-MS/MS metabolomics, we observed an accumulation of compounds associated with membrane breakdown, proteolysis, and energy metabolism. Concurrently, key antioxidants were progressively depleted, which correlated with color deterioration. Pathway analysis further highlighted significant alterations in glycerophospholipid, amino acid, and glutathione metabolism. These findings provide a scientific basis for developing improved preservation strategies aimed at maintaining the color and overall quality of donkey meat.

This study aimed to elucidate the dynamic changes in meat quality attributes and metabolomic profiles of Dezhou donkey longissimus lumborum (LL) during extended chilled storage. Donkey longissimus lumborum (LL) muscles (n = 4) were vacuum-packaged 24 h post mortem and stored at 0–4 °C for 0, 7, 14, or 21 days, followed by a 5-day aerobic display. Meat quality parameters and microstructural characteristics were evaluated, and untargeted metabolomics was performed using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Extended storage significantly improved tenderness (shear force decreased from 76.40 to 27.84 N; p < 0.05), concurrent with progressive muscle fiber degradation. However, color stability deteriorated markedly, with accelerated declines in redness observed beyond 14 days. Metabolomics analysis demonstrated that storage resulted in a substantial accumulation of lysophospholipids, free amino acids, and nucleotide degradation products, indicative of membrane deterioration, proteolysis, and ATP catabolism. Notably, acetyl-L-carnitine, cysteinylglycine (Cys-Gly), and nicotinamide exhibited progressive depletion, correlating with diminished antioxidant capacity and color deterioration. KEGG pathway enrichment revealed significant alterations in glycerophospholipid, amino acid, and glutathione metabolism. This study provides the first comprehensive metabolomic characterization of donkey meat during chilled storage, identifying potential biomarkers for freshness assessment and offering a scientific foundation for developing targeted preservation interventions.

Figures

Full text

1. Introduction

Donkey meat is acknowledged for its advantageous nutritional composition, which is marked by high protein content (23.56 g/100 g), low fat content (1.77 g/100 g), and low cholesterol content (66.70 g/100 g) [1]. Its polyunsaturated fatty acids (PUFAs) to saturated fatty acid ratio (0.73) is higher than that of beef (0.15), pork (0.29), and sheep (0.09) [1]. Additionally, donkey meat serves as a valuable source of essential amino acids and minerals, including iron, zinc, and potassium [2]. Donkey meat consumption in China is shifting from traditional ready-to-eat products toward chilled fresh formats [3]. This trend underscores the importance of understanding quality changes in donkey meat during chilled storage in order to preserve its nutritional and sensory attributes.

Meat color is a primary visual quality attribute that critically influences consumer purchase decisions, as it is often associated with freshness [4]. Color stability is largely determined by the concentration and redox state of myoglobin, which varies among animal species and breeds [5]. Previous findings demonstrated that donkey meat possesses higher myoglobin content (8.3–8.4 mg/g in donkey breeds) [6] compared to meats from other common food animals [7]. Additionally, donkey meat contains a greater proportion of PUFAs than conventional livestock and poultry species [1]. Given that lipid oxidation products, especially those originating from unsaturated fatty acids, can accelerate myoglobin oxidation [8], the high PUFA content in donkey meat suggests heightened susceptibility to lipid oxidation and, consequently, accelerated decline in color stability. Despite these insights, the dynamic changes and underlying mechanisms governing color stability in donkey meat during chilled storage remain poorly understood. This knowledge gap limits the development of targeted preservation strategies necessary to maintain color quality.

Metabolomics, a high-throughput analytical technique for comprehensively profiling small-molecule metabolites (molecular weight < 1500 Da) in biological systems, has emerged as a powerful tool for elucidating the mechanisms underlying meat quality changes [9]. Specifically, metabolomics enables systematic characterization of metabolic profiles and provides valuable insights into dynamic changes in meat quality during postmortem aging. Previous studies employing metabolomics approaches have evaluated various intrinsic and extrinsic factors affecting meat color, including muscle type [10], postmortem pH [11], packaging methods [12], and aging conditions [13]. Several metabolic pathways play a critical role in influencing meat color stability during postmortem aging, among which glycolysis and the tricarboxylic acid cycle are widely acknowledged as central biochemical processes [14]. Mitacek et al. (2019) [15] reported that metabolites such as fumaric acid, creatinine, fructose, and nicotinamide adenine dinucleotide (NADH) significantly decreased (p < 0.05) during beef aging, suggesting that the decline in color stability results from increased mitochondrial damage, depletion of metabolites capable of regenerating NADH, and elevated oxidative stress. Recent research evaluated the effects of various aging methods on beef quality attributes and metabolic profile changes during the aging process. These findings indicated that the progressive decline in glycerophospholipid levels with prolonged aging suggests substantial degradation of biological membranes in beef muscle tissue, which may influence meat color development and stability [16]. Overall, understanding metabolite changes during storage may provide enhanced insight into the biological processes that affect meat color.

However, metabolomic studies on donkey meat remain limited. Therefore, the objective of this study was to investigate meat quality characteristics and metabolic changes in donkey longissimus lumborum over a 21-day chilled storage period. The findings of this study aim to provide a scientific basis for developing improved preservation strategies to maintain the quality of donkey meat.

2. Materials and Methods

2.1. Experimental Design

Four male Dezhou donkeys (~24 months of age; body weight 250 ± 20 kg) were slaughtered at Dong-E-E-Jiao Co., Ltd. (Liaocheng, China) in accordance with the Agricultural Industry Standard of the People’s Republic of China [17] (NY/T 3743-2020). Prior to slaughter, all donkeys were reared under uniform dietary and management conditions. Within 24 h post-slaughter, the longissimus lumborum (LL) muscles were excised from the left side of each carcass and carefully trimmed of subcutaneous fat. Each LL muscle was subsequently sectioned into 5 cm thick portions, vacuum-packaged individually, and stored at 0–4 °C for 0 (2 d post-mortem), 7, 14, or 21 days. After completing their storage time, the vacuum packaging was removed and each portion was portioned into two 2.5 cm thick steaks. One steak were used for analysis of pH, purge loss, cooking loss, shear force, microstructure, and lipid oxidation. For metabolomic profiling, additional subsamples were collected, immediately snap-frozen in liquid nitrogen, transferred to cryogenic tubes, and stored at −80 °C until analysis. The other steak were placed on commercial plastic trays lined with absorbent pads and overwrapped with polyvinyl chloride film (material specifications as described previously [18]), and oriented with the freshly cut surface facing upward to simulate retail display. These samples were then aerobically displayed at 0–4 °C in the dark for 0, 1, 2, 3, 4, or 5 days. Instrumental color measurements were conducted at each designated display time point.

2.2. Meat Quality Measurements

2.2.1. pH Determination

The pH of donkey meat samples was measured according to the method reported previously [19]. A calibrated SevenGo pH probe (S2-Food Kit, Mettler-Toledo, Zurich, Switzerland) was employed, with calibration performed using standard buffers at pH 4.00 and 7.00. Measurements were obtained by inserting the probe directly into the meat; three replicate readings were recorded at distinct locations within each sample, and the mean value was calculated.

2.2.2. Purge Loss and Cooking Loss

The purge loss of donkey meat samples was measured according to the method reported previously [20]. Prior to storage, samples were weighed to record the initial weight (W_1_). During storage, samples were removed and reweighed on days 7, 14, and 21 to obtain the respective weight (W_2_). Purge loss was calculated as the percentage of weight loss using the following formula: Purge loss (%) = [(W_1_ − W_2_)/W_1_] × 100.

Cooking loss was determined according to the method as documented in previous study [20]. At each storage time (0, 7, 14, and 21 d), the samples were placed in heat-resistant bags and cooked in an 80 °C water bath to a core temperature of 70 °C, followed by overnight cooling at 0–4 °C. Cooking loss was calculated as the relative weight decrease using W_3_ (before cooking) and W_4_ (after cooking and cooling): Cooking loss (%) = [(W_3_ − W_4_)/W_3_] × 100.

2.2.3. Shear Force

Warner-Bratzler shear force (WBSF) was determined according to the method described by Hou et al. (2014) [21]. Following the cooking loss assessment, a minimum of four cylindrical cores were removed from each sample parallel to the muscle fiber orientation. Shear force was measured perpendicular to the fiber direction using a digital muscle tenderness analyzer (C-LM3B). The WBSF for each sample was calculated as the mean of the individual measurements obtained from all cores.

2.2.4. Histological Analysis of Microstructure

The microstructure of the histological analysis was examined according to Chen et al. (2022) [19] method. Muscle samples (0.4 × 0.4 × 0.8 cm) from each treatment group were fixed in 10% neutral-buffered paraformaldehyde solution for 24 h. Following routine dehydration through a graded ethanol series and paraffin embedding, the specimens were sectioned at a thickness of 10 μm. Sections were mounted on 3-aminopropyltriethoxysilane (APES)-coated slides, air-dried for 5 min, and stained with hematoxylin for 3 min. Following hematoxylin staining, slides were differentiated, rinsed, and subsequently stained with eosin for 30 s. Subsequent procedures included gradient ethanol dehydration, xylene clearing, and mounting with neutral balsam. The prepared slides were then examined and imaged under a light microscope.

2.2.5. Lipid Oxidation

Lipid oxidation was assessed based on thiobarbituric acid reactive substances (TBARS), following a procedure adapted from Chen et al. [22]. Minced meat (4 g) was homogenized under ice-cold conditions with 32 mL of 5% trichloroacetic acid (w/v) using an Ultra-Turrax T18 homogenizer (IKA, Staufen, Germany) operating at 13,000 rpm for 30 s. After filtration, 3 mL of filtrate was reacted with 0.75 mL of 60 mM TBA solution, followed by incubation at 80 °C for 90 min. The absorbance of the reaction solution was measured at 532 nm on a TU-1901 spectrophotometer (Beijing Puxi General Instrument Co., Ltd., Beijing, China) against a blank control. A calibration curve prepared with TEP was used for quantification, and results were reported as mg MDA/kg sample.

2.2.6. Instrumental Meat Color Measurement

The measurement method for meat color is referenced from Yang et al. (2024) [18]. Meat color was measured using a colorimeter (CR-400, Konica Minolta, Tokyo, Japan) calibrated with a white standard plate. After 0, 7, 14, and 21 days of storage, samples were placed on trays and allowed to bloom for 30 min at 0–4 °C. Following display periods of 1–5 days, the overwrap film was removed and surface color was immediately measured. CIE L* (lightness), a* (redness), and b* (yellowness) values were recorded at three locations per sample, and the mean values were used for statistical analysis.

Following colorimetric measurement, each sample was placed on a standardized white background inside a compact photo studio (65 cm side length; Shaoxing Ruiteng Photo Equipment Co., Ltd., Zhejiang, China). A Nikon D90 digital camera (Nikon Corporation, Tokyo, Japan) equipped with an AF-S DX Nikkor 18–105 mm lens was mounted vertically 25 cm above the sample. Illumination was provided by two 5600 K fluorescent lamps (Bull Group Co., Ltd., Zhejiang, China), and digital images were captured for visual color documentation.

2.3. Metabolomic Sample Processing and Detection

2.3.1. Metabolite Extraction

100 mg of each meat sample was homogenized in 400 μL of pre-chilled 80% methanol (4:1 methanol/water, v/v) by vortexing for 1 min, followed by incubation on ice for 5 min. The mixture was centrifuged at 4 °C for 20 min. The supernatant was diluted with purified water to a final methanol concentration of 53% (v/v) and centrifuged again under the same conditions. The resulting supernatant was transferred to an injection vial for subsequent LC-MS analysis.

2.3.2. Metabolomics Analysis

Chromatographic separation was performed using a Thermo Scientific Ultimate 3000 UHPLC system (Thermo Fisher, Bremen, Germany), equipped with a Hypersil Gold C18 column (Thermo Fisher, Waltham, MA, USA) and interfaced with a Q Exactive quadrupole-Orbitrap mass spectrometer (Thermo Fisher, Bremen, Germany). The mobile phase utilized 0.1% formic acid (Solvent A) and methanol (Solvent B). A specific gradient program was applied: initially 2% B for 1.5 min, increasing to 85% B by 3 min, followed by a hold at 100% B until 10 min. The gradient returned to 2% B at 10.1 min and equilibrated there until the run ended at 12 min. We used a 3 μL sample injection volume and a constant flow rate of 0.2 mL/min. For the electrospray ionization (ESI) source, settings included a spray voltage of 3.5 kV, an S-lens RF level of 60, and sheath/auxiliary gas flows of 35 psi and 10 L/min, respectively. Temperatures for the capillary and auxiliary gas heater were maintained at 320 °C and 350 °C.

2.4. Statistical Analyses

The effects of storage time on meat quality parameters (pH, purge loss, cooking loss, shear force, and TBARS) were evaluated by one-way analysis of variance (ANOVA) using the general linear model procedure in SPSS software (version 27.0). Two-way analysis of variance (ANOVA) of instrumental colorimetric values were carried out using a general linear model in SPSS software (version 27.0). Storage time (0, 7, 14, and 21 d), display period (0, 1, 2, 3, 4 and 5 days), and their interaction were fixed factors, and replicates (donkeys) was assigned as a random effect. Significant ANOVA results (p < 0.05) were followed by Fisher’s LSD test for multiple comparisons. Metabolites were annotated with reference to KEGG, HMDB, and LIPID MAPS. PCA and PLS-DA were conducted through the metaX software package (version 1.4.2). Metabolites were selected as differential if they exhibited an FC ≥ 2 or ≤0.5, reached a statistical significance of p < 0.05, and achieved a VIP score > 1 within the PLS-DA model. Volcano plots were generated using R software with the ggplot2 package (ggplot2 4.0.2.) to visualize metabolites based on their log_2_(fold change) and −log_10_(p-value). Functional enrichment analysis of the KEGG pathways associated with differential metabolites between day 0 and days 7, 14, and 21 samples was performed. A pathway was considered significantly enriched if the ratio of differential metabolites mapped to that pathway met the criterion of x/n > y/N and the hypergeometric test yielded a p-value < 0.05.

3. Results and Discussion

3.1. Changes in Meat Quality Attributes

The initial pH value on day 0 was 5.74, within the normal range for fresh meat. This value increased significantly (p < 0.05), peaking at 5.92 on day 14. This similarity in trends is consistent with findings in vacuum-packed beef muscles (Hughes et al., 2015) [23] and may be explained by bacterial-induced protein breakdown, as reported by Lan et al. (2016) [24]. During the period from day 14 to day 21, the pH value remained stable.

Concurrently, purge loss exhibited a progressive increase from 2.20% (day 7) to 2.29% (day 14) (p > 0.05), followed by a marked rise to 3.59% by day 21 (Table 1, p < 0.05). Cooking loss also increased slightly from 20.43% to 21.37% over the 21-day period (Table 1). These results collectively indicate a gradual decline in the water-holding capacity of donkey meat during storage, which is closely associated with protein denaturation and structural degradation of muscle fibers [25].

As shown in Figure 1A, shear force values decreased significantly (p < 0.05) with prolonged storage. A notable decline occurred by day 7 (44.00 N) compared to day 0 (76.40 N), approaching the tenderness range preferred by consumers [26], and reached a minimum by day 21 (27.84 N). In parallel, hematoxylin and eosin (H&E) staining revealed corresponding microstructural changes (Figure 1B). Fresh samples displayed tightly packed, well-defined muscle fibers with clear boundaries. After 7 days, inter-fiber spacing increased slightly while overall structure remained intact. By day 14, minor structural damage was evident, progressing to marked disintegration, including visible voids and areas of fiber lysis, by day 21. These morphological alterations are consistent with the observed reduction in shear force. Notably, TBARS values increased significantly (p < 0.05) throughout storage (Figure 1C), rising from 0.15 mg MDA/kg (day 0) to 0.28 mg MDA/kg (day 21). Despite this progressive increase, overall oxidation remained at a relatively low level (≤0.28 mg MDA/kg), likely due to the combined inhibitory effects of chilled temperature and vacuum packaging [27].

Meat color is a key sensory attribute that directly influences consumer purchasing decisions [28]. As shown in Figure 2A, the L* (lightness) value of all samples increased during chilled storage, likely due to increased free water content and enhanced light scattering within the meat matrix over time [29]. This trend is consistent with findings reported for beef under various chilling conditions [30]. During the subsequent display period, L* values generally rose initially before declining. Notably, samples stored for 0 days exhibited the highest L* values during display (days 2–5), followed by those stored for 7 days. The a* (redness) value gradually decreased in all groups throughout display (Figure 2B). Extended chilled storage accelerated this decline, with the 21-day group showing the most rapid and significant reduction, followed by the 14-day group. These results indicate that prolonged storage duration negatively impacts the color stability of donkey meat, a phenomenon that has been confirmed in meat from other species [31]. Regarding b***** (yellowness) values (Figure 2C), all groups peaked on the first display day before gradually decreasing. Samples stored for 0 and 7 days maintained higher b* values during display (days 1–5) compared to those stored for 14 and 21 days. As depicted in Figure 2D, samples stored for 0 and 7 days retained a bright red color preferred by consumers throughout the 5-day display. In contrast, slight browning was observed in 14-day samples after 5 days of display, while 21-day samples exhibited browning at an earlier display stage.

3.2. Metabolic Changes in Donkey Meat During Storage

3.2.1. Overview of Metabolomic Results

This study identified 689 metabolites using both positive and negative ion modes, with 312 detected in positive ion mode (Figure 3A) and 377 in negative ion mode (Figure 3B). The identified metabolites fell into major groups such as lipid-associated molecules; amino acid derivatives; organic acids and related compounds, oxygen-containing organics, and heterocyclic compounds; nucleosides, nucleotides, and analogues; organic nitrogen compounds; alkaloids and derivatives; organosulfur compounds; benzene derivatives; phenylpropanoids; and polyketides. PCA was conducted to assess overall differences in metabolic profiles among the four sample groups and to evaluate the clustering of quality control (QC) samples (Figure 3C,D). The PCA score plots revealed that, as storage time increased, the metabolic profiles progressively diverged from those at day 0. Notably, clear separation was observed between samples from day 0 and day 21 in both ionization modes, indicating substantial metabolic alterations over time. Moreover, the tight clustering of QC samples demonstrated method robustness and data reliability (Figure 3C,D). To further examine inter-group differences, PCA and PLS-DA were applied (Figure 4). In PCA, metabolic profiles from days 7 and 14 overlapped with those of fresh samples (day 0) in both ionization modes, suggesting limited metabolic variation at these earlier time points (Figure 4A). In contrast, day 21 samples were distinctly separated (Figure 4C,F). The paired PLS-DA comparisons revealed clear metabolic distinctions between day 0 and all subsequent time points in both ionization modes, with enhanced inter-group separation. Subsequently, permutation tests were carried out to avoid overfitting of the model (Figure S1). The cumulative value of the model parameter R2 (cum) was nearly 1, and the regression line of Q2 intersected the ordinate at a value less than 0, which indicated that the developed model was reliable (Figure S1). These findings indicate significant changes in the metabolic composition of donkey meat during storage. Differential metabolites were defined as those with a fold change (FC) > 2 or FC < 0.5 and a p-value < 0.05, and were visualized using volcano plots (Figure S2). The number of significantly altered metabolites compared to day 0 (D_0d) increased progressively with storage duration. Specifically, compared to D_0d, 127 differential metabolites were identified at day 7 (D_7d; positive ion mode: 59 up-regulated, 9 down-regulated; negative ion mode: 47 up-regulated, 12 down-regulated). This number increased to 143 by day 14 (D_14d; positive ion mode: 53 up-regulated, 19 down-regulated; negative ion mode: 49 up-regulated, 22 down-regulated), and further to 271 by day 21 (D_21d; positive ion mode: 97 up-regulated, 35 down-regulated; negative ion mode: 110 up-regulated, 29 down-regulated).

3.2.2. Identification of Differential Metabolites Between Two Groups

To identify representative metabolites, further screening was performed based on the following criteria: p < 0.05 in the PLS-DA model, FC ≥ 2 or ≤0.5, and a variable importance in projection (VIP) score > 1 (Tables S1–S3). Compared to day 0, the profile of lipid and lipid-like metabolites exhibited substantial alterations during storage. As fundamental components of cell membranes, phospholipids are critical for maintaining membrane integrity and function [32]. Notably, multiple lysophospholipids (LPLs), such as lysophosphatidylcholines (LPCs), lysophosphatidylethanolamines (LPEs), lysophosphatidic acids (LPAs), and lysophosphatidylinositol (LPI), demonstrated a time-dependent increase. This trend aligns with previous studies on Mongolian lamb meat, in which LPLs accumulated during wet-aging [33]. This also partly explains the reasons for the muscle fiber damage mentioned above. The rise in LPLs is largely attributed to phospholipase A_2_ (PLA_2_) activity, which catalyzes the hydrolysis of the sn-2 ester linkage of phospholipids to produce free fatty acids and LPLs [34]. Furthermore, phospholipid, which has a higher PUFA content to perform its function as a constituent of cellular membranes, serves as important substrates for lipid oxidation [35]. The resulting by-products, including free radicals (e.g., ROO•) and aldehydes, can subsequently accelerate heme protein oxidation and contribute to meat discoloration [36].

In chilled donkey meat, levels of various amino acids and their derivatives increased during storage (Tables S1–S3). Previous studies have indicated that most amino acids are generated through enzymatic hydrolysis of proteins and peptides over extended processing periods [37]. Taste-active peptides, amino acids, and amino acid derivatives are crucial in shaping meat flavor [38]. In the present study, key taste-related compounds, including L-norleucine, L-phenylalanine, and valine, were significantly elevated (p < 0.05) at day 7, 14, and 21 relative to day 0. This progressive accumulation of umami and sweet-tasting molecules, including certain dipeptides, reflects ongoing proteolysis and the formation of flavor precursors during storage, which together contribute to the characteristic savory and rich flavor of aged donkey meat. Beyond flavor, amino acids also influence other quality attributes such as tenderness and color. The association of meat tenderness with amino acid metabolites, particularly in relation to desmin degradation, has been documented [39]. Subsequent research has consistently reinforced the significant relationship between amino acids and meat tenderness [40]. As storage time increased, the appearance of chilled donkey meat shifted from initially fresh to visibly discolored. Ma et al. (2017) [41] discovered through metabolomics analysis of beef that phenylalanine or tryptophan exhibited a moderate negative correlation with color parameters (e.g., CIE a*, chroma, and visual lean color). Similarly, research on beef, lamb, and venison suggests that amino acid accumulation during storage may modulate biochemical pathways related to color stability [42]. Notably, glutathione levels decreased significantly after 7 days of storage. As an important antioxidant peptide in animal tissues, glutathione helps mitigate oxidative stress. Its antioxidant activity has been linked to slowed myoglobin oxidation, thereby improving color stability [43].

Nucleotide degradation serves as a well-established indicator of meat freshness during spoilage. Nucleotide derivatives in meat are mainly generated through postmortem ATP degradation, producing ADP, AMP, and IMP, which is subsequently converted by endogenous enzymes into hypoxanthine and xanthine [44]. In the present study, levels of hypoxanthine and xanthine [45] increased progressively after 7, 14, and 21 days of storage compared to 0 day (Tables S1–S3), supporting their proposed role as biomarkers of meat spoilage [16]. This accumulation is consistent with findings by Ma et al. (2017) [41], who identified xanthine and hypoxanthine as metabolites linked to oxidative instability in aged bovine muscles (longissimus lumbrum (LL), semimembranosus (SM), and psoas major (PM)), noting positive correlations with discoloration traits and non-heme iron accumulation. The underlying mechanism may involve the pro-oxidative activity of hypoxanthine, which can react with oxygen (O_2_) to generate superoxide anions (O_2_^−^), thereby promoting oxidative stress in tissues [46].

3.2.3. Metabolic Pathway Analysis

Based on the KEGG database, the top 20 enriched KEGG pathways in donkey meat that were significantly affected by storage time are illustrated in Figure 5. KEGG enrichment analysis indicated that differentially abundant metabolites were predominantly involved in lipid metabolism (e.g., glycerophospholipid metabolism, glycerolipid metabolism), amino acid metabolism (e.g., phenylalanine metabolism, tyrosine metabolism, alanine, aspartate and glutamate metabolism), and nucleotide metabolism pathways (e.g., purine metabolism and pyrimidine metabolism). In the early storage phase (7 days), changes in glycerophospholipid metabolism were observed, accompanied by the accumulation of lysophospholipids. The amino acid metabolism pathways corroborated the observed increase in amino acid content in donkey meat. By mid-storage (14 days), proteolysis intensified, leading to a marked increase in free amino acids and dipeptides, alongside products of nucleotide degradation. Furthermore, the significance of glutathione metabolism progressively intensified with prolonged storage. Notably, in the positive ion mode comparison between day 21 and day 0, glutathione metabolism ranked among the most enriched pathways. This finding is consistent with the marked downregulation of cysteinylglycine (Cys-Gly), a glutathione precursor and degradation product. These results suggest that, during the later stages of storage, the antioxidant defense mechanisms within donkey meat cells undergo substantial metabolic remodeling, leading to diminished capacity for free radical scavenging and thereby accelerating the oxidative browning of myoglobin.

3.2.4. Hierarchical Clustering Analysis of Differential Metabolites Across All Groups

To more clearly elucidate the overall evolution of the metabolic profile of donkey meat during storage, hierarchical clustering analysis was conducted on the significantly differential metabolites among the four groups, and the results were visualized via a heatmap (Figure 6). In the positive ion mode (Figure 6A), the primary cluster comprised compounds including acetyl-L-carnitine, creatine, nicotinamide, 3-phenyl-5-(trifluoromethyl)-4,5-dihydro-1H-pyrazol-5-ol, PC 19:2_18:5, PE O-17:1_18:2, 5′-S-methyl-5′-thioadenosine, S-adenosylmethionine, Cys-Gly, guanosine monophosphate, and N1-(2-cyanoethyl)-N1-cyclohexyl-4-bromobenzene-1-sulfonamide, which exhibited high abundance at the initial storage stage (day 0). Acetyl-L-carnitine is an endogenous factor involved in cellular energy regulation, promoting the mitochondrial transport of acyl-CoA (e.g., acetyl-CoA) required for fatty acid oxidation [47]. Ma et al. (2017) [41] identified metabolites that are associated with the oxidative stability of aged bovine muscles and found that acyl carnitines were positively associated with discoloration characteristics and non-heme iron accumulation. Zhang et al. (2010) [48] reported that dietary supplementation of Arbor Acres broilers with acetyl-L-carnitine improved meat color to some extent. The reduction in metmyoglobin (Met-Mb) to deoxy-Mb is driven by electron transfer from NADH, which can occur through enzymatic, non-enzymatic, and mitochondrial-mediated pathways [49,50,51]. As a precursor of NAD^+^, nicotinamide primarily delays the oxidative browning of myoglobin by preserving the activity of the NADH-dependent metmyoglobin reductase system in muscle tissue [52]. Esposito (2024) [53] also found that the nicotinamide content in chicken decreased significantly over storage time. This may indicate that the capacity of muscle cells to maintain NAD^+^/NADH homeostasis diminishes with prolonged storage, resulting in increased accumulation of metmyoglobin [15]. Notably, the concentration of hypoxanthine progressively increased with prolonged storage duration. Ma et al. (2017) [41] identified metabolites associated with the oxidative stability of aged bovine muscles and found that hypoxanthine was positively associated with discoloration characteristics.

In the negative ion mode (Figure 6B), the relative abundance of guanosine 5′-diphospho-D-mannose, inosine-5′-monophosphate (IMP), D-mannose 6-phosphate, α-D-mannose 1-phosphate, LNAPE 18:2/N-18:0, D-glucarate, cis-aconitic acid, and citric acid decreased over time. Conversely, concentrations of 4-oxoproline, elaidic acid, mycophenolic acid, and succinic acid increased. IMP is the principal flavor component in meat, but it is unstable and can be further degraded into hypoxanthine and inosine [54]. Abraham et al. [55] also observed a gradual increase in succinic acid during post-slaughter refrigeration of beef and speculated that it may be derived from glutamic acid or valine. The gradual decline in cis-aconitic acid and citric acid, key intermediates in the tricarboxylic acid (TCA) cycle, suggests impaired TCA cycle activity and reduced mitochondrial respiration. This metabolic alteration consequently hampers the regeneration of NADH (reduced nicotinamide adenine dinucleotide), thereby negatively impacting meat color stability.

4. Conclusions

This study characterized quality evolution and metabolomic changes in donkey longissimus lumborum during 21-day chilled storage and subsequent retail display. Extended storage enhanced tenderness but progressively compromised color stability, with samples stored beyond 14 days exhibiting accelerated discoloration. Metabolomic profiling identified 689 metabolites and revealed time-dependent accumulation of lysophospholipids, amino acids, and nucleotide degradation products, alongside progressive depletion of acetyl-L-carnitine, Cys-Gly, and nicotinamide. These metabolites showed consistent associations with color deterioration, indicating their potential as candidate biomarkers for freshness and quality assessment. Pathway analysis indicated that disrupted glutathione metabolism and depleted antioxidant reserves contribute to accelerated myoglobin oxidation during prolonged storage. Taken together, our findings suggest that the display period for donkey meat after 21 days of storage should not exceed 2 days to ensure acceptable quality. These findings provide mechanistic insights into post-mortem biochemical changes in donkey meat and offer a scientific basis for developing targeted preservation strategies to optimize shelf-life while maintaining desirable quality attributes. A limitation of this study is its focus on the longissimus lumborum muscle. Future research should validate these findings across different muscle types.