Comprehensive metabolite profiling in fruits of different Prunus cultivars

Abstract

Variations in metabolite profiles among Prunus cultivars shape fruit flavor and overall quality, and ultimately determine consumer preference. To comprehensively evaluate the metabolic quality of various Prunus cultivars, we adopted an integrated extraction strategy to obtain distinct metabolite fractions from identical samples. In this work, seven predominant Prunus cultivars from Xinjiang, China were investigated, 382 volatile metabolites and 417 non-volatile metabolites were identified, where phenols were detected as the dominant secondary metabolites and esters were abundant volatile metabolites. ‘Black’ displayed greater antioxidant activity relative to other cultivars, due to its more than four-fold higher content of pyroglutamic acid, glutamine, and γ-dodecalactone. Additionally, network pharmacology analysis confirmed that the metabolites maybe associated with diabetes, cancer, and Alzheimer's disease. Collectively, this study for the first time elucidates the full metabolite profile in Prunus, and provides practical guidance for consumer choices as well as a solid theoretical foundation for selective breeding of Prunus.

Highlights

• •417 non-volatile and 382 volatile metabolites were identified in Prunus fruits.
• •Phenols and esters were respectively most dominant non-volatile and volatile metabolite.
• •‘Black's top antioxidant is due to pyroglutamic acid, glutamine, γ-dodecalactone.
• •Metabolites were linked to diabetes and cancer via network pharmacology.
• •This study offers metabolic insights into consumer selection and Prunus breeding.

Prunus belongs to the subfamily Prunoideae of the Rosaceae family, representing one of the most crucial fruit crops in temperate regions. It comprises numerous economically valuable species such as plums, apricots, cherries, damsons, and jujubes, which are widely distributed across Eurasia and North America (Yu et al., 2021). Xinjiang is a primary region for premium fruit production in China, with a unique climate and environment favorable for Prunus growth (Yan et al., 2024). Fresh, dried, jammed, and smoked Prunus have different requirements for fruit quality. Driven by industrial upgrading and evolving consumer demands, the market evaluation criteria for Prunus fruits have shifted from appearance- and flavor-associated indicators (such as fruit size and sweetness) to nutritional value and functional attributes. This shift has driven research on Prunus to focus on the analysis of metabolic characteristics of specialty Prunus cultivars, identification of key functional components, and assessment of industrial application prospects (Abula et al., 2015; Xia et al., 2020).

Metabolomics, which is characterized by the capability of high sensitivity, high throughput, and comprehensive component profiling, is a powerful technology for deciphering the metabolic characteristics, differential metabolites, and functional associations in fruits (Hu et al., 2025; Cai et al., 2024; Chen et al., 2022). This technology also shows great promise in comprehensive metabolite profiling and differential metabolite screening, providing robust technical support for elucidating the metabolic characteristics and functional development potential of fruits. In a previous study, ultra performance liquid chromatograph-electro-spray ionization-tandem mass spectrometer (UPLC-ESI-MS/MS) was employed to identify 1439 metabolites (including amino acids, flavonoids, and phenolic acids) in Chinese cherries, among which anthocyanin components collectively determine the fruit color, while flavonoids and other components contribute to the bitter taste (Liu et al., 2024; Xiong et al., 2026). Comprehensive two-dimensional gas chromatography-time-of-flight mass spectrometry (GC × GC-TOFMS) facilitates the precise capture of trace volatile metabolites. It has been used to successfully identify 495 volatile compounds in ‘Bee Sugar’ plum, among which hydrocarbons, alcohols, ketones, and esters are significantly higher than those in ‘April’ plum (Hu et al., 2025).

Prunus fruits are rich in metabolites (such as amino acid, phenols, and lipids), exhibiting distinct advantages in flavor formation, nutritional provision, and health protection. Therefore, they represent a valuable resource for the development of functional foods and natural medicines (Yu et al., 2021). The flavor and storage quality of fruits are predominantly governed by primary metabolites: acidity is primarily modulated by organic acids, whereas sweetness and bitterness are mediated by amino acids. Metabolomic research on taste quality across diverse tomato cultivars has quantified their amino acid and organic acid profiles, establishing robust correlations between these metabolites and sensory taste perception (Cai et al., 2024). Multiple metabolites play important roles in the complex physiological processes of litchi pericarp development, such as color change and the establishment of defense mechanisms, ultimately influencing the quality of the fruit (Li et al., 2025). Antioxidant activity is a core functional trait of fruits and has emerged as a prominent research focus in recent years. The association between fruit antioxidant capacity and metabolites has been systematically elucidated and validated in multiple plant species. In a recent study, a total of 1133 secondary metabolites were identified in Siberian plums, and a comparative analysis revealed that 26 of these metabolites exhibited significant positive correlations with the antioxidant capacity (Hao et al., 2025; Yang et al., 2025). In addition, metabolomics combined with network pharmacology can be used to predict the targets of pivotal metabolites and infer functional components. This dual-omics approach has been applied to forecast antibacterial targets of metabolites in various citrus herbs (Li et al., 2025).

A wide range of Prunus cultivars are commercially available on the market at present, yet targeted utilization strategies for individual cultivars remain to be optimized. The endogenous metabolite profiles can serve as a preliminary basis for identifying the optimal application orientation of each cultivar. Metabolic profiles and antioxidant activity may vary across different Prunus cultivars, which in turn contributes to their divergent flavor and functional qualities. Currently, there are still inadequate metabolomics detection methods for Prunus, which can enable comprehensive and precise characterization of metabolites in Prunus fruits. Existing studies have only characterized a limited range of metabolites in Prunus fruits. Specifically, previous investigations have separately explored phenolic compounds in fresh plums (Rao et al., 2026) and volatile metabolites in smoked Prunus mume (Hu et al., 2025). Accordingly, comprehensive and systematic research delineating shifts in the material basis among diverse cultivars is still scarce. Against this backdrop, our study adopted a unified extraction strategy to achieve the comprehensive identification of the full spectrum of metabolites. In this study, we attempted to employ multi-metabolomics approaches to simultaneously extract and fully characterize the metabolites in fruits of seven Prunus cultivars, including primary metabolites, secondary metabolites, lipids and volatile metabolites, elucidate the mechanisms underlying the association between metabolic diversity and superior antioxidant/heath-promoting potential. The findings are expected to support the metabolomics basis for consumer preference and the breeding of Prunus, thereby contributing to high-quality development of Prunus industry.

1. Materials and methods

1.1. Experimental materials

Seven Prunus cultivars cultivated in Xinjiang, China were used as the test materials, including ‘Angeleno’, ‘France’, ‘Empress’, ‘Pluot’, ‘Bee Sugar’, ‘Qingcui’, and ‘Black’, all of which were sourced from the Western Plum Orchard of the Fourth Brigade, Twelfth Regiment, Aral City, Xinjiang Uygur Autonomous Region. Healthy, disease-free trees with consistent vigor and well-developed growth were selected for study. Ten uniformly ripe fruits were collected from the middle-outer fruit branches on the east, west, south, and north sides of each tree to form one biological replicate. Four biological replicates were established for each samples. Each Prunus cultivar was independently set as an experimental group for follow-up analysis. After separation of the peel and flesh, the samples were rapidly frozen in liquid nitrogen and stored at −80 °C. Upon completion of all sample collection, the samples were transported via dry ice to the lab for analysis. Every sample was equally divided into two portions. One portion was subjected to cryogenic grinding for volatile compounds analysis. The other portion underwent cryogenic grinding followed by freeze-drying for non-volatile metabolites and selected physicochemical parameters analysis.

2. Experimental methods

2.1. Extraction of non-volatile metabolites

Metabolites extraction was performed following the one-step protocol for polar and semi-polar metabolites from Arabidopsis samples (Salem et al., 2016). Layer separation was achieved after M1 (75% methyl tert-butyl ether (MTBE) + 25% pure methanol) and M2 (25% methanol aqueous solution) extraction, with the upper layer being nonpolar and the lower layer polar.

For primary and secondary metabolites detection, 400 μL lower layer was collected and evaporated to dryness under nitrogen. The residue was reconstituted with 200 (50% methanol aqueous solution containing lidocaine as internal standards), vortexed to mix uniformly, and then transferred to a sample vial with an insert tube.

For lipids detection, 500 μL upper layer was collected and evaporated to dryness under a nitrogen evaporator. The residue was reconstituted with 250 μL extraction solution (acetonitrile: isopropanol = 7: 3, containing 15:0–18:1 d7 PE (phosphatidylethanolamine) and 15:0–18:1 d7 PC (phosphatidylcholine) as internal standards), vortexed to mix uniformly, and then transferred to a sample vial with an insert tube.

2.2. Detection of non-volatile metabolites

All non-volatile metabolites (primary, secondary metabolites and lipids) analyses were performed using a Thermo Fisher U3000-Orbitrap Q Exactive Plus Liquid Chromatography-Mass Spectrometry (LC-MS) system.

2.2.1. Non-targeted primary metabolites detection

To detect primary metabolites, LC-MS was equipped with Waters ACQUITY UPLC BEH Amide column (100 × 2.1 mm, 1.7 μm) following the method described by Ding et al. (2021). The mobile phase consisted of water (containing 10 mM ammonium formate and 0.125% formic acid) (A) and acetonitrile/water (95:5, v/v, containing 10 mM ammonium formate and 0.125% formic acid) (B), with gradient elution: 0.0–2.0 min, 100% B; 7.7 min, 70% B; 9.5 min, 40% B; 10.25 min, 30% B; 12.75–17 min, 100% B. The flow rate was 0.3 mL/min, injection volume was 1 μL, and column temperature was 45 °C.

2.2.2. Non-targeted secondary metabolites detection

To detect secondary metabolites, LC-MS was equipped with Waters ACQUITY UPLC HSS T3 column (100 × 2.1 mm, 1.8 μm) following previously published method (Salem M et al., 2016). The mobile phase consisted of water (containing 0.1% formic acid) (A) and pure acetonitrile (B), with gradient elution: 0.0–2.0 min, 1% B; 16 min, 95% B; 20 min, 95% B; 20.5–23 min, 1% B. The flow rate was 0.4 mL/min, injection volume was 3 μL, and column temperature was 40 °C.

2.2.3. Non-targeted lipids detection

To detect lipids, LC-MS was equipped with Thermo Acclaim C30 column (150 × 2.1 mm, 3 μm) following our previously published method (Wang et al., 2022). The mobile phase consisted of water/methanol/acetonitrile (1:1:1, v/v/v, containing 5 mM ammonium acetate) (A) and acetonitrile/isopropanol (1:5, v/v, containing 5 mM ammonium acetate) (B), with gradient elution: 0.0–0.5 min, 20% B; 1.5 min, 40% B; 3.0 min, 60% B; 12.0 min, 80% B; 20.0 min, 90% B; 22.0 min, 98% B; 26.0–28.0 min, 20% B. The flow rate was 0.4 mL/min, injection volume was 3 μL, and column temperature was 60 °C.

All the above detections were performed with the same MS parameters. MS detection was performed with HESI ionization, using a sheath gas flow rate of 40 arb, auxiliary gas flow rate of 10 arb, auxiliary gas temperature of 350 °C, and ion transfer tube temperature of 320 °C. In positive ion mode, the spray voltage was 3.5 kV; in negative ion mode, the spray voltage was 3.0 kV. Compound identification was conducted using full MS/dd-MS² mode scanning in both positive and negative ion modes, while compound quantification was performed using full MS scanning in both ion modes.

2.2.4. Qualitative and quantitative analysis of non-volatile metabolites

Raw data were reviewed using Xcalibur 4.0 software to confirm that the instrument operated in good condition with stable signals throughout sample detection and analysis. Data processing and analysis were performed using Compound Discover 3.3 software, including peak identification, peak filtering, peak alignment, retention time correction, missing value imputation, sum ion merging, and metabolite identification. MS2 fragments were matched against the mzCloud online database and in-silico mzVault database for metabolite qualitative identification. Primary and secondary metabolites were quantified based on the added internal standard lidocaine (4 × 10⁻⁴ mg/mL), while lipids were quantified using the added phospholipid internal standards PC 15:0_18:1(d7) and PE 15:0_18:1(d7) (1 × 10⁻⁴ mg/mL).

2.3. Extraction and detection of volatile compounds

Samples were prepared following a previously described method with slight modifications (Du et al., 2026). First, 0.2 g thoroughly mixed sample and 3 mL 36% saline solution (containing the internal standard methyl nonanoate) were transferred into 20 mL headspace vial. The sample was equilibrated at 60 °C for 10 min, followed by absorption using a solid-phase microextraction fiber head for 40 min. Analysis was conducted using two-dimensional gas chromatography-time-of-flight mass spectrometry (GC × GC-TOFMS) (Pegasus BT4D, LECO, USA) equipped with 1D column Rxi-5 MS (30 m × 0.25 mm, 0.25 μm) and 2D column Rxi-17Sil MS (0.7 m × 0.25 mm, 0.25 μm). In 1D column temperature program, the initial temperature was ramped from 60 °C to 130 °C at 3 °C/min, followed by ramping from 130 °C to 240 °C at 30 °C/min, then ramping from 240 °C to 300 °C at 50 °C/min, and holding at 300 °C for 3 min. The 2D column temperature program followed the first dimension program with an additional 5 °C increase. Mass spectrometry analysis was performed using electron ionization (EI) at 70 eV. The ion source temperature was set to 250 °C, the acquisition rate was 200 spectra/s, the mass scan range covered 35–500 m/z, and the extraction frequency was 32 kHz.

Data processing was performed using ChromaTOF SYNC software with the following specific parameters: 1D peak width of 3 s; 2D peak width of 0.02 s; minimum signal-to-noise ratio of 10. Compound search was conducted against the NIST mainlib, with similarity 700, probability >800, and retention index deviation <30. Impurities and wrap around compounds were excluded through structural analysis.

2.4. Physicochemical assays

Total Antioxidant Capacity (T-AOC) was measured using the Beijing Sangon Reagent Kit (Cai et al., 2025).

The DPPH and ABTS Radical Scavenging Capacity employed the Beijing Sangon Reagent Kit (Zhong et al., 2025).

2.5. Data analysis

Statistical analysis was performed using GraphPad Prism 9 software. Principal component analysis (PCA) and supervised orthogonal partial least squares discriminant analysis (OPLS-DA) were conducted using SIMCA-P 14.1 software. Data were analyzed using analysis of variance (ANOVA), followed by t-tests to detect significant differences between sample means, with a significance level set at p < 0.05 (Chen et al., 2024; Zhao et al., 2024). Correlation and enrichment analyses were performed using Origin 2026 software. Cluster heatmap analysis was conducted using Web Bioinformatics (https://www.bioinformatics.com.cn/?p=6). Pathway analysis was completed using MetaboAnalyst (https://www.metaboanalyst.ca/). All raw images were processed using Adobe Photoshop 2025.

3. Results and discussion

3.1. Metabolomic characterization of seven Prunus cultivar fruits

The metabolite composition of plant fruits largely determines their flavor profiles and health-beneficial properties. This study primarily focused on seven Prunus cultivars widely cultivated in Xinjiang, namely ‘Angeleno’, ‘France’, ‘Empress’, ‘Pluot’, ‘Bee Sugar’, ‘Qingcui’, and ‘Black’ (Fig. 1A). A non-targeted metabolomics analysis was conducted to obtain the metabolomic profile of Prunus fruits. Non-volatile metabolites mainly comprise primary metabolites, secondary metabolites, and lipids. Given variations in polarity and chromatographic behavior across different metabolites, the extraction and LC-MS detection methods were optimized to enable comprehensive identification of the metabolites. Many common sugar and acid compounds exist as multiple isomers, which cannot be efficiently separated using conventional T3 columns. Therefore, we employed Amide columns, which are advantageous in separating highly polar compounds, for the detection and identification of these isomers. First, primary metabolites were detected and identified using Amide column, resulting in the identification of 82 compounds predominantly composed of amino acids, soluble sugars, and soluble acids. Subsequently, secondary metabolites were analyzed using a T3 column, resulting in the identification of 256 compounds belonging to 11 categories, including phenols, organic acids, and glycosides. Finally, a C30 column was employed for detection and identification of lipids, resulting in the identification of 79 compounds primarily encompassing glycerides, glycerophospholipids, glycolipids, and others. Volatile metabolites were detected and identified via GC × GC-TOFMS, and a total of 382 compounds were identified across ten major categories, including esters and alkanes (Fig. 1B).

Fig. 1.

Multidimensional detection of phenotypes and metabolites in seven Prunus cultivars. A: Phenotye of seven Prunus cultivars fruits, with four replicates per cultivar; B: Extraction and Detection outline of three strategies of Non-Volatile and Volatile metabolites.

Qualitative analysis of metabolites was performed based on the inherent characteristics of target compounds and available reference information. Amide column exhibit excellent performance in distinguishing isomers. Taking fructose and glucose as a typical example, these two soluble sugars possess identical molecular weights and secondary mass spectrometry (MS2) fragmentation patterns. Although the Amide column generates two distinct chromatographic peaks, it fails to distinguish the corresponding substance for each peak. Specifically, fructose and glucose were differentiated by matching their retention time with standards. T3 column enabled the simultaneous detection of diverse metabolites, which were initially identified via database retrieval; further confirmation was subsequently achieved through MS2 spectral alignment with data retrieved from PubChem. For lipids, identification was conducted using the database Lipidblast, followed by validation based on the combined fragment ion patterns acquired in both positive and negative ionization modes. As for volatile compounds, identification was mainly accomplished by matching their MS2 spectra with those in the National Institute of Standards and Technology (NIST) spectral library (Fig. S1A).

A principal component analysis (PCA) was performed on all the identified compounds. The PCA score plot revealed close clustering of biological replicates from each Prunus cultivar and tissue, indicating stable and reliable detection of the metabolites. Furthermore, the peel and flesh tissues were clearly separated, suggesting distinct metabolite compositions across different tissues (Fig. S1B). Hierarchical clustering analysis demonstrated that ‘Angeleno’, ‘France’, and ‘Empress’ were closely clustered, implying that these cultivars may possess similar metabolites or close domestication relationships, which is consistent with the clustering pattern observed in the PCA (Fig. S1C).

3.2. Analysis of primary metabolites in seven Prunus cultivars

The flavor of Prunus fruits is primarily determined by primary metabolites, such as soluble sugars and organic acids (Si et al., 2022). A qualitative analysis of the detected compounds confirmed 82 primary metabolites, including 28 amino acids, 15 soluble sugars, 17 soluble acids, and other compounds. A column proportion analysis on the contents of these compounds revealed that soluble sugars were dominant in the detected metabolites, representing over half of the identified primary metabolites (Fig. S2A). Clustering analysis demonstrated distinct inter-cultivar differences among the tested Prunus cultivars (Fig. 2B), while PCA showed tight clustering of the biological replicates for the same cultivar, indicating stable and reproducible sample detection. PCA was applied to reveal global metabolic variation in seven Prunus cultivars. The first principal component (PC1) explained 59.4% of total metabolic variance, and PC2 accounted for an additional 14.6% (Fig. 2A).

Fig. 2.

Analyse of primary metabolites detected in fruis of Prunus cultivars. A: Principal Component Analysis (PCA) plot of primary metabolites in fruis of seven Prunus cultivars; B: Clustering heatmap of primary metabolites in fruis of seven Prunus cultivars; C: Bar chart of highly abundant soluble sugars and acids in fruis of seven Prunus cultivars.

Malic acid, fructose, and sorbitol were the most abundant primary metabolites. Malic acid, a predominant organic acid, exhibited the lowest content in ‘Empress’ but higher contents in ‘Pluot’ and ‘Bee Sugar’. ‘France’ had significantly higher citric acid contents than all other cultivars (Fig. 2C). These results indicated that soluble acids could be play distinct roles in determining the flavor quality of Prunus fruits. A previous study has demonstrated that malic acid and citric acid dominate the organic acid composition in wild Prunus species (Mikulic-Petkovsek et al., 2016), where ‘Pluot’ was classified as a high-acid plum. High-acid plums are typically processed into dried fruits due to their textural properties, and therefore crisp-fleshed plums such as ‘Pluot’ may particularly suitable for this processing application (Hu et al., 2025). The high-acid cultivars ‘Pluot’ and ‘Bee Sugar’ in this study can also serve as processing alternatives. Additionally, ‘Pluot’ and ‘Bee Sugar’ exhibited higher fructose contents but significantly lower sucrose contents than other cultivars (Fig. 2C), probably due to the conversion of sucrose to fructose, which can enhance the sweetness flavor of fruits. Comprehensively considering the distribution patterns of sugars and acids, ‘Pluot’ and ‘Bee Sugar’ may be the two cultivars with the most desirable sweet-sour flavor among the seven studied cultivars. One study of apricot fruits has revealed that cultivars with superior flavor accumulate higher levels of sucrose, glucose, fructose, and sorbitol while lower levels of titratable acid, citric acid, and starch (Gou et al., 2023). Therefore, the low malic acid content and high sucrose content in ‘Empress’ may explain its high-sugar, low-acid quality characteristics, which are more appealing to consumers and thus commercially favored.

In addition to the detection of high levels of specific soluble sugars and acids, a variety of amino acids, sugar alcohols, and other metabolites were also identified as primary metabolites. A column analysis revealed no distinct difference in the distribution patterns of amino acids between the peel and flesh. In contrast, soluble sugars alcohols were present in relatively higher contents in the fruits of ‘Angeleno’, ‘France’, and ‘Empress’ (Fig. S2B). Among these soluble sugars alcohols, sorbitol, which is characterized by the highest content and recognized as a dietary component with certain health-promoting properties (Hu et al., 2025), was particularly enriched in the fruits of ‘Angeleno’, ‘France’, and ‘Empress’ (Fig. 2C), this phenomenon may partly explain why this fruit possess potential properties for supporting digestive function. Notably, ‘Angeleno’, ‘France’, and ‘Empress’ are three species belonging to the prune genus, which is consistent with prior research indicating that the high sorbitol content in plums is a major contributor to their laxative effects (Stacewicz-Sapuntzakis et al., 2001). Intriguingly, in contrast to the soluble sugar alcohols content, these three cultivars had relatively low detectable levels of soluble acids, suggesting that they may represent high-sugar and low-acid germplasm resources. Pathway enrichment analysis of the detected primary metabolites showed that the identified metabolites were predominantly enriched in amino acid biosynthesis and carbohydrate metabolism pathways (Fig. S2C). These results indicated that most of the identified primary metabolites were amino acids and carbohydrates, which is consistent with the metabolite identification results: amino acids were the most abundant compounds, while soluble sugars and sugar alcohols were the highest content compounds.

3.3. Analysis of secondary metabolites in seven Prunus cultivars

Secondary metabolites in fruits display substantial diversity, and their composition, content, and proportional distribution directly determine fruit color, flavor, texture, nutritional functionality, and storage stability. In this study, secondary metabolite profiling identified a total of 256 compounds, predominantly including 86 phenols, 51 organic acids, 22 glycosides, 16 terpenoids, 16 esters, 14 heterocycles, 10 aldehydes, 9 amino acids, 9 sugars, and 5 coumarins (Fig. 3A). Prunus fruits contain not only functional phenols but also a diverse array of other secondary metabolites, which collectively underpin the distinct flavor characteristics of Prunus fruits. A clustering analysis of the all identified secondary metabolites and phenols revealed no tissue-specific distribution pattern of these metabolites(Fig. 3C&S3A). Subsequent quantitative analysis of compound abundance using bar charts showed that organic acids and esters accounted for significantly high proportions of the secondary metabolites (Fig. 3B). Further characterization of these dominant compounds indicated that most high-abundance organic acids such as picolinic acid were predominantly accumulated in the flesh of ‘Pluot’, while esters such as methotrexate diethyl ester were more enriched in the flesh of ‘Angeleno’ (Fig. S3B). The heat map also showed that most compounds exhibited higher contents in ‘Pluot’, ‘Bee Sugar’, and ‘Qingcui’. These compounds were primarily organic acids, which is consistent with the bar chart results (Fig. 3B).

Fig. 3.

Secondary metabolites detected in fruits of seven Prunus cultivars. A: Number of secondary metabolite types across fruits of seven Prunus cultivars; B: Distribution of total detected secondary metabolites content across fruits of seven Prunus cultivars; C: Clustering heatmap of phenols within fruits of seven Prunus secondary metabolites; D: Bar chart of high-content phenols in fruits of seven Prunus cultivars.

As the predominant functional secondary metabolites in fruits, phenols not only directly regulate fruit astringency, color formation, antioxidant capacity and health-promoting properties but also significantly affect fruit storage stability in terms of their diversity and accumulation levels (Rao et al., 2026). A clustering analysis was further conducted on the 86 identified phenols compounds. Analysis of the relatively abundant phenolic compounds revealed that 4-hydroxybenzoic acid dominated in abundance. As a natural plant metabolite, it possesses antibacterial and antioxidant activities. It was detected at relatively high levels in both the peel and flesh of ‘Angeleno’ and ‘France’. Moreover, calendoflavoside exhibited higher contents in the peel of ‘Angeleno’. Calendoflavoside possesses potent anti-inflammatory and antioxidant activities. Notably, its content was significantly different between the peel and flesh. Moreover, ‘Qingcui’ showed a significantly higher cyclomammeisin content in both the peel and flesh than other cultivars. Two common phenols, procyanidin C1 and caffeic acid, were remarkably more abundant in ‘France’ than in other cultivars (Fig. 3D). Consistently, this study demonstrated considerable variations in phenols content among the seven Prunus cultivars and across different tissues. Furthermore, a higher phenols content has been shown to be correlated with enhanced antioxidant activity (Jarwan et al., 2023). Therefore, Prunus, by virtue of its wide variety of phenols, may also exert certain beneficial effects on human health, which warrants further in-depth investigation.

3.4. Analysis of lipids in seven Prunus cultivars

Lipids serve as crucial precursors for the biosynthesis of diverse aromatic and flavor compounds in fruits, and can also exert potent antioxidant activities. Glycerolipids can provide substrates for the synthesis of characteristic fruit flavor compounds through self-degradation or biotransformation (Chen et al., 2022). Here, a total of 79 lipids were identified across the fruits of seven Prunus cultivars, predominantly including 34 Glycerolipids (GL), 28 Glycerophospholipids (GP), 9 Saccharolipids (SL), and 8 Sphingolipids (SP) (Fig. S4A). Proportional bar chart analysis demonstrated that GP and GL are dominant lipids in Prunus fruits. Among these, GL accounted for the highest proportion (63%), followed by GP (43%) (Fig. 4A). Furthermore, a clustering analysis of the identified compounds via heatmap visualization revealed that the peel has higher contents of lipids than the flesh (Fig. 4B). Further characterization of the abundant compounds confirmed this peel-preferential accumulation pattern. Moreover TG (Triglyceride) 18:2_18:3_18:3 and PI (Phosphatidylinositol) 16:0_18:2 were abundant in both the peel and flesh of ‘France’, whereas PA (Phosphatidic acid) 16:0_18:1 was the lowest in this cultivar; ‘Qingcui’ exhibited relatively high levels of PG 16:0_18:1, PA 16:0_18:1, and MGDG (Monogalactosyldiacylglycerol) 18:2_18:2; ‘Bee Sugar’ contained higher levels of HexCe r(Hexosylceramides) 18:2;2O/16:1 (Fig. 4C). Collectively, lipids showed no distinct cultivar-specific distribution pattern but were consistently accumulated preferentially in the peel than in the flesh, which may be ascribed to the peel's requirement for enhanced antioxidant capacity to maintain fruit freshness. A waxy layer forms on the fruit surface to retain moisture, and lipids may accumulate as substrates or intermediates in the synthesis of certain waxes.

Fig. 4.

Lipids detected in fruits of seven Prunus cultivars. A: Proportional distribution of lipids content across fruits of seven Prunus cultivars; B: Clustering heatmap of lipids in fruits of seven Prunus cultivars; C: Bar chart of high-content lipids in fruits of seven Prunus cultivars.

Among the lipids identified in black wolfberries, glycosphingolipid exhibited a positive correlation with the anti-inflammatory activity. A correlation analysis revealed strong positive associations between major carbohydrates and the glycosphingolipid content (Chen et al., 2022). Additionally, we identified a significant amount of GL (such as TG 18:2_18:3_18:3), which is the primary energy storage form in fruits. During fruit ripening, TG will be hydrolyzed by lipases into free fatty acids and other small molecules to provide energy for fruit ripening (Zhu et al., 2024). Several GP such as PG 16:0_18:2, PG 16:0_18:1 and PA 18:0_18:1 at relatively high concentrations were further identified in Prunus fruits. Significant levels of GP were identified in seabuckthorn fruit oil, and prolonged storage led to a marked decrease in their content (Li et al., 2025).

Classification and pie chart analysis of the detected lipids indicated that GL and GP are dominant lipids in Prunus fruits. These two major lipids collectively accounted for over half of the identified lipids (Fig. 4A). Previous studies have demonstrated that certain GL undergo metabolic degradation to generate fatty acids, which are further converted into volatile aromatic compounds (such as aldehydes and esters) via the lipoxygenase pathway. GP components, such as phosphatidylcholine and phosphatidylethanolamine, can scavenge free radicals and chelate transition metal ions that readily catalyze oxidation reactions, thereby inhibiting lipid oxidation chain reactions (Wang et al., 2025). This may account for the abundant flavor quality and antioxidant activity observed in Prunus fruits. A pathway enrichment analysis of the identified lipids also revealed that these metabolites are primarily enriched in phospholipid biosynthesis pathways, which is consistent with the types of the detected compounds (Fig. S3B).

3.5. Analysis of volatile metabolites in seven Prunus cultivars

GC × GC-TOFMS was employed for the detection and identification of volatile compounds in fruits of seven Prunus cultivars. A total of 382 volatile metabolites were identified (Table S1). Esters (86 compounds) were the most abundant volatile metabolites, followed by alkanes (49), alcohols (44), aldehydes (41), and aromatic hydrocarbons (35) (Fig. 5A). A bar chart analysis of the total content of each type of volatile metabolites revealed that esters and aldehydes are dominant volatile metabolites, accounting for more than half of the detected metabolites. Thus, esters and aldehydes represent the core volatile metabolites in Prunus (Fig. 5B). As important volatile metabolites in Prunus fruits, esters exhibited significant cultivar variations. A previous study of apricot fruits has also indicated that ester content plays a crucial role in flavor formation (Li et al., 2021). A clustering analysis of the detected volatile metabolites, as visualized via heatmap (Fig. 5C) and volcano plot (Fig. 5D), revealed their higher contents in Prunus peel.

Fig. 5.

Volatile compounds in fruits of seven Prunus culivars detected by GCXGC-TOFMS. A: Proportional distribution of volatile compounds across fruits of seven Prunus culivars; B: Bar chart of the proportion of total volatile metabolite types in fruits of seven Prunus cultivars; C: Clustering heatmap of volatile compounds in fruits of seven Prunus culivars;D: Volcano plot of volatile compounds in different parts of fruits of seven Prunus culivars; E: Bar chart of high-content volatile compounds in fruits of seven Prunus culivars.

Further characterization of high-abundance compounds based on bar chart analysis (Fig. 5E) demonstrated that the peel has a larger number of high-abundance volatile metabolites than the flesh. However, these metabolites showed no distinct distribution patterns across seven Prunus cultivars. Their higher contents in the peel compared with those in the flesh may be attributed to the relatively abundant precursor substances (such as lipids and secondary metabolites) in the peel, which can undergo further degradation to generate more flavor compounds (Zhang et al., 2021). Among these, the most abundant compound, Octanoic acid,2-methyl-, methyl ester, exhibited no clear cultivar-specific distribution pattern in the flesh. Nevertheless, its content was significantly lower in the peel of ‘Bee Sugar’ (Fig. 5E). Although different compounds display distinct distribution patterns, similar distribution patterns were observed within the same compound class. For instance, toluene and styrene, two aromatic compounds, showed higher contents in the peel of ‘Empress’ and ‘Black’ (Fig. 5E), which may contribute to their enhanced aromatic profiles (Su et al., 2022). A large number of alcohols were identified in the seven Prunus cultivars. Previous studies have shown that alcohols are widely present in agricultural products, acting as important aromatic compounds and key contributors to the fermented flavor in fruit juices (Bi et al., 2022). They are also primary contributors to the abundant malt and roasted cocoa flavor compounds in dried tea samples (Hutasingh et al., 2024). Additionally, nine organic acids and some other acidic compounds were identified. It has been reported that fermentation of watermelon juice increases acid production, which can help extend the shelf life of food products (Shi et al., 2023). This phenomenon may explain the high storability and fragrant aroma of Prunus fruits.

In summary, among non-volatile metabolites, primary metabolites were predominantly composed of sugars and acids (such as malic acid, citric acid, sucrose and fructose), which are key determinants of the fruit flavor profiles. ‘Pluot’ and ‘Bee Sugar’ exhibited particularly distinct sweet-sour attributes. There were significant inter-cultivar differences in metabolic characteristics: ‘Angeleno’, ‘France’, and ‘Empress’ were clustered closely, which are characterized by high sugar content, low acidity, and abundant sorbitol accumulation. For secondary metabolites, phenols accounted for the highest proportion, and the cultivars such as ‘Angeleno’ and ‘France’ displayed significantly higher phenols contents, which endows the fruits with robust antioxidant and health-promoting properties. The lipids were dominated by GLs and GPLs, which were consistently accumulated at higher levels in the peel than flesh. These lipids serve as precursors for flavor compound biosynthesis and participate in antioxidant processes. A total of 382 volatile metabolites were identified, with esters as the most abundant class. The content of volatile metabolites in the peel was higher than that in the flesh, and inter-cultivar differences contribute to their unique flavor profiles. Overall, we constructed a comprehensive metabolic profile of seven Xinjiang Prunus cultivars with distinct metabolic characteristics, which have clear implications for consumer purchasing choices. ‘Angeleno’, ‘France’, and ‘Empress’, characterized by high sorbitol contents, are well-suited for fresh consumption and dietary use; ‘Pluot’ and ‘Bee Sugar’, which possess a pronounced sweet-tart flavor derived from their sugar-acid balance, are ideal for dried fruit processing; ‘Angeleno’ and ‘France’ are preferred as antioxidant-rich and health-beneficial cultivars due to their abundant phenols. Furthermore, analyses of lipids and volatile metabolites revealed that the peel of all seven Prunus cultivars exhibited significantly superior flavor quality compared with the flesh, which could be ascribed to the higher accumulation of flavor-precursor metabolites in the peel. The fruit flesh accumulates high levels of sugars and acids, while the peel accumulates high levels of bioactive compounds. Collectively, these findings systematically reveal the metabolic diversity of Xinjiang Prunus fruits, providing critical metabolomic evidence for evaluating cultivar characteristics and guiding consumer choices.

3.6. Antioxidant activity and network pharmacology of seven Prunus cultivars

Antioxidant activity represents a core functional trait of Prunus fruits, and multiple phenols with potent antioxidant capacity have been detected and identified. Here, the antioxidant activity of seven Prunus cultivars was assessed using T-AOC, DPPH and ABTS assays. The results indicated significant variations in antioxidant activity indices among different cultivars. The T-AOC level was the lowest (4.13 μmol/g) in the peel of ‘Bee Sugar’, whereas the highest (27.05 μmol/g) in that of ‘Black’. The T-AOC level in the flesh showed a consistent trend, but significantly lower than that in the peel (Fig. 6A), which may be due to the higher phenols content in peel than that in the flesh. ‘Black’ displayed higher DPPH and ABTS radical scavenging capacities in both the peel and flesh. ‘Bee Sugar’ exhibited weaker radical scavenging capacity in the peel, which is consistent with the trend of T-AOC level; in contrast, ‘Empress’ showed weak radical scavenging capacity in peel yet strong radical scavenging capacity in the flesh.

Fig. 6.

Antioxidant capacity indices and correlation anallysis with metabolites in fruits of seven Prunus cultivars. A: Antioxidant activity indices of seven Prunus cultivars. Different letters displays significant (P < 0.05) differences for each cultivars, according to the independent samples t-test. B: Proportion of metabolite categories highly correlated with antioxidant activity of sevenPrunus cultivars; C: Enrichment of metabolic pathways highly correlated with antioxidant activity in fruits of seven Prunus cultivars; D: Heatmap of metabolites highly correlated with antioxidant activity; E: Metabolites differentially expressed between ‘Black’ and other cultivars; F: Network Pharmacology Analysis of metabolites in fruits offseven Prunus cultivars; G: Distribution of Disease in the Human Body.

A considerable proportion of the detected metabolites displayed notable correlations with the antioxidant indices (|correlation coefficient| ≥ 0.5 with at least two antioxidant indices) (Fig. S5A-D), and the majority of them were positive correlations (Fig. 6B). These results indicated that the elevated accumulation of these metabolites might associated with the high antioxidant capacity in Prunus, which is the same as the mechanisms underlying the relationship of antioxidant properties and metabolites in jujube following steaming and sun-drying (Liu et al., 2025). Pathway enrichment analysis of differentially accumulated metabolites related to antioxidant activity revealed significant enrichment in key pathways, including nicotinate and nicotinamide metabolism, the pentose phosphate pathway, and glutathione metabolism. These pathways constitute the core of the cellular antioxidant defense system, providing NADPH and reducing equivalents to maintain redox homeostasis and scavenge reactive oxygen species, which is consistent with the superior antioxidant capacity observed in cultivars such as ‘Black’. Additionally, pathways involved in unsaturated fatty acid biosynthesis and tyrosine metabolism were also enriched, suggesting their roles possibly as precursors for volatile flavor compounds and phenolic antioxidants, respectively. Collectively, these enriched pathways may provide mechanistic insights into the metabolic basis of Prunus fruit antioxidant activity and flavor diversity (Fig. 6C) (Yue et al., 2025). Clustering heatmap analysis of these highly correlated compounds revealed their higher accumulation in the ‘Black’ with high antioxidant activity. These highly correlated compounds included phenols such as kaempferol and quercitrin, which is consistent with previous research findings (Fig. 6B-E). Previous studies have also indicated that during tomato ripening, increased accumulation of flavonoids, phenolic acids, and antioxidants, along with decreased sugar glycosides, modulates the remodeling of volatile aromatic compounds (Li et al., 2025).

Antioxidant activity analysis showed that the Prunus cultivar ‘Black’ exhibited significantly stronger antioxidant activity than the other tested cultivars. To further explore the underlying differences between ‘Black’ and other cultivars, orthogonal projections to latent structures discriminant analysis (OPLS-DA) was conducted on seven Prunus cultivars, and key differential metabolites were identified according to the variable importance in projection (VIP) values derived from the OPLS-DA model. The top three differential metabolites between ‘Black’ and other cultivars were pyroglutamic acid, glutamine, and γ-dodecalactone, which may contribute substantially to the distinct antioxidant activity observed in ‘Black’. Further quantitative determination confirmed that these three metabolites accumulated at markedly higher levels in ‘Black’, which may be consistent with its superior antioxidant capacity (Fig. 6E). Pyroglutamic acid has been reported to be positively correlated with antioxidant activity in probiotic vegetable juices from four crop species (Aloo et al., 2023); glutamine, as an important dietary supplement, can effectively enhance antioxidant capacity mainly by serving as a key precursor for glutathione, a major endogenous antioxidant molecule (Mahdavifard et al., 2022); and γ-dodecalactone is mainly biosynthesized from oleic acid, which together with linoleic acid has inherent antioxidant activity (Małajowicz et al., 2025), which may jointly explain the high accumulation of these three metabolites in the high antioxidant cultivar ‘Black’. Given its excellent antioxidant performance, ‘Black’ may serve as a valuable functional germplasm resource for further utilization and genetic improvement.

Moreover, To further explore the potential health benefits of metabolites in Prunus for humans, network pharmacology was employed to identify the molecular targets and perform disease pathway enrichment analysis. The results revealed significant enrichment of the metabolites in pathways related to diabetes, cancer, Alzheimer's disease, and various neurological disorders (Fig. 6E). These diseases may be involved in multiple neurological disorders and major organ diseases. Prunus fruits contain high levels of sugars, which further explains the enrichment of metabolites in diabetes-related pathways. Consistent with our findings, a previous study of Prunus has demonstrated that Prunus is rich in phenols capable of significantly enhancing the antioxidant capacity, and these compounds may exert substantial regulatory effects on vascular diseases (Cioni et al., 2024). Sugars and phenols could be exert their regulatory effects through glucose metabolism, exhibiting potential in preventing and intervening diabetes. They may also prevent cardiovascular diseases by regulating blood pressure and lipid metabolism (Liu et al., 2023). Furthermore, some metabolites in Prunus fruits were highly enriched in cancer-related pathways. It has been demonstrated that cancer cells can sense and utilize signals from a wide range of metabolites to promote tumorigenesis and metastasis, including intermediates of central carbon metabolism, lipids, amino acids, and nucleotides (Wang et al., 2020). Research on black rice has demonstrated that the abundant unsaturated fatty acids, such as α-linolenic acid and 11,14-eicosadienoic acid, inhibit amyloid pathology in Alzheimer's disease mouse models via cell-type-specific allosteric activation, thereby improving brain cognitive function and extending lifespan (He et al., 2025).

4. Conclusion

This study systematically characterized the metabolic profiles and antioxidant properties of seven major Prunus cultivars in Xinjiang using UPLC-ESI-MS/MS and GC × GC-TOFMS. A total of 417 non-volatile and 382 volatile metabolites were identified, which were classified into four categories. Prunus cultivars exhibit considerable variations in their metabolite profiles and antioxidant capacities. Among these, phenols and esters were most abundant non-volatile and volatile metabolites, respectively, and consistently displayed preferential accumulation in the peel. Moreover, GP and GL are dominant lipids in Prunus fruits; and sorbitol and malic acid are primary sugar and acid. Among these cultivars, ‘Black’ exhibited two-fold higher antioxidant activity than the other cultivars, which may be strongly associated with its more than four-fold higher content of pyroglutamic acid, glutamine, and γ-dodecalactone. Network pharmacology analysis revealed that the metabolites in Prunus fruits are enriched in pathways associated with diabetes, cancer, and Alzheimer's disease, which nay highlight their potential health benefits. The integrated analytical approach established in this study provides a valuable methodological reference for fruit metabolomics research. This method boasts distinct merits: it achieves broader coverage of detectable metabolites, yields superior consistent extraction outcomes, and supports the unified holistic evaluation of fruit samples. Collectively, consistent with previous reports, this study clarifies the metabolic diversity and functional characteristics of different Xinjiang Prunus cultivars, offers guidance for consumer choices and cultivar selection, and lays a theoretical foundation for Prunus quality improvement and functional food development. This study only investigated seven Prunus cultivars cultivated in Xinjiang. Other Prunus germplasms from different producing regions remain to be systematically explored. Meanwhile, postharvest metabolic changes in Prunus fruits represent another promising topic for further investigation.

Authors contribution

L.Z. and P.W. conceived and designed the research, M.W. designed and performed most of the experiments and analyzed the data, Y.C., X.W., Y.W., Z.D. and Y.Z. assisted the experiments, Y.C., G.L., F.Z., and Y-J.C. prepared materials, M.W. and L.Z. wrote the manuscript draft, G.L., X.Q., F.Z., Y-J. C., P.W. and L.Z. finalized the writing and revision of the manuscript.

CRediT authorship contribution statement

Mengxia Wu: Writing – review & editing, Writing – original draft, Methodology, Data curation. Ying Cao: Methodology, Data curation. Xuening Wu: Methodology, Data curation. Yifan Wang: Methodology, Data curation. Zhiwei Deng: Data curation. Yirui Zhao: Data curation. Gang Lu: Writing – review & editing, Resources. Xiaolu Qu: Writing – review & editing. Feng Zhu: Writing – review & editing, Resources. Yunjiang Cheng: Writing – review & editing, Resources. Ping Wang: Writing – review & editing, Writing – original draft, Conceptualization. Linlin Zhong: Writing – review & editing, Writing – original draft, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Fig. S1.

Cluster analysis and Metabolite identification informationof different Prunus cultivars and tissues. A: Identification information of volatile and non-volatile metabolites detected by different chromatographic columns and detection methods; B: Principal component analysis of fruit peel and flesh tissues of seven Prunus cultivars; C: Hierarchical cluster analysis of fruit peel and flesh tissues of seven Prunus cultivars.

Fig. S2.

Analysis of primary metabolites and their pathways in fruits of seven Prunus culivars. A: Analysis of all primary metabolite types and their contents in fruits of seven Prunus culivars; B: Classification of high-content compounds in fruits of seven Prunus culivars; C: Pathway enrichment analysis of all primary metabolites in fruits of seven Prunus culivars.

Fig. S3.

Clustering and pathway enrichment analysis of secondary metabolites in seven Prunus cultivars. A: Clustering analysis of all secondary metabolites in seven Prunus cultivars; B: Bar chart of the proportion of high-abundance organic acids and esters in fruits of seven Prunus cultivars.

Fig. S4.

Proportional distribution of lipids types and pathway enrichment analysis across fruits of seven Prunus cultivars. A: Pie chart of the proportion of different lipid types in fruits of seven Prunus cultivars; B: Pathway analysis of all lipids across fruits of seven Prunus cultivars.

Fig. S5.

Correlation analysis of all metabolites detected in fruits of seven Prunus cultivars. A: Correlation analysis of primary metabolites and antioxidant indices in fruits of seven Prunus cultivars; B: Correlation analysis of secondary metabolites and antioxidant indices in fruits of seven Prunus cultivars; C: Correlation analysis of lipid metabolites and antioxidant indices in fruits of seven Prunus cultivars; D: Correlation analysis of volatile metabolites and antioxidant indices in fruits of seven Prunus cultivars.

Table S1

Identification and Classification Information of Non-Volatile(including primary metabolites, secondary metabolites and lipids) and Volatile Compounds in fruits of seven Prunus cultivars

Data availability

The data that has been used is confidential.