Abstract Akebia species including, A. trifoliata (AKT), A. trifoliata ssp. australis (AKA), and A. quinata (AKQ) are popular for their sweet, aromatic fruits and pharmacological applications. Despite their commercial and medicinal importance, the metabolomic profiles of Akebia pulp remain largely unexplored. This study employs UPLC-MS/MS and GC–MS techniques to comprehensively analyze the chemical composition of the pulp from these three species. Among the 1429 metabolites detected and putatively identified, terpenoids, amino acids, flavonoids, and phenolics were predominant. AKT shows the richest bioactive compound accumulation, including significant levels of L-Isoleucyl-l-Aspartate, 6-C-Methylquercetin-3-O-rutinoside, and Madasiatic acid, suggesting its high nutritional and medicinal value. AKA has the lowest metabolite accumulation, while AKQ displays an intermediate profile with notable antioxidants like Quercetin-3-O-xyloside. These findings provide a foundation for optimizing the cultivation, harvesting, and utilization of Akebia fruits and underscore their potential for developing functional foods and nutraceuticals. Keywords: Akebia, Metabolites, Pulp, Terpenoids, Flavonoids Highlights * • We detected and putatively identified 1429 metabolites in Akebia species using UPLC-MS/MS and GC–MS. * • trifoliata showed the highest bioactive compound accumulation. * • Showed species-specific metabolic adaptations in diverse environments. * • Found significant KEGG enrichments in flavonoid and terpenoid pathways. 1. Introduction Akebia, a genus within the Lardizabalaceae family, stands out for its distinctive fruiting attributes, which significantly contribute to both ecological variety and agricultural interests. Species like A. trifoliata, A. trifoliata ssp. australis, and A. quinata are cultivated not only for their pleasing, sweet, and fragrant fruits but also for their notable medicinal properties ([33]Jiang et al., 2012; L. [34]Liu & Qian, 2002; [35]Nazir et al., 2024), long recognized in traditional Chinese medicine ([36]Huang et al., 2022; [37]Li et al., 2021; X. [38]Liu et al., 2023; [39]Maciąg et al., 2021). These fruits are prized for their complex metabolic profiles, including a variety of bioactive compounds like flavonoids, polyphenols, and vitamins, which are advantageous for health, offering anti-inflammatory, antioxidant, and anticancer benefits ([40]Zou et al., 2023). Recent studies have highlighted Akebia's potential as a promising new fruit crop ([41]Zou et al., 2023; [42]Zou, Yao, Zhong, Zhao, & Huang, 2018), with attributes such as high unsaturated fatty acids in seeds up to 39 %, useful in producing edible oils, and pectin in fruit peels for commercial food applications. Despite their increasing popularity and commercial potential, there is a scarcity of comprehensive studies on their chemical composition and comparative metabolomic analyses, which are essential for maximizing their agricultural and therapeutic capacities. The pulp of the Akebia fruit is highly valued for its flavor and health benefits ([43]Jiang et al., 2012). However, the chemical profiles of the pulp across different species have not been extensively studied. Previous studies have primarily focused on isolated bioactive compounds, such as flavonoids, saponins, and alkaloids, but a comprehensive metabolomic approach to examine the full range of metabolites in these fruits has been lacking ([44]Li et al., 2021). Metabolomics, utilizing advanced techniques like mass spectrometry and nuclear magnetic resonance (NMR), provides an effective tool to analyze the entire spectrum of metabolites in plant tissues, offering a more holistic view of their biochemical composition ([45]Deborde et al., 2017; [46]Eisenreich & Bacher, 2007; [47]Emwas et al., 2019). Moreover, recent genomic studies have enhanced our understanding of the molecular basis of Akebia species' traits. The chromosome-level genome of A. trifoliata has been sequenced, revealing valuable insights into the genetic makeup of this species and potentially guiding breeding efforts for improving fruit quality and medicinal properties ([48]Zhong et al., 2022). Such genomic advancements pave the way for more targeted metabolomic analyses, enabling a deeper understanding of how genetic variation influences the chemical profiles of Akebia fruit pulp. Metabolomics, the study of small molecules involved in cellular processes, offers a powerful approach to investigate the full spectrum of metabolites present in biological samples ([49]Dona et al., 2016). By employing techniques like mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectroscopy ([50]Dias et al., 2016; [51]Takis, Ghini, Tenori, Turano, & Luchinat, 2019), it is possible to gain a detailed and comprehensive understanding of the chemical composition of fruit pulp, which could significantly inform agricultural practices, enhance product development, and aid in the identification of bioactive compounds for potential therapeutic applications. Despite this, no study has yet explored the comparative metabolomic profile of Akebia pulp from these three important species. This research aims to fill this gap by conducting a thorough metabolomic analysis of the pulp from A. trifoliata (AKT), A. trifoliata ssp. australis (AKA), and A. quinata (AKQ) by utilizing advance chromatographic and mass spectrometric techniques. By identifying and comparing the metabolite compositions across these species, we aim to highlight their chemical diversity and provide a deeper understanding of the factors influencing their commercial and medicinal value. Such insights will contribute to optimizing the cultivation, harvesting, and utilization of Akebia fruits, paving the way for future studies on their nutritional and pharmacological potentials. 2. Materials and methods 2.1. Sample collection Pulp samples from three species of Akebia—A. trifoliata (AKT), A. trifoliata ssp. australis (AKA), and A. quinata (AKQ)—were collected at Lushan botanical Garden, Nanchang, China, grown under similar environmental conditions. The plants were established concurrently in 2022 and grown under identical field conditions, allowing for open pollination. Fruits were harvested in September 2023 at a consistent ripening stage. This stage was defined physiologically as the initial appearance of natural longitudinal cracking along the ventral suture, which typically occurs upon ripening in this genus (S.-Y. Zou, C. Feng, P.-X. Gao, T.-J. Li, T.-J. Jia, & H. Huang, 2023). For each species, nine representative fruits exhibiting this specific ripeness criterion were collected. The pulp tissue was carefully excised from these fruits. To prepare biological replicates, pulp material from three individual fruits was pooled. This resulted in a total of three independent biological replicates for each of the three species (n = 3 per group) used for subsequent metabolomic analysis. Fresh samples were immediately flash-frozen in liquid nitrogen and stored at −80 °C until further analysis. Frozen pulp samples were ground to a fine powder using a grinder (MM 400, Retsch) operated at 30 Hz for 30 s, with the grinding jar and steel beads pre-cooled in liquid nitrogen. Precisely 50 mg of the resulting powder was weighed using an electronic balance (MS105DΜ). To each weighed sample, 600 μL of pre-cooled 70 % methanol:water (v/v) solution, containing internal standard extract, was added. The mixture was vortexed vigorously for 30 s, repeated 6 times in total, and then allowed to stand at −20 °C for 30 min. Following incubation, the samples were centrifuged at 12,000 rpm for 3 min. The resulting supernatant was carefully aspirated and filtered through a 0.22 μm microporous membrane filter into an injection vial. The filtered extracts were then stored appropriately (e.g., at −80 °C or prepared for immediate analysis) prior to UPLC-MS/MS analysis. 2.2. Metabolite profiling The comprehensive metabolite profiling using UPLC-MS/MS and GC–MS platforms was carried out as a service by MetWare Biotechnology Co., Ltd. (Wuhan, China). Procedural blanks were included during sample preparation, and Quality Control (QC) samples (prepared by pooling aliquots of all extracts) were injected periodically (e.g., every 10 samples) throughout the analytical runs to monitor instrument stability, assess data quality, and check for contamination. 2.2.1. UPLC-MS/MS analysis Sample extracts were analyzed using a UPLC-ESI-MS/MS system comprising an ExionLC™ AD UPLC (SCIEX, Framingham, MA, USA) coupled to an Applied Biosystems 6500 Q TRAP MS/MS (SCIEX). Separation was performed on an Agilent SB-C18 column (1.8 μm, 2.1 mm * 100 mm). The mobile phase consisted of solvent A (pure water with 0.1 % formic acid) and solvent B (acetonitrile with 0.1 % formic acid). The gradient elution program was as follows: 5 % B maintained for 0–9 min with a linear increase to 95 % B, 95 % B held for 1 min (9–10 min), returned to 5 % B within 1.1 min (10–11.1 min), and equilibrated at 5 % B for 2.9 min (11.1–14 min). The flow rate was 0.35 mL/min, the column oven temperature was maintained at 40 °C, and the injection volume was 2 μL. The ESI source was operated in both positive and negative ion modes. Key parameters were: source temperature 500 °C, ion spray voltage (IS) +5500 V (positive) / -4500 V (negative), ion source gas I (GSI) 50 psi, gas II (GSII) 60 psi, and curtain gas (CUR) 25 psi. Collision-activated dissociation (CAD) gas was set to medium. Metabolite quantification was performed in Multiple Reaction Monitoring (MRM) mode using optimized declustering potential (DP) and collision energy (CE) for each precursor-product ion transition. The system monitored specific MRM transitions for metabolites based on the MetWare database (MWDB), which contains retention times and MS/MS spectra (secondary spectra) for numerous compounds, including common adducts (e.g., [M + H]+, [M + Na]+, [M + NH4]+, [M-H]-). Operating in MRM mode enhances precision, reproducibility, and sensitivity by selectively monitoring specific ion transitions. 2.2.2. GC–MS analysis Samples (500 mg powder in 20 mL headspace vial with saturated NaCl) were incubated at 60 °C for 5 min with continuous agitation (CTC Analytics Agitator). Volatiles were extracted using a 120 μm DVB/CWR/PDMS fiber (Agilent Technologies, Santa Clara, CA, USA) exposed to the headspace for 15 min at 60 °C, performed using a CTC Analytics SPME autosampler. GC–MS was conducted on an Agilent Model 8890 GC coupled to an Agilent 7000D mass spectrometer. The extracted volatiles were desorbed from the fiber in the GC injection port (250 °C) for 5 min in splitless mode. Separation was achieved on a DB-5MS capillary column (30 m × 0.25 mm × 0.25 μm, 5 % phenyl-polymethylsiloxane; Agilent J&W Scientific). Helium was used as the carrier gas at a constant linear velocity of 1.2 mL/min. The oven temperature program was: initial 40 °C held for 3.5 min, ramped at 10 °C/min to 100 °C, then at 7 °C/min to 180 °C, then at 25 °C/min to 280 °C, and held at 280 °C for 5 min. Electron ionization (EI) was performed at 70 eV. Mass spectra were recorded across a specified m/z range (e.g., 50–500 amu). The ion source, quadrupole mass detector, and transfer line temperatures were maintained at 230 °C, 150 °C, and 280 °C, respectively. Selected Ion Monitoring (SIM) mode was utilized for enhanced sensitivity and selectivity in the identification and quantification of target analytes where applicable. 2.3. Data processing and analysis Raw UPLC-MS/MS data files were processed using SCIEX Analyst software (v1.6.3). MRM peak areas were extracted and integrated using MultiQuant software (v3.0.3, SCIEX). Raw GC–MS data were processed using Agilent MassHunter Workstation software (v10.0). Peak detection, integration, and spectral deconvolution (where necessary) were performed. The areas of the peaks were normalized against internal standard (3-Hexanone-2,2,4,4-d4), and metabolite identification was based on the comparison of retention times and mass spectra with known libraries. Multivariate analysis was performed using R software (v.4.1.2, [52]www.r-project.org) to visualize clustering and differential accumulation patterns. Principal Component Analysis (PCA) was performed using the prcomp function from the R base stats package after Unit Variance scaling. Pearson correlation coefficients (PCC) were calculated between samples using the “cor” function in R to assess reproducibility within groups and distinctness between groups. Correlation heatmaps were generated using the ComplexHeatmap R package (v2.9.4). 2.4. Metabolite identification and classification Metabolite identification and annotation were performed using stringent criteria for both UPLC-MS/MS and GC–MS data, primarily leveraging the proprietary MetWare database (MWDB, MetWare Biotechnology Co., Ltd.) alongside publicly available databases and spectral libraries. 2.4.1. UPLC-MS/MS metabolite identification Non-volatile metabolites detected via UPLC-MS/MS were putatively identified (corresponding primarily to Metabolomics Standards Initiative [MSI] Level 2 or 3) by matching their characteristics against the MWDB database. The criteria included: Precursor Ion Mass: Matching the accurate m/z of the precursor ion within a specified tolerance (e.g., ≤ ±10 ppm). Retention Time (RT): Matching the chromatographic retention time within a narrow window (e.g., ± 0.1 min) compared to the database entry under identical analytical conditions. MS/MS Fragmentation Pattern: Comparing the acquired tandem mass spectra (MS/MS or secondary spectra) with the reference spectra in the MWDB, requiring a high similarity score (e.g., ≥ 0.7). Adduct information (e.g., [M + H]+, [M-H]-, [M + Na]+) was also considered. Multiple Reaction Monitoring (MRM) mode was employed for the quantification of selected metabolites. Peak areas were integrated from the resulting Extracted Ion Chromatograms (XICs Fig. S1), which were generated based on precursor-product ion transitions defined using the MetWare database (MWDB) and potentially confirmed with standards (corresponding to MSI Level 1 confidence for confirmed analytes). 2.4.2. GC–MS metabolite identification Volatile metabolites detected via GC–MS were putatively identified (MSI Level 2/3) based on: Mass Spectrum Matching: Comparing the acquired Electron Ionization (EI) mass spectrum against reference spectra in standard public libraries such as the National Institute of Standards and Technology (NIST) database and potentially the Fiehn RI library or other relevant mass spectral libraries. A high spectral similarity score (e.g., > 700 out of 1000) was typically required. Retention Index (RI): Comparing the experimental Kovats retention index (RI), calculated relative to a homologous series of n-alkanes, with RI values reported in the referenced libraries. A match within a defined tolerance window (± 20 RI units) was considered acceptable. 2.5. Differential accumulation of metabolites Differentially accumulated metabolites (DAMs) between the Akebia species (AKT, AKA, and AKQ) were identified through pairwise comparisons using Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA), a supervised multivariate statistical method. This analysis was performed using the R package ‘MetaboAnalystR’ (v1.0.1). Metabolites were considered significantly differentially accumulated if they met two criteria derived from the OPLS-DA model: a Variable Importance in Projection (VIP) score greater than 1, and an absolute Log[2] Fold Change (|Log2FC|) of 1.0 or greater, ensuring a substantial difference in abundance. Differentially accumulated metabolites were then mapped to metabolic pathways using KEGG pathway enrichment analysis, providing insight into the biological processes most impacted by these metabolic changes. The Differential Abundance Score (DA Score) was calculated for each pathway, reflecting the overall upregulation or downregulation of metabolites within that pathway. Pathway analysis further helped identify key metabolic networks that were distinctly regulated in each species, highlighting the unique metabolic profiles of AKT, AKA, and AKQ. Data quality was controlled using QC samples to ensure that only reliable metabolites were included in the analysis. 3. Results 3.1. Overview of metabolic profiling The metabolic profiling of three Akebia species—AKT, AKA, and AKQ—reveals significant differences in their metabolite compositions (Table S1). Class distribution of the 1429 compounds (1302 identified through UPLC-MS/MS and 126 through GC/MS) was identified showed that terpenoids (19.79 %) were the most abundant class, followed by amino acids and derivatives (8.39 %), flavonoids (8.16 %), and lipids (8.25 %). Other prominent classes included alkaloids (7.44 %), phenolic acids (7.21 %), and nucleotides and derivatives (3.38 %) ([53]Fig. 1A). Principal Component Analysis (PCA) was performed to assess the overall variation in the metabolic profiles of the three species. The PCA plot ([54]Fig. 1B and Fig. S2) clearly demonstrated that the metabolic profiles of the species were distinct, with samples from each species clustering separately along the first two principal components (PC1: 39.98 %, PC2: 38.04 %). This suggests that each species exhibits a unique metabolic signature, although variability was observed within each species. Furthermore, a correlation plot ([55]Fig. 1C) was generated to examine the relationships between metabolites within the samples. Strong positive correlations were observed among metabolites within each species, with correlation coefficients approaching 1. This indicates that each species follows a consistent metabolic pattern, while the inter-species correlations were lower, further supporting the distinct metabolic profiles of the three Akebia species. Fig. 1. [56]Fig. 1 [57]Open in a new tab Metabolic Profiling of Akebia Species. A) Class distribution of identified metabolites in the pulp of A. trifoliata (AKT), A. trifoliata ssp. australis (AKA), and A. quinata (AKQ). The most abundant classes include terpenoids (19.79 %), amino acids and derivatives (8.39 %), flavonoids (8.16 %), and lipids (8.25 %), among others. B) Principal Component Analysis (PCA) of the metabolic profiles of the three Akebia species. The first two principal components (PC1: 39.98 %, PC2: 38.04 %) clearly separate the species, indicating distinct metabolic profiles for each species. C) Correlation plot showing the relationships between metabolites within the samples. Strong positive correlations are observed within each species, suggesting consistent metabolic patterns within species, while inter-species correlations are lower, supporting the uniqueness of each species' metabolic profile. 3.2. Differential accumulation of metabolites The analysis of differentially accumulated metabolites (DAMs) across AKA, AKT, and AKQ revealed significant metabolic differences among the species ([58]Fig. 2, [59]Fig. 3). In the AKA vs. AKT comparison, a total of 1012 DAMs were identified, with 597 downregulated and 415 upregulated in AKA compared to AKT ([60]Fig. 2A). Similarly, in the AKQ vs. AKA comparison, 976 DAMs were detected, of which 377 were downregulated and 599 were upregulated in AKQ relative to AKA. In the AKQ vs. AKT comparison, 960 DAMs were identified, with an almost equal distribution of 481 downregulated and 479 upregulated metabolites in AKQ compared to AKT. These results highlight the significant biochemical diversity among the three Akebia species and reflect differences in their metabolic pathways and ecological adaptations. Fig. 2. [61]Fig. 2 [62]Open in a new tab Differentially Accumulated Metabolites (DAMs) across AKA, AKT, and AKQ Comparisons. A) Bar plot showing the number of DAMs identified in each pairwise comparison: AKA vs. AKT, AKQ vs. AKA, and AKQ vs. AKT. Blue bars represent the total number of DAMs, red bars indicate downregulated metabolites, and green bars represent upregulated metabolites in the respective comparisons. B) Venn diagram illustrating the overlap of DAMs among the three pairwise comparisons. Shared and unique metabolites are displayed for AKA vs. AKT (blue), AKQ vs. AKA (purple), and AKQ vs. AKT (yellow). C) Circular chart depicting the percentage distribution of common DAMs by their chemical categories, with Terpenoids (31.4 %) being the most abundant, followed by Flavonoids (11.7 %), Phenolic acids (10.0 %), and other classes such as Amino acids, Lipids, and Alkaloids. The chart highlights the diverse chemical nature of the DAMs across the comparisons. (For interpretation of the references to color in this figure legend, the reader is