From waste to wellness: pomelo peel derived exosome-like nanoparticles reshape gut microbiota and enhance beneficial metabolites Authors: Liyi Li, Chaofan Chen, Tiantian Yang, Ran Tuo, Yingjie Huang, Xiaodan Zhao, Le Chu, Fatao He, Ping Li, Mingdong Lu Published: 2026 ABSTRACT AbstractPomelo processing by-products, traditionally regarded as agricultural waste, are underutilized bioactive-rich candidate feedstocks for circular economy. This study optimized isolation protocol (differential centrifugation with PEG precipitation) for Pomelo Peel Exosome-Like Nanoparticles (PPELNs), uniform nanovesicles (195.8 ± 78.6 nm) with gastrointestinal stability. Proteomic analysis identified 1145 proteins enriched in carbohydrate, energy, amino-acid metabolism and stimulus response. In vitro human fecal fermentation model demonstrated PPELNs intervention altered gut microbiota, increased beneficial taxa (Bifidobacterium, Blautia, Phascolarctobacterium) and suppressed Escherichia-Shigella, suggesting potential prebiotic effects. Metabolomic profiling revealed elevated bioactive compounds (naringenin 7-rhamnoglucoside, auraptenol). PPELNs fermentation enhances short-chain fatty acids (SCFAs), including acetic, propionic, butanoic and valeric. Further analysis showed microbiota-metabolite correlations. Propionic, butyric and auraptenol positively correlated with Blautia but negatively with Escherichia-Shigella; naringenin 7-rhamnoglucoside positively correlated with Phascolarctobacterium and Ruminococcus. These findings upcycle pomelo peel into a high-value functional ingredient, provide the first evidence establishing specific metabolite-microbe correlations.Keywords: PPELNs, Gut microbiota, Fecal slurry fermentation, Metabolomics Introduction Plant exosome-like nanoparticles (PELNs) are heterogeneous nano-sized plant-derived vesicles isolated from botanical sources including pomelo, ginger, and lemons (Tembo et al., 2025). They offer advantages for scalable production without the risk of animal pathogen transmission. They possess inherent low immunogenicity, biocompatibility, cell-targeting capabilities, and enhanced penetration of biological barriers (D. Liu et al., 2025). PELNs have garnered attention for demonstrating multiple health benefits, including regulating the gut microbiome and balance, stimulating intestinal cellular renewal, and exhibiting tumor-suppressive and inflammation-inhibiting properties (Zhou et al., 2023). Recent investigations suggest that PELNs hold therapeutic potential across diverse conditions including autoimmune diseases (Han, Zhou, et al., 2025), dermatological disorders (Huang et al., 2024), hepatic pathologies (Huang et al., 2024; Zu et al., 2024), intestinal diseases (Yang et al., 2025), and metabolic disorders (Sundaram et al., 2022). Gut microbiota balance is crucial for nutrient absorption and digestion, as well as for the production and metabolism of bioactive substances. Certain PELNs exhibit favourable biological stability within the gut, modulating microbiota structure and activity, ultimately influencing host health and physiological functions. For instance, exosome-like nanovesicles derived from garlic may shield colonic tissue from DSS damage via TLR4/MyD88/NF-κB pathway suppression and gut microbiota regulation (Zhu et al., 2023). Exosome-like nanovesicles (ELNs) extracted from ginger harness their phospholipid acid content and specific microRNA (mdo-miR7267-3p) to enhance intestinal barrier function, reduce inflammatory responses in intestinal tissue, and optimise gut microbiota composition (Teng et al., 2018). Mulberry-bark sourced nanoparticles regulate the balance of gut immunity and integrity by activating the HSPA8 signalling pathway (Sriwastva et al., 2022). Pomelo (Citrus maxima), the largest citrus fruit, has a peel accounting for 30–50% of its fresh weight (Wei et al., 2025). However, between 2.8 and 4.7 million tonnes of pomelo peel waste are discarded annually during global processing operations. This exacerbates environmental pollution and represents a significant waste of biomass. Consequently, isolating PPELNs from pomelo peel offers a viable approach to upcycle this discarded resource. Despite the abundant resources of pomelo peel and the urgent need for its high-value utilization, research on PELNs derived from pomelo peel remains largely unexplored. Most studies on citrus exosome-like nanoparticles have used juice or pulp from lemons, grapefruit, sweet oranges and similar fruits, demonstrating promising applications in anti-inflammatory effects (Bruno et al., 2021), gut function regulation (Berger et al., 2020), among others. Research using only fruit peels as raw materials remains relatively limited, with existing literature focusing solely on the anti-inflammatory effects of tangerine peel. Suresh (Suresh et al., 2021) recently introduced a PEG-6000 precipitation method for isolating ginger-derived nanoparticles (PDNP), achieving high yields without ultracentrifugation while maintaining the integrity and functionality of critical molecular constituents. Therefore, this study employed a combined PEG and differential centrifugation approach to isolate and characterize ELNs (PPELNs) from pomelo peel, thereby achieving the reuse of waste resources. Their resistance to digestion was investigated using an in vitro simulated digestion system, while the gut microbiota-modulating capacity of PPELNs from the peel was explored using an in vitro fecal slurry fermentation model. Materials and methods Reagents Polyethylene glycol 6000 (PEG-6000) was procured from McLean Chemical Reagents Co. (Shanghai, China). Experimental water consisted of Milli-Q ultrapure water, with all other chemicals being of analytical grade. Preparation of PPELNs Fresh pomelos were purchased at a local supermarket in Hangzhou, China. Place the washed pomelo peel in an appropriate volume of PBS solution for soaking, then crushed with a juicer. Filter the resulting juice through four layers of gauze. Centrifuge at 6000g for 40 min at 4 °C, discard the pellet, then continue centrifuging at 10000g for 60 min at 4 °C. Combine the supernatant with a prepared 16% PEG solution in a 1:1 ratio, mix thoroughly, and incubate on a shaking incubator at 4 °C for 15 h. Centrifuge the mixture at 10,000g for 45 min at 4 °C. Resuspend the formed pellet in PBS, centrifuge at 1000g for 20 min at 4 °C and retain the supernatant. Sonicate at 4 °C for 30 min. Finally, filter the PPELNs solution using a 0.22 μm aqueous membrane. Store PPELNs samples at −80 °C for future use. Detection of particle size distribution, concentration and zeta potential of PPELNs The particle size distribution and concentration of PPELNs were determined using nanoparticle tracking analysis. This instrument tracks and analyses the Brownian motion of each particle, combining this with the Stokes-Einstein equation to determine nanoparticle hydrodynamic diameter and concentration. An appropriate volume of PPELNs solution was diluted to a suitable multiple prior to testing. Subsequently, 1 mL of the diluted solution was drawn into a syringe and slowly injected into the sample cell. Add the PPELNs solution to the ZetaView sample cell. The instrument applies an electric field within the cell, causing the charged nanoparticles to undergo electrophoretic motion under the influence of this field. Through laser scattering or video microscopy techniques, the instrument tracks the movement trajectories of individual particles and measures their electrophoretic mobility. Finally, based on the electrophoretic mobility and relevant physical parameters, the Zeta potential values of nanoparticles in the sample are calculated using the Henry equation, thereby determining the distribution of Zeta potentials. Electron microscopy examination of PPELNs PPELNs samples were fixed using glutaraldehyde and rose bengal solution, incubated in the fixative for 10–30 min; Dehydrate using 100% acetone. Following dehydration, embed the PPELNs in epoxy resin. Section the embedded material using an ultramicrotome. After sectioning, place the sections onto electron microscope grids and perform negative staining with uranyl acetate and lead citrate. Observe the morphology and size of the PPELNs using a transmission electron microscope and record the images. Proteomic analysis of PPELNs Protein extraction was performed on the obtained PPELNs samples. Following lyophilisation, an appropriate volume of SDT lysis buffer was added to each sample before transfer to Eppendorf tubes. Subsequently, samples were immersed in heated water for 3 min, followed by ultrasonic disruption for 2 min. Centrifugation at 16,000g for 20 min at 4 °C yielded the supernatant, which was subjected to protein quantification via the BCA assay. Take 15 μg of protein sample, add 5× loading buffer at a 5:1 (volume ratio), heat again in a boiling water bath for 300 s, then perform SDS-PAGE electrophoresis on an 8%–16% gel, staining with Coomassie Brilliant Blue. Subsequently, the extracted proteins underwent enzymatic digestion. The resulting peptide fractions were desalted, and their concentrations were measured for subsequent LC-MS analysis. An appropriate volume of peptide sample was analyzed via Vanquish Neo UHPLC chromatography. Finally, the mass spectrometry results were consolidated with DIA-NN software to facilitate database searching of DIA mass spectrometry data and quantitative analysis of proteins via DIA. In vitro digestion of PPELNs In vitro digestion protocols followed established methods (M. Zhang et al., 2016). To the purified PPELNs sample, add 1.34 μL of 18.5% (w/v) HCl (pH 2.0) and 24 μL of pepsin solution (80 mg/mL, 0.1 N HCl, pH 2.0) to the purified PPELNs sample. Incubate on a shaking water bath at 37 °C for 30 min to simulate gastric digestion conditions. Dilute an appropriate volume of the digested sample to a suitable concentration for particle size and zeta potential analysis. To the remaining sample, add 80 μL of 0.1 M NaHCO₃ solution including 24 mg/mL bile salts and 4 mg/mL pancreatic enzymes. Adjust the pH to 6.5 using 1 M NaHCO₃ to simulate the intestinal digestive environment. Incubate for a further 30 min under identical conditions. Again, take an appropriate amount of the digested sample and dilute to a suitable concentration for particle size and potential analysis. The effect of PPELNs on in vitro simulations of the human gut microbiome Fecal slurry fermentation model Fresh morning fecal samples were collected from six healthy young adult volunteers (three males and three females, aged 22–26 years) in Hangzhou, China. All subjects had no history of gastrointestinal diseases, had not taken probiotics within the past six months or antibiotics within the recent three months. They all followed a regular Chinese dietary pattern, characterized by high carbohydrate and plant-based food consumption with moderate protein and fat intake, without any special dietary restrictions such as vegetarianism. In addition, their dietary patterns had remained stable for at least four weeks prior to sample collection. This study has been approved by the Ethics Committee of Zhejiang Gongshang University (Ethics Committee Agreement Number: 250712-16). All volunteers participated with informed consent. Samples were collected using disposable sterile applicators. Precise 1.0 g portions of fecal matter were weighed and immediately transferred to containers holding 10 mL sterile PBS solution (pH 6.8), where they were thoroughly mixed. Subsequently, the mixture was passed through three layers of gauze to yield the healthy human fecal bacterial suspension. Transfer this to a new sterile test tube and store within an anaerobic workstation. Prior to fermentation, 16S rRNA gene sequencing was performed on fecal samples from six healthy volunteers to analyze the baseline gut microbiota composition. At the phylum level, the baseline microbiota was dominated by Bacillota, Bacteroidota, and Actinomycetota. At the genus level, the core microbiota included Bacteroides, Blautia, and Faecalibacterium. The preparation of the basal nutrient medium follows the method described by Ding Y (Ding et al., 2019). Its composition comprises: 2 g peptone, 2 g yeast extract, 0.1 g NaCl, 0.04 g K₂HPO₄, 0.04 g KH₂PO₄, 0.01 g MgSO₄·7H₂O, 0.01 g CaCl₂·6H₂O, 2 g NaHCO₃, 2 mL Tween 80, 0.02 g haem, 10 μL vitamin K1, 0.5 g bile salts, 0.5 g cysteine. This mixture is reconstituted using pH 6.8 sterile phosphate-buffered saline to prepare the basal nutrient medium (1 L). Add the supernatant from fecal samples to basal nutrient growth medium and conduct anaerobic fermentation at 37 °C for 24 h. The control group (Ctrl) comprised 13.5 mL medium and 1.5 mL fecal bacterial suspension; the low-dose pomelo peel exosome group (PPELNs-L) comprised 13.5 mL medium, 1.5 mL fecal bacterial suspension, and 162 μL PPELNs; The high-dose pomelo peel exosome group (PPELNs-H) comprised 13.5 mL medium, 1.5 mL fecal bacterial suspension, and 216 μL PPELNs. The low and high doses of PPELNs were 162 μL and 216 μL, accounting for 1.2% and 1.6% of the fermentation medium (13.5 mL), respectively. Following fermentation, samples were collected from each group for subsequent analysis. The Ori group consisted of the unfermented fecal bacterial suspension, serving as a baseline control to establish a reference for the initial microbial composition, thereby distinguishing between changes in the microbial community induced by the fermentation process and the specific regulatory effects of PPELNs. DNA extraction and 16S rRNA sequencing Total genomic DNA from the microbial community was found in the fecal fermentation samples using the QIAamp DNA Stool Mini Kit, adhering to the manufacturer's protocol. Determine quantification and assessment of DNA purity using the NanoDrop2000 (Thermo Scientific, USA). The DNA served as the template for PCR amplification focused on the V3-V4 hypervariable region of the 16S rRNA gene. Amplification products were separated by 2% agarose gel electrophoresis, processed via a DNA gel purification kit (PCR Clean-Up Kit, China Yuhua), and subjected to quantitative analysis. Library-prepared PCR isolates were generated with the NEXTFLEX® Rapid DNA-Seq Kit and analyzed via sequencing on the Illumina NextSeq 2000 platform (Shanghai Majorbio Bio-pharm Technology Co., Ltd). All data processing and figure generation completed on the Meiji Bio Cloud Platform. LC-MS untargeted metabolomics Aspirate 200 μL of the fecal fermentation sample into a 1.5 mL centrifuge tube. Add 400 μL extraction solvent. Vigorously mix the solution by vortexing for half a minute, followed by a 30-min low-temperature ultrasonic extraction at 5 °C and 40 kHz. Allow the sample to settle at −20 °C for 30 min, then at 4 °C. centrifuge at 13,000 ×g for 15 min. Transfer the supernatant, evaporate to dryness using nitrogen gas, redissolve in 100 μL of redissolution solvent, and perform low-temperature ultrasonic extraction for 5 min. Spin at 4 °C and 13,000g for 5 min, then pipette the supernatant into an analytical vial equipped with an insert tube. Samples were evaluated using UHPLC coupled with tandem FT ion-pair MS. The collected LC-MS data were analyzed using Progenesis QI metabolomics software, to generate a data matrix. Concurrently, MS and MS/MS spectral data were cross-referenced against public metabolite databases (HMDB, Metlin) and Meiji's proprietary repository to derive metabolite identifications. The final matched data was then transferred to the Meiji Cloud Platform for sophisticated analysis. The MS/MS spectra were matched against the HMDB and Metlin databases. Only metabolites with an MS/MS fragmentation score > 70 and classified as Level 2 (putatively annotated compounds) or higher according to the Metabolomics Standards Initiative (MSI) were retained for further analysis. Determination of short-chain fatty acids (SCFAs) Quantitative analysis of SCFAs in fermented samples was conducted using gas chromatography, with analysis conducted by Shanghai Majorbio Bio-pharm Technology Co., Ltd. Acetic, propionic, butyric, isobutyric, valeric and isovaleric acid served as standard reference materials. 500 μL filtered supernatant was combined with 100 μL crotonic acid. A 1 μL sample was injected onto a DB-WAX capillary column (Agilent Technologies, USA). The analysis programme was as follows: initial column temperature of 60 °C, ramped at 30 °C/min to 180 °C and held for 2 min; subsequently increased at 40 °C/min to 210 °C and maintained for 3 min. Sample volume 1 μL, detector at 230 °C, column flow 1 mL/min. Statistical analysis Results are shown as mean ± SD. Statistical analysis was conducted using GraphPad Prism. All sample analyses should be conducted at least three times. For determining significance, SPSS 19.0 was employed, considering a p-value below 0.05 as statistically significant. For multiple hypothesis testing, the false discovery rate (FDR) was controlled using the Benjamini-Hochberg procedure, and only metabolites with adjusted p-values (q < 0.05) were considered statistically significant. To validate the robustness of the PLS-DA model, a permutation test (n = 200) was performed, and the model was considered valid when permuted Q2 values did not exceed the original model's Q2. Results and discussion Characterization of PPELNs Characterization of PPELNs isolated and purified from pomelo peel. TEM images (Fig. 1B) and NTA were employed to assess PPELNs integrity, particle count, and size distribution (Fig. 1C). Results indicated that PPELNs exhibit a saucer-like structure comprising spherical particles with a distinct bilayer lipid bilayer configuration. The TEM images obtained in this study align with those reported by Jiaqi Wu (Wu et al., 2022), where PPELNs were similarly characterized as circular with a bilayer structure. The purified PPELNs averaged 195.8 ± 78.6 nm in diameter. The particle concentration was determined to be 5.9 × 109 particles/mL. Zeta potential analysis indicated that the PPELNs surface carries a negative charge, measured at −44.70 ± 0.53 mV (Fig. 1D). This substantial negative charge was crucial for the stability of the PPELNs, enabling them to resist aggregation reactions and reduce the likelihood of clearance, thereby prolonging their circulation time within the body. Isolation and characterization of PPELNs. (A) Extraction steps for PPELNs. (B) Transmission electron microscopy image of PPELNs (scale bar: 200 nm). (C) Particle size distribution analysis of PPELNs. (D) Zeta potential. SDS-PAGE results revealed (Fig. 2A) that PPELNs contained numerous proteins of varying molecular weights, predominantly centred around 25 kDa. A total of 1145 proteins were identified through proteomic analysis. The summary statistics for the proteins and peptides obtained through proteomic analysis are presented in Table 1. Subcellular localization analysis indicated that 381 proteins were localized to the cytoplasm, while 35 proteins were localized to the plasma membrane (Fig. 2B). Proteomics analysis. (A) SDS-PAGE electrophoresis results. (B) The total protein subcellular location identified in PPELNs. (C) GO annotation results. (D) KEGG pathway enrichment bar chart. Table 1Statistical analysis of DIA protein identification results.Sample nameNumber of proteinsUnique peptide segmentPPELNs-110683590PPELNs-210883559PPELNs-310093331Total11453768 Statistical analysis of DIA protein identification results. GO functional analysis serves as an effective tool for observing the functions of relevant proteins and exploring their biological significance. GO enrichment analysis is primarily categorized into biological processes, cellular processes, and molecular functions. GO analysis results indicate that PPELNs play significant roles in cellular processes, biological regulation, response to stimuli, localization, cellular anatomical structure, protein-containing complexes, catalytic, antioxidant, ATP-dependent, binding, molecular function regulatory and transporter activity (Fig. 2C). Specifically, when the gut is invaded by pathogens, intestinal cells initiate a series of defense mechanisms via response to stimulus to maintain gut health. These mechanisms involve interactions between biological regulation, cellular activities, responses to external stimuli, signal transduction, regulation of biological processes, and regulation of molecular functions, collectively participating in influencing and sustaining the equilibrium of the immune system within the organism. Moreover, KEGG pathway enrichment analysis revealed functional activities within protein metabolic pathways (Fig. 2D). Results indicated that the primary pathways involved were carbohydrate, energy, amino acid, lipid, cofactors and vitamins, glycan biosynthesis and metabolism, transport and catabolism, and other metabolic pathways. Notably, carbohydrate metabolism constitutes the core biochemical foundation of life processes. Through multi-step coordination, it converts dietary sugars into usable energy (ATP) and biomolecular precursors while dynamically regulating blood glucose homeostasis. Gut microbiota influence host energy extraction and inflammatory states via their carbohydrate metabolic activities, thereby affecting insulin resistance (Takeuchi et al., 2023). The carbohydrate metabolic capacity of gut microbiota alters with age, leading to diminished short-chain fatty acid (SCFA) levels. This, in turn, affects host expression of GPR41 and GPR43 receptors, ultimately influencing inflammatory states (You et al., 2022). Consequently, these KEGG pathways are intrinsically linked to metabolic processes and immune regulation. Collectively, they form the metabolic networks of cells and organisms, holding significant value for understanding disease pathogenesis and identifying potential therapeutic targets. Effects of PPELNs on gastrointestinal digestion The stability of PPELNs in an in vitro digestive environment was evaluated using a simulated gastrointestinal digestive system. Results indicated that after incubation in gastrointestinal fluids for 30 min, PPELNs exhibited no significant change in particle size compared to the initial digestion stage, demonstrating good digestive stability (Fig. 3). The lipid bilayer membrane structure of exosomes provided physical protection, enabling them to withstand exposure to gastric acid and digestive enzymes while maintaining their structure and function during digestion. Pomelo peel itself is rich in antioxidant components such as vitamin C and flavonoids. These constituents may exert antioxidant effects during digestion, shielding exosomes from oxidative damage and thereby preserving their particle size stability. Takakura H (Takakura et al., 2022) found that exosome-like nanoparticles derived from lemon and sweet orange exhibited no significant change in particle size within gastrointestinal digestive fluids, consistent with the findings of the present study. The stability of PPELNs against simulated gastrointestinal conditions was evaluated. PPELNs were dispersed in PBS (left panel); PPELNs were cultivated in simulated gastric fluid at 37 °C for 30 min (middle panel); PPELNs were first incubated in simulated gastric fluid at 37 °C for 30 min, then incubated in simulated intestinal fluid for 30 min (right panel). Particle size and zeta potential of PPELNs were measured (n = 3). The reduction in zeta potential (Fig. 3) may reflect alterations in particle surface charge, potentially arising from factors such as pH changes, variations in ionic strength. pH shifts in gastrointestinal fluids (gastric fluid being more acidic, intestinal fluid more alkaline) could influence the distribution of surface charges on particles, thereby causing changes in zeta potential. Results indicate that PPELNs exhibit favourable physical stability in simulated gastrointestinal fluids, with no significant aggregation or disaggregation observed. Surface negative charge helps inhibit protein conditioning, thereby slowing clearance by the reticuloendothelial system and prolonging circulation time in vivo (Ren et al., 2016). Consequently, PPELNs demonstrate robust gastrointestinal stability, effectively resisting pH fluctuations and digestive enzyme degradation to safeguard their internal contents. Effect of PPELNs on gut microbiota structure This study employed a fecal slurry fermentation model to simulate the human colonic environment. This model offers significant advantages in terms of experimental ethics, safety, and cost-effectiveness, enabling efficient and direct assessment of plant-derived exosome effects on gut health. To investigate the regulatory role of PPELNs on gut microbiota, we conducted 16S rRNA gene sequencing analysis on fermentation samples. Alpha diversity analysis (Fig. 4A) revealed no significant alterations in gut microbial richness or evenness following PPELNs intervention, indicating that PPELNs did not affect overall microbial diversity. However, PCoA analysis (Fig. 4B) revealed distinct separation of the PPELNs-treated group from the Ctrl group along the primary coordinate axis (PC1), indicating that PPELNs treatment altered the structural composition of microbial communities. Collectively, these findings suggest that PPELNs may achieve targeted regulation of the gut microbiome by selectively modulating the relative abundance of specific microbial populations, thereby altering community structure while preserving overall diversity. 16S rRNA analysis (n = 6 per group). (A) Alpha diversity. (B) PCoA. (C) Phylum-level and (D) genus-level composition of the gut microbiota. Ori denotes the unfermented original sample group, Ctrl denotes the fully fermented sample group, while PPELNs-L and PPELNs-H denote groups fermented after PPELNs addition. (E) The effects of different treatment groups on the single-bacterial abundance of Escherichia-Shigella, Blautia, Bifidobacterium and Phascolarctobacterium. (F) Phylogenetic tree. Circles radiating from the centre represent taxonomic levels from phylum to species. (G) LDA discriminant scatterplot. Further analysis of microbial community structure at the phylum level revealed that across all groups (Fig. 4C), the dominant phyla were primarily Bacillota and Pseudomonadota. Following fecal slurry fermentation, the microbial community structure of the PPELNs group exhibited a distinct shift, characterized by a significant increase in the relative abundance of Bacillota and a marked decrease in the relative abundance of Pseudomonadota. Among these, Bacillota, as one of the primary bacterial phyla in the human gut, exhibits increased abundance that plays a positive role in maintaining and improving intestinal health. This phylum encompasses numerous bacteria capable of efficiently fermenting dietary fiber and complex polysaccharides (Sun, Pei, et al., 2025). Their metabolic products, such as SCFAs, are crucial for sustaining intestinal barrier function and regulating immune homeostasis (Chenxiang et al., 2025). Pseudomonadota constitutes a complex microbial community. Certain members have been demonstrated to potentially harbor virulence factors such as type III secretion systems, exhibiting close associations with immune modulation and disease pathogenesis (Jarjees, 2020). Although studies have also identified that some Pseudomonas species (such as Pseudomonas fluorescens) may enhance host defenses by producing metabolites like sphingolipids (Peters et al., 2025), this phylum is generally regarded as a potential opportunistic pathogen within the gut. Consequently, PPELN intervention effectively suppresses excessive proliferation of the Pseudomonadota, thereby restoring a microecological environment conducive to host health. At the genus level, compared with the Ctrl group, the PPELNs intervention significantly increased the relative abundance of Bifidobacterium, Blautia and Phascolarctobacterium (p < 0.05), while the relative abundance of Escherichia-Shigella decreased significantly (p < 0.05) (Fig. 4D–E). Bifidobacterium constitutes a core beneficial microbiota in the human gut, synthesizing diverse beneficial metabolites including short-chain fatty acids, vitamins, and bacteriocins. It plays a pivotal role in enhancing intestinal barrier function, regulating immune balance, improving host mood, and facilitating nutrient absorption (J. Chen et al., 2021). Similarly, the significantly enriched Blautia represents a key butyrate-producing bacterium capable of effectively alleviating metabolic inflammation (J. Liu et al., 2024). Research indicates that specific strains within this genus (such as Blautia coccoides) metabolize tryptophan to produce indole-3-acetic acid, thereby improving insulin resistance and promoting adipose tissue browning, which is crucial for alleviating metabolic syndrome (Niu et al., 2024). Moreover, Phascolarctobacterium has been demonstrated to correlate with non-obese phenotypes and healthy metabolic states in humans (Liébana-García et al., 2025). Conversely, Escherichia-Shigella, as opportunistic pathogens, exhibit excessive proliferation closely linked to gut dysbiosis and heightened risks of multiple diseases including diabetes and inflammatory conditions (Shin et al., 2015). Linear discriminant analysis (LDA) was employed to examine the characteristic bacterial communities formed following PPELNs intervention. As depicted in Fig. 4F–G, after 24 h of fermentation, the relative abundance of multiple potentially beneficial bacteria present in the Ori group was significantly reduced in the control group (Ctrl), including c__Clostridia, f__Ruminococcaceae, and g__Faecalibacterium. These bacterial communities have been demonstrated to be closely associated with maintaining host health, such as enhancing the intestinal barrier, regulating immunity, and promoting short-chain fatty acid production. Among them, Clostridia has been identified as a key functional bacterial community in the gut associated with weight regulation (Petersen et al., 2019). In contrast, the addition of PPELNs (across both low- and high-dose groups) reversed this decline and demonstrated clear promotion of specific beneficial bacterial communities. For instance, the significantly enriched Flavonifractor plautii in the PPELNs group has been associated with inhibiting arterial stiffness, suggesting novel pathways for maintaining cardiovascular health through microbial intervention (Luo et al., 2023). Concurrently, Coriobacteriia, another promoted phylum, harbors members extensively involved in critical host physiological processes including glucose homeostasis, bile acid cycling, and lipid metabolism. These findings indicate that PPELNs maintain gut microbiota homeostasis and overall health by regulating the relative abundance of specific genera, inhibiting pathogenic bacterial growth, and promoting the proliferation of beneficial bacteria. Effect of PPELNs on metabolites in fecal slurry fermentation To investigate the effects of PPELNs on metabolites, we conducted an untargeted metabolomic analysis of the fermentation samples. Partial Least Squares Discriminant Analysis (PLS-DA) revealed distinct separation of sample points among the Ori, Ctrl, and PPELNs groups (Fig. 5A), indicating significant differences in metabolic profiles between the three groups. The PLS-DA model exhibited R2Y(cum) = 0.657 and Q2(cum) = 0.533, indicating good explanatory and predictive ability. Permutation testing (n = 200) confirmed model robustness, as none of the 200 permuted models (0/200) achieved a Q2 value exceeding that of the original model (p < 0.005), demonstrating no overfitting. Notably, PLS-DA was used only for visualizing global clustering patterns, not for variable selection; differential metabolites were strictly defined using fold change combined with FDR-adjusted p-values (q < 0.05). Volcano plot analysis (Fig. 5B) showed that compared to the Ctrl group, the PPELNs-L group exhibited 141 differentially identified metabolites (37 up-regulated, 104 down-regulated), while the PPELNs-H group showed 125 differentially identified metabolites (46 up-regulated, 79 down-regulated). This indicated that PPELNs intervention significantly altered the metabolic composition. Based on the weighted coefficients of the OPLS-DA model, the top 30 metabolites contributing to the distinction between the PPELNs and Ctrl groups were ranked using VIP scores (Fig. 5C). Results showed that all these metabolites had VIP values exceeding 2, indicating they were key differential metabolites distinguishing PPELNs-treated from control fermentation processes. Notably, compared to the Ctrl group, the PPELNs group exhibited significantly higher levels of naringenin 7-rhamnoglucoside and auraptenol (p < 0.05) (Fig. 5D). Untargeted Metabolomics (n = 6 per group). (A) PLS-DA analysis. (B) Volcano plot of differential metabolites between PPELNs and Ctrl group. (C) Metabolite VIP scores based on the PLS-DA model. (D) Relative expression levels of metabolites. (E) The correlation analysis for gut metabolites and microbiota indicators. Naringenin 7-rhamnoglucoside, a flavonoid glycoside, has been reported to exert anti-inflammatory effects (Wang et al., 2024) and reduce intestinal permeability (Xie et al., 2024); Auraptenol is a coumarin derivative belonging to a class of secondary metabolites found in citrus plants, exhibiting broad-spectrum antibacterial activity. The accumulation of these metabolites within the fermentation system may originate from the active constituents inherent to the PPELNs themselves. Furthermore, studies confirm that plant-derived exosome-like nanoparticles can carry diverse functional small-molecule metabolites such as flavonoids, saponins, polysaccharides, and alkaloids during their formation. These vesicle-encapsulated metabolites often exhibit greater stability and higher biological activity than their free forms, thereby influencing their uptake and accumulation within host cells (Han, Zheng, et al., 2025). For instance, Deng et al. (Deng et al., 2017) discovered that sulforaphane, enriched in broccoli vesicles, exhibits heightened biological activity through activation of the AMPK pathway, thereby alleviating colitis. Similarly, Chen et al. (Q. Chen et al., 2023) reported that flavonoids and polyphenols enriched in tea exosome-like nanoparticles exhibit distinct anticancer potential. These studies further substantiate the biological advantages of metabolites encapsulated within vesicles. Microbiome-metabolite correlation analysis (Fig. 5E) suggested significant positive correlations between auraptenol and Blautia(r = 0.7585, p = 0.0003), Bifidobacterium(r = 0.7589, p = 0.0003), and Gemmiger(r = 0.5253, p = 0.0252), as well as significant negative correlations with Escherichia-Shigella(r = −0.4716, p = 0.0482). Naringenin 7-rhamnoglucoside showed significant positive correlations with Phascolarctobacterium(r = 0.6150, p = 0.0066) and Ruminococcus(r = 0.6058, p = 0.0077). These findings indicate that the metabolites encapsulated within PPELNs may be associated with the enhanced abundance of the aforementioned beneficial bacteria. Research indicates that flavonoids can exert prebiotic effects, significantly enhancing the abundance of multiple beneficial gut bacteria including Blautia (Feng et al., 2024), Bifidobacterium (Pan et al., 2023), Planococcaceae, Bacteroides, lactobacillus (Ruiz-Álvarez et al., 2025; Wu et al., 2024). Furthermore, flavonoids exhibit antimicrobial functions by reducing the growth of opportunistic pathogens in the gastrointestinal tract, thereby maintaining a healthy gut microbiome (Sun, Shi, et al., 2025). Studies indicate that Auraptenol promotes the growth of Lactobacillaceae, Allobaculum and Oscillospira while reducing harmful Bacteroidetes (T. Chen et al., 2024). This beneficial microbial community reshaping, driven by PPELNs and key metabolites, may facilitate favourable shifts in the intestinal microbiome. Effect of PPELNs on SCFAs content in fecal slurry fermentation cultures SCFAs are primarily produced by gut symbiotic microorganisms through anaerobic fermentation of carbohydrates in the colon, with these microorganisms exhibiting highly heterogeneous SCFA production capabilities (Kim, 2021). This study examined variations in the levels of major SCFAs within fermented samples. Results indicated that compared to the Ctrl group, PPELNs-H intervention significantly elevated levels of acetate, propionate, butyrate, and valerate (p < 0.05), while PPELNs-L intervention significantly increased levels of propionate and valerate (p < 0.05) (Fig. 6A–D). Further Spearman correlation analysis suggested significant associations between specific bacterial genera and SCFA content (Fig. 6E). Blautia exhibited significant positive correlations with propionic acid (r = 0.4799, p = 0.0439), valeric acid (r = 0.5686, p = 0.0138), and butyric acid (r = 0.5748, p = 0.0126) levels; Escherichia-Shigella exhibited significant negative correlations with propionic acid (r = −0.6656, p = 0.0026) and butyric acid (r = −0.5088, p = 0.0311). This finding suggests that high-dose PPELN intervention may promote short-chain fatty acid production, which is consistent with changes in the relative abundance of Blautia. Previous studies have reported that Blautia supports intestinal mucosal barrier function by supplying energy to intestinal epithelial cells through acetate production (Zhao et al., 2025). Furthermore, Blautia activates the short-chain fatty acid receptor FFAR2 by generating propionic acid and acetic acid, thereby promoting colonic mucus secretion and further fortifying the intestinal mucosal barrier (Holmberg et al., 2024). Mass concentration of SCFAs produced during fermentation (n = 6 per group). (A) Acetic acid. (B) Propionic acid. (C) Butanoic acid. (D) Valeric acid. Fermentation duration: 24 h. (E) Correlation analysis of genus bacteria with changes in short chain fatty acids. The intervention with PPELNs exerted significant effects at multiple levels. On the one hand, it markedly elevated levels of active metabolites such as naringenin 7-rhamnoglucoside and auraptenol, either directly or indirectly. These metabolites synergise with PPELNs to selectively modulate gut microbiota composition, notably promoting the proliferation of beneficial bacteria such as Blautia. At the same time, the enrichment of these beneficial bacteria is accompanied by increased production of short-chain fatty acids, which are known to be associated with diverse health benefits, suggesting a potential link with gut microecological modulation (Fig. 7). Mechanisms by which PPELNs promote gut health through microbiota-metabolite interactions. Regarding the microbial community, unlike most plant-derived ELNs, which tend to modulate specific bacterial genera in a relatively limited manner, PPELNs can simultaneously and significantly enrich Bifidobacterium, Blautia, and Phascolarctobacterium. This suggests that PPELNs may influence the gut microbiome through more complex mechanisms. Regarding metabolites, PPELNs contain citrus-specific bioactive compounds such as naringenin 7-rhamnoglucoside, auraptenol, providing a unique material basis for their intestinal regulatory functions. In terms of stability and dose-response effects, PPELNs maintain stable particle size and zeta potential after simulated gastrointestinal digestion; their high negative zeta potential confers excellent colloidal stability. Notably, PPELNs exhibit clear dose-dependence: PPELNs-H intervention significantly elevated levels of acetate, propionate, butyrate, and valerate, while PPELNs-L intervention significantly increased levels of propionate and valerate. The PEG preparation method has certain limitations; it tends to coprecipitate free proteins and polysaccharides, and the absolute purity is not optimal. However, this method offers the advantages of scalability and low cost in waste-to-resource conversion. More importantly, the functional results in this study suggest that the functional effects of PPELNs are primarily attributable to the combined action of the particles and the active ingredients they carry, rather than being dominated by free impurities. This supports the development potential of PPELN formulations as functional food ingredients. In terms of raw materials, PPELNs are derived from pomelo peel agricultural waste, aligning with the principles of a circular economy; their raw material sustainability is superior to that of garlic and ginger. In summary, although there is currently a lack of direct comparative data between PPELNs and garlic or ginger ELNs within the same experimental system, a systematic cross-comparison with published literature reveals that PPELNs exhibit systematic and unique characteristics distinct from other plant-derived ELNs. Unlike jujube polysaccharides, whose regulatory mode hinges on fermentable carbon sources degraded by gut microbes to acidify fermentation systems(Ji et al., 2024; Li et al., 2026), PPELNs display divergent gut microecological regulatory properties. As carbon-source-type prebiotics, polysaccharides utilize carbohydrates as microbial energy substrates to fuel global microbial metabolism and proliferation. In this study, PPELN treatment induced no significant alterations in microbial α-diversity, with merely the community structure remodeled, hinting that PPELNs' modulatory capacity is unlikely to depend predominantly on carbohydrate provision. Combined with the metabolome-microbe correlation analysis outcomes, metabolites encapsulated in PPELNs are significantly correlated with the abundance of beneficial bacterial taxa. Collectively, these observations imply that plant-derived PPELNs and carbon-source-type prebiotics can be categorized as two classes of gut microecological regulators with divergent regulatory patterns. Conclusion This study successfully isolated and characterized PPELNs, confirming their excellent gastrointestinal stability. Proteomic analysis identified 1145 proteins enriched in carbohydrate, energy, amino acid metabolism and stimulus response. Using an in vitro human fecal fermentation model, we found that PPELNs can be efficiently utilized by the gut microbiota, selectively promoting the enrichment of beneficial taxa such as Bifidobacterium, Blautia, and Phascolarctobacterium, while simultaneously suppressing Escherichia-Shigella. This process is positively correlated with the production of short-chain fatty acids including acetic, propionic, butanoic and valeric acids. Additionally, the flavonoid (naringenin 7-rhamnoglucoside) and coumarin (auraptenol) components inherent to PPELNs were significantly enriched within the fermentation system. Microbiome-metabolite correlation analysis indicates that propionic acid, butyric acid and auraptenol were positively correlated with Blautia but negatively with Escherichia-Shigella; naringenin 7-rhamnoglucoside was positively correlated with Phascolarctobacterium and Ruminococcus. The dual regulation of microbial composition and metabolic function highlights the functional potential of PPELNs, providing a theoretical basis for their application in functional foods and targeted gut health products and enabling efficient utilization of biomass resources. Although static fermentation models serve as an effective starting point for functional screening, future research should employ more physiologically relevant models, such as in vivo models, alongside systematic characterization of their key bioactive compounds, as well as comprehensive lipidomic and RNA profiling of PPELNs, pending clarification with appropriate controls in future studies. 16S relative abundance does not equate to absolute quantification. This limitation will be addressed in future studies via qPCR based absolute quantification of key genera. Ethics and informed consent statement All procedures involving human subjects in this study were performed in compliance with relevant laws and institutional guidelines. This study was approved by the Ethics Committee of Zhejiang Gongshang University (Number: 250712–16; Approval Date: July 15, 2025). All participants participated with informed consent and authorized the use of their information. CRediT authorship contribution statement Liyi Li: Writing – review & editing, Writing – original draft, Methodology, Conceptualization. Chaofan Chen: Writing – review & editing, Writing – original draft, Methodology, Conceptualization. Tiantian Yang: Writing – review & editing, Visualization, Formal analysis. Ran Tuo: Writing – review & editing, Visualization, Formal analysis. Yingjie Huang: Writing – review & editing, Validation, Data curation. Xiaodan Zhao: Writing – review & editing, Validation, Data curation. Le Chu: Writing – review & editing, Validation. Fatao He: Writing – review & editing, Validation. Ping Li: Writing – review & editing, Project administration, Funding acquisition. Mingdong Lu: Writing – review & editing, Methodology, Formal analysis. Ethics declaration Informed consent and patient details Written informed consent to take part in the study and to publish the article has been obtained from all participants or their legal representatives. The privacy rights of participants have been observed. Human and biological material This study involved the collection of human fecal samples. Informed consent was obtained from all participants for the use of their specimens in this research and for publication. The samples were not sourced from executed prisoners or prisoners of conscience. Studies in Human This study was performed in compliance with relevant laws, regulatory frameworks and guidelines where the research took place. This study was approved by the The Ethics Committee of Zhejiang Gongshang University. (Approval No. 250712-16). 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. Acknowledgments This work was financially supported by the Pioneer Plan of Zhejiang Province (2025C01099), the Natural Science Foundation of Zhejiang Province (LTGD23H160005). Contributor Information Ping Li, Email: ping-biology@outlook.com. Mingdong Lu, Email: lumd@wmu.edu.cn. Data availability Data will be made available on request. References Associated Data Data Availability Statement Data will be made available on request.