Ultrasound-assisted extraction of total saponins from Ardisia mamillata Hance and their effects on the gut microbiota of mice fed a higher-fat and -sugar diet Authors: Jinhua Shao, Junyang Liu, Jiaxin Huang, Yuqing Zhu, Changhui Zhao Published: 2026 ABSTRACT AbstractArdisia mamillata Hance is a traditional Chinese medicinal plant rich in saponins and shows considerable potential for application in functional foods. However, systematic investigations into efficient extraction strategies and the biological activities of its constituents remain limited. To address this gap, the present study optimized an extraction process for total saponins from Ardisia mamillata Hance and evaluated their regulatory effects on gut microbiota dysbiosis induced by a high-fat, high-sugar diet in mice. Ultrasound-assisted extraction (UAE), combined with a Box–Behnken response surface design, was employed to optimize extraction parameters. The optimal conditions were determined to be an ethanol concentration of 71% (v/v), an ultrasonic time of 31 min, an ultrasonic temperature of 61 °C, and a solid–liquid ratio of 1:32 (g/mL). Under these conditions, the yield of total saponins reached (27.52 ± 0.16) mg/g, demonstrating superior extraction efficiency compared with conventional ethanol heating methods. In a diet-induced mouse model, administration of total saponins from Ardisia mamillata Hance partially inhibited body weight gain and alleviated excessive adipose accumulation, hepatocellular injury, and lipid metabolism disorders associated with the high-fat, high-sugar diet. Furthermore, the saponins contributed to the restoration of intestinal homeostasis by increasing the relative abundance of Muribaculaceae and Alloprevotella, enhancing short-chain fatty acid (SCFA) production, reducing the abundance of Lachnospiraceae_NK4A136 and Lactobacillus, and modulating the Firmicutes/Bacteroidetes (F/B) ratio. Overall, this study not only establishes an efficient extraction protocol for total saponins from Ardisia mamillata Hance but also highlights their potential to mitigate metabolic disorders through modulation of the “gut microbiota–SCFAs” axis. These findings provide a scientific foundation for the development of Ardisia mamillata Hance as a functional food ingredient or a promising prebiotic candidate.Keywords: Ardisia mamillata hance, Gut microbiota, Total saponins, Ultrasound-assisted extraction Introduction With ongoing shifts in dietary patterns, high-fat diets have emerged as a major driver of hepatocellular steatosis, which may progress to metabolic dysfunction–associated fatty liver disease (MAFLD) (Man et al., 2023) and increase the risk of obesity (Xu et al., 2024). Increasing evidence indicates that the gut microbiota, along with its collective genome, plays a pivotal role in nutrient absorption, energy homeostasis, and lipid storage (Li et al., 2019). As a key determinant of host health, the intestinal microbiota is intricately involved in immune regulation and the development of various diseases (Chen et al., 2023). Under physiological conditions, the gut microbiome exists in a dynamic state of equilibrium. However, microbial dysbiosis disrupts both its composition and functional capacity, thereby impairing normal physiological processes and contributing to disease onset (Ding et al., 2022). In this context, plant-derived bioactive compounds have attracted considerable attention for their potential to modulate gut microbiota and ameliorate metabolic disorders. For example, polysaccharides have been shown to partially restore microbial balance (Shi, 2023); saponins derived from quinoa husks alleviate hyperlipidemia in mice through modulation of gut microbial communities (Yu et al., 2023); total flavonoids extracted from hawthorn leaves reduce blood lipid levels in hyperlipidemic rats by influencing both lipid metabolism and microbiota composition (Hu et al., 2022); and berberine increases the relative abundance of the beneficial bacterium Akkermansia while decreasing the Firmicutes-to-Bacteroidetes (F/B) ratio, thereby regulating bile acid metabolism and exerting lipid-lowering effects (Jin, 2020). In addition, several medicinal plants, including rhubarb, Polygonum cuspidatum (Hu Zhang), and Alisma species, have been reported to remodel gut microbial structure, promote the growth of beneficial bacteria, and restore microbial homeostasis, collectively contributing to improvements in lipid metabolism disorders (Liu et al., 2023). Ardisia mamillata Hance, a species belonging to the genus Ardisia within the family Myrsinaceae (Ruxin, 2022), is known by several vernacular names, including maogaogang, hongdan, dwarf tea, shanzhupa, and hongzhantan (Ling et al., 2019a). It is widely distributed across multiple regions of China, including Guangdong, Fujian, Sichuan, Guangxi, Hunan, and Yunnan provinces (Tuo, 2013). The entire plant possesses medicinal value and, according to traditional Chinese medicine (TCM) theory, is associated with the Lung, Liver, Kidney, and Gallbladder meridians. In traditional and folk medicine, A. mamillata is commonly used for its therapeutic properties, such as clearing heat and promoting diuresis, activating blood circulation to achieve hemostasis, and facilitating tissue regeneration through the removal of necrotic material. Clinically, it has been employed in the treatment of conditions including pulmonary tuberculosis with hemoptysis, hepatitis, and cholecystitis (Lin et al., 2005). Modern phytochemical investigations have demonstrated that A. mamillata contains a wide range of bioactive constituents, including alkaloids (Ling et al., 2007), polysaccharides (Ling and Zeng, 2007), flavonoids (Lou et al., 2016), and saponins (Ling et al., 2019b). Among these, saponins are considered the principal pharmacologically active components. Saponins represent a structurally diverse class of naturally occurring compounds widely distributed in plants and are known to exhibit multiple biological activities, including anticancer, anti-inflammatory, anti-allergic, antileukemic, antiviral, hypoglycemic, and cardioprotective effects (Jun et al., 2024; Navarro del Hierro et al., 2018a). In addition, saponins have been reported to inhibit digestive enzyme activity (Zheng et al., 2007), regulate appetite (Karu et al., 2007), modulate lipid metabolism and energy expenditure (Han and Zhang, 2021), and alter the composition of the gut microbiota (Lee et al., 2015). Despite these advances, systematic studies investigating both the optimized extraction of saponins from A. mamillata and their potential regulatory effects on gut microbiota remain lacking. The efficient extraction of total saponins from Ardisia mamillata Hance represents both a key research objective and a prerequisite for subsequent pharmacological investigations. Currently, a variety of techniques are employed for the extraction of plant-derived saponins, including alkaline extraction, acid extraction, hot water extraction, microwave-assisted extraction (MAE), and enzyme-assisted extraction (EAE) (Mo et al., 2025). Among these approaches, ultrasound-assisted extraction (UAE) offers several notable advantages. Its cavitation effect disrupts plant cell structures, thereby enhancing the release of intracellular bioactive compounds, while simultaneously reducing extraction time and energy consumption. These features make UAE a highly efficient method for isolating plant-derived bioactive constituents (Zheng et al., 2024). In the present study, whole A. mamillata plants were used as the raw material, and UAE combined with a Box–Behnken response surface design was employed to optimize the extraction conditions. In parallel, male C57BL/6 mice were administered total saponins derived from A. mamillata via oral gavage to evaluate their effects on gut microbiota composition. This experimental design aimed to assess the potential of A. mamillata saponins in modulating gut microecology and in the prevention or mitigation of obesity. Overall, this study not only provides a scientific basis for the efficient extraction and utilization of A. mamillata resources but also offers new insights into the interactions between plant-derived bioactive compounds and the gut microbiota. The findings may contribute to the development of functional foods and natural product–based therapeutics, facilitate the modernization of TCM, and support innovation within the health industry, while also serving as a valuable reference for related fields of research. Materials and methods Materials Ardisia mamillata Hance was collected in Hunan Province, China, in September. The whole plants were vacuum-dried at 60 °C, ground into powder, passed through a 60-mesh sieve, and stored for subsequent use. Experimental animals: Specific pathogen-free (SPF) male C57BL/6J mice (6 weeks old, weighing 18–21 g) were obtained from Hunan Silaike Jingda Experimental Animal Co., Ltd. (certificate No. 430727231102079742). All experimental procedures complied with the Guide for the Care and Use of Laboratory Animals. The research project was ethically supervised by the Academic Ethics Committee of Hunan University of Science and Engineering. Methods Preparation of the standard curve Standard curve preparation: With slight modifications to the method described in reference (Xu et al., 2021; Man et al., 2023)0.4 mg of ginsenoside Rg1 standard was accurately weighed, dissolved in anhydrous ethanol using ultrasonication, and diluted to a final volume of 10.0 mL to obtain a stock solution. Aliquots ranging from 0.1 to 1.0 mL (i.e., 0.1, 0.2, 0.4, 0.6, 0.8, and 1.0 mL) were transferred into 10 mL stoppered test tubes and evaporated to dryness in an 80 °C thermostatic water bath. After complete solvent evaporation, 0.2 mL of freshly prepared 5% vanillin–glacial acetic acid solution and 0.8 mL of perchloric acid were added to each tube. The mixtures were thoroughly mixed and incubated at 65 °C for 20 min to allow color development. Subsequently, the reaction mixtures were cooled to room temperature, diluted to 10.0 mL with glacial acetic acid, and mixed thoroughly. A reagent blank was used as the control. Absorbance was measured at 581 nm. A standard curve was constructed by plotting absorbance (y-axis) against ginsenoside Rg1 concentration (x-axis), yielding the regression equation: y = 2.4369x + 0.18918 (R2 = 0.99413). Determination of total saponin content from Ardisia mamillata hance Determination of total saponin content: A 1.0 mL aliquot of the ethanol extract supernatant was transferred into a 10 mL stoppered test tube and evaporated to dryness in an 80 °C thermostatic water bath. Subsequently, 0.2 mL of freshly prepared 5% vanillin–glacial acetic acid solution and 0.8 mL of perchloric acid were added to each tube. The mixture was thoroughly mixed and incubated at 65 °C for 20 min to allow color development, followed by rapid cooling in an ice bath. After cooling, the reaction solution was diluted to 10.0 mL with glacial acetic acid and mixed thoroughly. The absorbance was then measured at 581 nm using a reagent blank as the control. The saponin concentration in the extract was calculated based on the standard curve described in Section 2.2.1. The extraction yield of total saponins from A. mamillata was subsequently calculated using Equation (1): where: Y represents the extraction yield of total saponins from A. mamillata (mg/g); c denotes the concentration of saponins in the extract solution (mg/mL); V is the total volume of the sample solution (mL); n refers to the dilution factor; and M is the mass of the plant material (g). Single-factor experiments Impact of ethanol concentration on the of total saponins’ yield from A. mamillata Single-factor experiment on ethanol concentration: A 20.0 g sample of A. mamillata dry powder was accurately weighed for extraction. Using the extraction yield of total saponins as the response variable, UAE was carried out under the following fixed conditions: ultrasonic power of 200 W, intermittent ultrasonication (3 s on/3 s off), extraction temperature of 70 °C, a solid–liquid ratio of 1:20 (g/mL), and an extraction time of 35 min. Ethanol concentration was varied at 60%, 65%, 70%, 75%, and 80% to evaluate its effect on total saponin yield. All experiments were performed in triplicate. Impact of ultrasonic time on the of total saponins’ yield from A. mamillata Single-factor experiment on ultrasonic time: A 20.0 g sample of A. mamillata dry powder was accurately weighed for extraction. Using the extraction yield of total saponins as the response variable, UAE was performed under the following fixed conditions: ultrasonic power of 200 W, intermittent ultrasonication (3 s on/3 s off), extraction temperature of 70 °C, solid–liquid ratio of 1:20 (g/mL), and ethanol concentration of 70%. Ultrasonic extraction time was varied at 25, 30, 35, 40, and 45 min to investigate its effect on total saponin yield. All experiments were conducted in triplicate. Effect of ultrasonic temperature on the yield of total saponins from A. mamillata Single-factor experiment on ultrasonic temperature: A 20.0 g sample of A. mamillata dry powder was accurately weighed for extraction. Using the extraction yield of total saponins as the response variable, UAE was performed under the following fixed conditions: ultrasonic power of 200 W, intermittent ultrasonication (3 s on/3 s off), solid–liquid ratio of 1:20 (g/mL), ethanol concentration of 70%, and extraction time of 30 min. Ultrasonic temperature was varied at 50, 55, 60, 65, and 70 °C to evaluate its effect on total saponin yield. All experiments were conducted in triplicate. Effect of material-to-liquid ratio on the yield of total saponins from A. mamillata Single-factor experiment on material-to-liquid ratio: A 20.0 g sample of A. mamillata dry powder was accurately weighed for extraction. Using the extraction yield of total saponins as the response variable, UAE was performed under the following fixed conditions: ultrasonic power of 200 W, intermittent ultrasonication (3 s on/3 s off), extraction temperature of 60 °C, ethanol concentration of 70%, and extraction time of 30 min. The material-to-liquid ratio was varied at 1:15, 1:20, 1:25, 1:30, and 1:35 (g/mL) to investigate its effect on total saponin yield. All experiments were conducted in triplicate. Response surface experimental design Response surface methodology (RSM) design: Based on the results of the single-factor experiments, RSM with a Box–Behnken design was employed to develop a mathematical model describing the relationship between four independent variables, ethanol concentration (A, %), ultrasonic time (B, min), ultrasonic temperature (C, °C), and liquid-to-material ratio (D, mL/g), and the extraction yield of total saponins. Model analysis was performed to assess the relative contribution of each factor to the total saponin yield from A. mamillata and to determine the optimal combination of extraction parameters. The coded levels of each factor are presented in Table 1. Table 1Factor level design.Table 1LevelFactorABCD−16525551:2507030601:3017535651:35 Factor level design. Comparative extraction experiments Ethanol heating (reflux) extraction: A 20.0 g sample of A. mamillata dry powder was accurately weighed and extracted with 71% (v/v) ethanol solution at a solid–liquid ratio of 1:32 (g/mL). Reflux extraction was performed at 90 °C for 1.5 h. All experiments were conducted in triplicate. The total saponin content was determined according to Sections 2.2.1–2.2.3. Ultrasound-Assisted Extraction (UAE): A 20.0 g sample of A. mamillata dry powder was accurately weighed and extracted with 71% (v/v) ethanol solution at a solid–liquid ratio of 1:32 (g/mL). UAE was performed under the following optimized conditions: extraction temperature of 61 °C, extraction time of 31 min, ultrasonic power of 200 W, and intermittent ultrasonication (3 s on/3 s off). All experiments were conducted in triplicate, and the total saponin content was determined according to Sections 2.2.1–2.2.3. Preliminary purification of total saponins from Ardisia mamillata hance Extraction of total saponins from A. mamillata: Whole-plant material of A. mamillata (vacuum-dried at 60 °C, ground, and passed through a 60-mesh sieve) was accurately weighed (20.0 g) and transferred into a round-bottom flask. A total of 640 mL of 71% (v/v) ethanol was added, and UAE was performed at 61 °C for 31 min (ultrasonic power of 200 W; intermittent ultrasonication, 3 s on/3 s off). The extract was vacuum-filtered, and the resulting supernatant was collected. The filtrate was then concentrated under reduced pressure until no ethanol odor remained. The residue was re-dissolved in distilled water and sequentially partitioned with petroleum ether (3 × 50 mL), and the organic phases were discarded. The aqueous phase was subsequently extracted with ethyl acetate (3 × 50 mL), and these organic fractions were also discarded. Finally, the aqueous phase was extracted with n-butanol (3 × 50 mL). The combined n-butanol fractions were concentrated under reduced pressure and vacuum-dried to obtain the total saponins of A. mamillata. UPLC-MS/MS Chromatographic conditions An ACQUITY UPLC® HSS T3 column (100 mm × 2.1 mm, 1.8 μm) was employed for analysis. The flow rate was set at 0.3 mL min−1, and the column temperature was maintained at 40 °C. The injection volume was 4 μL. Mobile phase A consisted of 0.1% formic acid in water, while mobile phase B consisted of 0.1% formic acid in acetonitrile. The gradient elution program was as follows: 0–1.5 min, 95% A; 1.5–2.5 min, A decreased from 95% to 90% and was held for 1.0 min; 2.5–14 min, A decreased from 90% to 60%; 14–22 min, A decreased from 60% to 5%; 22–25 min, 5% A; 25–26 min, A increased from 5% to 95%; and 26–30 min, 95% A. Mass spectrometry conditions An electrospray ionization (ESI) source was employed with the following parameters: nebulizer gas (GS1) pressure of 60 psi, auxiliary gas pressure of 60 psi, curtain gas pressure of 35 psi, ion source temperature of 550 °C, and spray voltage of 5500 V in positive ion mode and −4500 V in negative ion mode. Animal grouping and gavage Animal grouping and treatment: Eighteen SPF adult male mice were used in this study. After a 1-week acclimation period on a standard chow diet, the mice were randomly assigned to three groups (n = 6 per group). Animals were housed under controlled environmental conditions (22–26 °C, 55%–70% relative humidity) with ad libitum access to food and water. The experimental groups were as follows: Control group (Con): fed a standard maintenance diet for 13 weeks. High-fat and high-sugar diet group (Mod): fed a high-fat and high-sugar diet for 13 weeks. Treatment group (Ard): fed a high-fat and high-sugar diet for 12 weeks, followed by daily oral gavage of A. mamillata total saponins from week 9 to week 13 at a dose of 108 mg/kg for 5 weeks. Mouse feces collection Fecal sample collection: One day prior to the end of the experiment, fresh fecal samples were collected from each mouse into sterile Eppendorf (EP) tubes. The samples were immediately snap-frozen in liquid nitrogen and subsequently stored at −80 °C until further analysis. Measurement of mouse body weight The body weight of mice in the control group (Con), high-fat and high-sugar diet group (Mod), and treatment group (Ard) was measured once weekly, and longitudinal changes in body weight were recorded. Measurement of mouse organ fat weight At the end of the experiment, mice were fasted for 12 h, after which blood samples were collected and the animals were euthanized. Perirenal, inguinal, dorsal, and epididymal adipose tissues were rapidly excised, carefully trimmed to remove surrounding impurities, and weighed. Hematoxylineosin (HE) staining of liver and epididymal adipose tissue Fresh liver and epididymal adipose tissue samples from the control group (Con), high-fat and high-sugar diet group (Mod), and treatment group (Ard) were fixed in 4% paraformaldehyde for 48 h. The tissues were subsequently processed into paraffin blocks using standard histological procedures, including dehydration through a graded ethanol series, clearing in xylene, and paraffin infiltration. Paraffin-embedded tissues were sectioned at a thickness of 4 μm, followed by dewaxing and rehydration. Sections were then stained with hematoxylin and eosin (H&E). After staining, the sections were dehydrated, cleared, and mounted with coverslips. Morphological changes in liver and epididymal adipose tissues were examined under a light microscope, and representative images were captured for analysis. Detection of ALT, AST, TC, and TG levels in mouse serum Serum levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT), total cholesterol (TC), and triglycerides (TG) in the control group (Con), high-fat and high-sugar diet group (Mod), and treatment group (Ard) were measured according to the manufacturer's instructions provided with the respective assay kits. Effects of A. mamillata total saponins on the gut microbiota of mice 16S rRNA high-throughput sequencing PCR amplification of 16S rRNA genes: The V3–V4 hypervariable regions of the bacterial 16S rRNA gene were amplified using genomic DNA extracted from mouse fecal samples as the template. The following primers were used: forward primer (F), 5′-ACTCCTACGGGAGGCAGCAG-3′, and reverse primer (R), 5′-GGACTACHVGGGTWTCTAAT-3′. The PCR reaction mixture (25 μL total volume) consisted of 1.0 μL each of forward and reverse primers (5 μmol/L), 3.0 μL bovine serum albumin (BSA; 2 ng/μL), 12.5 μL of 2 × Taq Plus Master Mix, approximately 30 ng of template DNA, and sterile water to adjust the final volume. PCR amplification was performed under the following conditions: initial denaturation at 94 °C for 3 min; 28 cycles of denaturation at 94 °C for 30 s, annealing at 55 °C for 30 s, and extension at 72 °C for 1 min; followed by a final extension at 72 °C for 7 min. High-throughput sequencing and data processing: Each sample was amplified in triplicate, and the resulting PCR products from the same sample were pooled. Amplicons were verified by electrophoresis on a 2% agarose gel, after which the target bands were excised and purified using the AxyPrep DNA Gel Extraction Kit (Axygen). The purified DNA was subsequently quantified using the Qubit dsDNA HS Assay Kit (Invitrogen, USA). Sequencing libraries were constructed and subjected to high-throughput sequencing on the Illumina MiSeq PE300 platform. Raw FASTQ reads were initially processed using Trimmomatic for quality filtering, and paired-end reads were merged using PEAR. Chimeric sequences were identified and removed using UCHIME against a reference database, with additional de novo chimera detection performed to eliminate sequences not present in the reference. Low-quality and short reads were further filtered out to obtain high-quality sequences. Operational taxonomic units (OTUs) were clustered at a 97% sequence similarity threshold, and subsequent bioinformatic analyses were conducted using QIIME. Determination of short-chain fatty acid (SCFA) content Extraction of SCFAs from feces A 0.20 g sample of mouse feces stored at −80 °C was suspended in 1.60 mL of sterile deionized water. The mixture was thoroughly vortexed and allowed to stand at room temperature for 20 min, followed by centrifugation at 15,000 rpm for 15 min at 4 °C. The supernatant was carefully transferred to a new EP tube. An additional 1.60 mL of sterile deionized water was added to the remaining fecal pellet, and the extraction procedure was repeated. The resulting supernatants were combined and mixed thoroughly. After filtration through a 0.22 μm membrane filter, a 0.20 mL aliquot of the filtrate was mixed with 0.70 mL of sterile water and 0.10 mL of n-butanol solution (100 μg/mL) as an internal standard for subsequent GC analysis. Gas chromatography (GC) detection of SCFAs (1)Detection conditions Detection conditions A Nexis GC-2030 gas chromatograph (Shimadzu, Japan) equipped with an InterCap WAX highly polar capillary column (60 m × 0.32 mm × 0.25 μm) was used for analysis. The inlet temperature was set at 280 °C with a split ratio of 20:1. The oven temperature program was as follows: initial temperature of 60 °C held for 2.5 min, increased to 140 °C at a rate of 10 °C/min, and further increased to 170 °C at a rate of 3 °C/min (Xu et al., 2024). Detection was performed using a flame ionization detector (FID) maintained at 300 °C, and the injection volume was 1 μL. (2)Preparation of the standard curve Preparation of the standard curve Preparation of standard solutions and calibration curves: Analytical standards of acetic acid, propionic acid, n-butyric acid, isobutyric acid, n-valeric acid, and isovaleric acid were accurately weighed and dissolved in ultrapure water to prepare individual stock solutions at a concentration of 1000 μg/mL for each compound. The six standard solutions were then mixed at a volume ratio of 20:10:5:2:2:1 to obtain a mixed standard solution containing the six SCFAs at final concentrations of 500 μg/mL, 250 μg/mL, 125 μg/mL, 50 μg/mL, 50 μg/mL, and 25 μg/mL, respectively. This mixed standard solution was further diluted with ultrapure water to generate five concentration gradients (1 × , 0.6 × , 0.4 × , 0.2 × , and 0.1 × ). For example, the corresponding concentrations of acetic acid in these diluted solutions were 500, 300, 200, 100, and 50 μg/mL, respectively. For GC analysis, 200 μL of the mixed standard solution was combined with 700 μL of ultrapure water and 100 μL of n-butanol internal standard solution, resulting in a final n-butanol concentration of 100 μg/mL. The average peak area of n-butanol, obtained from multiple injections, was used to correct for inter-batch variation. Calibration curves were constructed by plotting the concentration of each standard acid (x-axis) against its corresponding peak area (y-axis). Data analysis Experimental data were processed and analyzed using Microsoft Excel 2007, Origin, and Design-Expert 8.0.6. One-way analysis of variance (ANOVA) was performed using SPSS 22.0, followed by Tukey's post hoc test for multiple comparisons. Statistical significance was set at P < 0.05. Results Optimization of UAE process for total saponins from Ardisia mamillata Hance Results of single-factor experiments As shown in Fig. 1(A), the extraction yield of total saponins from Ardisia mamillata Hance initially increased and then decreased with rising ethanol concentration, reaching a maximum at 70%. This behavior can be attributed to the amphiphilic nature of saponins, which contain both hydrophobic aglycone moieties and hydrophilic glycosidic chains. Changes in ethanol concentration alter the solvent polarity, and according to the principle of “like dissolves like,” 70% ethanol provides an optimal polarity that best matches the physicochemical characteristics of the saponins in A. mamillata, thereby facilitating their dissolution and release (Xing et al., 2024). When the ethanol concentration exceeded 70%, the extraction efficiency decreased, likely due to reduced solubility of more polar saponin fractions under lower solvent polarity conditions (Akbari et al., 2019). Effects of individual factors on the Extraction amount of total saponins from A. mamillata. As shown in Fig. 1(B), the extraction yield of total saponins increased progressively with increasing ultrasonic time. This trend is likely due to acoustic cavitation, which disrupts plant cell structures and enhances the release of intracellular compounds. In addition, ultrasonic energy is partially converted into heat, increasing the medium temperature and accelerating molecular motion, thereby further promoting mass transfer and saponin release (Fan et al., 2015). The yield reached its maximum at 30 min; however, further extension of extraction time resulted in a decline, possibly due to the degradation of saponin aglycone structures under prolonged ultrasonic exposure, leading to reduced overall yield (Belwal et al., 2019). As shown in Fig. 1(C), the extraction yield of total saponins increased with rising ultrasonic temperature but decreased when the temperature exceeded 60 °C. This trend may be explained by the fact that moderate temperature elevation enhances molecular motion and improves solute solubility, thereby facilitating saponin extraction. However, temperatures above 60 °C may induce thermal degradation of heat-sensitive saponin constituents, resulting in reduced yield (Ruilin et al., 2017). As shown in Fig. 1(D), increasing the solvent volume initially promoted the extraction yield of total saponins, whereas further increases led to a decline. This phenomenon may be attributed to changes in the effective polarity and mass transfer dynamics of the extraction system. At an optimal range, higher solvent volumes enhance solvent–matrix contact and improve saponin release. However, excessive solvent volume may promote the co-extraction of polyphenols, flavonoids, lipophilic impurities, carbohydrates, proteins, amino acids, and tannins (Liu et al., 2022; Wong et al., 2020; Ligor et al., 2018; Navarro del Hierro et al., 2018b). These compounds may form non-covalent complexes with saponins or encapsulate their molecular structures (Guo et al., 2024; Wu et al., 2025), thereby reducing their apparent solubility in the solvent system and ultimately decreasing the total saponin yield (Wen et al., 2019). Response surface optimization Based on the results of the single-factor experiments, four independent variables, ethanol concentration (A), ultrasonic time (B), ultrasonic temperature (C), and material-to-liquid ratio (D), were selected for further optimization. Using the extraction yield of total saponins from A. mamillata as the response variable, a RSM experiment was designed, as shown in Table 1. The experimental data (Table 2) were analyzed using multivariate polynomial regression with Design-Expert 8.0.5b software. A quadratic multiple regression model describing the relationship between total saponin yield (Y) and the four variables was established as follows: Y = 26.73 + 0.46A-1.78B + 1.92C + 3.34D-0.65AB+1.63AC-0.41AD-0.43BC-0.56BD-0.10CD-3.52 A2-3.65 B2-3.03C2-4.02D2. Table 2Response surface experimental design and results.Table 2NumberFactorExtraction amount/(mg/g)ABCD11−10018.61 ± 0.892−10−1021.35 ± 0.7530−11021.58 ± 0.674−110021.33 ± 0.595000025.21 ± 0.626−100−114.55 ± 0.51700−1−113.95 ± 0.718000027.16 ± 0.82900−1121.58 ± 0.7710000026.98 ± 0.8811011022.83 ± 0.6312100−116.73 ± 0.5513−1−10014.68 ± 0.4714−100121.88 ± 0.53150−10−114.38 ± 0.4116−101021.35 ± 0.8417101023.36 ± 0.79180−10121.62 ± 0.6919000027.22 ± 0.9320110022.66 ± 0.81210−1−1015.83 ± 0.742201−1018.79 ± 0.6123100122.42 ± 0.5424001124.71 ± 0.6525010−118.71 ± 0.9126000027.09 ± 0.8727001−117.48 ± 0.5828010123.69 ± 1.032910−1016.82 ± 0.72 Response surface experimental design and results. Based on the ANOVA results presented in Table 3, the regression model exhibited an F-value of 19.98 with P < 0.0001, indicating that the model was highly significant. The lack-of-fit term was not significant (P = 0.1796), suggesting that the model adequately fit the experimental data and that the experimental error was minimal. The coefficient of determination (R2 = 0.9523) and adjusted R2 (Adj. R2 = 0.9047) indicated that 90.47% of the variation in extraction yield could be explained by the selected variables, demonstrating strong model reliability and good predictive capability. The coefficient of variation (CV = 6.06%) further confirmed the high precision and reproducibility of the model. Regarding the regression coefficients, the linear terms (B, C, and D) and all quadratic terms (A2, B2, C2, and D2) were highly significant (P < 0.01), while the interaction term AC was significant (P < 0.05). In contrast, the remaining interaction terms (AB, AD, BC, BD, and CD) were not significant (P > 0.05). These results indicate that the effects of extraction parameters on total saponin yield are predominantly nonlinear rather than purely linear. Among the investigated factors, their influence on extraction yield followed the order: material-to-liquid ratio (D) > ultrasonic temperature (C) > ultrasonic time (B) > ethanol concentration (A). Table 3ANOVA of the regression model.Table 3SourceSum of SquaresdfMeanSquaresF-valueP-valueSignificanceMode446.771431.9119.98<0.0001∗∗A2.4812.481.560.2328B37.84137.8423.690.0002∗∗C44.05144.0527.570.0001∗∗D134.001134.0083.88<0.0001∗∗AB1.6911.691.060.3211AC10.69110.696.690.0215∗AD0.6710.670.420.5270BC0.7310.730.460.5098BD1.2811.280.800.3864CD0.04010.0400.0250.8765A280.32180.3250.28<0.0001∗∗B286.31186.3154.03<0.0001∗∗C259.66159.6637.34<0.0001∗∗D2105.031105.0365.75<0.0001∗∗Residual22.37141.69Lack of fit19.44101.942.660.1796Not significantPure error2.9340.73Total error469.1428 ANOVA of the regression model. According to the regression equation, the response surface and contour plots for the different factor interactions are presented in Fig. 2. The three-dimensional response surface plots and corresponding two-dimensional contour plots illustrate the variation of the response variable with changes in the independent factors. The steepness of the response surface reflects the sensitivity of the response to changes in each factor, while the contour shape indicates the strength of interaction between two variables. Circular contours suggest weak or negligible interactions, whereas elliptical contours indicate stronger interactions. In this study, the interaction between ethanol concentration (A) and ultrasonic temperature (C) produced a relatively steep response surface, and the corresponding contour plots tended toward an elliptical shape, suggesting a relatively strong interaction between these two variables. Contour plots and 3D response surface plots for optimization Validation of the extraction process for total saponins from A. mamillata By solving the quadratic polynomial regression equation, the optimal extraction conditions were determined as follows: ethanol concentration of 70.50% (v/v), ultrasonic time of 30.92 min, ultrasonic temperature of 61.26 °C, and material-to-liquid ratio of 1:31.96 (g/mL), with a predicted extraction yield of 27.89 mg/g. Considering practical operability, these conditions were adjusted to ethanol concentration of 71% (v/v), ultrasonic time of 31 min, ultrasonic temperature of 61 °C, and material-to-liquid ratio of 1:32 (g/mL). Three independent validation experiments were then conducted, each performed in triplicate. As shown in Table 4, the mean extraction yield of total saponins was (27.52 ± 0.16) mg/g, which closely matched the predicted value and demonstrated good reproducibility. This confirms a strong agreement between the response surface model and the experimental validation, indicating that the optimized extraction process is reliable and practically feasible. Table 4Experimental validation results.Table 4OptimalExperimental results total saponins(mg/g)Predicted value total saponins(mg/g)Ethanol concentration/%Ultrasonic time/minUltrasonic temperature/°CMaterial-to-liquid ratio/(g/mL)7131611:3227.52 ± 0.1627.89 Experimental validation results. Comparison of different extraction methods As shown in Table 5, ethanol heating (reflux) extraction yielded (23.56 ± 0.203) mg/g of total saponins, whereas UAE achieved a significantly higher yield of (27.32 ± 0.128) mg/g. Although both methods employed ethanol as the extraction solvent, UAE operated at a lower temperature and shorter extraction time, yet increased the extraction yield of total saponins from A. mamillata by 13.73%. These results are consistent with previous reports. Du et al. (2016) reported that under 74% (v/v) ethanol and an ultrasonic time of 33 min, the yield of quinoa saponins reached 23.37 mg/g. Liang et al. (Xiang et al., 2017) obtained a lower yield of 15.8 mg/g using ethanol reflux extraction. Similarly, Feng (2017) compared UAE, methanol reflux extraction, and ethanol heating extraction of quinoa bran and found that UAE was the most efficient method. Chen et al. (2021) further demonstrated that UAE increased the extraction yield of Ziziphus jujuba seeds by 14.25% compared with ethanol heating extraction. Collectively, these findings indicate that UAE offers higher extraction efficiency, shorter processing time, and lower energy consumption, while utilizing a relatively safe and environmentally friendly solvent system. This method not only enhances saponin yield but also demonstrates strong potential for industrial-scale application, providing a useful reference for the extraction of bioactive compounds from other medicinal plants. Table 5Effect of different extraction methods on the yield of total saponins from A. mamillata (mg/g).Table 5Extraction methodNumber of experiments123Average (‾x ± s)Ethanol heating extraction23.7823.6123.2923.56 ± 0.203bUltrasound-assisted method27.1527.3427.4627.32 ± 0.128a Effect of different extraction methods on the yield of total saponins from A. mamillata (mg/g). Chromatographic analysis and component identification Detection by UPLC–MS/MS revealed abundant chromatographic peaks in the positive ion mode for the total saponins. Through comparison of primary mass spectra and database matching, a total of 49 triterpenoid and steroidal compounds were identified in the saponin fraction (Table 6), including various saponins and sapogenins. Given that these compounds are widely reported to exhibit diverse biological activities, it is hypothesized that total saponins from A. mamillata may possess potential functional relevance in modulating gut microbiota, protecting hepatic function, regulating lipid metabolism, and exerting anti-inflammatory and anticancer effects. However, these proposed bioactivities require further systematic experimental validation. Table 6Triterpenoid and steroidal compounds.Table 6No.Compoundm/zIon modeRetentionTime (s)Molecular formula1Polyphyllin I855.4597[M+H]+26.23548C44H70O162Glycyrrhetinic acid471.3458[M+H]+13.79643C30H46O433-Oxo-olean-12-en-28-oic acid455.3526[M+H]+15.80213C30H46O34Oleanonic acid457.367[M+H]+11.91822C30H48O35Hederagenin473.3568[M+H]+15.80213C30H48O46Ginsenoside Rh3605.4426[M+H]+15.17085C36H60O77Soyasapogenol A475.3774[M+H]+14.02292C30H50O48Saikosaponin781.4713[M+H]+16.21812C42H68O139Bayogenin489.3569[M+H]+10.5718C30H48O510Ilexgenin A503.3393[M+H]+14.90868C30H46O611Ziyuglycoside I767.4586[M+H]+13.81543C41H66O1312Momordicoside L635.4076[M+H]+10.5533C36H58O913Araloside A927.5002[M+H]+10.5533C47H74O1814Momordin Ic765.4414[M+H]+10.5718C41H64O1315Akeboside D929.5093[M+H]+10.5718C47H76O1816Polygalasaponin F1091.561[M+2H]2+10.5718C53H86O23173-Hydroxy-12-oleanene-23,28-dioic acid487.3425[M+H]+10.86562C30H46O518Cycloastragenol491.3712[M+H]+10.9026C30H50O519Astragaloside IV785.4672[M+H]+10.9026C41H68O1420Polyphyllin VI739.4268[M+H]+10.93995C39H62O1321Gracillin885.4792[M+H]+10.93995C45H72O1722Pulsatilla saponin A751.4573[M+H]+11.29025C41H66O1223Raddeanin A897.5174[M+H]+11.29025C47H76O1624Ginsenoside Rh4621.4368[M+H]+11.91822C36H60O825Methyl protodioscin1063.568[M+H]+11.97537C52H86O2226Bayogenin489.3574[M+H]+12.10887C30H48O527Ginsenoside Rg6767.4881[M+H]+12.14653C42H70O1228Gypenoside A899.497[M+H]+12.14653C46H74O1729Deapioplatycodin D1093.541[M+H]+12.14653C52H84O2430Jujuboside B1045.56[M+H]+12.76367C52H84O2131Saikosaponin A781.4742[M+H]+13.10065C42H68O1332Bayogenin489.3529[M+H]+13.7776C30H48O533Astragaloside III785.4635[M+H]+13.81543C41H68O1434Protodeltonin1065.551[M+2H]2+13.85342C51H84O2335Ginsenoside Rd947.5477[M+H]+13.98592C48H82O1836Clinoposaponin A959.5179[M+H]+14.66703C48H78O1937Rosamultin651.4081[M+H]+14.98368C36H58O1038Notoginsenoside R1933.5421[M+H]+15.26417C47H80O1839Timosaponin B903.4975[M+H]+15.69815C45H74O1840Protopanaxadiol461.3988[M+H]+17.01573C30H52O341Gypenoside XLIX1047.57[M+H]+17.07305C52H86O21423-Oxo-oleanolic acid455.3502[M+H]+17.83835C30H46O343Phytolaccagenin533.3297[M+H]+18.44365C31H48O744Ginsenoside F3771.5049[M+H]+20.1472C41H70O1345Furostanol saponin577.3358[M+H]+20.64933C32H48O946Timosaponin BII921.5159[M+H]+20.94815C45H76O1947Dioscin869.4989[M+H]+20.98532C45H72O1648Pulsatilla saponin C943.512[M+H]+21.26513C48H78O1849Panaxatriol477.393[M+H]+21.37795C30H52O4 Triterpenoid and steroidal compounds. Effects of A. mamillata total saponins on body weight of mice fed a higher-fat and -sugar diet As shown in Fig. 3, no statistically significant differences in baseline body weight were observed among groups prior to modeling (P > 0.99). After modeling, body weight in both the Mod group (P < 0.0001) and the Ard group (P = 0.0001) was significantly higher than that in the Con group (P < 0.01). Following intervention, the Mod group maintained a significantly higher body weight compared with the Con group (P < 0.01), whereas the Ard group exhibited a significant reduction in body weight relative to the Mod group (P < 0.0001). These results suggest that total saponins from A. mamillata may, to some extent, suppress body weight gain in diet-induced obese mice. Effect of A. mamillata total saponins on body weight of mice fed a higher-fat and -sugar diet (n = 6, ∗∗∗ and above: p < 0.01). Effects of A. mamillata total saponins on fat weight of mice fed a higher-fat and -sugar diet The accumulation of perirenal, inguinal, epididymal, and dorsal adipose tissues is presented in Fig. 4. Compared with the control (Con) group, the Mod group exhibited a significant increase in perirenal fat weight (P = 0.0125), and highly significant increases in inguinal, epididymal, and dorsal fat weights (P < 0.0001), indicating that the high-fat and high-sugar diet induced marked adipose tissue accumulation in mice. In comparison with the Mod group, the Ard group showed a significant reduction in perirenal fat weight (P = 0.0014), as well as highly significant decreases in inguinal, epididymal, and dorsal fat weights (P < 0.0001 for inguinal and epididymal fat; P = 0.0001 for dorsal fat). These findings suggest that total saponins from A. mamillata may effectively alleviate diet-induced excessive fat accumulation in mice to a certain extent. Effect of A. mamillata total saponins on fat weight of mice fed a higher-fat and -sugar diet (n = 6, ∗: p < 0.05, ∗∗ and above: p < 0.01). Effects of A. mamillata total saponins on pathological changes in liver and epididymal adipose tissue of mice fed a higher-fat and -sugar diet HE staining of liver tissues from each group (Fig. 5) showed that hepatocytes in the control (Con) group exhibited normal histological architecture, including intact cellular morphology, uniform cell size, distinct cell boundaries, radially arranged hepatic cords, and densely and orderly organized cells, with clear nuclear staining and no evidence of congestion or intercellular edema (Fig. 5(A)). In contrast, compared with the Con group, liver tissues in the Mod group displayed disrupted cellular structure, characterized by disorganized hepatic cords, indistinct cell boundaries, and numerous cytoplasmic vacuoles of varying sizes, accompanied by a marked increase in vacuolar area (Fig. 5(B)). Following intervention, the Ard group showed substantial histological improvement relative to the Mod group, with more intact cellular architecture, clearer cell boundaries, a more orderly arrangement of hepatic cords, and a significant reduction in intracellular vacuolization (Fig. 5(C)). Effect of A. mamillata total saponins on pathological changes in liver and adipose tissue of mice fed a higher-fat and -sugar diet (20 × ). HE staining of epididymal adipose tissue from each group revealed that, compared with the control (Con) group (Fig. 5(D)), adipocytes in the Mod group (Fig. 5(E)) were markedly enlarged, with increased heterogeneity in cell size and morphology. In contrast, the Ard group exhibited noticeable improvement relative to the Mod group, as evidenced by a reduction in adipocyte size and a more uniform cellular morphology (Fig. 5(F)). Effects of A. mamillata total saponins on serum levels of ALT, AST, TC, and TG in mice fed a higher-fat and -sugar diet As shown in Fig. 6, compared with the control (Con) group, serum levels of ALT, AST, and TG were significantly elevated in the Mod group (P < 0.0001 for all), while TC was also significantly increased (P = 0.0184), indicating hepatic injury and lipid metabolism disorder in the model mice. In comparison with the Mod group, the Ard group showed markedly reduced serum ALT (P < 0.0001), AST (P < 0.0001), and TG levels (P = 0.0092), whereas TC showed no significant change (P = 0.5008). These results suggest that total saponins from A. mamillata may alleviate diet-induced hepatic injury and partially improve lipid metabolic disturbances in mice. Effect of A. mamillata total saponins on serum biochemical indices in mice fed a higher-fat and -sugar diet (n = 6, ∗: p < 0.05, ∗∗ and above: p < 0.01). Effects of A. mamillata total saponins on the gut microbiota of mice fed a high-fat diet OTU counts Fig. 7 presents the total OTU counts across the experimental groups, with the overlapping region indicating OTUs shared among groups (n = 585). Compared with the normal diet group, both the high-fat and high-sugar group and the treatment group exhibited pronounced alterations in OTU abundance. The numbers of unique OTUs in the normal diet, high-fat and high-sugar, and treatment groups were 296, 388, and 314, respectively. The high-fat and high-sugar group displayed the highest number of unique OTUs, whereas administration of A. mamillata total saponins reduced the number of unique OTUs in the treatment group to a level that did not significantly differ from that of the normal diet group (P > 0.05). These results suggest that A. mamillata total saponins may alleviate gut microbiota dysbiosis induced by a high-fat and high-sugar diet, thereby shifting the microbial composition of treated mice closer to that observed in mice fed a normal diet. Venn diagram of the gut microbiota. α-Diversity analysis α-Diversity reflects the overall diversity of the gut microbiota. The ACE and Chao1 indices quantify community richness, whereas the Shannon and Simpson indices reflect community diversity (Wang et al., 2023). Fig. 8 presents the α-diversity indices across all experimental groups. Regarding community richness, both the ACE (P = 0.0188, P < 0.05) and Chao1 (P = 0.0203, P < 0.05) indices were significantly higher in the high-fat and high-sugar group compared with the normal diet group, indicating that a high-fat and high-sugar diet markedly increased gut microbial richness in mice. The treatment group also exhibited higher richness values than the normal diet group; however, no significant difference was observed when compared with the high-fat and high-sugar group. These results suggest that administration of A. mamillata total saponins did not significantly reduce microbial richness, but also did not further exacerbate the diet-induced increase. Regarding community diversity, the Shannon index was significantly higher in the treatment group than in the high-fat and high-sugar group (P = 0.0371, P < 0.05). Although no significant difference was observed between the high-fat and high-sugar group and the normal diet group, the increase in the treatment group suggests that A. mamillata total saponins may have induced a modest improvement in microbial diversity. In contrast, the Simpson index showed no significant differences among groups, indicating that overall diversity remained relatively stable regardless of diet or saponin intervention. Taken together, these findings indicate that a high-fat and high-sugar diet significantly increased gut microbial richness, while treatment with A. mamillata total saponins exerted a limited but measurable effect on microbial composition. Specifically, it partially modulated richness and modestly enhanced diversity without inducing broad shifts in overall community diversity. α-diversity analysis; ACE index; Chao1 index; Shannon index; Simpson index. β-Diversity analysis As shown in the principal component analysis (PCA) plot (Fig. 9), samples from the normal diet group and the high-fat and high-sugar group were clearly separated, indicating distinct differences in gut microbial community structure. The treatment group clustered between these two groups and was positioned closer to the normal diet group, suggesting that A. mamillata total saponins exerted a modulatory effect by shifting the gut microbiota toward a composition more similar to that of the normal diet group. Principal coordinates analysis (PCoA) based on Bray–Curtis distance yielded consistent results. By projecting high-dimensional data into a lower-dimensional space while preserving pairwise sample dissimilarities, the PCoA plot clearly demonstrated differences in microbial community structure among the groups. The treatment group again clustered more closely with the normal diet group, further supporting the modulatory effect of A. mamillata total saponins. Similarly, non-metric multidimensional scaling (NMDS) analysis based on Bray–Curtis distance revealed comparable patterns. As a nonlinear ordination method well suited for ecological data, NMDS arranges samples according to their compositional similarity. In this analysis, the treatment group was located between the normal diet and high-fat and high-sugar groups, and closer to the normal diet group. This further confirms that A. mamillata total saponins contribute to the partial restoration of gut microbial composition disrupted by a high-fat and high-sugar diet. β-diversity analysis of gut microbiota. (A) PCA (B) PCoA based on Bray–Curtis distance; (C) NMDS based on Bray–Curtis distance. Analysis of gut microbial community structure As shown in Fig. 10(A), phylum-level analysis of the gut microbial community revealed that Firmicutes and Bacteroidetes were the dominant phyla across all groups. In the normal diet group, their relative abundances were 40% and 53%, respectively; in the high-fat and high-sugar group, 64% and 30%; and in the treatment group, 53% and 39%. The F/B ratios were 0.75, 2.13, and 1.36, respectively. The human gut microbiota contains approximately 104 bacterial species, with Firmicutes (Gram-positive) and Bacteroidetes (Gram-negative) together accounting for nearly 90% of the total microbial population (Shi, 2023). Previous studies have reported that an increased abundance of Firmicutes, a reduced abundance of Bacteroidetes, or an elevated F/B ratio is associated with obesity (Fei and Zhao, 2013; Cotillard et al., 2013; Cai et al., 2025). Following administration of A. mamillata total saponins, the relative abundance of Firmicutes decreased, while that of Bacteroidetes increased, resulting in a lower F/B ratio compared with the high-fat and high-sugar group. These findings suggest that A. mamillata total saponins may help regulate the F/B ratio by suppressing Firmicutes and promoting Bacteroidetes, thereby potentially alleviating high-fat diet–induced obesity in mice. Fig. 10(B) presents the genus-level microbial community structure. The dominant genera in the normal diet, high-fat and high-sugar, and treatment groups included Muribaculaceae (32%, 23%, and 28%, respectively), Lachnospiraceae (6%, 15%, and 11%), the NK4A136 group (2%, 17%, and 8%), Lactobacillus (7%, 9%, and 2%), and Alloprevotella (7%, 2%, and 4%). Administration of A. mamillata total saponins increased the relative abundance of Muribaculaceae and Alloprevotella, while decreasing Lachnospiraceae_NK4A136 and Lactobacillus. Muribaculaceae, belonging to the Bacteroidetes phylum (Ruxin, 2022), are dominant members of the rodent gut microbiota. They are highly oxygen-sensitive and require strictly anaerobic conditions for growth. Their taxonomy and functional roles have only recently been elucidated through genomic analyses. Muribaculaceae are associated with host metabolic health and have been linked to reduced risks of inflammatory bowel disease, obesity, and diabetes (Takeuchi et al., 2023). Alloprevotella can produce SCFAs, and its abundance is negatively associated with metabolic syndrome and diabetes (Zhang et al., 2015). In contrast, certain members of Firmicutes, such as Lachnospiraceae, have been implicated in the development of obesity and type 2 diabetes (Yang et al., 2025). Fig. 10(C) presents a heatmap of genus-level abundance. The color gradient (red to blue) represents increasing abundance, while branch lengths indicate similarity in abundance profiles among groups. The normal diet and high-fat and high-sugar groups clustered relatively closely, whereas the treatment group formed a distinct cluster, suggesting that A. mamillata total saponins significantly reshaped the genus-level gut microbial community structure in high-fat diet–induced mice. Microbial composition at the phylum level (A), genus level (B), and heatmap of genus-level gut microbiota abundance (C). Effects of A. mamillata total saponins on SCFA content in feces of mice fed a higher-fat and -sugar diet Changes in fecal SCFAs in mice are presented in Fig. 11. Compared with the control group, the production of acetate (P > 0.31), propionate (P = 0.0009, P < 0.01), n-butyrate (P = 0.0063, P < 0.01), isobutyrate (P < 0.0001), n-valerate (P = 0.0079, P < 0.01), and isovalerate (P > 0.20) by the gut microbiota was reduced in the model group. In comparison with the model group, the production of acetate (P < 0.0001), propionate (P = 0.0123, P < 0.05), n-butyrate (P < 0.0001), isobutyrate (P > 0.72), n-valerate (P = 0.0402, P < 0.05), and isovalerate (P > 0.60) was increased in the A. mamillata total saponin group. Among these metabolites, acetate and n-butyrate were significantly increased, while propionate and n-valerate showed a moderate but statistically significant increase in the A. mamillata group. The total SCFA content was highest in the A. mamillata group and lowest in the model group. Overall, these results indicate that A. mamillata total saponins enhance gut microbial SCFA production. Effect of A. mamillata total saponins on SCFA content in feces of mice fed a higher-fat and -sugar diet (n = 6, ∗: p < 0.05, ∗∗ and above: p < 0.01). Discussion Recent studies have demonstrated that plant-derived saponins exhibit a range of biological activities, including the scavenging of lipid peroxidation products, enhancement of hepatic energy storage, hepatocyte protection, and modulation of the gut microbiota (Liu et al., 2015; Su, 2024). Accordingly, the development of an efficient and rapid extraction method is a critical prerequisite for fully realizing the application potential of A. mamillata total saponins. Although conventional reflux extraction is operationally straightforward, its limitations, including prolonged extraction time, high energy consumption, and the potential degradation of thermolabile constituents under elevated temperatures, restrict its suitability for industrial-scale applications (Luo et al., 2025). In the present study, an UAE process was optimized. Under the optimal conditions (71% ethanol, solid–liquid ratio of 1:32 g/mL, 61 °C, 31 min), the yield of total saponins from A. mamillata reached 27.52 mg/g, representing a significant increase of 13.73% compared with the conventional ethanol reflux method. This result is consistent with the findings of Du et al. (Liu et al., 2022) and Chen et al. (Navarro del Hierro et al., 2018b) in studies on quinoa and Ziziphus jujuba seed saponins, further confirming that ultrasonic cavitation can effectively disrupt plant cell walls and facilitate the release of bioactive compounds. Importantly, the optimized process achieved a high extraction yield under relatively mild thermal conditions. Such low-temperature efficiency helps minimize the degradation of saponin glycosidic bonds induced by excessive heat, thereby better preserving their native structure and biological activity. In summary, the UAE method developed in this study demonstrates clear advantages over conventional approaches in terms of extraction efficiency, energy consumption, and preservation of active constituents, providing an efficient, energy-saving, and environmentally friendly strategy for the industrial utilization of A. mamillata resources. The gut microbiota comprises a complex community of microorganisms inhabiting the human intestinal tract and typically maintains a dynamic ecological balance with both the host and the external environment. Disruption of this equilibrium can impair multiple host physiological functions and contribute to the development of various diseases (Peng et al., 2022). Approximately 70% of hepatic blood flow is derived directly from the gastrointestinal tract via the portal vein, rendering the liver the first organ exposed to gut-derived metabolites and microbial products. Consequently, gut microbiota dysbiosis and impairment of the intestinal mucosal barrier may expose hepatocytes, Kupffer cells, and hepatic stellate cells to microbial-derived factors, thereby inducing the production of pro-inflammatory cytokines that play a central role in the initiation and progression of MAFLD (Barber et al., 2023). Recent studies have demonstrated that saponins can modulate the gut microbiota, restore intestinal microecological balance, and contribute to the prevention and treatment of MAFLD (Malnick and Michael, 2023). Microbial α-diversity reflects the richness and diversity within microbial communities, where the Chao index estimates species richness and the Shannon index reflects overall diversity. Zhong et al. (2022) reported that total saponins from Gynostemma pentaphyllum could ameliorate gut dysbiosis, restore microbial diversity, and improve MAFLD. Consistently, in the present study, total saponins from A. mamillata modulated microbial richness (Chao index) and diversity (Shannon index) in mice, restoring these parameters toward levels observed in the control group. Accumulating evidence supports a key role for the gut microbiota in metabolic disorders such as obesity and type 2 diabetes. At the phylum level, Firmicutes are often increased in obese individuals, whereas Bacteroidetes are reduced (Zhou et al., 2016). Certain members of Firmicutes, such as the Lachnospiraceae family, have been implicated in obesity and type 2 diabetes (Kahn and Davidson, 2014; Selph et al., 2025). Consistent with these findings, A. mamillata saponins increased the relative abundance of Bacteroidetes while decreasing that of Firmicutes, thereby reducing the F/B ratio. At the genus level, A. mamillata total saponins increased the relative abundance of Muribaculaceae and Alloprevotella, while decreasing Lachnospiraceae_NK4A136 and Lactobacillus. Muribaculaceae and Alloprevotella are important producers of SCFAs in the gut (Tang et al., 2024; Bai et al., 2025). Muribaculaceae have been shown to play a key role in maintaining intestinal homeostasis, suppressing inflammation, and alleviating metabolic syndrome (Takeuchi et al., 2023), while increased abundance of Alloprevotella is generally associated with a healthier metabolic profile (Zhang et al., 2015). The observed enrichment of these beneficial taxa provides a microbial basis for the subsequent increase in SCFA production. In contrast, a high-fat diet significantly increased the abundance of Lachnospiraceae_NK4A136, a group that has been associated with insulin resistance and obesity in several studies (Yang et al., 2025). Although Lactobacillus is commonly regarded as a probiotic genus, emerging evidence suggests that under conditions of diet-induced metabolic dysregulation, its overrepresentation may not be uniformly beneficial and could be associated with specific metabolic disturbances (Tang et al., 2021; Kwon et al., 2018; Lee et al., 2021). Collectively, these findings suggest that A. mamillata total saponins may contribute to the remodeling of gut microbial communities by suppressing taxa that become overabundant under pathological metabolic conditions. The gut microbiota ferments dietary fiber to produce SCFAs, which not only serve as the primary energy source for colonocytes but can also enter the liver via the portal vein, where they participate in systemic energy metabolism and immune regulation (Hu et al., 2018; Kim, 2023; Tan et al., 2023). In the present study, intervention with A. mamillata total saponins significantly increased the levels of acetate and butyrate, markedly elevated propionate and valerate, and restored total SCFA levels. As a natural product, A. mamillata total saponins are poorly digestible and absorbable by the host; however, they may function as potential prebiotic-like compounds. These saponins can be selectively utilized by beneficial bacteria, such as Muribaculaceae and Alloprevotella, thereby promoting their proliferation and enhancing SCFA-producing capacity (Chen et al., 2015; Luo et al., 2020). Notably, the significant increase in valerate was closely associated with the enrichment of Alloprevotella. As a potent histone deacetylase (HDAC) inhibitor, valerate may contribute to the maintenance of metabolic homeostasis through epigenetic regulation of gene expression (Paciolla et al., 2025). Among SCFAs, butyrate plays a key role in maintaining intestinal barrier integrity and attenuating hepatic inflammation by reducing endotoxin translocation into the circulation (Fan et al., 2021). Acetate and propionate, after hepatic uptake, may suppress de novo lipogenesis through activation of the AMPK signaling pathway or inhibition of key lipogenic enzymes (Li et al., 2013; Iannucci et al., 2016). These mechanisms are consistent with the observed reductions in serum TG and TC levels, as well as the improvement in hepatic steatosis in this study, suggesting that A. mamillata total saponins may exert lipid-lowering effects via modulation of the “gut microbiota–SCFAs–host metabolism” axis. Collectively, these findings provide a mechanistic foundation for future studies and may offer novel strategies for the prevention and treatment of metabolic diseases. Although this study provides meaningful insights, several limitations should be acknowledged. First, the effects of A. mamillata total saponins on the gut microbiota were evaluated exclusively in a murine model, and the underlying molecular mechanisms were not fully elucidated. In addition, the pharmacological activities of individual bioactive constituents within the total saponin fraction remain to be characterized in detail. Therefore, future studies should focus on elucidating the mechanistic pathways involved, isolating and purifying the active compounds, and conducting systematic pharmacological evaluations. Such investigations will provide stronger experimental evidence to support the potential development of A. mamillata total saponins as functional food ingredients or therapeutic agents. Conclusions In summary, UAE is an efficient method for obtaining total saponins from A. mamillata. The optimal extraction conditions were determined to be 71% ethanol, 31 min of ultrasonication, 61 °C, and a solid–liquid ratio of 1:32 (g/mL). Furthermore, A. mamillata total saponins were shown to inhibit body weight gain in obese mice to a certain extent and to alleviate hepatic fat accumulation, hepatocyte injury, and lipid metabolism disorders induced by a high-fat and high-sugar diet. Mechanistically, A. mamillata total saponins improved intestinal homeostasis by increasing the abundance of Muribaculaceae and Alloprevotella, elevating SCFA levels, suppressing Lachnospiraceae_NK4A136 and Lactobacillus, and modulating the F/B ratio. Collectively, these findings highlight the potential of A. mamillata total saponins in modulating the gut microbiota and provide important insights into their possible mechanisms for ameliorating metabolic disorders. CRediT authorship contribution statement Jinhua Shao: Writing – review & editing, Writing – original draft, Conceptualization. Junyang Liu: Writing – original draft, Data curation. Jiaxin Huang: Methodology, Investigation. Yuqing Zhu: Software, Resources. Changhui Zhao: Methodology, Investigation, Data handling, Formal analysis. Data availability statement All data generated or analyzed during this study are included in this article. Ethical approval All experimental procedures complied with the Guide for the Care and Use of Laboratory Animals. The research project was ethically supervised by the Academic Ethics Committee of Hunan University of Science and Engineering. Funding statement This article has received support from the Hunan Provincial Natural Science Foundation (2024JJ7196). 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. Handling Editor: Dr. Siyun Wang References Associated Data Data Availability Statement All data generated or analyzed during this study are included in this article.