Abstract As a multi-factorial disease, obesity has become one of the major health problems in the world, and it is still increasing rapidly. Konjac supplementation, as a convenient dietary therapy, has been shown to be able to regulate gut microbiota and improve obesity. However, the specific mechanism by which konjac improves obesity through gut microbiota remains to be studied. In this study, a high-fat diet (HFD) was used to induce a mouse obesity model, and 16S rDNA sequencing and an untargeted metabolomics were used to investigate the impact of konjac on gut microbiota and gut metabolites in HFD-induced obese mice. The results show that konjac can reduce the body weight, adipose tissue weight, and lipid level of high-fat diet induced obese mice by changing the gut microbiota structure and gut metabolic profile. Association analysis revealed that konjac supplementation induced changes in gut microbiota, resulting in the up-regulation of 7-dehydrocholesterol and trehalose 6-phosphate, as well as the down-regulation of glycocholic acid and ursocholic acid within the Secondary bile acid biosynthesis pathway, ultimately leading to improvements in obesity. Among them, g_Acinetobacter (Greengene ID: 911888) can promote the synthesis of 7-dehydrocholesterol by synthesizing ERG3. g_Allobaculum (Greengene ID: 271516) and g_Allobaculum (Greengene ID: 259370) can promote the breakdown of trehalose 6-phosphate by synthesizing glvA. Additionally, the down-regulation of glycocholic acid and ursocholic acid may be influenced by the up-regulation of Lachnospiraceae_NK4A136_group. In conclusion, konjac exerts an influence on gut metabolites through the regulation of gut microbiota, thereby playing a pivotal role in alleviating obesity induced by a high-fat diet. Keywords: Konjac supplementation, 16S rDNA sequencing, Untargeted metabolomics, Association analysis, High-fat diet, Obesity Graphical abstract Image 1 [51]Open in a new tab Highlights * • Konjac can alleviate obesity induced by high-fat diet. * • Konjac can alter the species composition of bacteria in the intestines of obese mice. * • Konjac can alter the structure of gut metabolites. * • Konjac may alleviate obesity through its effects on gut microbiota and metabolites. 1. Introduction Over two billion people worldwide suffer from obesity, a condition characterized by abnormal accumulation of white adipose tissue ([52]Caballero, 2019; [53]Chandrasekaran and Weiskirchen, 2024). Furthermore, obesity is a significant public health concern because of its association with diabetes, cardiovascular disease, and cancer ([54]Caruso et al., 2023). The pathogenesis and progression of obesity, as well as its impacts on human health, have been largely elucidated by global scientific researchers over the past few years ([55]Aron-Wisnewsky et al., 2019; [56]Breit et al., 2023; [57]Marcelin et al., 2019). Despite these advances, the complex etiology and multiorgan damage associated with obesity present challenges in developing a practical and efficient treatment. Therefore, it is imperative to conduct additional research on the causes and advancement of obesity, as well as to devise straightforward and efficient treatment and intervention approaches. This will not only improve the well-being of individuals and overall health outcomes but also reduce the financial strain on healthcare systems. Dietary therapy and traditional medicine are viable approaches to tackling obesity, with konjac powder emerging as a popular option owing to its natural health benefits. The primary component of konjac powder, konjac glucomannan (KGM), not only enhances the taste of food and keeps it fresh, but also regulates blood sugar, blood lipids, gut microbiota, oxidative stress, and immune function ([58]Ho et al., 2017; [59]Jian et al., 2017; [60]Rogovik and Goldman, 2009; [61]Zhao et al., 2020). Concurrently, KGM demonstrates the capacity to extend gastric emptying, enhance satiety, stimulate liver glycogen synthesis, enhance gut microbiota and metabolic functions via diverse molecular pathways, consequently influencing oxidative stress and immune inflammation regulation, and providing hepatoprotective and nephroprotective effects ([62]Tester and Al-Ghazzewi, 2016). However, the specific mechanism by which konjac exerts the bioactive role of anti-obesity is still controversial. While Xu et al. demonstrated that konjac may reduce obesity by increasing satiety and decreasing energy consumption (C. [63]Xu et al., 2023), conflicting results were reported by Hong et al. and Liu et al. who found that konjac supplementation did not lead to a reduction in energy intake ([64]Hong et al., 2023; [65]Liu et al., 2023). As a result, further investigation into the mechanisms behind konjac's beneficial effects is warranted to improve our understanding of obesity treatment. The gut microbiota is recognized as a significant contributor to the pathogenesis of metabolic disorders and functions as an endocrine organ that plays a role in regulating energy balance and immune responses ([66]Marchesi et al., 2016). The dysregulation of gut microbiota is a prominent characteristic of obesity, characterized by noticeable differences in microbial diversity between individuals with obesity and those without. This disturbance in the gut ecosystem can compromise the integrity of the gut barrier, causing changes in the levels of various gut metabolites including short-chain fatty acids (SCFAs), indole derivatives, and polyamines. These alterations have the potential to disrupt lipid metabolism and initiate inflammatory processes, ultimately contributing to the development of obesity ([67]Boulangé et al., 2016; [68]Cani et al., 2009; [69]Cuevas-Sierra et al., 2019). Mechanistic investigations have demonstrated that oral KGM supplementation enhances the growth of beneficial gut microbiota, modulates the gut microenvironment, and attenuates inflammation ([70]Jiang et al., 2018). These results highlight the substantial role of gut microbiota in mediating the beneficial effects of konjac on obesity. However, additional studies are necessary to explore the effects of konjac powder supplementation on the metabolic interactions between gut microbiota and its metabolites in the context of a high-fat diet. The aim of the study is to investigate what impact konjac powder might have on the gut microbiota and gut metabolites of mice who have been induced to become obese by high-fat diets. Utilizing 16S rDNA sequencing and non-targeted metabolomics, the relationship between konjac, gut microbiota, and gut metabolites was examined to elucidate the potential mechanism by which konjac may alleviate obesity. 2. Materials and methods 2.1. The preparation of konjac powder A white konjac (from Pingshan County, Yibin City, Sichuan Province, China) was cut into pieces and dried. It was then crushed by a portable grinder at room temperature and ground by a planetary ball mill. In 2 h, 200 rpm was used to grind the konjac, resulting in crude powder. Konjac glucomannan (KGM) was subsequently purified from crude konjac powder through a process involving dissolution of 60% alcohol and crude konjac powder at a volume ratio of 100:1, sonication for 20 min, microwave heating at 350 W for 20 s, vacuum filtration, and drying to yield konjac powder. The purified konjac powder contains a high concentration of KGM, exceeding 85% ([71]Dong Li et al., 2020). A solution of konjac powder was prepared by adding it to saline at a concentration of 0.08 g/ml for use in subsequent experiments involving mice. 2.2. Animal feeding and sample collection Six-week-old male C57BL/6J mice, bred under specific pathogen-free (SPF) conditions, were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. The mice were housed in the SPF barrier system of the Experimental Animal Center of Southwest Medical University, maintained at an ambient temperature of 22 ± 2 °C, humidity of 50%–60%, and a 12-h light-dark cycle. Throughout the experiment, the mice had ad libitum access to food and water. Following a one-week acclimation period, the C57BL/6 J mice were randomly assigned to either a control (CK) group or a high-fat diet (HFD) group. The high-fat diet (D12492) supplied by Research Diets is composed of 20% of calories derived from carbohydrates, 20% from protein, and 60% from fat. Detailed ingredients of the normal diet and the HFD can be found in [72]Supplement table 1. At the twelfth week of the study, mice that had been fed a high-fat diet were randomly assigned to either the model control (MK) group (n = 6) or the konjac (KON) group (n = 6), while mice on the control diet (CK) (n = 6) maintained their original diet. The KON group received a daily dose of konjac powder (0.8 g/kg) in addition to their high-fat diet for a duration of 5 weeks, as well as 0.1 ml of the konjac solution per 10 g of body weight via gavage. In contrast, CK and MK mice received an equivalent volume of normal saline daily via gavage, in addition to their regular diet. The konjac powder utilized in the study was supplied by Sichuan University of Science & Engineering. Following the completion of the experiment, all mice underwent an overnight fast and were anesthetized with 1% pentobarbital sodium (50 mg/kg) in the early morning of the subsequent day prior to euthanasia by cervical dislocation. Blood samples were promptly obtained and allowed to acclimate to room temperature. Adipose tissue surrounding the testis was harvested and weighed. The cecal contents situated distal to the ileocecal valve were collected and preserved at −80 °C for subsequent analysis using 16S rDNA sequencing and untargeted metabolome profiling. 2.3. Pathological section of adipose tissue Adipose tissue surrounding the epididymis was preserved in 4% paraformaldehyde, washed, dehydrated, embedded in paraffin, sectioned into 4 μm slices, and subsequently stained with hematoxylin-eosin (H&E). Photographs of the sample tissue were taken under a microscope for assessing adipose tissue histopathological features, and Image Pro Plus 6.0 was used for measuring and analyzing adipose cell area. 2.4. Serum biochemical analysis The collected blood samples were left at room temperature for 4 h, centrifuged at 4 °C and 3500r/min for 10 min, then the serum was collected. An equivalent volume of serum was extracted from each sample, and the concentrations of triglycerides (TG), total cholesterol (TC), low-density lipoprotein (LDL), and high-density lipoprotein (HDL) in the serum were measured using the TC220 automatic veterinary biochemical analyzer (manufactured by Tekang Technology Co., Ltd.). 2.5. 16S rDNA sequencing The genomic DNA of gut microbiota in cecum contents was extracted using HiPure Stool DNA Kit (D3141, Guangzhou Meiji Biological Co., LTD., China). 16S rDNA V3–V4 region was amplified using a specific primer (341F, CCTACGGGRBGCASCAG; 806R: GGACTACNNGGGTATCTAAT). Subsequently, the amplified products were quantified, purified, and utilized for library construction and sequencing. Following the sequencing process, raw reads were obtained and subsequently filtered for low quality using FASTP (V0.18.0, [73]https://github.com/OpenGene/fastp). Afterward, FLASH (V1.2.11, [74]http://www.cbcb.umd.edu/software/flash) was employed to splice reads to obtain tags. The tags acquired were subsequently subjected to filtration and elimination of chimeras to procure efficient tags. Abundance statistics for Operational Taxonomic Units (OTUs) were subsequently computed utilizing the filtered tags. After comparing the sequence information of the OTUs with the SILVA database, species annotation information for each OTU was obtained. Linear discriminant analysis effect size (LEfSe) analysis (V1.0, [75]http://huttenhower.sph.harvard.edu/lefse/) was used to detect biomarkers in gut microbiota, and species with LDA value > 4 were selected as biomarkers. Alpha diversity analysis was performed using QIIME (V1.9.1, [76]http://qiime.sourceforge.net/). R studio was used to conduct beta diversity analysis, including PCoA and NMDS analyses based on linear and nonlinear models. 2.6. Untargeted metabolomics The cecal contents were thawed at 4 °C and subsequently mixed with a solution consisting of methanol, acetonitrile, and water in a 2:2:1 ratio (v/v). The mixture was then subjected to ultrasound treatment at low temperature for 30 min, followed by incubation at −20 °C for 10 min. Subsequently, the supernatant was obtained by centrifugation at 14000g for 20 min at 4 °C. An equal volume was taken from each sample to be tested and mixed together as a quality control (QC) sample. Sample supernatant was placed into the Liquid Chromatograph-Mass Spectrometer (LC-MS) for analysis. Ultra-High performance liquid chromatography (UHPLC, Agilent 1290 Infinity LC) HILIC column was used for separation: column temperature 25 °C; flow rate 0.5 ml/min; sample size 2 μL; mobile phase A: water +25 mM ammonium acetate +25 mM ammonia water; and mobile phase B: acetonitrile. Random sequential sampling was employed to assess the stability of the system. Subsequently, the primary and secondary spectra of the samples were acquired using the AB Sciex TripleTOF 6600 mass spectrometer after elution (Shanghai Applied Protein Technology Co., Ltd.). The ESI source parameters were configured as follows: Ion Source Gas1 (Gas1) set to 60, Ion Source Gas2 (Gas2) set to 60, curtain gas (CUR) set to 30, source temperature set to 600 cquired using the AB Sciex TripleTOF 6600 mass spectrome During MS acquisition, the instrument was configured to capture mass-to-charge ratios ranging from 60 to 1000 Da, with an accumulation time of 0.20 s per spectra for the TOF MS scan. The production scan was performed utilizing information dependent acquisition (IDA) in high sensitivity mode, with specific parameters set as follows: a fixed collision energy (CE) of 35 V with a range of ±15 eV; declustering potential (DP) of 60 V (+) and −60 V (−); exclusion of isotopes within 4 Da, and monitoring of 10 candidate ions per cycle. Following this, the molecular characteristic peaks in secondary mass spectrometry were identified, and substance annotations were conducted in conjunction with Mass Bank, Metlin, MoNA, and other publicly available databases. Multivariate statistical analysis and cluster analysis were executed using R studio. Additionally, pathway enrichment analysis was performed utilizing the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. 2.7. MIMOSA2 analysis MIMOSA2 analysis ([77]http://elbo-spice.cs.tau.ac.il/shiny/MIMOSA2shiny/) ([78]Noecker et al., 2022) involves a regression analysis that utilizes a database to forecast microbial metabolic potential and subsequently establishes correlations between the predicted metabolic potential and observed metabolomic data. Following the annotation of OTU sequences in various databases such as KEGG, NCBI, EMBL-EBI, VMH, and others, MIMOSA2 developed a network prediction model of metabolic capacity utilizing OTU abundance and established metabolic functions. This model was employed to compute the Metabolic Potential (MP) score for each microbial taxon, facilitating the prediction of the influence of individual microbial units on metabolites present in the sample. Furthermore, a Community-level Metabolic Potential (CMP) score was generated by consolidating the metabolic potentials of all microbial units.Subsequently, the CMP scores were subjected to regression analysis against metabolite abundances determined through metabolomics, employing a rank prediction-based regression model to evaluate the predictive capability of CMP scores on metabolite levels. If P < 0.1, it was considered that the change in metabolite level was mediated by microorganisms. Finally, the overall model fit was decomposed into the contribution from each taxon, and the specific taxa that can explain variation in each metabolite was identified. 2.8. Statistical analysis This study used SPSS 25.0 statistical analysis software for data statistical analysis. The Shapiro-Wilk test was used to perform a normal test on the data, one-way ANOVA was performed on the data conforming to normal distribution. Among them, those with homogeneous variance were analyzed by Bonferroni, and those with heterogeneous variance were analyzed using by Tamhane's T2. The Kruskal-Wallis rank sum test was performed on the data not conforming to normal distribution. All values were expressed using mean ± SD. Spearman statistical analysis was used to analyze the correlation between different types of data in the sample, and R studio was used for matrix thermal mapping and hierarchical clustering. P < 0.05 shows that the experimental results in statistical significant difference, P < 0.01 shows that the result of the experiment is extremely remarkable difference in statistics. 3. Results 3.1. The effect of konjac on the phenotype of obese mice After 12 weeks of HFD, the body weight of HFD mice was significantly higher than that of CK mice ([79]Fig. 1A, P < 0.05). Before Konjac supplementation, there was no significant difference in body weight between KON and MK mice ([80]Fig. 1B). After konjac supplementation, KON mice continued to lose body weight, with a significant difference from MK mice at week 13 (P < 0.05), and the significant difference from CK mice disappeared at week 16 ([81]Fig. 1B). There was no significant difference in total food intake and energy intake between KON and MK mice over the course of the 5-week konjac supplementation period ([82]Supplement Figs. 1a–b). In addition, the adipose tissue weight and the percentage of adipose tissue weight to body weight of MK mice were significantly higher than that of CK mice after HFD (P < 0.01), but after konjac supplementation, the adipose tissue weight as well as the percentage of adipose tissue weight to body weight of KON mice were significantly lower than that of MK group ([83]Fig. 1C–D, P < 0.05). From the H&E staining, we observed that konjac supplementation also had the effect of reducing the area of adipocytes ([84]Fig. 1E), but after calculating the average area of adipocytes in the sections, we found that this effect was not significant ([85]Fig. 1F). After HFD, the serum levels of TC, LDL, and HDL in MK mice were significantly higher than those in CK mice ([86]Fig. 1G, P < 0.01). After konjac supplementation, serum TC, TG, and HDL levels in KON mice were significantly lower than those in MK mice ([87]Fig. 1G, P < 0.05). These results suggest that konjac glucomannan (KGM) in konjac powder could significantly reduce the HFD-induced increase in body weight, adipose tissue weight, and serum lipid levels in mice. Fig. 1. [88]Fig. 1 [89]Open in a new tab Konjac supplementation can significantly reduce the increase of body weight, adipose tissue weight, and serum lipid level induced by HFD in mice (n = 6 per group). (A) Body weight of mice in the high-fat diet induced obesity model (0–12 weeks); (B) body weight of mice during konjac supplementation (12–17 weeks); (C) adipose tissue weight; (D) the percentage of adipose tissue weight to body weight; (E) H&E stained pathological sections of adipose tissue ( × 400); (F) average adipocyte area; and (G) serum TC, TG, LDL and HDL levels. Different lowercase letters indicate significant differences between groups. “*” indicates P < 0.05 and “**” indicates P < 0.01. 3.2. The effects of konjac on gut microbiota in obese mice 3.2.1. Alpha diversity and beta diversity The Alpha diversity analysis revealed that the ACE, Chao1, and Shannon indices of gut microbiota in CK mice were significantly higher compared to those in MK and KON mice (P < 0.05). However, there was no significant difference observed between MK and KON mice in this regard ([90]Fig. 2A). The Simpson index also did not differ significantly among the three groups of mice ([91]Fig. 2A). This indicates that konjac supplementation did not significantly alter the Alpha diversity of the gut microbiota. However, in the Beta diversity analysis, PCoA analysis and NMDS analysis both showed a clear separation trend between gut microbiota of CK, MK, and KON mice ([92]Fig. 2B). The stress value in NMDS analysis was less than 0.2, indicating that the model could accurately reflect the differences between samples. These results indicate that KGM in konjac powder can significantly affect the gut microbiota structure of HFD-induced mice. Fig. 2. [93]Fig. 2 [94]Open in a new tab Konjac supplementation significantly alters the structure and species composition of gut microbiota in mice after HFD (n = 6 per group). (A) Alpha diversity analysis, including ACE, Chao1, Shannon, and Simpson values; (B) beta diversity analysis, including PCoA and NMDS analysis; (C) Score diagram and (D) evolutionary branching diagram of bacteria with LDA score greater than 4 in LEfSe analysis; (E) venn diagram of the number of OTUs in each group; (F) phyla level bacteria with the top 10 relative abundances; and (G) genera level bacteria with the top 20 relative abundance. “#” indicates a significant difference between CK and MK mice, “&” indicates a significant difference between CK and KON mice, and “*” indicates a significant difference between MK and KON mice, P < 0.05. 3.2.2. LEfSe analysis We screened the biomarkers of genus level in each group. The biomarkers in CK mice were Marvinbryantia, Candidatus_Saccharimonas, and Allobaculum. The biomarkers of MK mice were Coriobacteriaceae_UCG_002, Colidextribacter, Faecalibaculum, Blautia, and Dubosiella. The biomarker in KON mice was Lachnospiraceae_NK4A136_group ([95]Fig. 2C–D). 3.2.3. Species composition analysis of gut microbiota Venn diagrams show that the three groups of mice shared 283 OTUs, accounting for 22% of the total OTUs. The number of OTUs unique to CK, MK, and KON mice was 505, 136, and 175, respectively, accounting for 39%, 10%, and 13% of the total OTUs, respectively ([96]Fig. 2E). Stacked plots show the top 10 phyla level bacteria by relative abundance, among which Bacteroidota was significantly downregulated in both MK and KON mice relative to CK mice, and Desulfobacterota and Patescibacteria were significantly up-regulated and down-regulated in KON mice compared with CK mice, respectively ([97]Fig. 2F, P < 0.05). In addition, Actinobacteriota was down-regulated in KON mice relative to MK mice after konjac supplementation ([98]Fig. 2F, P < 0.05). At the genus level, it is noteworthy that eight of the top 20 genera in relative abundance were biomarkers in the LEfSe analysis ([99]Fig. 2G). Among them, the relative abundance of MK mice biomarkers Coriobacteriaceae_UCG_002, Colidextribacter, Faecalibaculum, Blautia, and Dubosiella was significantly up-regulated in MK mice compared with CK mice (P < 0.05). Lachnospiraceae_NK4A136_group, a biomarker in KON mice, was significantly up-regulated as compared to MK mice (P < 0.05). Biomarkers of CK mice, Candidatus_Saccharimonas and Allobaculum, were both significantly up-regulated in CK mice relative to MK and KON mice ([100]Fig. 2G, P < 0.05). 3.3. The effects of konjac on gut metabolites in obese mice 3.3.1. Multivariate statistical analysis and screening of differential metabolites PCA, PLS-DA and OPLS-DA all show a significant trend of separation between gut metabolites in both positive and negative ion modes in the three groups of mice ([101]Supplement Figs. 2a–c). The R2Y values in PLS-DA and OPLS-DA were greater than 0.7, and the permutation test based on OPLS-DA show that the intersection point of Q2 regression curve and ordinate of all models was less than 0 ([102]Supplement Fig. 2d). This indicates that the model predictions are reliable and that no overfitting occurred. Metabolites with VIP >1 and P < 0.05 were classified as differential metabolites. A total of 2778 differential metabolites were detected across the three groups. Specifically, 970 metabolites were found to be significantly up-regulated and 982 metabolites were significantly down-regulated in MK mice, while 944 metabolites were significantly up-regulated and 828 metabolites were significantly down-regulated in KON mice compared to CK mice ([103]Supplement Fig. 2e, P < 0.05). A statistical analysis revealed that 515 metabolites exhibited significant up-regulation, while 395 metabolites exhibited significant down-regulation in KON mice compared to MK mice ([104]Supplement Fig. 2e, P < 0.05).The cluster heatmap shows the distribution of these differential metabolites ([105]Fig. 3A). Among them, 7-dehydrocholesterol and trehalose 6-phosphate were up-regulated after HFD, while glycocholic acid and ursocholic acid were down-regulated ([106]Fig. 3A). Fig. 3. [107]Fig. 3 [108]Open in a new tab Konjac supplementation can significantly change the gut metabolites of mice after HFD (n = 6 per group). (A) Differential metabolite cluster heat map; (B) KEGG pathway enrichment bubble map of CK vs MK; (C) KEGG pathway enrichment bubble map of CK vs KON; (D) KEGG pathway enrichment bubble map of MK vs KON; (E) abundance of glycocholic acid and ursocholic acid; (F) abundance of trehalose 6-phosphate and trehalose; and (G) abundance of 7-dehydrocholesterol. “*” indicates P < 0.05 and “**” indicates P < 0.01. 3.3.2. KEGG pathway enrichment analysis and obesity-related differential metabolites The KEGG pathway enrichment analysis was performed to explore the changes in metabolic pathways. The results show that the differential metabolites between CK and MK mice and between CK and KON mice were significantly enriched in Steroid hormone biosynthesis, Steroid biosynthesis, Primary bile acid biosynthesis, Phenylalanine metabolism, and ABC transporters ([109]Fig. 3B–C, P < 0.05). In addition, the differential metabolites between KON and MK mice were significantly enriched in Secondary bile acid biosynthesis, Starch and sucrose metabolism, HMG-CoA reductase inhibitors and Vitamin B6 metabolism ([110]Fig. 3D, P < 0.05). Meanwhile, the differential metabolites between KON and CK mice were also significantly enriched in Secondary bile acid biosynthesis ([111]Fig. 3C, P < 0.05). Notably, two differential metabolites enriched in the Secondary bile acid biosynthesis pathway, glycocholic acid and ursocholic acid, are both widely believed to be positively associated with the development of obesity. In our study, both glycocholic acid and ursocholic acid were significantly upregulated in MK mice, but this was significantly reversed by konjac supplementation ([112]Fig. 3E, P < 0.01). Trehalose 6-phosphate and trehalose, which were enriched in Starch and sucrose metabolism pathways, are both thought to be negatively correlated with obesity, and they were both significantly up-regulated after konjac supplementation ([113]Fig. 3F, P < 0.05). In addition, when other differential metabolites were analyzed, we found that 7-dehydrocholesterol, which has the effect of alleviating obesity, was significantly down-regulated in MK mice, while konjac supplementation significantly up-regulated their levels in KON mice relative to MK mice ([114]Fig. 3G, P < 0.01). 3.4. Association analysis between gut 16S rDNA sequencing and untargeted metabolome 3.4.1. Correlation analysis between gut microbiota and differential metabolites To investigate the effect of konjac on gut metabolites through gut microbiota, a correlation analysis was done on eight biomarkers from the top 20 genera in relative abundance and differential metabolites in four KEGG metabolic pathways that were significantly enriched in KON mice compared to MK mice. The results show that both glycocholic acid and ursocholic acid in Secondary bile acid biosynthesis were significantly positively correlated with the biomarkers of MK mice, Coriobacteriaceae_UCG_002, Faecalibaculum, Blautia, and Dubosiella ([115]Fig. 4A, P < 0.05). Glycocholic acid and ursocholic acid were also significantly negatively correlated with the biomarkers of CK and KON mice, Candidatus_Saccharimonas and Lachnospiraceae_NK4A136_group ([116]Fig. 4A, P < 0.05). In addition, trehalose 6-phosphate was significantly negatively correlated with Coriobacteriaceae_UCG_002, a biomarker of MK mice, and positively correlated with Candidatus_Saccharimonas, a biomarker of CK mice ([117]Fig. 4A, P < 0.05). Fig. 4. [118]Fig. 4 [119]Open in a new tab Analysis of the correlation between gut microbiota and differential metabolites (n = 6 per group). (A) Heat map of correlation analysis of differential metabolites and biomarkers in LEfSe analyses; (B) regression scatter plots of CMP values and actual abundances of seven metabolites regulated by gut microbiota in MIMOSA2 analysis; and (C) heat map of the contribution of each gut microbiota taxon to alterations in these seven metabolites. “*” indicates P < 0.05 and “**” indicates P < 0.01. 3.4.2. MIMOSA2 analysis To further explore the specific role of gut microbiota on gut metabolites, we performed a MIMOSA2 analysis. The results show that there were seven differential metabolites with significant regression relationships between CMP values and the actual detected abundance in the metabolome, including 5alpha-androstane-3,17-dione, 7-dehydrocholesterol, trehalose 6-phosphate, D-glucosamine 6-phosphate, D-glucuronate, FMN, and SN-38 ([120]Fig. 4B, P < 0.1). This suggests that the gut microbiota mediated the changes in these seven metabolites. Subsequently, we identified the key contributors in the gut microbiota responsible for these metabolite changes. Among them, g_Acinetobacter (Greengene ID: 911888) can mediate the synthesis of 7-dehydrocholesterol by synthesizing delta7-sterol 5-desaturase (ERG3, [121]K00227). It also had a positive contribution to the change in the level of 7-dehydrocholesterol ([122]Fig. 4C). In addition, g_Allobaculum (Greengene ID: 271516) and g_Allobaculum (Greengene ID: 259370) can mediate the decomposition of trehalose 6-phosphate by synthesizing maltose 6′-phosphate glucosidase (glvA, [123]K01232). Both had a positive contribution to the change in the level of trehalose 6-phosphate ([124]Fig. 4C). This suggests that both the downregulation of 7-dehydrocholesterol and trehalose 6-phosphate in MK mice and the upregulation in KON mice were mediated by the gut microbiota. 3.4.3. Correlation analysis between gut microbiota, differential metabolites, and obesity-related phenotypes In order to delve deeper into the connection between the markedly changed gut microbiota, gut metabolites, and obesity-related phenotypes, the biomarkers within the top 20 genera and obesity-related differential metabolites were analyzed in relation to obesity-related phenotypes. The results show that biomarkers Candidatus_Saccharimonas, Allobaculum, and Lachnospiraceae_NK4A136_group of CK and KON mice were significantly negatively correlated with adipose tissue weight ([125]Fig. 5A, P < 0.05). Five biomarkers of MK mice, Coriobacteriaceae_UCG_002, Colidextribacter, Faecalibaculum, Blautia, and Dubosiella, were significantly positively correlated with body weight, adipose tissue weight, and average adipocyte area ([126]Fig. 5A, P < 0.05). Among them, Coriobacteriaceae_UCG_002, Blautia, and Dubosiella were also significantly positively correlated with serum TC ([127]Fig. 5A, P < 0.05). In addition, trehalose 6-phosphate and 7-dehydrocholesterol, which are regulated by gut microbiota, were both significantly negatively correlated with adipose tissue weight and mean adipocyte area ([128]Fig. 5A, P < 0.05). Trehalose 6-phosphate was also significantly negatively correlated with all serum lipid parameters ([129]Fig. 5A, P < 0.05). Glycocholic acid and ursocholic acid in Secondary bile acid biosynthesis are both widely associated with obesity-related phenotypes. They were significantly positively correlated with all obesity-related phenotypes except serum TG ([130]Fig. 5A, P < 0.05). Fig. 5. [131]Fig. 5 [132]Open in a new tab Correlation analysis and Sankey diagram between gut microbiota, differential metabolites, and obesity-related phenotypes (n = 6 per group). (A) Heat map of correlation analysis between biomarkers from LEfSe analysis, some differential metabolites, and obesity-related phenotypes; and (B) Sankey diagram of gut microbiota-differential metabolism-obesity-related phenotypes. The red nodes in the Sankey diagram represent the biomarkers of MK mice in LEfSe analysis, obesity-related phenotypes and the differential metabolites that are positively correlated with obesity; yellow nodes represent KON mice biomarkers in LEfSe analysis; blue nodes represent CK mice biomarkers in LEfSe analysis and differential metabolites that are negatively correlated with obesity; purple nodes represent key contributors in MIMOSA2 analysis. The red band indicates a positive correlation; yellow and blue bands indicate negative correlations; and purple bands indicate positive contributions. (For interpretation of the references