A gut microbial odd-chain fatty acid alleviates atherosclerosis in mice

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

The gut microbiota plays a pivotal part in human health, yet the molecular mechanisms that underlie its effects are largely unexplored. Bacteroides, a dominant genus in the human gut microbiota, is depleted in patients with atherosclerosis1, but its causal relationship with disease remains unclear. Here, using a mouse model, we show that administration of Bacteroides uniformis alleviates atherosclerosis through the upregulation of hepatic low-density lipoprotein receptor expression. Bioactivity-guided screening revealed pentadecanoic acid (PA, C15:0), a saturated odd-chain fatty acid, as a principal bioactive metabolite. PA supplementation reduced atherosclerotic plaque burden by around 50% and significantly improved plasma lipid profiles, a result that underscores its therapeutic potential. Mechanistically, PA enhances cholesterol clearance by directly inhibiting HMG-CoA reductase, suppressing hepatic cholesterol biosynthesis and promoting plasma low-density lipoprotein cholesterol removal. Analyses of 100 gut bacterial strains revealed that PA production occurs across multiple Bacteroidota genera. Notably, PA is markedly depleted in patients with dyslipidaemia. In summary, a Bacteroidota-derived odd-chain fatty acid regulates gut–liver crosstalk, and modulation of the microbial–metabolic axis has atheroprotective potential.

This is a preview of subscription content, access via your institution

Access Nature and 54 other Nature Portfolio journals

Get Nature+, our best-value online-access subscription

Receive 52 print issues and online access

Prices may be subject to local taxes which are calculated during checkout

Fig. 2: B. uniformis enhances plasma LDL-C clearance through the SREBP2–LDLR axis.Fig. 3: Isolation and identification of B. uniformis-derived metabolites that enhance LDL-C uptake.Fig. 4: PA suppresses hepatic cholesterol biosynthesis by inhibiting HMGCR activity.Fig. 5: Bacteroidota are major gut microbial producers of PA. Similar content being viewed by others Integrative metagenomic and metabolomic analyses reveal the potential of gut microbiota to exacerbate acute pancreatitis Article Open access 21 March 2024 Dietary fat alters goblet cell function and microbial bile acid metabolism to promote intestinal lipid absorption in mice Article 10 June 2026 Unveiling the oral-gut connection: chronic apical periodontitis accelerates atherosclerosis via gut microbiota dysbiosis and altered metabolites in apoE−/− Mice on a high-fat diet Article Open access 13 May 2024 Explore related subjects Discover the latest articles and news in related subjects. Bacterial host response Dyslipidaemias Probiotic Modulation of Lipid Metabolism Data availability All data supporting the findings of this study are available in the Article, its Supplementary Information and the public repositories described below. The RNA-seq data generated for this study are available at the NCBI under the accession number PRJNA1213157. Metagenomic datasets and human gut metagenomes and metagenome-assembled genomes from publicly available databases were used in this study, and raw data were accessed under the BioProjects PRJNA615842 (‘Alteration in gut microbiota composition and functional relevance in subclinical carotid atherosclerosis in the general population’), PRJEB21528 (‘The gut microbiome in atherosclerotic cardiovascular disease’) and in the European Nucleotide Archive under accession number ERP116715 for the UHGG catalogue. The human HMGCR structure used for molecular docking is available from the RCSB PDB under accession 1HW8. Source data are provided with this paper.

Jie, Z. et al. The gut microbiome in atherosclerotic cardiovascular disease. Nat. Commun. 8, 845 https://doi.org/10.1038/s41467-017-00900-1 (2017).

Article ADS PubMed PubMed Central CAS Google Scholar

Chakaroun, R. M., Olsson, L. M. & Backhed, F. The potential of tailoring the gut microbiome to prevent and treat cardiometabolic disease. Nat. Rev. Cardiol. 20, 217–235 https://doi.org/10.1038/s41569-022-00771-0 (2023).

Wasen, C. et al. Bacteroidota inhibit microglia clearance of amyloid-beta and promote plaque deposition in alzheimer’s disease mouse models. Nat. Commun. 15, 3872 https://doi.org/10.1038/s41467-024-47683-w (2024).

Article ADS PubMed PubMed Central CAS Google Scholar

Oliphant, K. et al. Bacteroidota and Lachnospiraceae integration into the gut microbiome at key time points in early life are linked to infant neurodevelopment. Gut Microbes 13, 1997560 https://doi.org/10.1080/19490976.2021.1997560 (2021).

Article PubMed PubMed Central Google Scholar

Ryan, F. J. et al. Colonic microbiota is associated with inflammation and host epigenomic alterations in inflammatory bowel disease. Nat. Commun. 11, 1512 https://doi.org/10.1038/s41467-020-15342-5 (2020).

Article ADS PubMed PubMed Central CAS Google Scholar

Puri, P. et al. The circulating microbiome signature and inferred functional metagenomics in alcoholic hepatitis. Hepatology 67, 1284–1302 https://doi.org/10.1002/hep.29623 (2018).

Article PubMed PubMed Central CAS Google Scholar

Baragetti, A. et al. Gut microbiota functional dysbiosis relates to individual diet in subclinical carotid atherosclerosis. Nutrients 13, 304 https://doi.org/10.3390/nu13020304 (2021).

Article PubMed PubMed Central CAS Google Scholar

Lim, B., Zimmermann, M., Barry, N. A. & Goodman, A. L. Engineered regulatory systems modulate gene expression of human commensals in the gut. Cell 169, 547–558 https://doi.org/10.1016/j.cell.2017.03.045 (2017).

Article PubMed PubMed Central CAS Google Scholar

Nie, Q. et al. Gut symbionts alleviate MASH through a secondary bile acid biosynthetic pathway. Cell 187, 2717–2734 https://doi.org/10.1016/j.cell.2024.03.034 (2024).

Zhang, S. et al. Bacteroides uniformis degrades β-glucan to promote Lactobacillus johnsonii improving indole-3-lactic acid levels in alleviating colitis. Microbiome 12, 177 https://doi.org/10.1186/s40168-024-01896-9 (2024).

Article PubMed PubMed Central CAS Google Scholar

Yan, Y. et al. Bacteroides uniformis-induced perturbations in colonic microbiota and bile acid levels inhibit TH17 differentiation and ameliorate colitis developments. NPJ Biofilms Microbiomes 9, 56 https://doi.org/10.1038/s41522-023-00420-5 (2023).

Article PubMed PubMed Central CAS Google Scholar

Morita, H. et al. Bacteroides uniformis and its preferred substrate, α-cyclodextrin, enhance endurance exercise performance in mice and human males. Sci. Adv. 9, eadd2120 https://doi.org/10.1126/sciadv.add2120 (2023).

Article PubMed PubMed Central CAS Google Scholar

Stiekema, L. C. A. et al. Impact of cholesterol on proinflammatory monocyte production by the bone marrow. Eur. Heart J. 42, 4309–4320 https://doi.org/10.1093/eurheartj/ehab465 (2021).

Article PubMed PubMed Central CAS Google Scholar

Luo, J., Yang, H. & Song, B. L. Mechanisms and regulation of cholesterol homeostasis. Nat. Rev. Mol. Cell Biol. 21, 225–245 https://doi.org/10.1038/s41580-019-0190-7 (2020).

Somanathan, S. et al. AAV vectors expressing LDLR gain-of-function variants demonstrate increased efficacy in mouse models of familial hypercholesterolemia. Circ. Res. 115, 591–599 https://doi.org/10.1161/CIRCRESAHA.115.304008 (2014).

Article PubMed PubMed Central CAS Google Scholar

Kim, G. et al. Gut–liver axis calibrates intestinal stem cell fitness. Cell 187, 914–930 https://doi.org/10.1016/j.cell.2024.01.001 (2024).

Article PubMed PubMed Central CAS Google Scholar

Goedeke, L. et al. MMAB promotes negative feedback control of cholesterol homeostasis. Nat. Commun. 12, 6448 https://doi.org/10.1038/s41467-021-26787-7 (2021).

Article ADS PubMed PubMed Central CAS Google Scholar

Goedeke, L. et al. MicroRNA-148a regulates LDL receptor and ABCA1 expression to control circulating lipoprotein levels. Nat. Med. 21, 1280–1289 https://doi.org/10.1038/nm.3949 (2015).

Article PubMed PubMed Central CAS Google Scholar

Goldstein, J. L. & Brown, M. S. Regulation of the mevalonate pathway. Nature 343, 425–430 https://doi.org/10.1038/343425a0 (1990).

Muta, Y. et al. Enhanced SREBP2-driven cholesterol biosynthesis by PKCλ/ι deficiency in intestinal epithelial cells promotes aggressive serrated tumorigenesis. Nat. Commun. 14, 8075 https://doi.org/10.1038/s41467-023-43690-5 (2023).

Article ADS PubMed PubMed Central CAS Google Scholar

Canfran-Duque, A. et al. Macrophage-derived 25-hydroxycholesterol promotes vascular inflammation, atherogenesis, and lesion remodeling. Circulation 147, 388–408 https://doi.org/10.1161/CIRCULATIONAHA.122.059062 (2023).

Xu, D. et al. Cholesterol sulfate alleviates ulcerative colitis by promoting cholesterol biosynthesis in colonic epithelial cells. Nat. Commun. 13, 4428 https://doi.org/10.1038/s41467-022-32158-7 (2022).

Article ADS PubMed PubMed Central CAS Google Scholar

Che, L. et al. Cholesterol biosynthesis supports the growth of hepatocarcinoma lesions depleted of fatty acid synthase in mice and humans. Gut 69, 177–186 https://doi.org/10.1136/gutjnl-2018-317581 (2020).

Yu, H. et al. GPR146 deficiency protects against hypercholesterolemia and atherosclerosis. Cell 179, 1276–1288 https://doi.org/10.1016/j.cell.2019.10.034 (2019).

Article PubMed PubMed Central CAS Google Scholar

Istvan, E. S. & Deisenhofer, J. Structural mechanism for statin inhibition of HMG-CoA reductase. Science 292, 1160–1164 https://doi.org/10.1126/science.1059344 (2001).

Pfeuffer, M. & Jaudszus, A. Pentadecanoic and heptadecanoic acids: multifaceted odd-chain fatty acids. Adv. Nutr. 7, 730–734 https://doi.org/10.3945/an.115.011387 (2016).

Article PubMed PubMed Central CAS Google Scholar

Wu, Y. et al. Galactooligosaccharides and Limosilactobacillus reuteri synergistically alleviate gut inflammation and barrier dysfunction by enriching Bacteroides acidifaciens for pentadecanoic acid biosynthesis. Nat. Commun. 15, 9291 https://doi.org/10.1038/s41467-024-53144-1 (2024).

Article ADS PubMed PubMed Central CAS Google Scholar

Qin, N., Li, L., Wang, Z. & Shi, S. Microbial production of odd-chain fatty acids. Biotechnol. Bioeng. 120, 917–931 https://doi.org/10.1002/bit.28308 (2023).

Zhan, C. et al. Improved polyketide production in C. glutamicum by preventing propionate-induced growth inhibition. Nat. Metab. 5, 1127–1140 https://doi.org/10.1038/s42255-023-00830-x (2023).

Wang, X. et al. Bacteroides methylmalonyl-CoA mutase produces propionate that promotes intestinal goblet cell differentiation and homeostasis. Cell Host Microbe 32, 63–78 https://doi.org/10.1016/j.chom.2023.11.005 (2024).

Qu, Z. et al. Selective utilization of medicinal polysaccharides by human gut Bacteroides and Parabacteroides species. Nat. Commun. 16, 638 https://doi.org/10.1038/s41467-025-55845-7 (2025).

Article ADS PubMed PubMed Central CAS Google Scholar

Liu, S. et al. Gut microbes exacerbate systemic inflammation and behavior disorders in neurologic disease CADASIL. Microbiome 11, 202 https://doi.org/10.1186/s40168-023-01638-3 (2023).

Article PubMed PubMed Central CAS Google Scholar

He, Z. et al. Microbiota in mesenteric adipose tissue from Crohn’s disease promote colitis in mice. Microbiome 9, 228 https://doi.org/10.1186/s40168-021-01178-8 (2021).

Almeida, A. et al. A unified catalog of 204,938 reference genomes from the human gut microbiome. Nat. Biotechnol. 39, 105–114 https://doi.org/10.1038/s41587-020-0603-3 (2021).

Jenkins, B. J. et al. Odd chain fatty acids; new insights of the relationship between the gut microbiota, dietary intake, biosynthesis and glucose intolerance. Sci. Rep. 7, 44845 https://doi.org/10.1038/srep44845 (2017).

Article ADS PubMed PubMed Central CAS Google Scholar

Weitkunat, K. et al. Odd-chain fatty acids as a biomarker for dietary fiber intake: a novel pathway for endogenous production from propionate. Am. J. Clin. Nutr. 105, 1544–1551 https://doi.org/10.3945/ajcn.117.152702 (2017).

Venn-Watson, S., Lumpkin, R. & Dennis, E. A. Efficacy of dietary odd-chain saturated fatty acid pentadecanoic acid parallels broad associated health benefits in humans: could it be essential? Sci. Rep. 10, 8161 https://doi.org/10.1038/s41598-020-64960-y (2020).

Article ADS PubMed PubMed Central CAS Google Scholar

Trieu, K. et al. Biomarkers of dairy fat intake, incident cardiovascular disease, and all-cause mortality: a cohort study, systematic review, and meta-analysis. PLoS Med. 18, e1003763 https://doi.org/10.1371/journal.pmed.1003763 (2021).

Article PubMed PubMed Central CAS Google Scholar

Chooi, Y. C. et al. Effect of an Asian-adapted Mediterranean diet and pentadecanoic acid on fatty liver disease: the TANGO randomized controlled trial. Am. J. Clin. Nutr. 119, 788–799 https://doi.org/10.1016/j.ajcnut.2023.11.013 (2024).

Brevik, A., Veierod, M. B., Drevon, C. A. & Andersen, L. F. Evaluation of the odd fatty acids 15:0 and 17:0 in serum and adipose tissue as markers of intake of milk and dairy fat. Eur. J. Clin. Nutr. 59, 1417–1422 https://doi.org/10.1038/sj.ejcn.1602256 (2005).

Smedman, A. E., Gustafsson, I. B., Berglund, L. G. & Vessby, B. O. Pentadecanoic acid in serum as a marker for intake of milk fat: relations between intake of milk fat and metabolic risk factors. Am. J. Clin. Nutr. 69, 22–29 https://doi.org/10.1093/ajcn/69.1.22 (1999).

Ampong, I., Watkins, A., Gutierrez-Merino, J., Ikwuobe, J. & Griffiths, H. R. Dietary protein insufficiency: an important consideration in fatty liver disease? Br. J. Nutr. 123, 601–609 https://doi.org/10.1017/S0007114519003064 (2020).

Yoshida, N. et al. Bacteroides vulgatus and Bacteroides dorei reduce gut microbial lipopolysaccharide production and inhibit atherosclerosis. Circulation 138, 2486–2498 https://doi.org/10.1161/CIRCULATIONAHA.118.033714 (2018).

Wei, W. et al. Parabacteroides distasonis uses dietary inulin to suppress NASH via its metabolite pentadecanoic acid. Nat. Microbiol. 8, 1534–1548 https://doi.org/10.1038/s41564-023-01418-7 (2023).

Article PubMed PubMed Central CAS Google Scholar

Virani, S. S., Smith, S. C. Jr., Stone, N. J. & Grundy, S. M. Secondary prevention for atherosclerotic cardiovascular disease: comparing recent US and European guidelines on dyslipidemia. Circulation 141, 1121–1123 https://doi.org/10.1161/CIRCULATIONAHA.119.044282 (2020).

Ambrosy, A. P. et al. Triglyceride levels and residual risk of atherosclerotic cardiovascular disease events and death in adults receiving statin therapy for primary or secondary prevention: insights from the KP REACH study. J. Am. Heart Assoc. 10, e020377 https://doi.org/10.1161/JAHA.120.020377 (2021).

Article PubMed PubMed Central CAS Google Scholar

Lloyd-Jones, D. M. et al. Use of risk assessment tools to guide decision-making in the primary prevention of atherosclerotic cardiovascular disease: a special report from the American heart association and American college of cardiology. Circulation 139, e1162–e1177 https://doi.org/10.1161/CIR.0000000000000638 (2019).

Kim, B. K. et al. Long-term efficacy and safety of moderate-intensity statin with ezetimibe combination therapy versus high-intensity statin monotherapy in patients with atherosclerotic cardiovascular disease (RACING): a randomised, open-label, non-inferiority trial. Lancet 400, 380–390 https://doi.org/10.1016/S0140-6736(22)00916-3 (2022).

Jiang, S. Y. et al. Discovery of a potent HMG-CoA reductase degrader that eliminates statin-induced reductase accumulation and lowers cholesterol. Nat. Commun. 9, 5138 https://doi.org/10.1038/s41467-018-07590-3 (2018).

Article ADS PubMed PubMed Central CAS Google Scholar

Preiss, D. et al. Risk of incident diabetes with intensive-dose compared with moderate-dose statin therapy: a meta-analysis. JAMA 305, 2556–2564 https://doi.org/10.1001/jama.2011.860 (2011).

Armitage, J. The safety of statins in clinical practice. Lancet 370, 1781–1790 https://doi.org/10.1016/S0140-6736(07)60716-8 (2007).

She, J. et al. Statins aggravate insulin resistance through reduced blood glucagon-like peptide-1 levels in a microbiota-dependent manner. Cell Metab. 36, 408–421 https://doi.org/10.1016/j.cmet.2023.12.027 (2024).

Reagan-Shaw, S., Nihal, M. & Ahmad, N. Dose translation from animal to human studies revisited. FASEB J. 22, 659–661 https://doi.org/10.1096/fj.07-9574LSF (2008).

van der Vaart, J. I., van Eenige, R., Rensen, P. C. N. & Kooijman, S. Atherosclerosis: an overview of mouse models and a detailed methodology to quantify lesions in the aortic root. Vasc. Biol. 6, e230017 https://doi.org/10.1530/VB-23-0017 (2024).

Article PubMed PubMed Central Google Scholar

Hu, X. et al. A gut-derived hormone regulates cholesterol metabolism. Cell 187, 1685–1700 https://doi.org/10.1016/j.cell.2024.02.024 (2024).

Wood, D. E., Lu, J. & Langmead, B. Improved metagenomic analysis with Kraken 2. Genome Biol. 20, 257 https://doi.org/10.1186/s13059-019-1891-0 (2019).

Article PubMed PubMed Central CAS Google Scholar

Uritskiy, G. V., DiRuggiero, J. & Taylor, J. MetaWRAP-a flexible pipeline for genome-resolved metagenomic data analysis. Microbiome 6, 158 https://doi.org/10.1186/s40168-018-0541-1 (2018).

Article PubMed PubMed Central Google Scholar

Wang, X. et al. Candida albicans accelerates atherosclerosis by activating intestinal hypoxia-inducible factor 2α signaling. Cell Host Microbe 32, 964–979 https://doi.org/10.1016/j.chom.2024.04.017 (2024).

Mitsche, M. A., McDonald, J. G., Hobbs, H. H. & Cohen, J. C. Flux analysis of cholesterol biosynthesis in vivo reveals multiple tissue and cell-type specific pathways. eLife 4, e07999 https://doi.org/10.7554/eLife.07999 (2015).

Article PubMed PubMed Central Google Scholar

Zhao, S. et al. Dietary fructose feeds hepatic lipogenesis via microbiota-derived acetate. Nature 579, 586–591 https://doi.org/10.1038/s41586-020-2101-7 (2020).

Article ADS PubMed PubMed Central CAS Google Scholar

Hendrix, S. et al. Hepatic SREBP signaling requires SPRING to govern systemic lipid metabolism in mice and humans. Nat. Commun. 14, 5181 https://doi.org/10.1038/s41467-023-40943-1 (2023).

Article ADS PubMed PubMed Central CAS Google Scholar

Lu, X. Y. et al. Feeding induces cholesterol biosynthesis via the mTORC1–USP20–HMGCR axis. Nature 588, 479–484 https://doi.org/10.1038/s41586-020-2928-y (2020).

Kuntz, S. et al. Metabolic fate and distribution of 2′-fucosyllactose: direct influence on gut microbial activity but not on brain. Mol. Nutr. Food Res. 63, e1900035 https://doi.org/10.1002/mnfr.201900035 (2019).

Article PubMed PubMed Central CAS Google Scholar

Rudloff, S. et al. Fucose as a cleavage product of 2′ fucosyllactose does not cross the blood–brain barrier in mice. Mol. Nutr. Food Res. 65, e2100045 https://doi.org/10.1002/mnfr.202100045 (2021).

Yogev, Y. et al. Limb girdle muscular disease caused by HMGCR mutation and statin myopathy treatable with mevalonolactone. Proc. Natl Acad. Sci. USA 120, e2217831120 https://doi.org/10.1073/pnas.2217831120 (2023).

Article PubMed PubMed Central CAS Google Scholar

Morales-Rosado, J. A. et al. Bi-allelic variants in HMGCR cause an autosomal-recessive progressive limb-girdle muscular dystrophy. Am. J. Hum. Genet. 110, 989–997 https://doi.org/10.1016/j.ajhg.2023.04.006 (2023).

Article PubMed PubMed Central CAS Google Scholar

Trub, A. G. et al. Statin therapy inhibits fatty acid synthase via dynamic protein modifications. Nat. Commun. 13, 2542 https://doi.org/10.1038/s41467-022-30060-w (2022).

Article ADS PubMed PubMed Central CAS Google Scholar

Liu, Y. et al. Artemisinins ameliorate polycystic ovarian syndrome by mediating LONP1–CYP11A1 interaction. Science 384, eadk5382 https://doi.org/10.1126/science.adk5382 (2024).

Zheng, L. et al. CRISPR/Cas-based genome editing for human gut commensal Bacteroides species. ACS Synth. Biol. 11, 464–472 https://doi.org/10.1021/acssynbio.1c00543 (2022).

We are grateful to C. Wu for his guidance and help with animal experimental designs; S. Li for assistance with the purification of HMGCR protein; S. Liu and X. Li for discussions and support; staff at the Multi-Omics Mass Spectrometry Core of the Biomedical Research Core Facilities at Shenzhen Bay Laboratory for assistance with the MS and NMR experiments; staff at the Shenzhen Bay Laboratory Supercomputing Center for providing the platform for meta-omic data analyses; and staff at OE Biotech Co., Ltd. for technical support with scRNA-seq. Part of the computational analysis work was supported by the High-performance Computing Public Platform (Shenzhen Campus) of Sun Yat-sen University.

This work has been supported by the National Key Research and Development Program of China (grant no. 2020YFA0907800 to W.Z. and X.M.), the Shenzhen Medical Research Fund (B2502007 to W.Z. and X.T.), the Major Program of Shenzhen Bay Laboratory (C1012523006 to X.T. and W.Z.), the Shenzhen Science and Technology Programs (RCJC20231211085944057 to W.Z., and ZDSYS20220606100803007 to W.Z. and X.M.), the Shenzhen Science and Technology Programs (KQTD20200820145822023 to W.Z. and X.M.), Shenzhen Bay Laboratory Start-up Funds (21230051 to X.T.), the Shenzhen Bay Scholar Fellowship (to X.T.), and the Guangdong Basic and Applied Basic Research Fund (2514050002740 to W.Z.).

These authors contributed equally: Chao Yin, Youzhe Chen, Gan Lin, Mingwei Cai

Shenzhen Key Laboratory for Systems Medicine in Inflammatory Diseases, Zhongshan School of Medicine, Shenzhen Campus of Sun Yat-sen University, Shenzhen, China

Chao Yin, Gan Lin, Xianzun Xiao, Kaining Han, Chaoxiong Mei, Peizhi Fan, Yibo Zhao, Lihong Du, Yanqin Xie, Yudan Mao, Xiangting Zhou, Xue Gao, Li Jin, Peijie Li, Xiangyu Mou & Wenjing Zhao

Institute of Chemical Biology, Shenzhen Bay Laboratory, Shenzhen, China

Youzhe Chen, Mingwei Cai, Miaomiao Qin, Jiang Wang, Ruolan Sun & Xiaoyu Tang

CAS Key Laboratory of Quantitative Engineering Biology, Shenzhen Institute of Synthetic Biology, Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China

Department of Clinical Laboratory, The Seventh Affiliated Hospital of Sun Yat-sen University, Shenzhen, China

Shenzhen Medical Academy of Research and Translation (SMART), Shenzhen, China

Department of Geriatrics, Shenzhen Key Laboratory of Bone Tissue Repair and Translational Research, The Seventh Affiliated Hospital, Sun Yat-sen University, Shenzhen, China

Institute of Molecular Physiology, Shenzhen Bay Laboratory, Shenzhen, China

X.M., X.T. and W.Z. conceived, designed and supervised the project. C.Y., G.L., X.X., K.H., C.M., M.Q., P.F., L. Du, Y.X., Y.M., X.Z., X.G., L.J., P.L. and R.S. performed the mouse experiments. C.Y., Y.C. and G.L. participated in the bioassay tests and compound isolation and identification. M.C. performed the bioinformatic analyses. C.Y., Y.C., G.L. and J.W. participated in culturing bacteria and LC–MS analyses. C.Y., Y.C., G.L., M.C., C.P. and C.M. performed molecular, biochemical and cellular experiments. Y. Zhang performed the computational simulation. C.Y., G.L., X.X. and Z.Q. participated in the gene editing of B. uniformis under the supervision of L. Dai, X.M., X.T. and W.Z., and C.Y., G.L., P.F., Y. Zhao, L. Du, Y.X. and Z.T. were responsible for patient recruitment and the collection of clinical samples, performed under the supervision of X.L., Z.L., X.M. and W.Z. X.M., X.T. and W.Z. provided primary financial support for the study. The manuscript was prepared by C.Y., Y.C., G.L., M.C., X.M., X.T. and W.Z. with contributions from all authors. All authors reviewed and provided their approval for the final manuscript.

Correspondence to Xiangyu Mou, Xiaoyu Tang or Wenjing Zhao.

X.M. is an inventor on a pending patent application (CN202510500448.8) covering the use of PA for the prevention or treatment of atherosclerosis. The remaining authors declare no competing interests.

Nature thanks Arash Haghikia, Robert Quinn, Federico Rey, Soraya Taleb and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

a. Genus-level aggregation of the top ten healthy-enriched bacterial species from Fig. 1a. (n = 223 for case; n = 189 for control). Data are presented as box plots of relative abundance (%). Boxes indicate the 25th and 75th percentiles, central lines indicate the median, whiskers extend to the most extreme data points within 1.5 × the interquartile range calculated on the log10-transformed relative-abundance scale, and points beyond the whiskers represent outliers. A pseudocount of 0.00001% was added to zero relative-abundance values before plotting on a log scale. Statistical significance was determined by two-sided Wilcoxon rank-sum tests with Benjamini–Hochberg correction for multiple comparisons across the 7 aggregated genera: **P < 0.01; ****P < 0.0001. The adjusted P values (left to right): <0.0001, <0.0001, 0.0011, 0.0013, <0.0001, <0.0001, <0.0001. b, g and h. Representative images of Oil Red O-stained plaques in whole aorta with quantifications (n = 5 for b; n = 7 for h). For b, P values (left to right): 0.0281, 0.0003. For h, P values (left to right): 0.0065. c, d and i. Representative images of Oil Red O-stained plaques in the sections of aortic roots with quantifications (n = 5 for d; n = 7 for i). For d, P values (left to right): 0.0056. For i, P values (left to right): 0.0013. Scale bar: 200 μm. e. Absolute quantification of B. uniformis abundance in mouse faeces collected at multiple time points after a single oral gavage, determined by TaqMan qPCR based on the copy numbers of the 16S rRNA-encoding gene (n = 6). f. Absolute quantification of B. uniformis abundance in mouse faeces collected at multiple time points during PBS or BU treatment, determined by TaqMan qPCR based on the copy numbers of the 16S rRNA-encoding gene (n = 6). P = 0.0011. j and k. OGTT and ITT indexes (n = 7). l. Body weight changes (n = 7). m. Ratios of liver mass to body mass (n = 7). P < 0.0001. n and p. Representative flow cytometric plots (n) and quantification (p, n = 8) of blood CD11B+ Ly6C+ monocytes (total, and separately as Ly6Chi and Ly6Clo). P values (left to right): 0.0189, <0.0001, 0.0376. o and q. Representative flow cytometric plots (o) and quantification (q, n = 8) of blood CD11B+ F4/80+ macrophages. P = 0.0067. r. Relative quantification of Bacteroides in faeces (n = 7). P = 0.0017. b-d and f-r. HFD-fed SPF Apoe−/− mice were gavaged with PBS and BU (Low dose: 108 CFU/mouse; High dose: 109 CFU/mouse) three times per week for BU − 1 month or BU − 3 months (12 weeks). Data are presented as mean ± SEM (b, d-f, h-m and r), or the median with the first and third quartiles in box and whiskers (p and q). Statistical significance was determined using one-way ANOVA with Dunnett’s test (b, d, h and i), two-tailed t-tests with Welch’s correction (f), or two-tailed Student’s t test (j-m and p-r): *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Related to Fig. 2. a. Gene ontology (GO) biological process analysis in HFD-fed SPF Apoe−/− mice treated with PBS (n = 4) and BU (n = 4) three times per week for 12 weeks. b. Diagram depicting the negative feedback regulation of cholesterol biosynthesis and its impact on hepatic Ldlr expression and plasma LDL-C levels. c. Schematic illustration of the SaCas9-sgRNA strategy and experimental design. HFD-fed SPF Apoe−/− mice were administered a single intravenous injection of adeno-associated virus (AAV8-sgNTC or AAV8-sgLdlr). After 7 days, the mice were gavaged with PBS or BU for 12 weeks. d and e. Protein levels (d) and quantification analysis (e) of Ldlr in liver tissues, with n = 3 per group. P values (up to down, left to right): 0.0354, 0.0019, 0.0002. f-i. Analysis of plasma levels of triglycerides (f, TG), total cholesterol (g, TC), high-density lipoprotein cholesterol (h, HDL-C), and low-density lipoprotein cholesterol (i, LDL-C) was performed (n = 7 mice). P values (left to right): f, 0.0002, 0.0025; g, 0.0011, 0.0020; i, 0.0002, 0.0005. j. Left: Representative images of CD68 immunofluorescence staining used to detect macrophages in the sections of aortic roots. Right: Quantification of CD68-positive area (n = 7 mice). P values (left to right): 0.0015, <0.0001. Scale bar: 200 μm. k and l. Representative flow cytometric plots (k) and quantification (l) of blood CD11B+ Ly6C+ monocytes (total, and separately as Ly6Chi and Ly6Clo) (n = 7 mice). P values (left to right): Ly6Chi, <0.0001, <0.0001. m and n. Representative flow cytometric plots (m) and quantification (n) of blood CD11B+ F4/80+ macrophages (n = 7 mice). P values (left to right): 0.0004, 0.0072. Data are presented as mean ± SEM (e-j), or the median with the first and third quartiles in box and whiskers (l and n). Statistical significance was determined using one-way ANOVA with Tukey’s post hoc test (e-j, l and n): *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Schematics in b and c created in BioRender; Mou, X. https://biorender.com/1x3qxol (2026).

a. Schematic illustration of the B. uniformis crude extract screening platform. b and c. Representative images (b) of DiI–LDL uptake in HepG2 cells incubated with the extracts (1 mg/ml) prepared from fresh medium and BU culture supernatant using three solvents of varying polarity (dichloromethane, ethyl acetate, or n-butanol), with the quantification (c, n = 3 independent experiments) of the mean intensity of DiI–LDL. P values (left to right): 0.0005, 0.0268, 0.0073, 0.0334, 0.0311, 0.0070. Yellow, DiI–LDL; blue, nuclei stained with Hoechst. Scale bar, 100 μm. d. Relative mRNA (d) of genes involved in LDL-C uptake in HepG2 cells incubated with the indicated extracts (1 mg/ml) from fresh medium and BU culture supernatant (n = 3 independent experiments). P values (left to right): SREBF2, 0.0011, 0.0003, 0.0004, 0.0002, 0.0359; LDLR, 0.0014, 0.0011, 0.0446, 0.0086, 0.0111, 0.0117. e and f. Protein levels (e) and quantification analysis (f) of genes involved in LDL-C uptake in HepG2 cells incubated with the indicated extracts (1 mg/ml) from fresh medium and BU culture supernatant (n = 3 independent experiments). P values (left to right): LDLR, 0.0455, 0.0173, 0.0005, 0.0039; pre-SREBP2, <0.0001, 0.0060, 0.0045, 0.0482; nSREBP2, 0.0043, 0.0128, 0.0418. Data are presented as mean ± SEM (c, d and f). Statistical significance was determined using two-tailed t-tests with Welch’s correction (c, d and f): *P < 0.05; **P < 0.01; ***P < 0.001, ****P < 0.0001. Schematic in a created in BioRender; Mou, X. https://biorender.com/1x3qxol (2026).

a. Free total cholesterol in HepG2 cells treated with Control 3 or Extract 3 (1 mg/ml; n = 3 independent experiments). P = 0.0052. b. Representative images of Oil Red O-stained plaques in whole aorta with quantifications (n = 8 for PBS and BU; n = 7 for HKBU and Extract 3; BU = Live B. uniformis; HKBU = Heat-killed B. uniformis; Extract 3 = ethyl acetate extract of B. uniformis culture). P values (left to right): 0.0039, 0.0207. c. Representative images of Oil Red O-stained plaques in the sections of aortic roots with quantifications (n = 6). P values (left to right): 0.0269, 0.0212. Scale bar: 200 μm. d-g. Analysis of plasma levels of triglycerides (d, TG), total cholesterol (e, TC), high-density lipoprotein cholesterol (f, HDL-C), and low-density lipoprotein cholesterol (g, LDL-C) was performed (n = 7). P values (left to right): d, 0.0007, 0.0486; e, 0.0406, 0.0025; g, 0.0002, 0.0169. h. Left: Representative images of CD68 immunofluorescence staining used to detect macrophages in the sections of aortic roots. Right: Quantification of CD68-positive area (n = 7). P values (left to right): <0.0001, <0.0001. Scale bar: 200 μm. i and j. Representative flow cytometric plots (i) and quantification (j, n = 7) of blood CD11B+ Ly6C+ monocytes (total, and separately as Ly6Chi and Ly6Clo). P values (left to right): total Ly6C+, 0.0330, 0.0218; Ly6Chi, 0.0016, 0.0062. k and l. Representative flow cytometric plots (k) and quantification (l, n = 7) of blood CD11B+ F4/80+ macrophages. P values (left to right): 0.0123, 0.0086. m. Relative mRNA levels of genes involved in hepatic cholesterol metabolism (n = 6). P values (left to right): Ldlr, 0.0417, <0.0001; Srebf2, 0.0267, 0.0004; Hmgcr, 0.0044, <0.0001; Mvk, 0.0006; Pcsk9, 0.0360, <0.0001. n and o. Protein levels (n, n = 3 per group, representative data of three independently repeated experiments) and quantification analysis (o) of genes involved in hepatic cholesterol metabolism. P values (left to right): LDLR, 0.0292, 0.0003; pre-SREBP2, 0.0012, 0.0007; nSREBP2, 0.0068, 0.0002. b-o. HFD-fed SPF Apoe−/− mice were gavaged with PBS, live BU, heat-killed BU (HKBU) and Extract 3 three times per week for 12 weeks. Data are presented as mean ± SEM (a-h, m and o), or the median with the first and third quartiles in box and whiskers (j and l). Statistical significance was determined using two-tailed Student’s t test (a), one-way ANOVA with Dunnett’s test (b, f-h, j, l, m and o), or Kruskal-Wallis test followed by Dunn’s post hoc test (c-e): *P < 0.05; **P < 0.01; ***P < 0.001, ****P < 0.0001.

Related to Fig. 3. a and b. Representative images (a) of DiI–LDL uptake in HepG2 cells incubated with the indicated BU fractions (100 μg/ml), with the quantification (b, n = 3 independent experiments) of the mean intensity of DiI–LDL. P values (left to right): <0.0001, 0.0004, <0.0001, <0.0001, 0.0006, <0.0001. Yellow, DiI–LDL; blue, nuclei stained with Hoechst. Scale bar, 100 μm. c. Relative mRNA levels of genes involved in LDL-C uptake in HepG2 cells incubated with the indicated BU fractions (100 μg/ml; n = 3 independent experiments). P values (left to right): SREBF2, 0.0062, 0.0018, 0.0091; LDLR, 0.0134, 0.0118, 0.0013, 0.0280. d, e and f. Protein levels (d) and quantification analysis (e and f) of genes involved in LDL-C uptake in HepG2 cells incubated with the indicated BU fractions (100 μg/ml; n = 3 independent experiments). For e, P values for DMSO versus F1, F2 and F5-F8 were <0.0001, 0.0409, <0.0001, <0.0001, <0.0001 and <0.0001, respectively. For f, P values for pre-SREBP2 (DMSO versus F1-F3 and F5-F9) were <0.0001, 0.0042, 0.0176, 0.0005, 0.0060, 0.0006, <0.0001 and 0.0065, respectively; P values for nSREBP2 (DMSO versus F1-F8) were <0.0001, <0.0001, 0.0002, 0.0468, 0.0008, 0.0010, 0.0068 and <0.0001, respectively. Data are presented as mean ± SEM (b, c, e and f). Statistical significance was determined using one-way ANOVA with Dunnett’s test (b, e and f), or Kruskal-Wallis test followed by Dunn’s post hoc test (c): *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

a. Relative mRNA expression levels of the key genes involved in LDL-C uptake (80 μM; n = 3 independent experiments). P values (left to right): SREBF2, <0.0001, 0.0035, <0.0001, <0.0001; LDLR, <0.0001, <0.0001, <0.0001, 0.0009, <0.0001. b-d. Protein levels (b and c) and quantification analysis (d) of genes involved in LDL-C uptake in HepG2 cells incubated with the indicated compounds (n = 3 independent experiments). For d, P values for LDLR (DMSO versus F1-1, Cpd 2, Cpd 3 and Cpd 6-Cpd 8) were 0.0469, 0.0393, 0.0018, <0.0001, <0.0001 and 0.0007, respectively. P values for pre-SREBP2 (DMSO versus Cpd 2, Cpd 3, Cpd 7, Cpd 9 and Cpd 10) were <0.0001, <0.0001, 0.0497, 0.0036 and 0.0007, respectively. P values for nSREBP2 (DMSO versus Cpd 2, Cpd 3, Cpd 6 and Cpd 7) were 0.0197, 0.0035, 0.0286 and 0.0009, respectively. e. HRMS of Cpd 3 in negative mode produces a reporter ion at m/z 241.2188. f. HPLC-MS assays of BHIS, BU culture, and PA standards: TIC chromatograms are under negative ion mode. Data are presented as mean ± SEM (a and d). Statistical significance was determined using one-way ANOVA with Dunnett’s test (a and d): *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

a. Experimental design. HFD-fed SPF Apoe−/− mice were gavaged with corn oil, PA (dissolved in corn oil) and ATO for 8 weeks. b-e. Analysis of plasma levels of triglycerides (b, TG), total cholesterol (c, TC), high-density lipoprotein cholesterol (d, HDL-C), and low-density lipoprotein cholesterol (e, LDL-C) was performed (n = 7). P values (left to right): c, 0.0035, 0.0294; d, 0.0266; e, 0.0002, 0.0004. f. Left: Representative images of CD68 immunofluorescence staining used to detect macrophages in the sections of aortic roots. Right: Quantification of CD68-positive area (n = 8). P values (left to right): 0.0002, <0.0001. Scale bar: 200 μm. g and i. Representative flow cytometric plots (g) and quantification (i, n = 8 mice for corn oil and ATO; n = 7 mice for PA) of blood CD11B+ Ly6C+ monocytes (total, and separately as Ly6Chi and Ly6Clo). P values for i (left to right): total, 0.0077, 0.0010; Ly6Chi, 0.0383, 0.0133; Ly6Clo, 0.0225, 0.0029. h and j. Representative flow cytometric plots (h) and quantification (j, n = 8 mice for corn oil and ATO; n = 7 mice for PA) of blood CD11B+ F4/80+ macrophages. P values (left to right): 0.0145, 0.0162. k. Relative mRNA levels of genes involved in hepatic cholesterol metabolism (n = 6). P values (left to right): 0.0004, 0.0019, <0.0001, 0.0002, 0.0004. l and m. Protein levels (l, n = 3 per group, representative data of three experiments) and quantification analysis (m) of genes involved in hepatic cholesterol metabolism. P values (left to right): 0.0499, 0.0003, 0.0241. b-m. HFD-fed SPF Apoe−/− mice were gavaged with Corn oil, PA (25 mg/kg) or ATO (25 mg/kg) three times per week for 8 weeks. Data are presented as mean ± SEM (b-f, k and m), or the median with the first and third quartiles in box and whiskers (i and j). Statistical analysis was performed using one-way ANOVA with Dunnett’s test (b-f and j), two-sided Welch ANOVA with Dunnett’s T3 test (i), or two-tailed t-tests with Welch’s correction (k and m): *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Schematic in a created in BioRender; Mou, X. https://biorender.com/1x3qxol (2026).

HFD-fed SPF Apoe−/− mice were administered a single intravenous injection of adeno-associated virus (AAV8-sgNTC or AAV8-sgLdlr). After 7 days, the mice were gavaged with corn oil or PA for 12 weeks. a and b. Protein levels (a) and quantification analysis (b) of Ldlr in liver tissues, with n = 3 per group. P values (left to right): <0.0001, <0.0001. c. Representative images of Oil Red O-stained plaques in whole aorta with quantifications (n = 7). P values (left to right): 0.0005, <0.0001. d. Representative images of Oil Red O-stained plaques in the sections of aortic roots with quantifications (n = 7). P values (left to right): 0.0374, 0.0017. Scale bar: 200 μm. e-h. Analysis of plasma levels of triglycerides (e, TG), total cholesterol (f, TC), high-density lipoprotein cholesterol (g, HDL-C), and low-density lipoprotein cholesterol (h, LDL-C) was performed (n = 7). P values (left to right): e, 0.0304, 0.0006; f, 0.0004, 0.0006; h, <0.0001, <0.0001. i. Left: Representative images of CD68 immunofluorescence staining used to detect macrophages in the sections of aortic roots. Right: Quantification of CD68-positive area (n = 7). P values (up to down, left to right): 0.0095, 0.0474, <0.0001. Scale bar: 200 μm. j and l. Representative flow cytometric plots (j) and quantification (l, n = 7) of blood CD11B+ Ly6C+ monocytes (total, and separately as Ly6Chi and Ly6Clo). P values (left to right): total, 0.0187, 0.0011; Ly6Chi, <0.0001, 0.0150; Ly6Clo, 0.0099. k and m. Representative flow cytometric plots (k) and quantification (m, n = 7) of blood CD11B+ F4/80+ macrophages. P values (left to right): 0.0001, <0.0001. n. PA binding to HMGCR was illustrated by SPR assay of purified HMGCR (catalytic portion) with PA. a-m. HFD-fed SPF Apoe−/− mice, with adeno-associated virus injection (AAV8-sgNTC or AAV8-sgLdlr), were gavaged with Corn oil or PA (25 mg/kg) three times per week for 12 weeks. Data are presented as mean ± SEM (b-i), the median with the first and third quartiles in box and whiskers (l and m). Statistical significance was determined using one-way ANOVA with Tukey’s post hoc test (b, c, e-i and m), Kruskal-Wallis test followed by Dunn’s post hoc test (d), or two-sided Welch ANOVA with Dunnett’s T3 test (l): *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Related to Fig. 5. a and b. Isotopologue distributions of PA from [1-13C] acetate (a) and [1-13C] propionate (b). c. Schematic representation of the biosynthetic pathway of PA in Bacteroidota. d. Putative gene operon containing ack/pta in C. glutamicum and mutA/scpA and pta/ack in B. thetaiotaomicron were highly conserved in BU DSM6597. e. Phylogenetic tree of 100 gut microbial isolates with corresponding PA production levels. f and g. Prevalence and presence of PA synthesis genes in the human gut microbiome. Prevalence and presence of ack/pta and mutA/scpA/mce/mcd genes (f) phylum level and (g) genus level. Given that the mutA and mce genes are most abundant in Bacteroidota, only the genus-level data for phylum Bacteroidota are displayed in (g). The presence of genes for PA biosynthesis is detailed in Supplementary Table 7.

a. PCR analysis of colonies counter-selected for the deletion of pta or mutA genes in B. uniformis, showing data from individual PCR reactions. b. SEM visualization of WT, ΔmutA, and Δpta. SEM images captured at a magnification of 30,000×. Scale bar: 300 nm. The experiments in a and b were independently repeated three times with similar results. c. Quantification of growth for WT and mutant strains in BHIS medium under anaerobic conditions (n = 3 independent experiments, each from a separate colony). OD600 were determined at the indicated time points. d-g. Analysis of plasma levels of triglycerides (d, TG), total cholesterol (e, TC), high-density lipoprotein cholesterol (f, HDL-C), and low-density lipoprotein cholesterol (g, LDL-C) was performed (n = 7). P values (left to right): e, 0.0011, 0.0172, 0.0036; g, <0.0001, 0.0005, 0.0021. h. Left: Representative images of CD68 immunofluorescence staining used to detect macrophages in the sections of aortic roots. Right: Quantification of CD68-positive area (n = 7). P values (left to right): 0.0029, 0.0344. Scale bar: 200 μm. i and j. Representative flow cytometric plots (i) and quantification (j, n = 7) of blood CD11B+ Ly6C+ monocytes (total, and separately as Ly6Chi and Ly6Clo). P values (up to down, left to right): total Ly6C+, 0.0307, 0.0233; Ly6Chi, 0.0123, 0.0319, 0.0207, <0.0001, 0.0078. k and l. Representative flow cytometric plots (k) and quantification (l, n = 7) of blood CD11B+ F4/80+ macrophages. P values (left to right): 0.0006, 0.0282, 0.0420. m. Relative mRNA levels of genes involved in hepatic cholesterol metabolism (n = 6). P values (up to down, left to right): Ldlr, 0.0020, 0.0019, 0.0030; Srebf2, 0.0242, 0.0025, 0.0028; Hmgcr, 0.0140, 0.0005, 0.0312; Mvk, 0.0297, 0.0129, <0.0001, 0.0068; Pcsk9, 0.0081, 0.0004, 0.0081. n and o. Protein levels (n, n = 3 per group, representative data of three experiments) and quantification analysis (o) of genes involved in hepatic cholesterol metabolism. P values (left to right): LDLR, <0.0001, <0.0001, <0.0001; pre-SREBP2, 0.0008, 0.0263, 0.0064; nSREBP2, 0.0003, 0.0003, 0.0007. d-o. HFD-fed SPF Apoe−/− mice were gavaged with PBS, WT, Δpta and ΔmutA three times per week for 12 weeks. Data are presented as mean ± SEM (c-h, m and o), or the median with the first and third quartiles in box and whiskers (j and l). Statistical significance was determined using one-way ANOVA with Dunnett’s test (c), one-way ANOVA with Tukey’s post hoc test (d-h, j, m and o), or Kruskal-Wallis test followed by Dunn’s post hoc test (l): *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

a. Correlation of serum PA with parameters associated with CVD. Q represents the quartile of PA proportions (replotted from published data) b. Relative abundance of bacterial genes involved in PA synthesis among case and control individuals (n = 223 for case; n = 189 for control). RPKM values for each gene were calculated using CoverM (v0.6.1). Data are presented as violin plots with embedded box plots; violin width indicates the distribution density, boxes indicate the 25th and 75th percentiles, central lines indicate the median, whiskers indicate the minimum and maximum values within 1.5 × the interquartile range, and points beyond the whiskers represent outliers. Statistical significance was determined using a two-sided Wilcoxon rank-sum test with Benjamini–Hochberg correction for multiple comparisons. c. Summary diagram depicting the role of PA in the modulation of hepatic cholesterol metabolism and atherosclerosis progression. d. The levels of C17:0 in faecal samples (n = 90 per group). P = 0.0004. e. HPLC-MS analysis of BHIS, BU culture, and C17:0 standards. Extracted ion chromatograms (EICs) were obtained at m/z 269.2461 ± 0.01 [M − H]− for C17:0. f. HRMS of C17:0 BU fermentation supplemented with 13C-labeled acetate (Left) and propionate (Right). g. C17:0 levels in faeces, liver tissues and serum. HFD-fed SPF Apoe−/− mice after 12-week gavage with PBS or BU (n = 7). h. FTC in HepG2 cells treated with DMSO and long chain fatty acids (C15:0, C16:0 or C17:0; 80 μM; n = 3 independent experiments). P values (left to right): 0.0127, <0.0001, 0.0004. i. Inhibition curve of C16:0, C17:0 and ATO against HMGCR (representative data of two independent experiments). Data are presented as mean ± SEM (d, g and h). Statistical significance was determined using two-tailed Mann-Whitney U test (d), two-tailed t-tests with Welch’s correction (g), or two-sided Welch ANOVA with Dunnett’s T3 test (h): *P < 0.05; ***P < 0.001; ****P < 0.0001. Schematic in c created in BioRender; Mou, X. https://biorender.com/s4a1t8m (2026).

Supplementary Methods and Supplementary Figs. 1–7 providing additional experimental procedures and supporting data for the study.

Bacterial species enriched in healthy controls compared with individuals with ASCVD.

Genus-level aggregation of the top ten healthy-enriched bacterial species.

Comparative transcriptomic analysis of lipid-metabolism-related genes in mouse liver.

Protein sequence of human HMGCR (residues 426–888).

Human gut bacterial strains used for the analysis of PA production.

Homologous genes involved in PA synthesis in the human gut microbiome.

Prevalence and presence of PA synthesis genes in the human gut microbiome at the phylum level.

Prevalence and presence of PA synthesis genes in the human gut microbiome at the genus level.

sgRNA and primer sequences used in this study.

Source data for Supplementary Figs. 1, 3 and 5.

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Yin, C., Chen, Y., Lin, G. et al. A gut microbial odd-chain fatty acid alleviates atherosclerosis in mice. Nature (2026). https://doi.org/10.1038/s41586-026-11142-x

DOI: https://doi.org/10.1038/s41586-026-11142-x

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Close Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.