Fig. 6

Ihubacter massiliensis was identified as a key bacterium for TMA/TMAO production in the humanized gnotobiotic mice (hGM) model. a A diagram illustrates the experiments of TMAO productivity-based hGM model. Four hGM groups were conducted through fecal microbiota transplantation (FMT) from two high-TMAO producers and two low-TMAO producers defined by oral carnitine challenge test (OCCT). b The d9-carnitine challenge test was performed in all groups of hGM through gavage. Only mice in the O15 hGM group demonstrated remarkable TMA/TMAO-producing ability, whereas others demonstrated undetectable d9-TMA and d9-TMAO in the plasma. The three TMA/TMAO-nonproducer hGM groups also showed increasing d9-γ-butyrobetaine (d9-γBB) level, whereas the d9-γBB in O15 hGM group was increased in the beginning and dropped latter. Time reflects hours after oral gavage of d9-carnitine. Bars represent the mean ± S.E.M for the indicated number of mice. c A Venn diagram showed 3 OTUs (Ihubacter massiliensis OTU#: LT576391.1.1479, Bifidobacterium bifidum OTU#: S83624.1.1532, and unclassified Ruminococcaceae OTU#: HQ769937.1.1427) were identified from O15 hGM group after excluding the intersections with other hGM groups. d I. massiliensis was significantly enriched in high-TMAO producers than low-TMAO producers defined by OCCT. Data were analyzed by paired Wilcoxon test; bars represent the mean ± S.E.M for the indicated groups. e The relative abundance of E. timonensis plus I. massiliensis was significantly associated with TMAO productivity (p < 0.0001, r = 0.41). f Among the 51 high-TMAO producers, either E. timonensis or I. massiliensis were detected in 22 subjects (43%) while I. massiliensis was detected in only 2 among 61 low-TMAO producers (3.3%). Using the presence/absence of these two bacteria in feces contributes to 43% sensitivity and 97% specificity in identifying high-TMAO producer