560 in T2D onsequence after TCMs whic dulat thr i Baica whi onstrated n et al. 002 in diabetic rats (Waisundara eta enriched by 20O2l.howmldiosedependontynegiceectswi staphylococcus aureus and other B-lactam-tes et al. tion with nla lower abundance of F et a bloods t on gut microbiota d yo a functionally mportant phylo in ntered 2008).Additionally beer oiot-du correlated with HOMA-B 2012h mplying that tsin GOD tha d th The bohydrate compo microbiota Wa: very ifidobacteria dosage differences across the four deco ions.Othe starch easily ddit n.se These results indicate that mod of gut of "pea0d ，migh be i homeos asis bv r 。。 (Saulnier et al..2011).One ex mple from this genu microbiota-targeted diabetes 883ctgeteaeP8aicieRanwtioeal 2000 aula GQD grid BALB/c mic bacte ngss that dtsen et al., 2012.hes res um spp.in the gut.A gh t is stil of 2500 in China Conflict of Interest ties and unknown mechan The authors declare no conflict of interest. nents from being identified.Our study suggests that Acknowledgements avenue for identifying chemical components in The ISME Journal in T2D patients rather than a mere consequence after the symptoms have been alleviated. Several putative beneficial genera that responded to GQD treatment were identified using redundancy analysis, including Faecalibacterium, Bifidobacterium and Gemmiger. Faecalibacterium, a butyrate￾producing bacterial group (Duncan et al., 2002), which was profoundly enriched by GQD in our study, shows anti-inflammatory effects partly through reducing colonic cytokine synthesis and increasing anti-inflammatory cytokine secretion (Sokol et al., 2008). Diabetic obese patients have lower abundance of F. prausnitzii compared with non-diabetic obese patients, which demonstrates a negative correlation with inflammatory cytokines C-reactive protein and interleukin-6 (Furet et al., 2010). Another study also showed that the gut of T2D patients is characterized by a reduction of F. prausnitzii compared with that of healthy people (Qin et al., 2012). Furthermore, F. prausnitzii has been shown to be a functionally important phylo￾type because it was associated with eight urinary metabolites (Li et al., 2008). In our study, the relative abundance of F. prausnitzii was negatively corre￾lated with HbA1c, FBG and 2h-PBG, and positively correlated with HOMA-b, suggesting that F. praus￾nitzii might be a pivotal phylotype associated with the improvement of T2D. Two other genera, Bifidobacterium and Gemmiger, which are also reported to confer beneficial effects (Gossling and Moore, 1975; Sokol et al., 2008; Fukuda et al., 2011), were significantly enriched by GQD in T2D patients. A selective increase of Bifidobacteria induced by a prebiotic (namely oligofructose) improved gut permeability and inflammation in ob/ob mice (Cani et al., 2009). In addition, several genera were significantly inhibited by GQD, such as Alistipes and Odoribacter. Higher abundance of several taxa of the genus Alistipes were associated with greater frequency of abdominal pain in irritable bowel syndrome pediatric patients (Saulnier et al., 2011). One example from this genus, Alistipes putredinis, was isolated from inflamed and non-inflamed intestinal tissues of children with suspected acute appendicitis (Rautio et al., 2000, 2003). Alistipes and Odoribacter were significantly increased in grid floor stress-induced BALB/c mice (Bangsgaard Bendtsen et al., 2012). These results suggest that the enrichment of beneficial bacteria, particularly Faecalibacterium spp., and reduction of pathogen-like bacteria might be involved in the amelioration of T2D by GQD. TCM formula, a form of polypharmacy, has been developed and advocated for use in the treatment of many diseases for over 2500 years in China. However, the complexities and unknown mechan￾isms of TCMs prevent the active chemical compo￾nents from being identified. Our study suggests that gut microbiota might be involved in the effect of a widely used TCM formula, GQD. This opens an avenue for identifying chemical components in TCMs, which can modulate gut microbiota structure as a potential mechanism for disease alleviation. Baicalin, puerarin and berberine were identified as the three most abundant chemical components in GQD. Baicalin, which demonstrated antioxidant properties together with the glucose-lowering effects in STZ-induced diabetic rats (Waisundara et al., 2009), showed dose-dependent synergic effects with b-lactam antibiotics against methicillin-resistant Staphylococcus aureus and other b-lactam-resistant strains of S. aureus in vitro (Liu et al., 2000). Puerarin can dose-dependently increase glucose utilization in STZ-induced diabetic rats (Hsu et al., 2003). However, it showed poor absorption into the bloodstream after oral administration in vivo (Luo et al., 2011) and was mainly excreted via feces in its intact form (Zhu et al., 1979). It is possible that puerarin could have an impact on gut microbiota, but no study about this has been reported yet. Berberine showed significant glucose-lowering effects in a multicentered, randomized, double￾blinded and placebo-controlled clinical trial (Zhang et al., 2008). Additionally, berberine has been reported to modulate gut microbiota during prevention of high-fat-diet-induced obesity and insulin resistance in rats (Zhang et al., 2012b), implying that berberine could be one of the major active ingredients in GQD that modulated the gut microbiota in our study. The carbohydrate compo￾nents of GQD likely exerted few effects on gut microbiota because insoluble dietary fiber was undetectable and soluble dietary fiber was very low. Additionally, these components showed no dosage differences across the four decoctions. Other carbohydrates, such as starch, were easily digested and absorbed before they reached large intestine. These results indicate that modulation of gut microbiota by chemical components, such as ber￾berine, might be involved in improving glucose homeostasis by GQD, suggesting that TCMs may serve as a new source for drug leads in gut microbiota-targeted diabetes management. In conclusion, our study suggests that structural alterations of gut microbiota, induced by Chinese herbal formula GQD, are associated with the anti￾diabetic effects of GQD. In particular, this treatment enriched the number of beneficial bacteria, such as Faecalibacterium spp. in the gut. Although it is still unclear whether changes of gut microbiota by GQD directly contribute to the improvement of glucose homeostasis, our clinical study provides circum￾stantial evidence that gut microbiota might be involved. Conflict of Interest The authors declare no conflict of interest. Acknowledgements This work was supported by Project 2010CB530601 of the National Basic Research Program of China (973 Program), Microbiota shift in alleviation of type 2 diabetes J Xu et al 560 The ISME Journal