International Journal of Obesityhttps://doi.org/10.1038/s41366-018-0187-x

ARTICLE

Animal Models

Konjaku flour reduces obesity in mice by modulating thecomposition of the gut microbiota

Yongbo Kang1,2,3● Yu Li1,2 ● Yuhui Du1,2

● Liqiong Guo1,2● Minghui Chen1,2

● Xinwei Huang1,2● Fang Yang4

●

Jingan Hong4● Xiangyang Kong1,2

Received: 13 April 2018 / Revised: 23 May 2018 / Accepted: 2 July 2018© Springer Nature Limited 2018

AbstractBackground Changes in the intestinal flora composition is referred to as dysbiosis, which is related to obesity devel-opment, thus supporting the potential roles of nutrients acting on intestinal flora to exert salutary effects on energeticmetabolism of host. Dietary fiber has been known to affect the composition of intestinal flora. The aim of the present studywas to investigate the functional effects of konjac flour (KF) on obesity control in respect to improving inflammation,metabolism, and intestinal barrier function, and the possible association of the effects with intestinal flora compositionchanges.Methods Mice (n= 30) were randomly divided into control group (n= 10), high-fat-diet (HFD) group (n= 10), and KFintervention group (n= 10), followed by feeding for 12 weeks and with adding a KF daily supplementation for the treatmentgroup. Body weight, fat accumulation, inflammation, and energetic metabolism markers in multiple tissues and the gutmicrobiota of the mice were examined at the end of the experiment.Results The KF supplementation significantly reduced the gains in weight, fat mass, as well as adipocyte size of HFDmice and lowered the serum TC, leptin (LEP), thiobarbituric acid-reacting substance (TBARS), IL-6, and lipopoly-saccharide (LPS) levels in HFD mice. KF also upregulated the expression of intestinal mucosa protein gene Intection andtight junction ZO-1 in HFD mice, as well as upregulate the expression of energy metabolism genes PPARα and CPT-1 aswell as the fat metabolism gene HLS in livers and fat tissues, and downregulate that of fat synthesis gene PPARγ (p <0.05). The KF treatment increases the α-diversity and change the β-diversity of the intestinal microflora in HFD mice andboosted the abundances of some obesity-related beneficial microorganisms (such as Megasphaera elsdenii) in theintestinal microflora of HFD mice, while reduced those of harmful microorganisms (such as Alistipes, Alloprevotella,Bacteroides acidifaciens, and Parabacteroides goldsteinii). The abundance of Alistipes was positively correlated withweight, fat mass, serum TC, TG, LEP, IL-6, and LPS contents as well as PPARγ gene expression; while notably andnegatively related to the expression of CPT-1 and HLS genes (p < 0.01). KF remarkably increased the abundance ofAerococcaceae, while reduced that of Alistipes finegoldii (p < 0.01).Conclusions Supplementation with KF achieves favorable effects on treating obesity, improving inflammatory response,metabolism, and intestinal barrier function, by regulating intestinal microfloral structure in HFD-fed mice.

* Yongbo Kang657151276@qq.com

Xiangyang Kong1745982615@qq.com

1 Medical Faculty, Kunming University of Science and Technology,Kunming, Yunnan, China

2 Genetics and Pharmacogenomics Laboratory, Kunming Universityof Science and Technology, Kunming, Yunnan, China

3 School of Basic Medical Sciences, Shanxi Medical University,Taiyuan, Shanxi, China

4 Nutrition Department, The First People’s Hospital of YunnanProvince, Kunming, Yunnan, China

Electronic supplementary material The online version of this article(https://doi.org/10.1038/s41366-018-0187-x) contains supplementarymaterial, which is available to authorized users.

1234

5678

90();,:

1234567890();,:

Introduction

Obesity is defined as a disease related to a number ofhealth problems and shortened life expectancy [1]. Anincreasing amount of evidence demonstrates that obesity isclosely correlated with chronic mild inflammation (it caninduce insulin resistance, type 2 diabetes mellitus, cardio-vascular disease, obstructive sleep apnea, fatty liver disease,and cancer) [2, 3]. Currently, the high prevalence ofobesity is the major threat to public health, and the worldhas witnessed an approximate of 0.5 billion obese and1.4 billion overweight people [4]. Therefore, preventingobesity is the major challenge in the modern society.

Growing evidence has suggested that change in theintestinal flora composition is related to obesity and theassociated metabolism disturbance [5]. The intestinalflora includes trillions of bacteria, which are beneficial tonutrient collection and energy regulation [6, 7]. Theincreased proportions of the major phyla Firmicutes/Bacteroidetes, as well as alterations of several bacterialspecies are likely to contribute to obesity development indietary and genetic model mice [8, 9]. The animal modelstudy indicates that intestinal dysbiosis induced by obesityresulted from environment or genetic factors will damagethe intestinal integrity [10, 11]. Such process will resultin the release of Gram-negative bacteria endotoxin lipopo-lysaccharide (LPS) from the intestine to the blood flow [12],thus resulting in insulin resistance and metabolic inflam-mation in obese mice [13]. Furthermore, long-term LPSinjection can result in mild obesity and insulin resistancein mice [13].

A large number of studies suggest that regulatingthe intestinal flora may be an effective strategy to treatobesity. Therefore, dietary intervention is the potentialmeasure for regulating intestinal flora and improvinghost health status [14, 15]. The therapeutic effect of pre-biotics on obesity and the associated metabolic disturbancehas been verified in human study and animal model.Prebiotics are non-digestible, fermentable carbohydratesand fibers, which can decrease the body weight andplay the anti-inflammatory effect through elevating theabundance of specific salutary bacteria discovered inthe intestinal tract [16, 17]. Prebiotics can not onlychange the intestinal flora but can also improve the integrityof gut tight junction and reduce LPS-induced bloodendotoxemia [10]. Consequently, prebiotics may showbeneficial anti-inflammatory effect on the obesity-inducedinflammation.

Konjac flour (KF) is abundant in the konjac tuber(Amorphophallus konjac), which mainly contains konjacglucomannan (KGM) [18]. Over the past several centuries,KGM is consumed in the manner of rubber jelly, noodles,and other foods in Asia. KGM is a complex carbohydrate

constituted of D-glucose and D-mannose units and connectedby the β-1,4 glucosidic bond [19]. KGM shows high visc-osity and swelling capacity, which can withstand thedigestion in small intestine [20]. Besides, it has been ver-ified to change the colon/stool flora compositions in rats[21], adults [18], and sows [22]. However, a little infor-mation is available regarding the effect of KF on bodyweight, obesity-related disease, and intestinal flora, as wellas the correlation of intestinal flora with metabolic para-meters after dietary supplementation of KF so far.

The study aimed to evaluate the antiobesity and anti-inflammatory effects of KF treatment, as well as its roles inmaintaining intestinal integrity and regulating the intestinalflora in diet-induced obese mouse model.

Materials and methods

Animals and experimental design

Thirty male C57BL/6J mice aged 3 weeks bought fromHunan SJA Laboratory Animal Co., Ltd. were fed intransparent plastic feeding cages (five mice per feedingcage), and kept in a controlled environment (12-h light-dark cycle, 22 ± 1 °C as well as 55 ± 5% relative humid-ity) provided with food and water ad libitum. After1 week of acclimatization, the mice were divided in tothree groups (n= 10 per group): a control group, fed witha control diet (TP23302, fat 10% kcal, carbohydrate 71%kcal, protein 19% kcal; TROPHIC Animal Feed High-Tech Co.) (CT group); a group fed a high-fat diet (HFD)(TP23300, fat 60.00% kcal, carbohydrate 20.60% kcal,protein 60% kcal; TROPHIC Animal Feed High-TechCo.) (HFD group); and a group fed the same HFD, butsupplemented with KF supplied by Engineering ResearchCenter of Sustainable Development and Utilization ofBiomass Energy Ministry of Education [23] (HFD-KFgroup) (KF was administered in the drinking water allibitum at the final concentration of 67 mg/ml). Since theaverage water intake per rat was approximately 4.5 ml/day, the dose of KF administered was 300 mg/day). Thedrinking water was changed every day. Body weight wasrecorded weekly. Food intake was recorded once in2 days. After 12 weeks of KF treatment, blood and feceswere collected. Liver and adipose tissues (subcutaneous,epididymal, retroperitoneal, perirenal mesenteric, andinterscapular) were removed and weighed. For real-timePCR analysis, the liver, epididymal adipose as well ascolon tissue were immersed in RNA Keeper TissueStabilizer (Vazyme, China) before keeping at −80 °C.The study was given permission by the ethical committeeof Medical School, Kunming University of Science andTechnology.

Y. Kang et al.

Next-generation sequencing and sequence pre-processing

For the analysis of the gut microbiota, total genomic DNAwas isolated from the stool sample using the QIAamp DNAStool Mini Kit (Qiagen, USA) in accordance with themanufacturer’s instructions. The bacterial 16S rRNA geneV3–V4 region was amplified by forward primer F-5′-CCTACGGGRSGCAGCAG-3′ and reverse primer R-5′-GGACTACVVGGGTATCTAATC-3′. PCR method andDNA purification were performed in accordance with theIllumina Miseq 16S Metagenomic Sequencing LibraryPreparation protocol. For the first round of PCR reaction,PCR mixtures (25 µl) contained 12.5 µl of 2× KAPAHiFiHotStartReadyMix (KAPA Biosystem, Woburn, MA,USA), 2.5 µl of Microbial Genomic DNA(5 ng/µl), and 5 µlof amplicon PCR forward (1 μM) and reverse (1 μM) pri-mers, respectively. The PCR protocol was as follows: initialdenaturation at 95 °C for 3 min, 25 cycles of denaturation at95 °C for 30 s, annealing at 55 °C for 30 s, as well asextension at 72 °C for 30 s, followed by a final extension at72 °C for 5 min. The products of PCR were visualized bygel electrophoresis to confirm the size of the amplicon (550bp). For the second round, the products of PCR were pur-ified by the AMPure XP beads. Index PCR attaches dualindices and Illumina sequencing adapters by the NexteraXT Index Kit. Index PCR was carried out on all purifiedDNA by a MJ Research PTC-200P Thermal Cycler (MJResearch, Waltham, USA). PCR mixtures (50 μl) included25 μl of 2× KAPA HiFiHotStartReadyMix, 5 μl of NexteraXT Index 1 Primers (N7XX) (1 μM), 5 μl of Nextera XTIndex 2 Primers (S5XX) (1 μM), 10 μl of PCR Grade Water,and 5 μl of purified DNA. The PCR protocol was as fol-lows: initial denaturation at 95 °C for 3 min, 8 cycles ofdenaturation at 95 °C for 30 s, annealing at 55 °C for 30 s, aswell as extension at 72 °C for 30 s, followed by a finalextension at 72 °C for 5 min. The products of Index PCRwere visualized by gel electrophoresis to confirm the size ofthe amplicon (630 bp). The products of Index PCR werethen purified by the AMPure XP beads. Purified index PCRproduct concentration was measured using a NanoDropND2000 spectrophotometer (Thermo Scientific, Wilming-ton, DE, USA), diluted to a working concentration of 20 ng/μl by 10 mM Tris (pH 8.5), and equimolar amounts (100 ng)pooled. Following separation of products through 2%agarose gel electrophoresis, properly sized PCR productswere recycled and concentrated by a Zymo clean Gel DNArecovery Kit (Zymo Research Corp., Irvine, CA, USA).Final pooled DNA was quantified through the InvitrogenQubit platform (52.4 ng/μl).

A great number of reads of 16S rRNA gene generated byIllumina Miseq sequencer was initially trimmed for qualityusing the standard software tools from Illumina. The 16S

rRNA reads were processed in the light of a previouslydescribed QIIME pipeline [24]. Briefly, sequences wereclustered into operational taxonomic units using 97%homology with USEARCH and NCBI’s 16SMicrobialdataset (https://github.com/mtruglio/QIIME_utilities) andtaxonomy.

Bioinformatic analysis

A Venn diagram was adopted to show core intestinal floraspecies among 78% of the mice in all groups despite oftreatment. Principal coordinate analysis (PCoA) was adop-ted to assess the beta diversity among the groups throughade4 package, cluster packages, fpc packages, as well asclusterSim package in R software (Version 2.15.3). Theanalysis of molecular variance was applied to the weightedUniFrac distance matrix in order to define if the overallcomposition of the intestinal flora was significantly differentbetween different groups. T test was used to compare therelative levels of intestinal flora among the groups. In orderto differentiate the abundant bacterial taxa, we applied theLefSe analytic method for biomarker discovery. Pearson’scorrelation analyses were used to assess correlationsbetween taxonomy levels and obese indexes, blood para-meters as well as gene expression.

Histological analysis

Subcutaneous and epididymal adipose tissues were washedwith sterile phosphate-buffered saline (PBS), fixed in 10%v/v formalin/PBS, embedded in paraffin, sliced into 3 μmsections, as well as stained with hematoxylin and eosin.Images were obtained under a microscope (AxioObser Z1,Germany) at a magnification of ×200. Cell sizes weremeasured through DIXI3000 (Leica, Wetzlar, Germany).

Blood parameters

Plasma total cholesterol (TC), triglyceride (TG), leptin(LEP), adiponection (ADP), thiobarbituric acid-reactingsubstances (TBARSs), tumor necrosis factor-α (TNF-α),interleukin-6 (IL-6) as well as LPS concentrations weredetermined using an ELISA kit (Meimian, Jiangsu KeteBiotechnology Co., Ltd, China).

Real-time PCR analysis

Total RNA was isolated from tissues (adipose and livertissues) through Trizol (Ambion, USA) in accordance withthe manufacturer’s protocols. cDNA was prepared with theMaxima H Minus First Strand cDNA Synthesis Kit(Thermo Scientific, USA). Real-time PCRs were carried outon a 7500 Real Time PCR System (Applied Biosystems

Konjaku flour reduces obesity in mice by modulating the composition of the gut microbiota

Inc., USA) using SYBR® Premix Ex TaqTM II (Tli RNaseHPlus) (Takara Bio Inc.) in accordance with the manu-facturer’s instructions. β-actin gene was used as reference.Primer sequences of the targeted mouse genes are listed inTable S1.

Statistical analysis

All analyses were carried out by applying the IBM SPSSStatistics software 19.0 for Windows, and one-way analysisof variance was adopted to assess inter-group differences.

Results

Effects of KF treatment on body weight, liverweight, fat mass, and adipocytes in HFD-inducedobese mice

Weight gains induced by HFD for 12 weeks were 40.79 ±4.26 and 28.05 ± 2.45 g, respectively, in the HF group andCT group (Fig. 1a). However, after KF treatment, the HFD-induced weight gain was reduced by 47% in the HFDgroup (p < 0.05), and that in the HFD-KF group (11.83 ±

2.61 g) was less than that in the HFD group (22.18 ± 4.74g) (Fig. 1b). Compared with the CT group, the food effi-ciency ratio (FER; weight gain divided by food consump-tion weight) in the HFD group was remarkably increased,revealing higher efficiency of HFD on weight gain. On thecontrary, FER was decreased by 41% (p < 0.05) in the KFtreatment group (Fig. 1c) suggesting lower weight gain perg food consumption. Moreover, compared with the HFDgroup (1.50 ± 0.46 g), the liver weight in the HFD-KFgroup (1.03 ± 0.13 g) was reduced by 31% (p < 0.05)(Fig. 1d).

The fat mass in the HFD group 12 weeks later wasnotably higher than that in the CT group (Fig. 1e). Micereceiving KF treatment had low fat accumulation, sincethe total fat gain in HFD-KF group (3.89 ± 0.44 g) was46% lower than that in the HFD group (7.18 ± 0.75 g) (p< 0.05) (Fig. 1e). Compared with the CT group, the adi-pose tissue depots analysis in the HFD group suggestedhigh fat mass in superior mesenteric, intra-scapular, sub-cutaneous, epididymis, retroperitoneal and perirenalregions (p < 0.05) (Fig. 1f). Compared with the HFDmice, the adipose tissue depots in the mesenteric, intra-scapular, subcutaneous, epididymis, retroperitoneal andperirenal regions of HFD-KF group was markedly

Fig. 1 HFD and KF effect on body weight, fat mass, and adipocytemorphology. Effects of KF treatment on a body weight evolution, bbody weight gain, c food efficiency ratio, d liver weight, e total whitefat mass, f partial white fat mass, and g subcutaneous and epididymal

adipocyte morphology. Subcutaneous and epididymal adipose tissuemorphology shown at ×200 magnification. h subcutaneous adipocytemean area (μm2), i epididymal adipocyte mean area. *p < 0.05 versusCT and §p < 0.05 versus HF

Y. Kang et al.

reduced by 28% (p < 0.05) (Fig. 1f). Additionally, histo-logical analysis of epididymis and subcutaneous adiposetissues suggested that the adipocytes size outstandinglyincreased in HFD group compared with that in CT group,but that in KF-treated mice was markedly reduced (p <0.05) (Fig. 1g–i).

Effects of KF treatment on blood parameters in HFD-induced obese mice

Compared with the CT group, the serum TC, TG, LEP,TBARS, TNF-α, IL-6, and LPS levels in the HFD groupwere notably elevated (p < 0.05), while the serum ADPlevel was dramatically reduced (p < 0.05) (Table S2).Conversely, serum TC, LEP, TBARS, IL-6, and LPS levelswere markedly decreased in the HFD-KF group comparedto the HFD group (p < 0.05), while serum TG, ADP, andTNF-α levels were no significantly different (Table S2).

Effects of KF treatment on gene expression in thecolon tissue of diet-induced obese mice

We assayed the expression of gene relating gut barrierfunction in colon tissue. HFD uptake would dramaticallydownregulate Intetion, ZO-1, and ZO-2 gene expression(p < 0.05) (Fig. 2). However, HFD intake would not lead tomarkedly different Occludin gene expression (Fig. 2). Wefound that Intection and ZO-1 gene expression in colontissue significantly elevated in the HFD-KF group com-pared with the HFD group (p < 0.05) (Fig. 2). However,Occludin and ZO-1 gene expression was not significantlydifferent in the HFD-KF group compared with the HFDgroup (p < 0.05) (Fig. 2). Altogether, these results suggestedthat KF treatment can improve the intestinal barrierfunction.

Effects of KF treatment on gene expression in theadipose tissue of diet-induced obese mice

Subsequently, we had analyzed the expression of energymetabolism-related genes in adipose tissues. HFD uptakecould markedly upregulate the expression of PPARα,CPT-1, LPL, and PPARγ genes (p < 0.05) (Fig. 3). None-theless, HFD uptake would outstandingly downregulatethe expression of HLS and FAS (p < 0.05) (Fig. 3). Wediscovered that the HLS gene expression in adipose tissuesof HFD-KF group was notably higher than that in HFDgroup (p < 0.05) (Fig. 3), whereas the expression of LPLand PPARγ genes in HFD-KF group was notably lower thanthat in HFD group (Fig. 3). In addition, PPARα, CPT-1, andFAS gene expression was not significantly different in theHFD-KF group compared with the HFD group (Fig. 3).Altogether, these results suggested that KF treatment canimprove energy metabolism in the adipose tissue.

Effects of KF treatment on gene expression in thehepatic tissue of diet-induced obese mice

We also analyzed the expression of energy metabolism-related genes in the liver. HFD uptake markedly elevatedthe expression of PPARα, LPL, HLS, and PPARγ genes,while HFD intake markedly reduced CPT-1 gene expres-sion (p < 0.05) (Figure S1). In addition, HFD uptake did notdiffer outstandingly the expression of FAS (Figure S1).Conversely, the expression of PPARα, CPT-1, HLS as wellas FAS genes remarkably elevated in the HFD-KF groupcompared with the HFD group, while LPL gene expressionwas markedly lower (p < 0.05) (Figure S1). In addition,PPARγ gene expression was not remarkably different in theHFD-KF group compared with the HFD group (Figure S1).In conclusion, these results suggested that KF treatment canimprove energy metabolism in the hepatic tissue.

Effects of KF treatment on gut microbial diversity indiet-induced obese mice

To construct the intestinal flora alterations resulted fromdiet-induced obesity as well as the influence of KF therapyon intestinal flora, we had carried out new-generationsequencing of the stool 16S rRNA. Figure 4a indicates thatquantity of common species when applying 78% standard;there were altogether 833 shared species among the 3groups, and there were more species between two groups orin each group.

To further assayed whether the overall composition ofthe intestinal flora was dynamically changed, we first usedPCoA to cluster samples along orthogonal axes of maximalvariance based on weighted UniFrac distance. Difference inthe overall composition in the intestinal flora among

Fig. 2 HFD and KF effect on gene expression in colon. *p < 0.05versus CT and §p < 0.05 versus HF

Konjaku flour reduces obesity in mice by modulating the composition of the gut microbiota

different groups was statistically remarkable upon analysisof variance test (p < 0.05) (Fig. 4b).

Effects of KF treatment on the gut microbiotacomposition of diet-induced obese mice

The intestinal flora composition among different groups wasmarkedly different.

At the phylum level, compared with the CT group, therelative level of Bacteroidetes, Actinobacteria, Acid-obacteria, Gemmatimonadete, Verrucomicrobia, Chloro-flexi, Synergistetes, Thermomicrobia, Planctomycetes,Chlorobi, Spirochaetes, BRC1, and Deinococcus-Thermuswere markedly lower in the HFD group, while therelative level of Firmicutes was notably higher (p < 0.05,<0.01, or <0.001) (Fig. 5a). Moreover, difference betweenthe HFD group and HFD-KF group was not significant(Fig. 5a).

At family level, the relative abundance of Bacteroi-dales S24-7 group, Prevotellaceae, Erysipelotrichaceae,Pasteurellaceae, Streptococcaceae, Carnobacteriaceae,Neisseriaceae, Clostridiaceae 1, Christensenellaceae,Comamonadaceae, Micrococcaceae, Peptostrepto-coccaceae, Clostridiales vadinBB60 group, JG34-KF-361, and Haliangiaceae was remarkably lower in the HFDgroup than that in the CT group, while the relative level of

Lachnospiraceae and Rikenellaceae was markedly higher(p < 0.05, <0.01, or <0.001) (Fig. 5b and Supplementarytable 1). Moreover, the relative level of ClostridialesvadinBB60 group notably reduced in the HFD-KF com-pared with the HFD group, while the relative abundanceof Nitrospiraceae, JG34-KF-361, Haliangiaceae, andAerococcaceae was remarkably higher in the HFD-KFthan that in the HFD group (p < 0.05, <0.01, or <0.001)(Fig. 5b and Table S3).

At genus level, the relative level of Rikenellaceae RC9gut group, Lachnospiraceae NK4A136 group, Alistipes,unidentified Ruminococcaceae, Tyzzerella, Lachnospir-aceae UCG-006, and Ruminiclostridium 6 was remarkablyhigher in the HFD group than that in the CT group, whilethe relative abundance of Alloprevotella, Faecalibaculum,Rikenella, Parabacteroides, Prevotella 1, Streptococcus,Ruminococcaceae UCG-009, Ruminococcaceae UCG-014,Aeromicrobium, and Haliangium was markedly lower in theHFD group than that in CT group (p < 0.05, <0.01, or<0.001) (Fig. 5c and Table S3). In addition, the relativeabundance of Alistipes, Alloprevotella, and Peptococcuswas markedly lower in the HFD-KF than that in the HFDgroup, while Aeromicrobium, unidentified Nitrospiraceae,Haliangium, and Aliivibrio were notably higher in the HFD-KF than in the HFD group (p < 0.05, <0.01, or <0.001)(Fig. 5c and Table S3).

Fig. 3 HFD and KF effect ongene expression in epididymalfat. a Energy metabolism, blipolysis, c lipid synthesis. *p <0.05 versus CT and §p < 0.05versus HF

Y. Kang et al.

At the species level, the relative level of Alistipes fine-goldii markedly elevated in the HFD group compared withthe CT group, while the relative level of Megasphaeraelsdenii remarkably reduced (p < 0.05, <0.01, or <0.001)(Fig. 6). In addition, the relative abundance of Bacteroidesacidifaciens, A. finegoldii, Parabacteroides goldsteinii, andClostridium sp. ND2 was significantly lower, while therelative abundance of Aliivibrio fischeri and M. elsdeniiwas markedly higher in the HFD-KF group than that inthe KF group (p < 0.05, <0.01, or <0.001) (Fig. 6).

Next, we further analyzed the phylum, genus, and spe-cies (Figure S2) with notable differences among the CTgroup, the HFD group, and the HFD-KF group (Figure S2).Bacteroidetes, Bacteroidia, Bacteroidales, BacteroidalesS24-7 group, Prevotellaceae, and Alloprevotella playedimportant roles in the CT group. In addition, Rikenellaceae,Lachnospiraceae NK4A136 group, Alistipes, and B. acid-ifaciens played important roles in the HFD group. However,Firmicutes, Clostridia, Clostridiales, Lachnospiraceae, andRikenellaceae RC9 gut group played important roles in theHFD-KF group.

Gut microbiota-associated obesity index, bloodparameters, and gene expression

At the phylum level, 10 phyla were significantly associatedwith obesity index, blood parameters, and gene expression(p < 0.05 or <0.01) (Figure S3a). Firmicutes, Bacteroidetes,and Proteobacteria were related to obesity. For obesityindex, Firmicutes was notably and positively related tobody weight, fat mass as well as liver weight, while Bac-teroidetes was remarkably and negatively related to bodyweight and fat mass (p < 0.05 or <0.01). For blood para-meters, Firmicutes was significantly and positively corre-lated with serum TC, TG, LEP, TBARS, TNF-α, IL-6, andLPS level, while Firmicutes was notably and negativelyrelated to serum ADP level (p < 0.05 or <0.01).

Bacteroidetes and Proteobacteria were significantly andnegatively associated with serum L-6 and LPS level(p < 0.01). For gene in the colon, Firmicutes was notablyand negatively related to the expression of ZO-2 (p < 0.01).For gene in the hepatic tissue, Firmicutes was notably andpositively correlated with the expression of PPARα, LPL,HLS, and PPARγ (p < 0.05 or <0.01). Bacteroidetes wasnotably and negatively related to the expression of PPARα,HLS, and PPARγ (p < 0.05 or <0.01). For gene in theadipose tissue, Firmicutes was significantly and positivelyrelated to the expression of PPARα, CPT-1, LPL, andPPARγ (p < 0.05 or <0.01). Bacteroidetes was notably andnegatively related to CPT-1(p < 0.05 or <0.01).

At family level, 31 families were significantly associatedwith obesity index, blood parameters, and gene expression(p < 0.05 or <0.01) (Figure S3b). Enterobacteriaceaewas correlated with obesity, while ClostridialesvadinBB60 group was related to KF treatment. Forobesity index, Enterobacteriaceae was significantly andnegatively related to body weight, fat mass, and liverweight (p < 0.01). Enterobacteriaceae and ClostridialesvadinBB60 group were significantly and negatively relatedto serum TC, TG, LEP, TBARS, TNF-α, IL-6, and LPSlevels (p < 0.05 or <0.01). For gene in the colon, Clos-tridiales vadinBB60 group was notably and negativelyassociated with the expression of ZO-1 (p < 0.01). For genein the hepatic tissue, Clostridiales vadinBB60 group wasremarkably and negatively correlated with the expressionof CPT-1, HLS, and PPARγ (p < 0.05 or <0.01). Enter-obacteriaceae and Clostridiales vadinBB60 group werenotably and negatively associated with the expressionof LPL (p < 0.01). For gene in the adipose tissue, Clos-tridiales vadinBB60 group was remarkably and positivelycorrelated with the expression of PPARγ and FAS, whileClostridiales vadinBB60 group was notably and negativelyassociated with the expression of CPT-1 and LPL (p < 0.05or <0.01).

Fig. 4 HFD and KF effect on gutmicrobiota structure. a Venndiagram of overlap in speciesshared by the mice from eachgroup, b β-diversitydemonstrated that samplestended to cluster togetheraccording to the dietaryconditions

Konjaku flour reduces obesity in mice by modulating the composition of the gut microbiota

At genus level, 27 genera were notably associated withobesity index, blood parameters, and gene expression (p <0.05 or <0.01) (Figure S3c). Alistipes and Alloprevotellawere related to KF treatment. For obesity index, Alistipeswas significantly and positively associated with body

weight, fat mass as well as liver weight, while Alloprevo-tella was remarkably and negatively associated with bodyweight, fat mass as well as liver weight (p < 0.01). Forblood parameters, Alistipes was notably and positivelyassociated with serum TC, TG, LEP, TBARS, TNF-α, IL-6,

Fig. 5 Phylum, family, andgenera strikingly different acrossgroups. a Relative abundance ofthe top 35 different phylum; brelative abundance of the top 35different family; c relativeabundance of the top 35different genera. The abundanceprofiles are transformed in to Zscores by subtracting theaverage abundance and dividingthe standard deviation of allsamples. Z scores are negative(shown in blue) when the rowabundance is lower that themean. *p < 0.05 or **p < 0.01versus CT and §p < 0.05 or §§p <0.01 versus HF

Y. Kang et al.

and LPS levels, while Alloprevotella was notably andnegatively associated with serum TC, TG, LEP, TBARS,TNF-α, IL-6, and LPS (p < 0.05 or <0.01). Alistipes wasremarkably and negatively related to serum ADP, whileAlloprevotella was notably and positively associated withserum ADP level (p < 0.01). For gene in the colon, Alistipeswas notably and negatively associated with the expressionof Intectin, ZO-1, ZO-2, and Occludin (p < 0.05 or <0.01).

For gene in the hepatic tissue, Alloprevotella was remark-ably and negatively associated with the expression ofPPARα, CPT-1, LPL, HLS, and PPARγ (p < 0.05 or <0.01).Alistipes was significantly and positively related to theexpression of LPL and PPARγ, while Alistipes was notablyand negatively related to the expression of CPT-1 and FAS(p < 0.05 or <0.01). For gene in the adipose tissue, Allo-prevotella was significantly and negatively related to the

Fig. 6 Relative abundance at species level across groups. a Bacteroides acidifaciens, b Alistipes finegoldii, c Parabacteroides goldsteinii, dClostridium sp. ND2, e Aliivibrio fischeri, f Megasphaera elsdenii. *p < 0.05 or **p < 0.01 versus CT and §p < 0.05 or §§p < 0.01 versus HF

Konjaku flour reduces obesity in mice by modulating the composition of the gut microbiota

expression of PPARα, LPL, and PPARγ, while Alloprevo-tella was remarkably and positively associated with theexpression of FAS (p < 0.05 or <0.01). Alistipes was notablyand positively associated with the expression of LPL andPPARγ, while Alistipes was remarkably and positivelyassociated with the expression of HLS (p < 0.05 or <0.01).

At species level, 27 species were significantly associatedwith obesity index, blood parameters, and gene expression(p < 0.05 or <0.01) (Figure S3d). Clostridium leptum 1,Bacteroides fragilis, and Akkermansia muciniphila wererelated to obesity, while A. finegoldii and B. acidifaciensgroup were related to KF treatment. For obesity index, A.finegoldii and C. leptum 1 were significantly and positivelyrelated to body weight, fat mass as well as liver weight,while B. fragilis and A. muciniphila were significantly andnegatively associated with body weight, fat mass as well asliver weight (p < 0.05 or <0.01). For blood parameters, A.finegoldii and C. leptum 1 were significantly and positivelyrelated to serum TC, TG, LEP, TBARS, TNF-α, IL-6,and LPS levels, while B. fragilis was notably andnegatively associated with serum TC, TG, LEP, TBARS,TNF-α, IL-6, and LPS levels (p < 0.05 or <0.01). C. leptum1 was remarkably and negatively related to serum ADPlevel, while B. fragilis was notably and positively asso-ciated with serum ADP level (p < 0.01). For gene in thecolon, A. finegoldii was notably and negatively associatedwith the expression of Intectin and ZO-2 (p < 0.01). B.acidifaciens and A. finegoldii were notably and negativelyassociated with the expression of ZO-1 (p < 0.01). For genein the hepatic tissue, B. acidifaciens and B. fragilis wereremarkably and negatively associated with the expression ofPPARα (p < 0.01). B. acidifaciens and A. finegoldii werenotably and negatively associated with the expression ofCPT-1 (p < 0.01). A. finegoldii was remarkably and posi-tively associated with the expression of LPL, while B. fra-gilis was notably and negatively associated with theexpression of LPL (p < 0.01). C. leptum 1 was remarkablyand positively associated with the expression of HLS andPPARγ, while B. acidifaciens, B. fragilis, and A. mucini-phila were notably and negatively associated with theexpression of HLS (p < 0.05 or <0.01). B. acidifaciens wasremarkably and negatively associated with the expression ofPPARγ (p < 0.01). A. finegoldii was notably and negativelyassociated with the expression of FAS (p < 0.01). For genein the adipose tissue, C. leptum 1 was remarkably andpositively associated with the expression of PPARα andCPT-1, while B. acidifaciens, B. fragilis, and A. muciniphilawere significantly and negatively related to the expressionof PPARα as well as CPT-1 (p < 0.05 or <0.01). A. muci-niphila was remarkably and negatively associated withexpression of LPL (p < 0.01). A. finegoldii was notably andnegatively associated with expression of HLS (p < 0.01).C. leptum 1 was remarkably and positively associated with

expression of PPARγ, while B. acidifaciens and A. muci-niphila were significantly and negatively related toexpression of PPARγ (p < 0.05 or <0.01). These findingsdemonstrated that these phyla, families, genera, and speciesplay vital roles in obesity development.

Discussion

This study evaluated the effects of KF on the obesity,inflammation, metabolism, intestinal barrier function, andintestinal flora. KF treatment reduced body weight gain,fat mass gain as well as liver weight gain in diet-inducedobese mice. The relationship between specific diet-based food intake and weight gain is of great importanceto nutritional intervention research. On this account,FER is used as the index of the specific food efficiency.Compared with HFD mice, the KF-treated mice hadlower FER. Meanwhile, compared with HFD mice, theKF-treated mice had markedly decreased weight gainand total fat gain. These findings suggested that KF playsa certain role in suppressing HFD-induced weight gain.Noteworthily, fat accumulation remarkably reduced in themesenteric fat depot for the KF treatment group. There issome evidence to indicate other oligose also reduce thesize of adipocytes in the rat mesenteric adipose depot,which may be partially attributed to its adjacency to thegastrointestinal tract [25, 26]. Notably, a longer oligosetherapy duration may also result in greater alterations invisceral adipose depot.

In addition, KF treatment improved serum lipid, meta-bolic markers, and inflammatory factors in diet-inducedobese mice. Mice treated with KF had a greater reduction inserum TC level than the HFD mice, which is consistent withthe results from previous studies in the human and mice[27]. Furthermore, mice treated with KF showed a greaterreduction in serum LEP and TBARS level compared to theHFD mice. LEP is a protein hormone secreted by whiteadipose tissue, which can enter the blood circulation tomodulate the satiety, energy consumption, inflammatoryand immune response, carbohydrate and lipid metabolism,and intestinal nutrition absorption [28]. Serum LEP levelhas been found to be increased in obese patients [29].Serum TBARS reflect lipid peroxidation that is the oxida-tion of polyunsaturated fatty acids and lipid [30]. Lipidperoxidation can seriously damage the cell membrane,lipoprotein, and other lipid-containing structures, such asmaking the membrane fluidity and permeability change,damaging DNA and proteins, and thus affect the normal cellfunction [30]. A large number of studies suggested thatobese patients have significantly increased lipid peroxida-tion [30]. These results suggested that KF treatment canimprove lipid metabolism under HFD.

Y. Kang et al.

Chronic mild inflammation is a feature of obesity.Numerous studies suggested that obese patients had sig-nificantly increased inflammatory cytokine TNF-α as wellas IL-6 level [31]. The same results were found in thisstudy. In addition, mice treated with KF led to a greaterreduction in serum IL-6 level compared to the HFD mice.Several potential mechanisms may account for the reductionin pro-inflammatory cytokines in KF-treated mice.According to report, the pathogenic intestinal flora canirritate the LPS production and release from the gut epi-thelial cells, which can subsequently bind with the cytokinereceptor that can trigger the release of pro-inflammatorycytokine [32]. Reports suggest that the intestinal barrierfunction can be improved through consuming oligose, thusreducing the secretion of LPS from gut epithelial cells,resulting in decrease in the production of pro-inflammatorycytokines in the blood [33]. Nevertheless, it is important toevaluate whether oligose will result in the reduction ininflammatory cytokine production due to intestinal barrierfunctional change or other factors. Not surprisingly, in thisstudy HFD mice had significant increase of serum LPS levelcompared to the CT group. However, mice treated with KFled to a greater reduction in serum LPL level compared withthe HFD mice. Moreover, KF treatment improved intestinalbarrier function in the diet-induced obese mice. Mice treatedwith KF had significant increase of expression of colonicIntetion and ZO-1 genes compared with the HFD mice.

Meanwhile, KF therapy improved the expression ofgenes taking part in energy metabolism in the diet-inducedobese mice. Mice treated with KF had significant increase inthe expression of hepatic PPARα as well as CPT-1 genescompared to the HFD mice. These results suggested that KFtreatment may reduce fat accumulation by increasingmitochondrial oxidation of long-chain fatty acids as well asfatty acid oxidation in liver [34, 35]. In adipose tissues andliver tissues, lower expression of LPL gene as well as higherexpression of HSL gene in the KF-treated mice demon-strates that oligose therapy may decrease fatty acid intake aswell as enhance lipolysis [36, 37]. Mice treated with KF hadsignificant decrease in the expression of adipose PPARγgene compared with the HFD mice. These findings indi-cated that KF treatment may promote the deposition of fattyacids in adipose tissue and the differentiation of adipocytes[38]. However, further research is required to determinewhether oligose has direct influence on the energy meta-bolic genes, or that the oligose effect is mainly mediated byalterations in intestinal flora.

Oligose can regulate gut microbiota structure, which mayexplain that KF treatment improved obesity-related indexes.KF treatment remarkably improved the diversity, richness,and evenness of the gut flora in diet-induced obese mice.Meanwhile, KF also regulate intestinal flora structurein diet-induced obese mice. At family level, the relative

abundance of Nitrospiraceae JG34-KF-361, Haliangiaceae,and Aerococcaceae remarkably elevated in the HFD-KFgroup compared with the HFD group, while the relativeabundance of Clostridiales vadinBB60 group notablyreduced in the HFD-KF group compared with the HFDgroup. At genus level, the relative abundance of Aero-microbium, unidentified Nitrospiraceae, Haliangium, andAliivibrio markedly enhanced in the HFD-KF group com-pared with the HFD group, while the relative abundance ofAlistipes, Alloprevotella, and Peptococcus significantlydecreased in the HFD-KF group compared to the HFDgroup. Clarke et al. [39] also found that the relative abun-dance of Alistipes notably elevated in obese mice. Obesemice treated with vancomycin or probiotic Lactobacillussalivarius UCC118 producing bacteriocin not only reducedweight gain but also reduced the relative abundance Alis-tipes and Peptococcus [39]. In addition, some studies foundacquired immune deficiency syndrome patients showedremarkably elevated relative abundance of Alistipes com-pared to healthy subjects [40, 41]. Recent studies hadshown that Alistipes was associated with health risks [42–44]. Therefore, Alistipes may be an important genus asso-ciated with obesity. Louis et al. [45] also found that weightand the relative abundance of Alloprevotella significantlyreduced in obese patients after weight loss treatment. Atspecies level, the relative abundance of A. fischeri and M.elsdenii notably elevated in the HFD-KF group comparedwith the HFD group, while the relative abundance of B.acidifaciens, A. finegoldii, P. goldsteinii, and Clostridiumsp. ND2 remarkably decreased in the HFD-KF groupcompared with the HFD group. Petriz et al. [46] also dis-covered that compared with Wistar, the relative abundanceof B. acidifaciens in obese rats is higher. Recent studiesdemonstrated that B. acidifaciens plays a vital role in pro-ducing the immune globulin (IgA) in mouse large intestine[47]. Therefore, it may exert the adaptive function in theintestinal mucosa immune system [48]. IgA is increasedduring metabolic disturbance; as a result, the relative levelof B. acidifaciens may be related to the role of intestinalflora in obesity-specific inflammatory signal transductionin obese rats [49, 50]. Liu et al. [51] also found thatP. goldsteinii was significantly positively correlated withobesity-related indexes. Some studies found that bodymass index and weight significantly decreased, whileM. elsdenii remarkedly increased in obese patient afterRoux-en-Y gastric bypass surgery [52, 53]. M. elsdeniiis the most important lactate utilization bacteria in therumen. Recent studies have shown that Megasphaeraspp. can reduce lactic acidosis by using lactate to producepropionate, acetate, and butyrate [54, 55].

The abundances of Clostridiales vadinBB60 group,Alistipes, Alloprevotella, A. finegoldii and B. acidifacienswere significantly associated with body weight, fat mass,

Konjaku flour reduces obesity in mice by modulating the composition of the gut microbiota

blood lipids, serum metabolic factors, serum inflammatoryfactors, serum endotoxins, and the expression of genestaking part in intestinal barrier function, and energy meta-bolism. Clostridiales vadinBB60 group was positivelycorrelated with serum TG and LEP, ZO-1, PPARα, andCPT-1 gene expression as well as LPL and HLS geneexpression, while Clostridiales vadinBB60 group wasnegatively correlated with FAS gene expression. Alistipeswas positively related to body weight, fat mass, serum TC,TG, and LEP as well as PPARγ gene expression, andnegatively with serum ADP and CPT-1 gene expression.Alloprevotella was positively correlated with FAS geneexpression, negatively with PPARα, CPT-1, LPL, and HLSgene expression. A. finegoldii was positively related to bodyweight, fat mass, and serum LPS, and negatively related toIntectin, ZO-1, ZO-2, CPT-1, and LPL gene expression. B.acidifaciens was negatively related to PPARα, CPT-1, andHLS gene expression. These results suggested that thesebacteria play important roles in the development of obesity.

Conclusions

KF treatment could improve the obesity-related indexes,inflammatory state, intestinal barrier function, and energymetabolism in diet-induced obese mice. Furthermore, KFtreatment could also increase the diversity and improve thestructure of intestinal microbiota. KF treatment could pro-mote the production of beneficial bacteria and inhibit thegrowth of harmful bacteria. Taken together, these findingssuggested that KF treatment might regulated the intestinalflora and supply a natural alternative for antiobesity.

Acknowledgements We thank Engineering Research Center of Sus-tainable Development and Utilization of Biomass Energy Ministry ofEducation for supplying KF.

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflict ofinterest.

References

1. Hoyt CL, Burnette JL, Austergussman L. “Obesity is a disease”:examining the self-regulatory impact of this public-health mes-sage. Psychol Sci. 2014;25:997.

2. Yoshimoto S, Loo TM, Atarashi K, Kanda H, Sato S, OyadomariS, et al. Obesity-induced gut microbial metabolite promotes livercancer through senescence secretome. Nature. 2013;499:97–101.

3. Osborn O, Olefsky JM. The cellular and signaling networkslinking the immune system and metabolism in disease. Nat Med.2012;18:363–74.

4. Moustafa ES, Froguel P. From obesity genetics to the future ofpersonalized obesity therapy. Nat Rev Endocrinol. 2013;9:402.

5. Ridaura VK, Faith JJ, Rey FE, Cheng J, Duncan AE, Kau AL,et al. Gut microbiota from twins discordant for obesity modulatemetabolism in mice. Science. 2013;341:1241214.

6. Dethlefsen L, Mcfall-Ngai, Amp M, Relman DA. An ecologicaland evolutionary perspective on human-microbe mutualism anddisease. Nature. 2007;449:811–8.

7. Bäckhed F, Ding H, Wang T, Hooper LV, Koh GY, Nagy A, et al.The gut microbiota as an environmental factor that regulates fatstorage. Proc Natl Acad Sci USA. 2004;101:15718.

8. Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, Mardis ER,Gordon JI. An obesity-associated gut microbiome with increasedcapacity for energy harvest. Nature. 2006;444:1027–31.

9. Goodman AL, Kallstrom G, Faith JJ, Reyes A, Moore A, DantasG, et al. Extensive personal human gut microbiota culturecollections characterized and manipulated in gnotobiotic mice.Proc Natl Acad Sci USA. 2011;108:6252–7.

10. Cani PD, Possemiers S, Wiele TVD, Guiot Y, Everard A, RottierO, et al. Changes in gut microbiota control inflammation in obesemice through a mechanism involving GLP-2-driven improvementof gut permeability. Gut. 2009;58:1091.

11. Cani PD, Bibiloni R, Knauf C, Waget A, Neyrinck AM, DelzenneNM, et al. Changes in gut microbiota control metabolicendotoxemia-induced inflammation in high-fat diet–induced obe-sity and diabetes in mice. Diabetes. 2008;57:1470–81.

12. Cani PD, Amar J, Iglesias MA, Poggi M, Knauf C, Bastelica D,et al. Metabolic endotoxemia initiates obesity and insulin resis-tance. Diabetes. 2007;56:1761.

13. Wellen KE, Hotamisligil GS. Inflammation, stress, and diabetes.J Clin Invest. 2005;115:1111–9.

14. Kovatcheva-Datchary P, Arora T. Nutrition, the gut microbiomeand the metabolic syndrome. Best Pract Res Clin Gastroenterol.2013;27:59–72.

15. Dewulf EM, Cani PD, Claus SP, Susana F, Puylaert Philippe GB,Neyrinck AM, et al. Insight into the prebiotic concept:lessons from an exploratory, double blind intervention studywith inulin-type fructans in obese women. Gut. 2013;62:1112–21.

16. Nakamura YK, Omaye ST. Metabolic diseases and pro- andprebiotics: mechanistic insights. Nutr Metab. 2012;9:60.

17. Everard A, Lazarevic V, Derrien M, Girard M, Muccioli GG,Neyrinck AM, et al. Responses of gut microbiota and glucose andlipid metabolism to prebiotics in genetic obese and diet-inducedleptin-resistant mice. Diabetes. 2011;60:2775–86.

18. Chen HL, Cheng HC, Liu YJ, Liu SY, Wu WT. Konjac acts as anatural laxative by increasing stool bulk and improving colonicecology in healthy adults. Nutrition. 2006;22:1112–9.

19. Li B, Xia J, Wang Y, Xie B. Grain-size effect on the structure andantiobesity activity of konjac flour. J Agric Food Chem.2005;53:7404–7.

20. Sun HQ, Tan CQ, Wei HK, Zou Y, Long G, Ao JT, et al. Effectsof different amounts of konjac flour inclusion in gestation diets onphysio-chemical properties of diets, postprandial satiety in preg-nant sows, lactation feed intake of sows and piglet performance.Anim Reprod Sci. 2015;152:55–64.

21. Young W, Roy NC, Lee J, Lawley B, Otter D, Henderson G, et al.Bowel microbiota moderate host physiological responses to diet-ary konjac in weanling rats. J Nutr. 2013;143:1052–60.

22. Tan CQ, Wei HK, Sun HQ, Long G, Ao JT, Jiang SW, et al.Effects of supplementing sow diets during two gestations withkonjac flour and Saccharomyces boulardii on constipation inperipartal period, lactation feed intake and piglet performance.Anim Feed Sci Technol. 2015;210:254–62.

23. Wu CF, Dong YY, Li JJ, Tang XH, Huang ZX. Study on thepreparation of Konjac Oligo-glucomannan by β-mannanase. Bio-technol Bull. 2010;2010:118–22.

Y. Kang et al.

24. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, BushmanFD, Costello EK, et al. QIIME allows analysis of high-throughputcommunity sequencing data. Nat Methods. 2010;7:335–6.

25. Sato M, Uzu K, Yoshida T, Hamad EM, Kawakami H, Mat-suyama H, et al. Effects of milk fermented by Lactobacillusgasseri SBT2055 on adipocyte size in rats. Br J Nutr. 2008;99:1013.

26. Neyrinck AM, Hée VFV, Piront N, Backer FD, Toussaint O, CaniPD, et al. Wheat-derived arabinoxylan oligosaccharides withprebiotic effect increase satietogenic gut peptides and reducemetabolic endotoxemia in diet-induced obese mice. Nutr Diabetes.2012;2:e28.

27. Keithley JK, Barbara S, Mikolaitis SL, Mark DM, Zeller JM, LouF, et al. Safety and efficacy of glucomannan for weight loss inoverweight and moderately obese adults. J Obes. 2013;2013:610908.

28. SáInz N, Barrenetxe J, Morenoaliaga MJ, MartíNez JA. Leptinresistance and diet-induced obesity: central and peripheral actionsof leptin. Metabolism. 2015;64:35–46.

29. Correia ML, Haynes WG. Leptin, obesity and cardiovasculardisease. Curr Opin Nephrol Hypertens. 2004;13:215–23.

30. Halliwell B, Chirico S. Lipid peroxidation: its mechanism, mea-surement, and significance. Am J Clin Nutr. 1993;57:715S.

31. Kang Y, Zhang X, Cai Y, Su J, Kong X. Gut microbiota andmetabolic disease: from pathogenesis to new therapeutic strate-gies. Allergol Immunopathol. 2017;28:1.

32. Musso G, Gambino R, Cassader M. Interactions between gutmicrobiota and host metabolism predisposing to obesity and dia-betes. Annu Rev Med. 2011;62:361–80.

33. Ohland CL, Macnaughton WK. Probiotic bacteria and intestinalepithelial barrier function. Am J Physiol Gastrointest Liver Phy-siol. 2010;298:807–19.

34. Zhang Y, Ma K, Song S, Elam MB, Cook GA, Park EA. Per-oxisomal proliferator-activated receptor-gamma coactivator-1alpha (PGC-1 alpha) enhances the thyroid hormone inductionof carnitine palmitoyltransferase I (CPT-I alpha). J Biol Chem.2004;279:53963–71.

35. Finck BN, Kelly DP. Peroxisome proliferator–activated receptor γcoactivator-1 (PGC-1) regulatory cascade in cardiac physiologyand disease. Circulation. 2007;115:2540.

36. Preiss-Landl K, Zimmermann R, Hämmerle G, Zechner R.Lipoprotein lipase: the regulation of tissue specific expression andits role in lipid and energy metabolism. Curr Opin Lipidol.2002;13:471.

37. Lafontan M, Langin D. Lipolysis and lipid mobilization in humanadipose tissue. Prog Lipid Res. 2009;48:275–97.

38. Wang J, Jiang M, Xia JH, Li X, Yi GH, Yuan CG. The role ofPPARδ in lipid metabolism. Chem Life. 2008;28:175–7.

39. Clarke SF, Murphy EF, O’Sullivan O, Ross RP, O’Toole PW,Shanahan F, et al. Targeting the microbiota to address diet-induced obesity: a time dependent challenge. PLoS ONE. 2013;8:e65790.

40. Dinh DM, Volpe GE, Duffalo C, Bhalchandra S, Tai AK, KaneAV, et al. Intestinal microbiota, microbial translocation, andsystemic inflammation in chronic HIV infection. J Infect Dis.2015;211:19–27.

41. Mchardy IH, Li X, Tong M, Ruegger P, Jacobs J, Borneman J,et al. HIV Infection is associated with compositional and

functional shifts in the rectal mucosal microbiota. Microbiome.2013;1:1–12.

42. Xu J, Lian F, Zhao L, Zhao Y, Chen X, Zhang X, et al. Structuralmodulation of gut microbiota during alleviation of type 2 diabeteswith a Chinese herbal formula. ISME J. 2015;9:552–62.

43. Saulnier DM, Riehle K, Mistretta TA, Diaz MA, Mandal D, RazaS, et al. Gastrointestinal microbiome signatures of pediatricpatients with irritable bowel syndrome. Gastroenterology. 2011;141:1782.

44. Rautio M, Eerola E, Väisänen-Tunkelrott ML, Molitoris D,Lawson P, Collins MD, et al. Reclassification of Bacteroidesputredinis (Weinberg et al., 1937) in a new genus Alistipes gen.nov., as Alistipes putredinis comb. nov., and description ofAlistipes finegoldii sp. nov., from human sources. Syst ApplMicrobiol. 2003;26:182–8.

45. Louis S, Tappu RM, Damms-Machado A, Huson DH, BischoffSC. Characterization of the gut microbial community of obesepatients following a weight-loss intervention using whole meta-genome shotgun sequencing. PLoS ONE. 2016;11:e0149564.

46. Petriz BA, Castro AP, Almeida JA, Gomes CP, Fernandes GR,Kruger RH, et al. Exercise induction of gut microbiota mod-ifications in obese, non-obese and hypertensive rats. BMCGenomics. 2014;15:511.

47. Yanagibashi T, Hosono A, Oyama A, Tsuda M, Suzuki A,Hachimura S, et al. IgA production in the large intestineis modulated by a different mechanism than in the smallintestine: Bacteroides acidifaciens promotes IgA production in thelarge intestine by inducing germinal center formation andincreasing the number of IgA+B cells. Immunobiology. 2013;218:645–51.

48. Peterson DA, Mcnulty NP, Guruge JL, Gordon JI. IgA response tosymbiotic bacteria as a mediator of gut homeostasis. Cell HostMicrobe. 2007;2:328.

49. Rodriguezsegade S, Camiña MF, Carnero A, Lorenzo MJ, AlbanA, Quinteiro C, et al. High serum IgA concentrations in patientswith diabetes mellitus: agewise distribution and relation to chroniccomplications. Clin Chem. 1996;42:1064–7.

50. Hotamisligil GS, Hotamisligil GS. Inflammation and metabolicdisorders. Nature. 2007;444:860–7.

51. Liu Z, Chen Z, Guo H, He D, Zhao H, Wang Z, et al. Themodulatory effect of infusions of green tea, oolong tea, and blacktea on gut microbiota in high-fat-induced obese mice. Food Funct.2016;7:4869–79.

52. Patrone V, Vajana E, Minuti A, Callegari ML, Federico A,Loguercio C. et al. Postoperative changes in fecal bacterial com-munities and fermentation products in obese patients undergoingbilio-intestinal bypass. Front Microbiol. 2016;7:200

53. Federico A, Dallio M, Tolone S, Gravina AG, Patrone V, RomanoM, et al. Gastrointestinal hormones, intestinal microbiota andmetabolic homeostasis in obese patients: effect of bariatric sur-gery. In Vivo. 2016;30:321.

54. Shetty SA, Marathe NP, Lanjekar V, Ranade D, Shouche YS.Comparative genome analysis of Megasphaera sp. reveals nichespecialization and its potential role in the human gut. PLoS ONE.2013;8:e79353.

55. Marounek M, Fliegrova K, Bartos S. Metabolism and somecharacteristics of ruminal strains of Megasphaera elsdenii. ApplEnviron Microbiol. 1989;55:1570–3.

Konjaku flour reduces obesity in mice by modulating the composition of the gut microbiota