Novel role of hesperidin improve obesity in HFD mice by modulating the composition of the gut microbiota

Background Hesperidin is a plant-derived dihydroavone derivatives with multiple pharmacological function. Obesity is associated with low-grade chronic inammation and intestinal dysbiosis. We examined the possibility that hesperidin may prevent diet-induced obesity by modulating the composition of the gut microbiota. High-fat diet (HFD)-fed mice were treated with hesperidin. Its effects on the gut microbiota were assessed by horizontal faecal microbiota transplantation (FMT) and 16S rDNA-based microbiota analysis. Results Gut microbiota analysis revealed that hesperidin selectively promoted the growth of benecial Lactobacillus salivarius and harmful Staphylococcus sciuri, Desulfovibrio C21_c20 and inhibiting benecial Bidobacterium pseudolongum, Mucispirillum schaedleri and harmful Helicobacter ganmani , Helicobacter hepaticus . in HFD-fed mice. However, hesperidin reverses obesity, inammation and improves gut integrity in HFD-fed mice. The anti-obesity effects and hesperidin-modulated Lactobacillus salivarius, Desulfovibrio C21_c20, Mucispirillum schaedleri and Helicobacter hepaticus were transmissible via horizontal faces transfer from hesperidin-treated mice to HFD-fed mice. The anti-obese effects and most hesperidin modied gut microbiota horizontal fecal transplantation. Our data demonstrate that hesperidin has role to reduce body weight and reverse HFD-related disorders in HFD-fed mice by enriching benecial and inhibiting harmful microbes. Staphylococcaceae family; Desulfovibrio C21_c20 in Proteobacteria phylum, Deltaproteobacteria class, Desulfovibrionales order, Desulfovibrionaceae family and Akkermansia muciniphila in Verrucomicrobia phylum, Verrucomicrobiae class, Verrucomicrobiales order, Verrucomicrobiaceae family were decreased in HFD-fed mice. The other four including Helicobacter ganmani and Helicobacter hepaticus in the Proteobacteria phylum, Epsilonproteobacteria class, Campylobacterales order, Helicobacteraceae family; Bidobacterium pseudolongum in Actinobacteria phylum, Actinobacteria class, Bidobacteriales order, Bidobacteriaceae family and Mucispirillum schaedleri in Deferribacteres phylum, Deferribacteres class, Deferribacterales order, Deferribacteraceae family were increased HFD-fed mice. closer look hesperidin from the phylum to species levels.


Background
Obesity is considered to be a disease condition associated with high risk of numerous health problems.
The increasing prevalence of obesity has becoming a major threat to public health, which makes administration of obesity a main challenge for modern societies. 1 Obesity is characterized by fat mass accumulation, chronic subclinical in ammation and imbalanced gut microbiota. Un t life style especially high fat diet and inadequate exercises, neuronal and hormonal factors, genetic and epigenetic mechanisms all contribute to obesity development. 2 Gut microbiota plays roles in obesity development. 3 A number of bioactive chemicals have been reported to alleviate disease symptoms by modulating gut microbiota. Hesperidin is a avanone glycoside (a subclass of avonoids) that is rich in citrus fruits and quite a few vegetables. 4 Previous studies have shown that hesperidin has various biological activities including vitamin-like activity, antioxidant, anti-in ammatory, anticarcinogenic, anti-hyperglycemic, antihypolipidemic and antiallergic properties. 4−6 Although a large number of studies have been published describing its new pharmacological activities, molecular targets and mechanisms of action. None reported its' effects on HFD induced obesity and gut microbiota.
In brief, gut microbiota is a potential target for hesperidin to intervene HFD resulted obesity. In the present study, we examined whether hesperidin can decrease obesity in HFD-fed mice. Our results indicate that hesperidin reduced obesity, in ammation, improved gut integrity and modi ed a few gut microbiota species in HFD-fed mice. The anti-obese effects and most hesperidin modi ed gut microbiota species were transmissible through horizontal fecal transplantation. Our data demonstrate that hesperidin has a role to reduce body weight and reverse HFD-related disorders in HFD-fed mice by enriching some bene cial and inhibiting harmful microbes.

Results
Hesperidin prevents HFD-induced obesity in mice.
HFD feeding for 10 weeks led to signi cant increases in body weight, epididymal and visceral fat accumulation, plasma total cholesterol, triglyceride, high-density lipoprotein, low-density lipoprotein and slight increases in liver weight ( Figure 1A-1I). 2% hesperidin did not produce any signi cant effects in normal diet mice except for decreasing plasma low-density lipoprotein ( Figure 1A-1I). Supplementation with hesperidin decreased weight gain, fat accumulation and plasma lipids in a dose-dependent manner in HFD-fed mice ( Figure 1A-1I). The effects of hesperidin on body weight and obesity parameters were not due to reduced food consumption or energy extraction according to our weekly feeding records. These results implied that hesperidin reduced weight gain, fat accumulation and plasma lipids in HFD-fed mice.
Hesperidin reduced in ammation in HFD-fed mice.
Studies have shown that obese was characterized by low grade in ammation with higher proin ammatory cytokines including tumor necrosis factor-alpha (TNF-α), interleukin-1-beta (IL-1β), interleukin-6 (IL-6). 7 We measured plasma levels of TNF-α and IL-6 proteins and colonic messenger RNA (mRNA) expression levels of these cytokines after 10 weeks of HFD feeding with or without hesperidin supplementation. IL-1β, TNF-α and IL-6 levels were higher in plasma and colons of HFD-fed mice compared with normal diet feeding mice (Figure 2A-2E). while the expression level of these cytokines was reduced in a dose-dependent manner by hesperidin treatment (Figure 2A-2E). Inducible NO-synthase (iNOS) is a key pro-in ammatory mediator. iNOS mRNA expression increased in colons of HFD-fed mice compared to ND-fed mice but decreased following treatment with hesperidin ( Figure 2F). These results indicate that hesperidin reduced in ammation in HFD-fed mice.
Hesperidin maintains intestinal integrity in HFD mice.
Previous studies have shown that gut microbiota dysbiosis caused by HFD increased gut permeability and subsequently resulted in releasing of bacterial endotoxin into the circulation. 8 We examined effects of hesperidin on gut integrity. Colon length, lipid binding protein (LBP) and intestinal fatty acid binding protein (iFABP) are key markers of intestinal integrity, muc2 is an important indicator of gut barrier function, and claudin2, occluding and zonula occludens-1 (ZO-1) are three main tight junction components. HFD feeding reduced colon length, expression of the tight junction components, increased plasma LBP and iFABP, while all these effects were reversed by hesperidin supplementation (Figure 3A-3G). These results suggested that hesperidin improved intestinal barrier integrity in HFD-fed mice.
The gut microbiota of obese humans and HFD-fed mice is characterized by an increased Firmicutes-to-Bacteroidetes ratio, elevated endotoxin producing Proteobacteria, reduced immuno-homeostatic Akkermansia muciniphila. 9 10 We examined the effects of hesperidin on gut microbiota composition by performing a pyrosequencing-based analysis of bacterial 16S rRNA (V3-V4 region) in caecal feces. A total of 36211258 effective reads were obtained from all fecal samples. Based on 99% similarity level, the reads were clustered into 343273 OTUs. HFD reduced OTUs compared to ND-fed mice. Hesperidin reversed HFD-induced OTU decreasing in a dose dependent manner ( Figure S1A). Microbiota richness and evenness were increased by hesperidin as indicated by α-diversity analysis ( Figure S1B). UniFracbased principal coordinates analysis (PCoA) showed a distinct clustering of microbiota composition for each treatment group ( Figure S1C). Hesperidin also decreased Firmicutes-to-Bacteroidetes ratio ( Figure   S1D).
The OTUs can be annotated to 8 phylums, 13 classes, 15 orders, 22 families, 29 genuses and 19 species ( Figure 4). We detected 8 species that were signi cantly different between ND-fed and HFD-fed mice. Four of them including Lactobacillus salivarius in the Firmicutes phylum, Bacilli class, Lactobacillales order, Lactobacillaceae family; Staphylococcus sciuri in the Firmicutes phylum, Bacilli class, Bacillales order, Staphylococcaceae family; Desulfovibrio C21_c20 in Proteobacteria phylum, Deltaproteobacteria class, Desulfovibrionales order, Desulfovibrionaceae family and Akkermansia muciniphila in Verrucomicrobia phylum, Verrucomicrobiae class, Verrucomicrobiales order, Verrucomicrobiaceae family were decreased in HFD-fed mice. The other four including Helicobacter ganmani and Helicobacter hepaticus in the Proteobacteria phylum, Epsilonproteobacteria class, Campylobacterales order, Helicobacteraceae family; Bi dobacterium pseudolongum in Actinobacteria phylum, Actinobacteria class, Bi dobacteriales order, Bi dobacteriaceae family and Mucispirillum schaedleri in Deferribacteres phylum, Deferribacteres class, Deferribacterales order, Deferribacteraceae family were increased in HFD-fed mice. A closer look at the microbial community revealed speci c in uence of hesperidin from the phylum to species levels. Lactobacillus salivarius, Staphylococcus sciuri and Desulfovibrio _C21_c20 were enriched in the Hesperidin supplemented HFD-fed mice ( Figure 4, Figure S2A-2C); Helicobacter ganmani, Helicobacter hepaticus, Bi dobacterium pseudolongum and Mucispirillum schaedleri were decreased in the hesperidin supplemented HFD-fed mice ( Figure 4, Figure S2D-2G). While Akkermansia muciniphila failed to be changed by hesperidin (Figure 4, Figure S2H). Those results implied that hesperidin modi ed composition of the gut microbiota and reverses part of HFD-induced gut dysbiosis.
The bene cial effects of hesperidin were transferable by fecal transplantation.
It was reported that diet-induced obesity and associated metabolic disorders may resulted from gut microbiota. 3 Anti-obesogenic effects of Chinese herbs such as polysaccharides from Ganoderma lucidum and Hirsutella sinensis were mediated by the gut microbiota. 11 12 We tested whether the bene cial effects of hesperidin may also be mediated by the gut microbiota. Fecal microbiota from NDfed mice treated with saline, hesperidin was transplanted into HFD-fed recipients. To further con rm that our method of FMT works, one more control was conducted: fecal microbiota from HFD-fed mice treated with saline were transplanted into ND-fed recipients ( Figure 5A). FMT from HFD-fed mice increased obesity traits, in ammation and gut integrity in ND recipients, though most of the indicators were not signi cant ( Figure 5B-5G, Figure 6A-6F, Figure 7A-7G). On the contrary, FMT from ND-fed groups reduced obesity traits, in ammation and gut integrity in HFD recipients compared with the controls ( Figure 5B-5G, gure 6A-6F, gure 7A-7G). Furthermore, FMT from hesperidin treated ND-fed groups had more signi cant effects in reducing obesity traits, in ammation and gut integrity in HFD recipients ( Figure 5B-5G, Figure  6A-6F, Figure 7A-7G). These results proved that the gut microbiota mediates the bene cial effects of hesperidin.
FMT transmitted speci c intestinal microbial taxa.
To check whether the bene cial effects of hesperidin are resulted from the speci c microbe it regulated and whether the modi ed microbiota can be transmitted to recipients by FMT, we sequenced the gut microbiota after FMT. FMT from hesperidin treated ND-fed groups increased OTUs in HFD recipients ( Figure S2A). FMT increased total species diversity as indicated by Chao1 value, but decreased richness and evenness of the main microbiota as shown by Shannon and Simpson value ( Figure S2B). UniFracbased principal coordinates analysis (PCoA) showed a distinct clustering of microbiota composition for each treatment group ( Figure S2C). FMT failed to reverse HFD induced increasing of Firmicutes-to-Bacteroidetes ratio ( Figure S2D). Among the seven microbiome that speci cally regulated by hesperidin, Lactobacillus salivarius, Staphylococcus sciuri, Desulfovibrio C21_c20, Mucispirillum schaedleri and Helicobacter hepaticus can be transmitted from donor to recipient mice ( Figure 8, Figure S4A, 4B, 4C, 4E, 4G) while Helicobacter ganmani and Bi dobacterium pseudolongum can not ( Figure 8, Figure S4D, 4F). Akkermansia muciniphila was also failed to be transmitted from ND-fed donor to HFD-fed recipient mice ( Figure 8, Figure S4H).

Discussion
Polyphenols have been reported to modulate the metabolism and/or in ammation related to obesity. 13 As a bioactive chemical belong to polyphenols, hesperidin was extensively studied for its effects on cancer and cardiovascular diseases but not obesity. 4 5 In vitro studies indicated that citrus polyphenols including hesperidin caused a reduction in adipocyte differentiation, lipid content in the cell and adipocyte apoptosis, which showed positive role in the management of obesity. 14 Animal evidence were not entirely consistent, but most of them indicated a reduction in adipose tissue, increased genes expression resulted in stimulation to β-oxidation, improved lipid pro le and glycemia as well as improved in ammatory status. 15 Our experiments on HFD-fed mice also indicated a reduction in adipose tissue, improved lipid pro le and in ammatory status. Moreover, hesperidin improved intestinal barrier function in HFD-fed mice. Role of hesperidin on decreasing intestinal in ammation and restoring intestinal barrier function was also proved in DSS-induced colitis mice. 16 However, solid clinical evidences were very limited. A systematic review and meta-analysis concluded that hesperidin might not affect lipid pro le and blood pressure based on 10 randomized controlled clinical trials. 17 Therefore, well-designed trials on human are still needed to con rm anti-obesity effects of hesperidin.
The trillions of gut microbiota play important roles in ingesta digestion, immunity regulation and energy equilibrium. Innumerable studies indicated that changes in the composition of the gut microbiota were related to the development of various diseases including obesity. 18 Lesser diversity and richness, increased ratio of the major phyla Firmicutes/Bacteroidetes and changes in several bacterial species are common features of both obesity mice and human fecal samples. 10 Moreover, obese animals with gut dysbiosis often have impaired intestinal integrity. 19 Gut microbiota that has 10 times the number of human cells and 150 times number of genes of the human genome was considered a "hidden organ". 20 New ndings from this eld and their importance for human health has providing a new frontier to understand occurrence and development of various diseases, as well as mechanism of drugs, traditional herb medicines, bioactive chemicals and functional foods. 21 22 For obesity, research found that the gut microbiota of obese humans and HFD-fed animals were different from lean and ND-fed animals, and different studies on obesity or weight losing subjects found quite a few microbes from phylum to species level that were positively or negatively associated with obesity, however, there are not many consistent conclusions for an obese microbiota pro le except for an increased Firmicutes-to-Bacteroidetes ratio, elevated abundance of Akkermansia muciniphila. 23 Different studies found different obesity related microbes probably because of different experimental objects and background diet since gut microbiota was different between species and diet was the most important factor shaping it. 24 25 Second, not all microbes that in a lower taxon like genera and species play the same role within a given higher taxon like phylum. Different bacterial species present different characteristics, which may be related to bene cial or harmful traits. For example, Lactobacillus such as Lactobacillus plantarum and paracasei have been associated with thinness, while species such as Lactobacillus reuteri that have been associated with obesity. 26 This suggests that the physiological effects of microbes are detailly dependent on the strain. Therefore, it may be inaccurate to conclude traits related microbe that at higher than species or strains level. Third, as members in a complicated ecological system, one microbe may be regulated by another according to speci c physiology conditions, for example, both Bi dobacterium pseudolongum and Akkermansia muciniphila were bene cial microbes, however, rats fed with Bi dobacterium pseudolongum strain Patronus led to a large increase of mucus thickness associated with a decrease of Akkermansia muciniphila which was lost to bi dobacteria in the competition for the mucus niche. 27 Therefore, when state a traits related microbe, it is under the speci c physiology conditions with a speci c microbial community characteristic. Forth, experiment procedures including sampling, sequencing and bioinformatic analysis may also contribute to the inconsistency among reports.
In the present study using 16S rRNA sequencing, we detected 8 species that were signi cantly different between ND-fed and HFD-fed mice. Desulfovibrio C21-c20 was reported to be positively related to cisplatin-induced mucositis of Male Wistar rats 28 and negatively related to antihyperlipidemic effects of Rhizoma Coptidis alkaloids in high-fat and high-cholesterol induced hyperlipidemic B6 mice. 29 Staphylococcus sciuri has been reported as a human opportunistic pathogen in nosocomial diseases and related infections. 30 Helicobacter hepaticus is a pathogen that can cause typhlitis, colitis, and hepatitis. 31 Helicobacter ganmani may also be a pathogen since H. ganmani infection was associated with a signi cant increase in the expression of the proin ammatory cytokine IL12/23p40 in IL10-de cient mice. 31 Those four microbes were considered to be harmful microbes. Lactobacillus salivarius is a promising probiotic since it has been reported to possess certain abilities such as enhancement of the immune system, attenuation of gut in ammation and to have antimicrobial activity against pathogenic bacteria like Staphylococcus aureus. 32 33 Akkermansia muciniphila has been proved to have a negative correlation to overweight, obesity, untreated type 2 diabetes mellitus or hypertension studies and its bene cial effects on obesity have been reported by a clinical trial. 34 35 Bi dobacterium pseudolongum is a bene cial microbe that has a role to protect gut barrier, however it competes mucus nich with Akkermansia muciniphila. Mucispirillum schaedleri was reported to confers protection against Salmonella colitis in mice by competing for anaerobic respiration substrates in the gut. 36 Those four were considered to be bene cial microbes.
In our study, we found bene cial Lactobacillus salivarius, Akkermansia muciniphila and harmful Staphylococcus sciuri, Desulfovibrio C21_c20 were decreased in HFD-fed mice, while bene cial Bi dobacterium pseudolongum, Mucispirillum schaedleri and harmful Helicobacter ganmani, Helicobacter hepaticus were increased in HFD-fed mice compared to ND-fed controls. Bene cial Lactobacillus salivarius and harmful Staphylococcus sciuri, Desulfovibrio C21_c20 were enriched in the Hesperidin supplemented HFD-fed mice, while bene cial Bi dobacterium pseudolongum, Mucispirillum schaedleri and harmful Helicobacter ganmani, Helicobacter hepaticus were decreased in the hesperidin supplemented HFD-fed mice. Beni cial Akkermansia muciniphila failed to be changed by hesperidin. Those results implied that HFD did not enrich all harmful or inhibit all bene cial microbes and hesperidin did not enrich all the bene cial or inhibit all harmful microbes. This double role of hesperidin on both bene cial and harmful microbes may explain the unstable and inconsistent anti-obesity results in the animal and clinical tests.
To further study causal relationship between gut microbiota and disease, FMT (fecal microbiota transplantation) was usually conducted. Human fecal microbiota transplants from obese twins to germfree mice resulted in the increase of body fat, compared to those mice receiving FMT from lean twins, which proved that gut microbiota could be the cause of obesity. 37 FMT is also the way to study causal relationship between gut microbiota and effects of drugs, traditional herb medicines, bioactive chemicals and functional foods. A traditional Chinese medicine Ganoderma lucidum mycelium was reported to reduce body weight, in ammation and insulin resistance in HFD-fed mice. Fecal microbiota transplants from mice treated with water extracts of Ganoderma lucidum mycelium to HFD-fed mice transmitted the anti-obesity effects coincidentally, which proved that anti-obesity effects of Ganoderma lucidum mycelium was mediated by gut microbiota. 11 Our research also indicated that FMT transmitted donors' traits to the receptors, which provided one more evidence for gut microbiota mediated effects of bioactive chemicals and FMT as an effective therapy for diseases.
Interestingly, although FMT from healthy donors often bring about a good result to receipts, when comparing receipts gut microbiota before and after FMT and check the similarity to the donors' gut microbiota, it found that the receipts gut microbiota was obviously changed by FMT, however, the similarity of the receipts' gut microbiota after FMT to the donor's was not as high as expected, especially when the receipts were not germ-free subjects like human or SPF animals. 38 This may attribute to colonization ability of the microbes, some may easy to win out in the competition with intrinsical microbes or set down in the receipts' gut, some may not easy to win an ecological niche. The colonization ability of gut microbes has not been well studied and worth more research when exploring the mechanism of bene t effects of FMT therapy for virous health problems.
In our study, taking a close look to the speci c microbes that be transmitted to recipients by FMT, Lactobacillus salivarius, Staphylococcus sciuri, Desulfovibrio C21_c20, Mucispirillum schaedleri and Helicobacter hepaticus can be transmitted from donor to recipient mice while Helicobacter ganmani, Bi dobacterium pseudolongum and Akkermansia muciniphila can not. Akkermansia muciniphila was known to set down in the mucosa layer, Bi dobacterium pseudolongum also shown high adhesion to porcine colonic mucin. 39 It may be harder for them to replace the original microbes that occupied the mucus niche.
Since there are thousands of species in the gut microecosystem, scientists believe there are key players among the gut microbiota and they have been screening the driving species that contribute to the development of disease and bene cial effects of intervention. 40 It's a common goal for scientists in this eld to nd the key players and genetic or environment factors that regulate the key microbes. For example, Parabacteroides goldsteinii was found to be enriched by Hirsutella sinensis mycelium that produce anti-obesogenic and antidiabetic effects in obese mice. Oral treatment of obese mice with live P.
goldsteinii bacteria thus produce anti-obesogenic and antidiabetic effects. 12 So P. goldsteinii was the key microbe that contribute to bene cial effects of Hirsutella sinensis mycelium. Further on, impacts of the microbes on the host rely on mostly their metabolites. Lactobacillus, Bi dobacterium, Faecalibacterium prausnitzii are negatively correlated with cardiovascular disease and type 2 diabetes because they are SCFA-producing species. 41 42 When the strain Enterobacter was screened and isolated from a morbidly obese human and inoculated to germfree mice, it induced obesity and insulin resistance because it's an endotoxin-producing bacterium. 43 Those delighting ndings provide potential application of gut microbiota and their metabolites as novel biomarkers for disease diagnosis and new probiotics for disease therapy. The species that were screened to be signi cantly changed by HFD and hesperidin in the present study, especially Staphylococcus sciuri, Desulfovibrio C21_c20, Mucispirillum schaedleri, Helicobacter hepaticus and Helicobacter ganmani that have not been reported to be obese related in previous literatures need further veri cation by live strain supplementation. Their metabolites and molecular mechanisms hinted for obese need further exploration.

Conclusions
In conclusion, our results indicate that hesperidin reduced obesity, in ammation, improved gut integrity and modi ed a few gut microbiota species in HFD-fed mice (Fig. 9). The anti-obese effects and most hesperidin modi ed gut microbiota species were transmissible through horizontal fecal transplantation.
Our data demonstrate that hesperidin has a role to reduce body weight and reverse HFD-related disorders in HFD-fed mice by enriching some bene cial and inhibiting harmful microbes.

Murine.
Animal experiments were approved and performed in accordance with the guidelines of Laboratory animal center of Guangzhou Medical University. Eight-week-old male mice of the C57BL/6 were purchased from Guangdong medical laboratory animal center (GDMLAC) and kept under controlled temperature and light conditions (25℃,12h light-dark cycle), with free access to food and water. Mice were randomly distributed into eight groups containing six animals each. Mice were housed in groups of three animals per cage, and were fed with either a normal diet (13.5% of energy from fat; D12450; GDMLAC, China) or a high-fat diet (40% of energy from fat; D12451; GDMLAC, China). The formula of the diet was shown in Supplementary Table 1. Each group of mice was fed for 10 weeks with chow diet or HFD, with free access to either water or saturated hesperidin (Aladdin, CAS#520-26-3) solution at 0.1 or 0.2% (w/v).
Mice were supplemented every other day with sterile saline (vehicle), hesperidin (100, 200 mg/kg BW) by intragastric gavage from the fth week of feeding. Fecal microbiota transplantation (FMT) was started from the fth week of feeding. At the tenth week, animals were fasted for 12 hours before killing. Mice were deeply anaesthetized with 1% pentobarbital sodium (50 mg/kg BW) and whole blood was withdrawn through ventral aorta in tubes containing anticoagulant KEDTA. Visceral adipose tissues, epididymal white adipose tissues and the liver were removed and weighed. Colorectum were removed and its length was measured. Fecal in cecum was squeezed out. All samples were immersed in liquid nitrogen and stored at −80°C for further analysis.

Fecal microbiota transplantation.
Stools from donor mice of each diet group were collected under a laminar ow hood in sterile conditions and 100 mg was suspended in 3ml of sterile saline. The solution was vigorously mixed and centrifuged at 2000g for 3 min. The deposit was resuspended in 3ml of sterile saline and used as transplant material. Fresh transplant material was prepared on the same day of transplantation within 10 min before oral gavage (10ml/kg BW) to prevent changes in bacterial composition. Recipient mice were inoculated every other day with fresh transplant material by oral gavage for 6 weeks before being killed for subsequent analysis.

Caecal microbiota analysis
Caecal microbiota DNA was extracted using a Stool DNA Kit (Guangzhou IGE biotechnology, China) and applied to ampli cation of V3-V4 regions of 16S rDNA. Caecal microbiota composition was assessed using Illumina2500 sequencing of 16S rDNA amplicon and QIIME-based microbiota analysis. Highquality reads for bioinformatics analysis were selected and all of the effective reads from all samples were clustered into OTUs based on 99% sequence similarity according to Qiime Uclust. OTUs were annotated through RDP Classi er (Version 2.2), con dence cutoff 0.8 according to the GreenGene database, then composition and abundance information of each sample at different classi cation levels were statistically summarized.
Total RNA was isolated using UNIQ-10 column trizol total RNA isolation kit (Sangon Biotech, China).
Equal amounts of total RNA were used to synthesize cDNA with the PrimeScript TM RT reagent kit with gDNA Eraser (Cat# RR047A, TAKARA, Japan). qRT-PCR was performed in triplicate using TB Green TM premix Ex Taq TM (Cat# RR820A, TAKARA, Japan), 96-well plates and the 7500 Real-Time PCR System (Applied Biosystems). The Applied Biosysterm software (life technologies) was used for data analysis.
Relative quanti cation was done using the 2 -ΔΔc(t) method. Expression was normalized against the housekeeping gene β-actin. Mean expression levels of ND-fed mice were set as 100%. The primers used are shown in Supplementary Table 2.

Statistical analysis
Statistical analyses of data were performed using GraphPad Prism Version 7.00. Unless otherwise indicated, comparisons of two groups in which both groups passed a Shapiro-Wilke normality test were compared by two-tailed Student t-test. Those in which one or both groups did not pass a Shapiro Wilke normality test were compared by a nonparametric Mann-Whitney U-test.

Declarations
Ethics approval and consent to participate Not applicable Consent for publication Not applicable Availability of data and material Please contact author for data requests.

Figure 4
Hesperidin changes relative abundance of speci c intestinal microbial taxa. Phylogenetic tree created manually showing speci c changes in intestinal microbial community at different taxonomic levels caused by hesperidin supplementation to HFD mice. Nodes represent taxa, and the size of each node represents its relative abundance. The color red indicates an increase, blue represents a decrease and black means no change of relative abundance in HFD-Hes200 compared with HFD mice. The full color of the nodes indicates the statistically signi cant difference and the hollow nodes indicate the statistically non-signi cant difference by unpaired two-tailed student's t-test. See also additional Fig. S2. Figure 5 Body weight, fat accumulation and plasma lipid were reversed by fecal transplantation from hesperidintreated mice to HFD-fed mice. (A) ND-and HFD-fed mice were treated every other day with 10ul/g BW of either saline or faecal bacteria from donor mice at 2% (w/v) by intragastric gavage for 6 weeks (n=6 for each group).(B) Body weight gain (C) Liver weight (D) Epididymal fat (E) visceral fat (F) Plasma total cholesterol (G) Plasma triglyceride. Data are expressed as mean ± SEM. All differences were analysed using unpaired two-tailed student's t-test(*P<0.05, **P<0.01) .  (G) Relative mRNA expression levels of ZO-1 in colon. Data are expressed as mean ± SEM. All differences were analyzed using unpaired two-tailed student's t-test(n.s. not signi cant, *P<0.05, **P<0.01).

Figure 8
FMT changes relative abundance of speci c intestinal microbial taxa. Phylogenetic tree created manually showing speci c changes in intestinal microbial community at different taxonomic levels caused by FMT from ND hes200 to HFD mice. Nodes represent taxa, and the size of each node represents its relative abundance. The color red indicates an increase, blue represents a decrease and black means no change of relative abundance in HFD-Hes200 compared with HFD mice. The full color of the nodes indicates the statistically signi cant difference and the hollow nodes indicates the statistically non-signi cant difference by unpaired two-tailed student's t-test. See also additional Fig. S4. Figure 9