Small intestinal microbiota composition altered in obesity-T2DM mice with high salt fed

Obesity has become a global concern because of increasing the risk of many diseases. Alterations in human gut microbiota have been proven to be associated with obesity, yet the mechanism of how the microbiota are altered by high salt diet (HSD) remains obscure. In this study, the changes of Small Intestinal Microbiota (SIM) in obesity-T2DM mice were investigated. High-throughput sequencing was applied for the jejunum microbiota analysis. Results revealed that high salt intake (HS) could suppress the body weight (B.W.) in some extent. In addition, significant T2DM pathological features were revealed in high salt-high fat diet (HS-HFD) group, despite of relatively lower food intake. High-throughput sequencing analysis indicated that the F/B ratio in HS intake groups increased significantly (P < 0.001), whereas beneficial bacteria, such as lactic acid or short chain fatty acid producing bacteria, were significantly decreased in HS-HFD group (P < 0.01 or P < 0.05). Furthermore, Halorubrum luteum were observed in small intestine for the first time. Above results preliminary suggested that in obesity-T2DM mice, high dietary salt could aggravate the imbalance of composition of SIM to unhealthy direction.

www.nature.com/scientificreports/ alterations of the bacterial population in small bowel and intestinal abnormalities in mouse or rat [9][10][11] . Previous investigations illustrated a correlation between SIM balance with diseases, dietary nutrition uptake, and function regulations 12,13 . However, the relationship between HSD and SIM alteration in obesity individuals remains to be unclear 14,15 .
In this study, a high salt-high fat diet feeding mouse model and a high-throughput sequencing technology were conducted to investigate the changes of microbial flora in the proximal small intestine of the mice. Moreover, the connection between high dietary salt, SIM, and metabolic syndrome were tentatively discussed based on obtained results.

Results
Body weight (B.W.), food and water intake. Before feeding experimental diets, average B.W. of experimental animals were 23 ± 1.75 g. After 8 weeks feeding, the B.W. of four different groups were remarkably different. The most obvious increases were observed in HFD group from 23 to 44 g in 8 weeks (2.63 g/week). The B.W of HS-HFD group was increased from 23 to 34 g (1.76 g/week). Moreover, the B.W of HS-ND and ND groups raised slightly from 23 g to about 28 g (Fig. 1A). Significant statistical differences were revealed between both HFD and HS-HFD groups and their paired ND groups (P < 0.001), while no remarkable statistical differences were appeared between HSD-ND and ND groups (P = 0.62, Fig. 1B).
For food intake, significantly less food intake was observed in HS-HFD group than HFD group (P < 0.05), while a same dietary intake was examined in HSD-ND and ND groups (Fig. 1C). Accordingly, the energy intake for HS-HFD group was less than HFD group (P < 0.05, Fig. 1D). In addition, HS-ND group has relatively higher water intake than paired ND and HF-HSD groups (P < 0.05), and almost same amount of water intakes were measured in HFD and HS-HFD groups (Fig. 1E).

Fasting blood glucose (FBG), glucose tolerance (GTT), insulin tolerance (ITT), and serum insulin level.
Comparatively higher FBG levels were investigated in both HFD and HS-HFD groups than their correspondent ND groups (P < 0.001, and P < 0.05), and HFD samples had a higher FBG than HSD group (P < 0.05, Fig. 1F). Glucose tolerance tests (GTT) showed that GTT in both HFD and HS-HFD groups were evidently higher than that of two ND groups (P < 0.001), and the GTT of HFD group was relatively higher than that of HS-HFD group (P < 0.001, Fig. 1G,H). For insulin tolerance, differences were not obvious among the groups (Fig. 1I,J), however, the serum insulin concentrations were significantly increased in HFD group (P < 0.001, or P < 0.01, Fig. 1K). ND (n = 10), HFD (n = 9), HS-ND (n = 10), HS-HFD (n = 10). *, # P < 0.05, **P < 0.01, ***P < 0.001. Panel (E) was measured by Student's t test followed by unpaired t test. Other statistics results were measured by one-way ANOVA followed by Tukey's multiple comparison test. www.nature.com/scientificreports/ Triglyceride (TG), high density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C) concentration in liver tissue. It was observed that lipid profile of HFD group were greatly impacted by HFD. In addition, the TC value was significantly increased in HFD group compared with ND (P < 0.05), and HS-HFD groups (P > 0.05, Fig. 2A). HDL level of HFD group were significantly lower than ND group (P < 0.05, Fig. 2B). LDL level were significantly higher in HFD group than ND group, and also higher than HS-HFD group (P < 0.01, P < 0.05, Fig. 2C). It is notable that there was no significant difference between HS-ND and HS-HFD group mice TG, HDL, LDL level in liver.
Mouse jejunum microbiota analysis. The diversity, richness, and composition of jejunum microorganisms were analyzed in normal and obesity-T2DM mice, respectively. Total of 501,322 effective 16S rRNA raw reads were obtained from mice jejunum microflora samples. Effective tags, optimized sequence, other parameters were shown in Table 1. Obtained results were further confirmed with species accumulation curves analysis (Fig. 3A), and Venn diagrams showed the interrelationship of OTUs in jejunum samples among the different groups (Fig. 3B). Analysis indicated that the sequencing and analyzing methodologies in this study were appropriate and acceptable. Total of 277 OTUs in all samples were observed. Among them, 57 OTUs were common for all four groups, and some characteristic OTUs were obtained, such as 4 OTUs were unique for ND group, 9 were found in HFD group only, 7 were special in HS-ND group, and 3 were particular in HS-HFD group, respectively. As shown in Table 1 and Fig. 4, dynamic changes of jejunum bacterial diversity were evaluated for each different group. Alpha diversity was been applied to express richness, diversity, and evenness, respectively. Richness indexes Ace and Chao indicated that total OTU numbers of HS-ND were significantly higher than that of ND group (P < 0.05, Fig. 4A,B). Sobs, observable OTUs number, has also been used to illustrate species richness. Results showed that HS-ND group had significantly higher richness than ND group (P < 0.01). In addition, PD-whole tree measurement indicated that the diversity level of OTUs in HS-ND was higher than ND group (Fig. 4D). No remarkable difference was observed in Shannon and Simpson diversity analyses (Fig. 4E,F). Above results illustrated that high salt intake increased richness of the microbiomes.
Principal coordinate analysis (PCoA) was conducted with Bray-Curtis distance normalization method, which have been used to illustrate the distribution of different samples based on the bacterial community structure  www.nature.com/scientificreports/ ( Fig. 5A). ANOSIM of Weighted UniFrac distances (Fig. 5B) revealed a significant difference in the bacterial communities among the groups (R = 0.608, P = 0.004). These results indicated that the diversity of bacterial community within the same group were less obvious than that of different groups, which elucidated that bacterial community structures were variable enough among different groups.

Microorganism community composition in the mouse small intestines.
To investigate the effects of high dietary salt on SIM composition of obesity-T2DM mice, this work compared the jejunum microorganisms in different groups. At the phylum level, microorganisms mainly belonged to five phyla, which were  . PD-whole-tree was Faith's phylogenetic diversity, reflects the difference in the preservation of the evolutionary history of the species in the sample. n = 3/group, # P < 0.05, ## P < 0.01. All statistics results were measured by Student's t test followed by unpaired t test. www.nature.com/scientificreports/ Firmicutes, Bacteroidetes, Proteobacteria, Verrucomicrobia, and Actinobacteria, respectively. Firmicutes was the most dominated phylum in all groups with total occupation of 64%, 88%, 45%, and 51.6% in SIM of ND, HFD, HS-ND, and HS-HFD groups, respectively (Fig. 6A). The ratio in HFD was significantly higher than that in ND group (P < 0.05). According to Fig. 6B, Bacteroidetes was dramatically decreased in both HFD groups (total as 5% and 0.2%, respectively) compared to two ND groups (22% and 13%, respectively) and for HS-HFD vs. HS-ND, P < 0.01. The F/B ratio increased greatly in HS-HFD group than other groups (P < 0.001) (Fig. 6C). It was worth to mention that another known obesity-related major phylum, Proteobacteria, increased in both HSgroups, while no obvious changes were reveled in HFD group (Fig. 6D). Figure 6E-O showed the changes in class, order, family, genus levels of bacteria, respectively. The lactic acid producing bacteria, Bacilli (Fig. 6E) and Lactobacillales, (Fig. 6I) markedly increased in ND than HFD and HS-ND groups, which consist with previous reports (Ferguson et al. 38 ). Furthermore, HS treatment increased Rickettsiales and Alphaproteobacteria significantly between HS-ND and ND groups (Fig. 6G,K), and same trend for HS-HFD vs. HFD was observed as well. The relative abundance at species level was assessed as well (Fig. 7A). Among them, two HS groups had relatively lower abundance of Lactobacillus reuteri and Clostridium leptum, and there was a statistical difference only between HS-ND and ND groups (P < 0.05, Fig. 7B,C). Clostridium leptum, an obligate anaerobe, appeared significantly less in HS-ND than in ND group (Fig. 7C). In addition, within HS groups, Clostridiales bacterium in HS-HFD less than in HS-ND group (P < 0.05, Fig. 7D). In the human infection associated bacteria, Cupriavidus pauculus, analysis, HS-HFD group showed significantly increased value compared to HFD and HS-ND groups (P < 0.01, Fig. 7E).

Discussion
Body weight (B.W.), food and water intake. The association between high salt intake with increased risk of obesity and other metabolic syndromes is still a controversial topic in related fields 16 . In the guts, microorganism community has been playing important roles in regulating several metabolic syndromes, such as obesity and diabetes 17,18 . In this work, obesity-T2DM mouse model was constructed to investigate the effect of high dietary salt on the small intestine microbiota to understand more on the relationship among high-salt, SIM and the metabolic diseases.
Analysis indicated that high salt intake could restrict the B.W. Through comprehensive comparing among intake patterns of food, water and energy, it was found that the increasing degree of B.W for HS-ND and HS-HFD were relatively lower than that of ND and HFD groups. Possible reasons for this phenomenon were discussed based on the results of published works. Some previous studies pointed that HFD or HFD-HS could increase the B.W. of animal [19][20][21] . Meanwhile, several research results also revealed that high salt diet suppresses B.W. no matter fed with ND or HFD 22 . Several workers have tried to explain the effect of high salt on B.W. from reninangiotensin system 22 , uncoupling protein 1 (UCP-1) and L-thyroxine mechanisms 23,24 . The affecting reason could also be analyzed from other points. It was preliminary considered that high dietary salt might affect the appetite on different extent in different groups. Moreover, the affect was more remarkable in HS-HFD group. Nevertheless, the reason could be more complicated for different groups and samples 25 .   www.nature.com/scientificreports/ Fasting blood glucose (FBG), glucose tolerance (GTT), insulin tolerance (ITT), and serum insulin level. In HS-HFD group, the B.W., food, and energy intake were much lower than HFD group, and however, this group still showed clear T2DM pathological characteristics compared to HS-ND group. It was reported that NaCl could affect glucose absorbing in intestinal lumen, and might improve glucose tolerance 26 . HFD mice presented remarkably severer glucose tolerance than HS-HFD mice, and relatively higher serum insulin level than HS-HFD groups. Previous works also suggested that the obesity contributed much more than high dietary salt for the T2DM development 27 .

Triglyceride (TG), high density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C) evaluation in liver.
It is well known that HFD treatment disturbs serum lipid profile with higher TG, LDL-c and lower HDL-c concentration 28 .TG, HDL, LDL result revealed in the present study similar with many previous reports. However, HS-HFD group has showed higher TG level with no significance compared to HS-ND group, and similar level both HDL, LDL with HS-ND group. previous reports have concluded that TG, LD markedly increased in high fat and high salt SD rat 29 . HDL, LDL result of the present study revealed as relatively different with some previous studies, and further studies on this point will be conducted in our future works.
Mouse jejunum microbiota analysis. In this work, the density of the bacteria in jejunum was 1.96 × 10 5 ± 1.3 × 10 5 to 2.53 × 10 6 ± 1.46 × 10 5 copies/μg·extracted DNA. The bacterial communities in different regions of human intestine exist in various types, and they are crucial for human health 30 . Contrast to the fecal microorganisms represented large intestine microbiota, studies were inadequate for small intestine symbionts due to the comparably lower accessibility, lower abundance, and anaerobic environment 31 . Regardless of the large surface area of its mucosa, it was believed that the SIM had lower-density compared to large intestine microbiota. This was mainly for the existence of bactericidal substances, such as bile acids, and the rapid luminal flow rate 30 . It was reported that the density of the bacteria in the region of SI was about 10 3 in duodenum, 10 4 -10 5 in jejunum, and 10 7 -10 8 in ileum 18 .

Microorganism community composition in the mouse small intestines. The effect of SIM on ani-
mal health still remains to be explored. Moreover, the effect of HS on the SIM in individuals suffering T2DM is also unclear 32 . Combined with earlier studies 33,34 , it was assumed that HS intake could increase the richness yet decrease the diversity of SIM. In addition, HS intake might cause the microbiota compositions switched into an unhealthy direction. The increased value of F/B ratio, which was mainly from the two most predominant phyla, Firmicutes and Bacteroidetes, in healthy individuals 34,35 , has been considered as a risk factor for obesity, hypertension, and metabolic disorders 36,37 . In this study, HS consumption decreased both Firmicutes and Bacteroidetes, while HFD increased Firmicutes and decreased Bacteroidetes as in large intestines. It was worth to mention that HS-HFD reduced Bacteroidetes in jejunum dramatically, and this change caused a significant increase of F/B ratio (P < 0.001). Results indicated a trend that high dietary salt could decrease Firmicutes in SIM, which was opposite to the change of Firmicutes in large intestines 37,38 . Proteobacteria, other diseases related bacteria, was elevated by HS in the jejunum, which been identified as a possible marker for predisposing to disease onset 39 , such as autism, IBD, CVDs, and T2DM. Results revealed that in obesity-T2DM mice, unhealthy microbiota composition had been aggravated by high dietary salt.
Those bacteria, which were affected or changed significantly in HS-HFD group (Bacteroidia, Coriobacteriia, Alphaproteobacteria, Coriobacteriales, Coriobacteriaceae, & Enterorhabdus), were all known to be crucial bacteria for fat-intestinal microbiota-diseases. It was noteworthy that compared with the HFD and ND groups, the HS treatment aggravated the deterioration caused by HFD in small intestines. To the best of our knowledge, this result was reported for the first time. Rickettsiales belongs to Alphaproteobacteria class, and Alphaproteobacteria belongs to phylum Proteobacteria. It was assumed that these subgroups contributed more on the changes of the Proteobacteria phylum by HS treatment 40 . Further studies were expected to explore the changing types and mechanisms on Proteobacteria phylum of these subgroups.
It was observed that HS intake led to decrease of Lactobacillus species in intestines, which was consistent with published work 41 . Published works suggested that probiotics have beneficial effects on disorders including reduced insulin resistance, hepatic steatosis 42 , and prevention of obesity 43 . Non-pathogenic Clostridia could construct an anti-inflammatory environment by regulating T cells' activity and secreting fatty acid. Onal et al. reported that HS-HFD significantly decreased the Clostridia bacteria strains, and, relative reduction of Clostridium leptum related with inflammatory bowel disease 36 . In this study, HS decreased C. leptum significantly in jejunum, which might induce a derangement on anti-inflammatory system of SIM.
In obesity-T2DM mice, HS significantly increased the Gram-negative bacterium, depleted the Bacteroidetes, elevated F/B ratio, reduced the beneficial gut bacteria (such as Lactobacillus sp., Clostridium sp., and increased the pathogenic bacteria Rickettsiales). Jejunum is an important segment of the small intestines. For dwelling of micro-symbionts, jejunum contains relatively more bacteria than duodenum. It supports the bacterial colonization, decides the bacterial diversity and density, especially the growth of health-related Gram-positive aerobics and facultative anaerobics like Lactobacilli 22,32 . Considering the lower abundance of microorganisms in small intestine, HS intake could induce a more rapid and profound imbalance in SIM in an unhealthy animal than that in healthy ones. However, the effect and mechanisms of SIM imbalance on the bacterial colonization, dynamics, richness and diversity in large intestine are still a challenging topic in related research fields.
Effect of HS treatment on archaea and halophilic microorganisms in jejunum were also investigated. The results showed that there were Euryarchaeota and Thaumarchaeota phyla, and a species Halorubrum luteum www.nature.com/scientificreports/ found in SIM. However, no obvious differences were observed among groups. Euryarchaeota was considered the predominate archaea in intestine and widely distributed in colon 44 . Halorubrum luteum was a halophilic archaeon belong to Halorubrum genus 45 . Currently, some new species were detected in fecal and colonic contents, and classified to be a member of Halorubrum genus 46 . To our best of knowledge, Halorubrum luteum were observed in small intestine for the first time.
In summary, this work was conducted to study the behavior and changes of microbiota in the small intestine affected by HS-HFD in obesity-T2DM animals. The results were concluded as follows. Firstly, the B.W could be restricted by HS, and this trend was more obvious in HS-HFD group. Secondly, significant T2DM pathological characteristics were observed in HS-HFD group, regardless of relatively lower food and energy intake, and this probably from the influence of NaCI on glucose absorbing. Thirdly, jujunum microbiota analysis indicated that SIM had relatively lower density compared with large intestine microbiota. Finally, microorganism community composition study suggested that HS intake could increase the richness and yet decrease the diversity of SIM. Moreover, high dietary salt could aggravate the SIM composition imbalance, and further lead to an unhealthy direction. Additionally, some limitations still existed in this study, such as the sample size was relatively small, and the experiments needed to be designed more comprehensively. Future studies will focus on deeper researches related to the connections between high salt diet and obesity, and more other related disorders from the prospective of microbiota alteration in the small intestine.

Materials and methods
Animal ethics. All experimental animal procedures were approved by the Ethics Committee for Animal Research, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences (protocol number: SIAT-IRB-170401-YGS-RPG-A0312-01). All experimental methods were carried out in accordance with relevant guidelines and regulations. All methods are reported in accordance with ARRIVE guidelines.
Experimental protocol. In this study, 6-weeks old specific pathogen-free (SPF) level 40 male C57BL/6 J mice were obtained from Hunan SJA Laboratory Animal Co., Ltd. Mice were acclimated to housing conditions (22 °C, 55% humidity, 12 h: 12 h light: dark cycle) in SPF level animal center, Peking University Shenzhen Graduate School with ad libitum access to water and diets for 1 week before formal study. After the adaptation period, the mice were divided randomly into 4 diet groups: Fed normal diet (ND, n = 10), Fed high fat diet (HFD, contains 45% fat; D12451; Research Diets, New Brunswick, New Jersey, n = 10), Fed normal diet and high salt drinking (HS-ND, 2% NaCl, n = 10) and Fed high fat diet and high salt drinking (HS-HFD, 45% fat, 2% NaCl, n = 10) groups 47 , respectively (Supplementary Table S1) for 8 weeks 4 . The filtered to sterile drinking water with or without salt was supplemented with 0.01% aspartame (Nantong Changhai Food Additive Co. LTD, China) to give a better palatability. Each group were treated for 8 weeks. B.W., food intake and water intake were recorded weekly; energy intake was calculated approximately according to the net amount of food intake and the physiological fuel values of the ingredients. FBG, GTT, and ITT. FBG level, GTT, and ITT were applied to evaluate the murine obesity-T2DM model.
Mice tail snip blood samples were used for FBG measuring (blood glucose strips, and Accu-Check glucometer, Roche, Switzerland). The animals were undergone fasting for 5 h with free water before the FBG test. For GTT, the mice were pre-fasted for 16 h, and then glucose solution (2 g/kg·B.W.) was injected intraperitoneally (i.p.). Blood glucose was measured at the following different time points, 0, 15, 30, 60, 90, and 120 min (min), respectively. The ITT was performed after 4 h fasting, i.p. injected insulin (0.75 U/kg, Humulin, Lilly, Country), and the blood glucose measured at the indicated time points as in GTT. www.nature.com/scientificreports/ High-throughput sequencing. Microbiota analysis was performed by Tiny Gene Bio-Tech Co., Ltd.

Scientific
(Shanghai, China). The genomic DNA of the jejunum bacteria was extracted. Specific primers (see Supplementary Table S2) were used to amplify the V3-V4 region of bacterial and archaea 16S rRNA gene, and then each target fragment was recovered as template for secondary round PCR amplification. After that, PCR products were recovered (AxyPrep DNA gel recovery kit, AXYGEN, USA, CB55766755), and quantified with a FTC-3000™ real-time PCR instrument (Funglyn Maple ridge, Canada). The samples were mixed according to the equal mole ratio to complete library preparation before sequencing. The 2 × 300 bp based library was high-throughput sequenced (Illumina Miseq, Illumina, California) and carried out bioinformatics analyses. The effective sequences obtained were controlled by Trimmomatic software (v.0.35), and sequences splicing by FLASH software (v.1.2.11, Adobe, USA). MOTHUR software v.1.33.3 was used for the quality of reads after merge. Using UPARSE (usearch, version v8.1.1756), the representative sequences of full name Operational Taxonomic Units (OTU) were obtained by clustering with 97% similarity. The obtained OTU representative sequences were then compared with the progressive object of the database by MOTHUR (classify. seqs) for species annotation; OTUs without annotation were filtered out. Finally, the OTUs were clustered based on sequence similarity, taxonomic analysis, alpha diversity, beta diversity and inter-group difference analysis. Statistical analysis. All data were presented as the mean ± standard deviation. All in vivo experimental data statistical differences among four groups and microbiota results were analyzed with one-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test, or Student's t-test. The P values less than 0.05 were considered significant statistically. Marked as *P < 0.05, **P < 0.01, ***P < 0.001 for ANOVA, and # P < 0.05, ## P < 0.01 for t test, respectively.

Data availability
The datasets presented in this study can be found in NCBI with the accession number PRJNA864077.