**3. Discussion**

Seaweed polysaccharides have various biological activities [11,13,21–25] and also show potential to be developed as prebiotics that promote the healthy growth of gu<sup>t</sup> bacteria. Previous studies have compared the similarities and differences between the cecal and fecal microbiota of animals and human volunteers [6–8]. However, the regulatory effects of polysaccharides extracted from *S. fusiforme* (a common and widely eaten seaweed) on the cecal and fecal microbiota of HFD-fed mice, which would benefit the study of seaweed polysaccharide-based gu<sup>t</sup> microbiota regulators, have not been compared. Here, the impacts of 16 weeks of water and acid extracted *S. fusiforme* polysaccharides (SfW and SfA) on the cecal and fecal microbiota of HFD-fed mice were investigated. We found that the HFD significantly altered the dominant phyla Bacteroidetes and Actinobacteria, and the dominant genera *Coriobacteriaceae*, *S24-7*, and *Ruminococcus*, but did not affect the abundance of Firmicutes, *Clostridiales*, *Oscillospira*, and *Ruminococcaceae* in cecal microbiota and the Simpson's index of fecal microbiota. Co-treatments with SfW and SfA partially reversed the dysbiosis of Firmicutes, *S24-7*, *Ruminococcus*, *Clostridiales*, and *Ruminococcaceae*. The administration of SfW and SfA also altered the abundance of some genes encoding

carbohydrate-metabolizing enzymes between cecal and fecal microbiota in 16-week HFDfed mice and demonstrated that cecal microbiota was more significantly regulated by *S. fusiforme* polysaccharides.

Our recent study reported that five *S*. *fusiforme* polysaccharides prepared through hot-water and acid extraction showed a wide range of molecular weight (10–698.3 kDa), and the HCl-extracted polysaccharide (Sf-A) mainly consisted of glucose, fucose, and galactose [16]. Cheng et al. prepared a 205.8 kDa *S*. *fusiforme* polysaccharide (SFF) by acid extraction and the polysaccharide was mainly composed of fucose and galactose [17]. The chemical structures of SfW and SfA were more similar to Sf-A than SFF, which may be because they are sourced from the same supplier.

HFD has been demonstrated to adversely affect gu<sup>t</sup> microbiota composition through increasing the abundance of Firmicutes and Proteobacteria and decreasing Bacteroidetes [26–29]. The present study demonstrated that 16 weeks of HFD significantly increased the relative abundance of Actinobacteria and decreased the abundance of Bacteroidetes and Verrucomicrobia in both the cecal and fecal microbiota but only enriched Firmicutes and Proteobacteria in fecal microbiota. The discrepancy between the previous studies with our findings indicates that HFD consistently decreases the abundance of Bacteroidetes in both the cecal and fecal microbiota, but the impacts on other phyla depend on the sampling position and the specific conditions of the animal models. Our recent study demonstrated that four weeks of HFD feeding significantly increased the relative abundance of *Coriobacteriaceae* and *Oscillospira* in cecal microbiota [11], but in the present study, the abundance of *Oscillospira* in cecal microbiota was not enriched by HFD, which may be explained by the prolonged HFD treatment. These findings further deepen our understanding about the impact of HFD on gu<sup>t</sup> microbiota composition.

It is well known that microbiota composition varies in different parts of the gastrointestinal tract, and there are significant differences in the quantity and quality of microorganisms in the cecum contents and feces [8]. According to a recent report, the cecal and fecal microbiota exhibits different taxonomic structures, functional activities, and metabolic pathways [30]. Guo et al. also reported that the abundance of Firmicutes in the cecum of mice fed a normal diet was much higher than that in the feces, and *S24-7* was much lower in the cecal contents, which was highly consistent with the present study [9]. Here, we comprehensively compared the similarities and differences between the cecal and fecal microbiota in 16-week HFD-fed mice, including the abundance of genes encoding carbohydrate-metabolizing enzymes in cecal and fecal microbiota. HFD showed a more significant influence on the α-diversity of cecal microbiota than that of fecal microbiota, suggesting that cecal microbiota may be more suitable to represent the gu<sup>t</sup> microbiota composition. Some bacteria (Actinobacteria and *Coriobacteriaceae*) were significantly enriched in the cecal contents, while others (Firmicutes, *Oscillospira*, and *Ruminococcaceae*) were enriched in the fecal samples, which may be associated with the different function of the cecum and colon. For example, *Coriobacteriaceae*, a family within the phylum Actinobacteria, are strictly anaerobic bacteria and contribute to the metabolism of bile salts, steroids, and dietary polyphenols [31]. Ariangela et al. reported that the TNBS colitis severity was most closely correlated with the composition of colonic mucus microbiome, but not fecal or cecal microbiome [32], suggesting that the choice of sampling site depends on the experimental design. Interestingly, SfA was mainly composed of glucose, fucose, and galactose, and administration of SfA significantly decreased the abundance of genes encoding α-galactosidase and β-glucosidase, which may be ascribed to the interactions between the polysaccharide and gu<sup>t</sup> microbiota.

Previous studies reported the effects of several *S. fusiforme* polysaccharides on gu<sup>t</sup> microbiota composition [11,17,33]. Chen et al. reported that 12 months of oral administration of an *S. fusiforme* polysaccharide decreased the abundance of the phyla Firmicutes, Proteobacteria, and the genera *Lactobacillus* and *Helicobacter* in small intestinal microbiota [33]. Cheng and colleagues found that 6 weeks of oral administration of an *S. fusiforme* polysaccharide significantly decreased the relative abundances of several diabetes-related microbiota in fecal samples, including *Bilophila*, *Oscillibacter*, and *Mucispirillum* [17]. Our previous study reported that 4 weeks of *S. fusiforme* polysaccharide treatment significantly altered the relative abundance of the phylum Bacteroidetes, and the genera *Oscillospira*, *Mucispirillum*, and *Clostridiales* in cecal microbiota [11]. In the present study, 16 weeks of SfW administration partially reversed HFD-induced alterations of *Clostridiales* and *Ruminococcaceae* in both the cecal and fecal microbiota. The differences in the regulatory effects of *S. fusiforme* polysaccharides on gu<sup>t</sup> micobiota may be explained by the structural difference of polysaccharides, or difference in the animal model, and sampling position. For example, the microbiota composition in the intestinal contents from the ICR mice was investigated by Chen et al., while in the present study, we determined the microbiota composition in fecal and cecal samples from the ICR mice.

#### **4. Materials and Methods**
