The environmental conditions in the GI tract varied in a location-specific manner
The realization of the variable nature of environmental factors in the GI tract prompted us to investigate the possibility of the location-specific environmental variations in the GI tract by using male SPF C57BL/6 mice. The whole GI tracts of mice of ten weeks’ old were divided into six parts (Figure S1), and the GI contents from each location as well as feces were collected and analyzed. As expected, the pHs and water contents of the GI contents were very different from each other in a location-specific manner along the GI tract and those of feces were not similar to the GI contents at any location (Table S1, S2), indicating heterogeneous environments along the GI tract. These results clearly showed that the environmental conditions in the GI tract vary reflecting the local function in the GI tract. The environmental condition of feces was not similar to those of any part of the GI contents, nor the overall GI content.
Metagenome sequencing unveiled the location-specific diversity of gut microbiome in the GI tract
We next investigated the diversity of gut microbiome at different locations within the same mouse by 16S rDNA-based metagenome analyses. The V3-V4 sites of the 16S rRNA genes of the isolated genomic DNAs of the gut microbiome of the GI contents were sequenced using the MiSeq™ platform (Illumina). The sequence reads containing incorrect primer, barcode sequences, sequences with more than one ambiguous base, low-quality sequences or chimeras were 2.2%, and these sequence reads were removed. The filtered 16S rDNA sequences were used to identify individual microbes by matching the 16S rDNA sequences with the SILVA reference (region V3-V4) database (https://www.arb-silva.de/). All of the identified 16S rDNA sequences were able to be classified into 13 different phyla; Bacteroidetes (51.5%), Firmicutes (35.88%), Proteobacteria (8.29%), Epsilonbacteraeota (1.26%), Cyanobacteria (0.94%), Actinobacteria (0.63%), Patescibacteria (0.5%), Deferribacteres (0.17%), Tenericutes (0.62%), Verrucomicrobia (0.08%), Planctomycetes (0.04%), Fusobacteria (0.03%), and Gemmatimonadetes (0.01%) (Figure. 1A). Interestingly, the abundance of the two most abundant groups of microbes was reversed from the stomach to feces along with the GI tract (Figure 1B), suggesting that microbial flora change reflecting the environmental change in the GI tract. The abundance of Firmicutes gradually decreased from the stomach to feces while the abundance of Bacteroidetes gradually increased (Figure 1B).
Alpha-diversity analysis showed that microbiomes in the different locations of the GI tract completely differed from each other
The gross microbiome analysis at the phylum level along the GI tract indicated that the microbiome was ever-changing along the GI tract reflecting their various environments. The microbiome in the GI tract was very different from the fecal microbiome (Figure 1A, 1B, Figure S2~S4), and the discrepancy depending on locations became more evident at lower taxonomic levels (Figure S5). Also, interestingly enough, the microbiome in the upper GI tracts and small intestines completely differed from those of the lower GI tracts within the same mouse and the degree of differences gradually decreased from the stomach to feces (Figure S5, Table S3~S7). It should be noted that the microbiome differences of large intestines among different mice were significantly decreased, demonstrating quite similar microbiome compositions of large intestine and feces among different mice. The microbiome analysis at the class level demonstrated that Bacteroidia was unanimously abundant along the GI tracts while most abundance was observed with Bacilli and Clostridia in the stomach, with Bacilli and Erysipelotrichia in the small intestine, and with Clostridia in the large intestine and feces (Table S4). Likewise, the GI tract was unanimously abundant with the order of Bacteroidales followed by Lactobacillales and Clostridiales in the stomach, Lactobacillales and Erysipelotrichales in the small intestine, and Clostridiales in the large intestine and feces, respectively (Table S5). At the family levels, Muribaculaceae was unanimously abundant followed by Lactobacillaceae and Lachnospiraceae in the stomach, Lactobacillales in the small intestine, and Lachnospiraceae and Ruminococcaceae in the large intestine and feces, respectively (Table S6). At the genus levels, there was a clearly distinguished pattern along with the GI locations despite the presence of unidentified groups (Table S7). Helicobacter was in the stomach as well as large intestine but not in the small intestine. Lactococcus, Dubosiella, Parasutterella, and Turicibacter were specifically observed in the small intestine while Helicobacter, Bacteroides, Alloprevotella, Odoribacter, and Alistipes in the large intestine and feces (Table S7).
Our initial comparison of the microbiome compositions at locations along the GI tracts was followed by a thorough diversity analysis of the microbiome. The taxonomic α-diversities measured by ACE richness, Shannon diversity, and Fisher’s alpha all indicated that the diversities of gut microbiome at colon and cecum were much higher than that of feces (Figure 2, Table S8). The diversities of microbiome compositions were lowest in jejunum and ileum while highest in cecum and colon. It should be noted that the diversities of the fecal microbiome were lower than those of the cecum and colon. Clearly, the α-diversity analyses indicated that the fecal microbiome did not represent the microbiome in the GI tract of its host, contrary to the general baseline assumption.
Beta-diversity analysis confirmed that microbiomes in different locations of the GI tract completely differed from each other
The discrepancy of the composition of gut microbiome along the GI tract became more evident with the ratio analysis between location and local species (Figure 3, Figure S6-S8). To compare the diversities of the microbiomes at different locations, β analysis method was applied. The NMDS plots based on Bray-Curtis distances showed that the microbiome compositions were very diverse in location-specific manners in all of the tested three mice and that, more significantly, the fecal microbiome did not represent the microbiome of the GI tracts (R2=0.49, P=0.003 ADONIS) (Figure 3A, Figure S6A). We transformed the OTUs of each microbiome into principal components using an unweighted UniFrac metric for Principal coordinates analysis (PCoA). Eigenvalues of each microbiome in different locations of the GI tracts were very different from each other (Figure 3B, Figure S6B, Figure S6C). PCoA confirmed again that the fecal microbiome communities in all of the tested three mice did not represent any part of the microbiome communities in the guts. Other ordination plot methods also clearly confirmed our result (Figure S7). The correlation analysis of OTU values with respect to the locations of the GI tract by drawing a heat map of the top-ranked OTUs defined at the bray curtis distance level revealed that feces had a distinguished microbial profile compared with any locations of the GI tracts (Figure 3C, Figure S8). Unsupervised hierarchical clustering clearly partitioned the samples into two distinguished groups, and this pattern was observed repeatedly over a wide range of phylogenetic levels (Figure S8).
Alpha-diversity analysis showed that the fecal microbiome did not represent the microbiome of the whole GI tract
Location-specific analysis on microbiome clearly indicated that microbiome in the GI tract varied on its location under varied physical and chemical environments, and that fecal microbiome might not represent the actual microbiome in the GI tract (Figure 1~2, Figure S2~S8). To investigate that possibility, we directly compared fecal microbiome compositions with the microbiome composition of the whole GI tracts in each mouse. As expected, the gross microbiome analyses revealed that the microbiome composition of the GI tracts was clearly different from the the composition of the fecal microbiome (Figure 4A, B, Figure S9A, S9B). The microbiome discrepancy between feces and the GI tract became more evident at lower taxonomic levels (Figure 4B, Figure S9C). The most abundant microbial families in the GI tracts were Muribaculaceae, Lactobacillaceae, Lachnospiraceae Ruminococcaceae, and Erysipelotrichaceae in the decreasing order while Muribaculaceae, Ruminococcaceae, Lachnospiraceae, and Prevotellaceae were in fecal microbiomes (Figure 4B). At the genus level, Lactobacillus, Lactococcus, Dubosiella, and Turicibacter were highly represented in the GI tract but not in feces (Figure S9D and Table S7).
After noting the microbiome discrepancy between feces and the GI tract by direct comparison, thorough taxonomic α-diversity analyses were performed. The taxonomic α-diversities measured by ACE richness, Shannon diversity, and Fisher’s alpha all indicated that the diversities of the microbiome of the GI tracts were much higher than the fecal microbiome (Figure 4C). Also, the fecal microbiome did not represent the microbiome of the GI tracts in all of the three mice. Shannon diversity (p <0 .05), ACE richness (p < 0.01), and Fisher’s alpha (p <0.01) concluded that the microbiome of the GI tract was statistically different from the fecal microbiome and the fecal microbiome does not represent the microbiome of GI tract. j
Beta-diversity analysis confirmed that fecal microbiome did not represent the microbiome of the whole GI tract
Comparative analysis on the microbiome of feces and the GI tracts by β diversity analyses (community structure: R2 = 0.1, p <0 .05 ADONIS) further solidified that the fecal microbiome did not represent the microbiome in the GI tract of its host. Both NMDS and RDA plots showed that the fecal microbiome was distinctly different from the microbiome of the GI tracts in all of the tested mice (Figure 5A, 5B). Interestingly, the microbial community of the fecal microbiome was closer to each other in individual mice than to that of the GI microbiome within the same mice. The distinct difference of microbiome compositions between feces and the GI tract within a mouse became more evident with a correlation analysis of total OTUs with respect to feces and the GI tract. The heat map of all OTUs defined at the Bray-Curtis distance level revealed that the fecal microbiome was completely different from that of the GI tracts as demonstrated in a distinguished pattern of microbial profile among feces and also among the GI tracts in all tested mice rather than between the microbiome compositions of feces and the GI tract within same mice (Figure 5C).