Sequence evaluation and assessment of microbial diversity
For all four samples, a total of 1446597 raw sequences were collected, with 1378928 good quality sequences recovered after quality and length filtering. Following that, the sequences were divided into 974 OTUs, with 339, 254, 328, and 225 OTUs covering the MCA, MTA, MCP, and MTP, respectively. Unclassified bacterial OTUs account for 5.6 % and 8.26 % of total OTUs in MCA and MTA, respectively; while 8.54 % and 8.0 % of bacterial OTUs remain unclassified in MCP and MTP, respectively.Rarefaction curves for all four samples revealed that the number of observed species initially increased with sequencing depth, and then plateaued, resembling general rarefaction curves tapering off at the end. The saturation plateau on the rarefaction curves indicated that the amount of sequencing was sufficient for further analysis (Fig. S1).
The microbial community in C. mrigala's foregut and hindgut is more abundant and diverse under control (DO2 = 7±0.5ppm) conditions. In response to low dissolved oxygen concentrations, the anterior gut shows a 3.6 times (Chao1 value =33) and posterior gut shows a 2.9 times (Chao1 value =40) drop in gut bacterial population abundance (Chao1 value =120 or 114 in the foregut or hindgutof the control)(Fig. 1). The presence of changes in the bacterial community in C. mrigala anterior and posterior gut samples was revealed by beta diversity analysis (Fig. 2 a).In both control and experimental conditions, gut samples have clearly distinct microbial compositions. The Venn diagram depicts sharing of 107 OTUsby all four samples (Fig. 2 b). Control and experimental anterior gut samples shared 10 more OTUs with these 107 OTUs, and control and experimental posterior gut samples shared 22 more OTUs. In the control sample, the anterior guts of C.mrigala contained 143 unique OTUs, but only 47 in the experimental sample. The posterior gut of control fish, on the other hand, contained 112 unique OTUs, whereas the experimental sample contained 20 unique taxa.
Differences in gut microbial community among control and experimental samples
In the study of actual abundance, nine phyla were found to be significant contributors to gut bacterial populations. Proteobacteria, Bacteroidetes, Firmicutes, Fusobacteria, and Actinobacteria were the most common phyla found in all four intestinal samples. According to a phylum-based study, when the dissolved oxygen content dropped from 7±0.5ppm to 0.35±0.05 ppm, C. mrigala had a lower population of Proteobacteria, Bacteroidetes, Firmicutes, and Actinobacteria in both the anterior and posterior guts(Fig. 3). Surprisingly, the population of Fusobacteria increases in both the anterior and posterior guts under the experimentalconditions described above. Dendrogram analyses revealed that the anterior gut microbiota of control C. mrigala is more similar to that of its posterior gut microbiota (Fig. S2).
Other bacterial classes, besides Fusobacteria, Bacteroidia, Erysipelotrichia, and Cyanobacteria, showed a decrease in bacterial abundance during experimental conditions compared to control conditions in both gut areas, according to a class-based analysis (Fig. 4). Furthermore, only MTAdid Chloroplast and Mollicutes show a small increase in bacterial population abundance. MTA and MTP had significantly lower concentrations of Betaproteobacteria, Gammaproteobacteria, Flavobacteriia, and Alphaproteobacteria than MCA and MCP control samples. The population of Sphingobacteria, Cytophagia, and Deinococci was reduced to nil under dissolved oxygen stress conditions. Verrucomicrobiae were only found in experimental conditions. Under experimental conditions, the abundance of Clostridia and Bacilli decreases in both guts, while Erysipelotrichia increases.
OTUs were distributed in 32 different orders in MCA, with Flavobacteriales and Burkholderiales dominating at the top, followed by Fusobacteriales, Clostridiales, Enterobacteriales, and Actinomycetales, Neisseriales, and Pseudomonadale (all with the same number of OTUs); in the experimental anterior gut (MTA), Fusobacteriales were dominant, followed by Erysipelotrich (in descending order). OTUs in MTP, on the other hand, were distributed in 30 different orders, with Flavobacteriales dominating the list followed by Burkholderiales, Fusobacteriales, Actinomycetales, Pseudomonadales, Clostridiales, and Bacillales at the bottom, while Enterobacteriales, Bacteroidales, Actinomycetales, Erysipelotrichales, and Clostridiales dominated(Fig. S3).
Flavobacteriaceae and Fusobacteriaceae were almost equally prevalent among OTUs distributed across 49 families in the MCA sample, whereas Flavobacteriaceae was most prevalent among OTUs classified across 51 families in the MCP sample. Fusobacteriaceae was found to be the most prevalent family among OTUs sorted from MTA and MTP sequence reads, respectively, for 25 and 27 families.
As shown in Fig. 5, the abundance of the gut microbial community was significantly influenced by the amount of dissolved oxygen in the water. Both the anterior and posterior guts had a low abundance of microbial diversity under dissolved oxygen stress (0.35±0.05 ppm). All other genera, with the exception of Cetobacterium, Pseudomonas, Aeromonas, and Staphylococcus, vanished completely in the experimental situations, as shown in Fig. 5. Mycobacteriumand Turibacter were found in both the anterior and posterior guts of C. mrigala during the DO2-stress experiment. Cetobacterium was the most abundant genus in both the anterior and posterior guts of C.mrigala, with a dramatic increase in abundance under DO2 stress conditions. In addition, the relative abundance of Aeromonasin MTA was higher than that of MCA.
Community profiling, Clustering and correlation of control vs. experimental gut microbiota
According to the microbiome investigation of C. mrigala, less availability of DO2 concentration in water altered bacterial groups constituting key contributors of core intestinal microbiota during control conditions. DO2 stress caused the main contributing groups to shift, as well as the appearance and disappearance of a few microbial populations in the core intestinal microbiome (Fig. 6).
At the family level, heatmap analysis of all four samples was performed to reflect microbiological complexity and measure similarity or dissimilarity among and among samples. The heatmap pattern shows that bacterial diversity and relative abundance patterns for C. mrigalagut samples (both anterior and posterior) diverged significantly under normal and stressed DO2concentrations (Fig. 7). In terms of the overall heatmap abundance pattern, the anterior and posterior gut samples of C. mrigala (MCA and MCP) do not closely resemble the gut sample under DO2 stress (MTA and MTP).The bacterial complexity of the control fish's anterior gut (MCA) resembled that of the fish's posterior gut (MCP). Similarly, the bacterial complexity of the anterior gut (MTA) of the DO2 stressed fish was more similar to that of the posterior gut (MTP). Furthermore, many of the bacterial species found in MCA and MCP were reduced to zero in MTA and MTP (under DO2 stress conditions). According to a pattern study in MicrobiomeAnalyst, Erysipelotrichaceae has a positive relationship with Mycobacteriaceae, Enterobacteriaceae (Plesiomonas), Fusobacteriaceae, and Aeromonadaceae, but a negative relationship with other family members (Fig. 8). Similarly, the genus Cetobacterium had a positive relationship with Turicibacter and Mycobacterium but a negative correlation with the remaining genera.