The diversity and abundance of the gut bacterial community in carp at the early developmental stage vary with the developmental process
In this study, common carp at the fertilized egg stage, yolk sac stage, and 1, 3, 14, 21, 28, 42 and 63 days after initial feeding were selected as samples (Fig. S1A-K). Genomic DNA was extracted from common carp, feed and pond water samples under different treatments (Table 1). The concentration and quality of genomic DNA in the samples was measured and met the requirements (Table S1). The final effective sequences were obtained by removing primers and linker sequences, bases with mass values less than 20 at both ends, sequences with lengths less than 200 bp, and chimeric sequences (Table S2). The length statistics of effective sequences showed that most of the effective sequences ranged from 430-470 bp, among which the number of sequences between 435-445 bp and 455-465 bp was the highest (Fig. S2A).
We classified the valid data as OTUs, and the number of OTUs varied greatly among different samples. The number of species in feed samples was the lowest, followed by fish samples, and the number of species in pond water samples was the highest (Table S3). The rarefaction curve was plotted by taking the number of effective sequences of samples as the abscissa and the OTU type classified by effective sequences as the ordinate (Fig. S2B). With the increase in the number of samples, the rarefaction curve of all samples tended to flatten off, indicating that the amount of sequencing data of samples was reasonable and could reflect the composition of bacteria in the samples. Based on the OTU analysis results, informatics analysis was conducted, and a rank– abundance plot (Fig. 1A) was constructed. The curve width of the feed group was narrower, and the declining trend was faster. The curve width of pond water samples was larger, and the declining trend was slower. The carp samples were intermediate between the feed group and the pond water group in terms of declining trend and curve width. The results showed that the species abundance of pond water samples was the highest, the species evenness of the feed group was the highest, and the carp samples were in the intermediate position. We also calculated alpha-diversity index statistics for different samples (Table S4) and constructed boxplots of Shannon index and Chao1 index differences among sample groups (Fig. 1B and Fig. 1C). The results showed that the diversity and abundance of the samples varied with the development process.
The gut microbiota of juvenile carp changed during development
To investigate the changes in the gut microbiota during the early development of common carp, we analyzed the taxonomic changes in the gut microbiota at the phylum, genus and species levels during development from fertilized eggs to 63 days after initial of feeding. At the taxonomic level of phyla (Fig. 2A), Proteobacteria, Cyanobacteria, Firmicutes and Actinobacteria occurred in all fish samples, but their relative abundances changed with the development process. For example, Proteobacteria and Firmicutes increased first and then decreased twice in a row with common carp development. The species with high abundances also changed at different developmental stages. For example, from Bacteroidetes and Proteobacteria at the CCE and CCY stages to Proteobacteria at CC1, CC3, CC14, from Proteobacteria and Cyanobacteria at CC21, to Proteobacteria and Fusobacteria at CC28. There were also some differences in the number of taxa in different developmental stages. The CC1 and CC3 stages had the most taxa (20 taxa each), the CC14 and CC28 stages had the fewest taxa (12 and 11 taxa, respectively), and the remaining stages had 13-15 taxa.
At the genus level (Fig. 2B), only Pseudomonas occurred in all fish samples, but the relative abundance of Pseudomonas varied greatly at different developmental stages, showing a first increasing and then decreasing trend two consecutive times. Excluding unclassified and ambiguous taxa, the relative abundances of most genera varied with developmental stage.
At the species level (Fig. 2C), there were different types of gut microbiota at different stages, with relatively few species at CCY, CC63 and CC28. Moreover, the species with high abundance varied in different developmental stages. For example, Sphingomonas sp. LYH-20 and Chroococcopsis gigantean SAG 12.99 occurred in the CCE stage. The species found in the CCY stage was Gyrodactylus salaris (gyrodactylosis fluke), and those found in the CC1 stage were Shewanella putrefaciens, Comamonas testosteroni, Comamonas aquatica, Oryza meyeriana and Exiguobacterium undae. The results showed that the composition of the gut microbiota changed with development.
The gut microbiota of juvenile common carp was more affected by host selection pressure
Next, the similarities and differences of the samples were analyzed. The heatmap analysis was conducted based on the 30 OTUs with the highest abundance (Fig. 3A). It was found that the similarity among pond water samples was high, and the similarity between CCE and CCY samples was high, and these samples clustered together with pond water samples. The similarity among feed samples was also high, and CC1 and feed samples were clustered together. The remaining fish samples were more similar and clustered together. Subsequently, we conducted UPGMA tree analysis (Fig. S3), weighted UniFrac distance matrix analysis (Fig.3B), principal coordinate analysis (PCoA) (Fig. 4A), and non-metric multidimensional scaling (NMDS) analysis (Fig. 4B) for the common carp-related samples. UPGMA tree analysis showed that the difference between the microbiota in the gut, pond water and feed was reflected in the fact that the common carp samples were clustered with pond water samples and then with feed samples. Except for 9 samples from CC28, CC42 and CC63, the common carp samples were clustered first within groups and then between groups. On the one hand, the differences between groups were greater than the differences within groups. On the other hand, the common carp samples were different from the pond water samples and feed samples, reflecting the difference between the carp samples and other nonbiological samples. The weighted UniFrac distance matrix heatmap (Fig.4C) showed that the differences between the common carp samples and other samples were significantly higher than the differences among the common carp samples, indicating that the differences between the common carp samples and other samples were more significant than the differences among different developmental stages. PCoA and NMDS analysis showed that pond water samples were clustered together, feed samples were clustered together, and carp samples were basically clustered together. However, the CCE, CCY and CC1 samples were more concentrated and closer to the pond water samples, reflecting the difference between the carp samples and environmental samples. The most likely reason for this difference was host selection pressure, which is consistent with previous reports [29].
The comparative analysis between the common carp samples and the pond water samples showed that the gut microbiota of larval and juvenile common carp was different from that of pond water (Table S5). Common carp samples, pond water samples and feed samples were further combined, and intergroup analysis of similarities (ANOSIM) showed that the composition of microbiota between common carp samples and pond water samples and between carp samples and feed samples were significantly different (Fig. 5A-C), indicating that the construction of the gut microbiota of fish was more influenced by the host during development. Thus, the microbiota structure was related to the host development stage, and environmental factors, including the pond water and feed, had less influence on the fish gut microbiota than the host. This result was consistent with the research of gibel carp [36]. To further explore the differences between groups of common carp samples at adjacent developmental stages, we analyzed the differences in the five bacterial genera with the largest differences between groups at two adjacent developmental stages (Fig. 6A-H).
Core microbiota in the intestinal tract of juvenile common carp during development
A petal diagram of common bacteria in all common carp samples were made (Fig. 7A). There was only one common OTU in all common carp samples, which was annotated to the class Gammaproteobacteria, but the abundance in different samples varied significantly (one-way ANOVA, P<0.0001). This bacterium played an important role in juvenile common carp, but it could not be identified to the species level, so it was difficult to determine its specific role in the development, nutrition and immunity of common carp. As the differences within groups were significantly smaller than the differences between groups, the results of the petal diagram of carp samples at each sampling time point (Fig. 7B) showed that there were 26 OTUs, among which 3 were annotated to species, 10 to genus, 11 to family, 1 to order, and 1 to class. The number of reads with common OTUs accounted for 49.1% of the total. Shewanella putrefaciens [37]was one of the common microorganisms with different abundances in samples at different times. CC1 and CC3 had the highest S. putrefaciens abundances. However, this bacterial species was not detected in feed samples, and only pond water sample W63 contained a very small amount of this bacterium, indicating that S. putrefaciens occurs in fish samples, and its relative abundance changed continuously during development. As S. putrefaciens was present as the core bacterium in this study, it must play an important role in the early development of juvenile common carp. The petals diagram of OTUs in the pond water samples (Fig. 7C) showed that there are 62 species of OTUs in all the pond water samples. The average number of reads contained in 62 OTUs accounted for 48.5% of the total reads, and there were some differences among different samples, ranging from 28.1% to 68.1%. There are 15 common OTUs annotated to species, 21 to genus, 20 to family, 4 to order, and 2 to class.