Annotation, sequencing depth and α-diversity analysis of effective sequences
In this study, 739979 sequences were obtained from 20 samples (Table S1), including 328124021bp of base pairs, with an average sequence length of 443.42 bp. Then, the quality filtering and further processing of the obtained reads were carried out, and a total of 1673 OTUs were identified, with significant differences in the number of OTUs in different samples (Table S2). In the annotation process, due to the limited amount of data in the database and the length of the sequencing fragments, many OTUs cannot be annotated to species, and some can only be annotated to genera, families, orders, classes, phyla, or even only to the bacterial communities. Rarefaction curves and Rank-abundance distribution curve were used to evaluate the sequencing depth and species richness of the samples. The results showed that the dilution curves of 20 samples tended to be flat with the increase of sample size, among which the curve of pond sediment was the highest, followed by pond water, and the other 18 intestinal samples were lower, but the curve height of different intestinal samples also had great differences (Fig. 1A). The distribution curve of pond sediment is relatively gentle and wide, while that of pond water is similar to other intestinal samples, with relatively narrow distribution curve width and fast decline (Fig. 1B). The α diversity index (Chao1, Ace, Shannon and Simpson) was further analyzed to estimate the richness and diversity of intestinal flora in different samples (Table 2), and the values of each index in different samples were different. Then, we constructed the distribution diagram of Shannon index among different samples (Fig 1C), and the results showed that the diversity index of pond bottom mud was the highest, followed by pond water, and then common carp, crucian carp and grass carp.
Analysis of microflora composition of samples
All samples were analyzed at the classification level of phylum, family and genus, and the results showed that there were certain differences in the types of flora in different samples. At the phylum level (Fig. 2), 15 different phyla were detected in 20 samples. Proteobacteria, Firmicutes and Fusobacteria were the dominant bacteria groups in the intestinal samples of three fish, but their relative contents varied according to different species. The content of Proteobacteria in the intestinal samples of carp was higher than that of crucian carp and grass carp, and in the intestinal samples of grass carp, there were significantly fewer categories with higher content than that of carp, while the sediment flora in the pond belonged to the most phylum. In addition to the high content of Cyanobacteria, Proteobacteria, Bacteroidetes, Chloroflexi, Verrucomicrobia, Acidobacteria and Firmicutes, there are still many kinds of phyla, and in addition to the high content of phyla and unannotated phyla, the content of others is relatively high. In pond water, Actinobacteria, Bacteroidetes, Proteobacteria and Cyanobacteria were the main bacterial phyla, and other phyla accounted for a small proportion.
At the level of family classification, pie charts of sample flora content were made (Fig. 3), it can be seen that the high content families in the intestinal samples of carp were different. In general, the contents of Streptococcus, Aeromonadaceae, Pseudomonadaceae, Flavobacteriaceae, Comamonadaceae, Mycoplasmataceae, Bacillus, Micrococcaceae, Moraxellaceae and Fusobacteriaceae are relatively high. In addition, the contents of Shewanellaceae, Xanthomonadaceae, Hados. Sed. Eubac and Brevinemataceae are also higher. The contents of Porphyromonadaceae, Chitinophagaceae, Erysipelotrichaceae, Vibrionaceae, CK-1C4-19 norank and Peptostreptococcaceae were higher in some samples. Compared with other samples, the pond sediment has the most species of families, among which the families with higher content are Anaerolineaceae, Chloroplast norank, Subsection III Family1, SB 1 norank, Hydrogenophilaceae, VadinHA17 norank, Aeromonadaceae, Flavobacteriaceae, Comamonadaceae and Chitinophagaceae. In pond water, the contents of Ichthyosporaceae, Flavobacteriaceae, Microbacteriaceae, Chitinophagaceae, Burkholderiaceae, Comamonadaceae and Cryomorphaceae were higher.
At the level of genus classification (Fig. 4A), Aeromonas, Lactococcus and Candidatus Bacilloplasma were high in the six samples of carp, but the relative contents of these genera were different in various samples. Aeromonas, Cetobacterium, Lactococcus and Pseudomonas were higher in crucian carp; the contents of Aeromonas, Cetobacterium, shewanella and Brecinema were higher in grass carp. The high content of Chloroplast norank, Planktothricoides, SB-1 norank and Thiobacter were found in pond sediment; and in pond water, the contents of Sporichthyaceae norank, Flavobacteria, Candidatus Limnoluna and hgcI clade were higher.
At species level (Fig. 4B), Lactococcus pisciumd, Pseudomonas reactans and Stenotrophomonas maltophilia K were higher in carp and crucian carp samples, while the bacterial species with higher content in grass carp were quite different from carp and crucian carp. The contents of Fusobacterium varium, uncultured anaerobic bacterium, Sebaldella termitidis ATCC 33386 and Lactococcus piscium were higher in grass carp. Neorickettsia risticii Potomac horse fever agent is an endemic species of carp samples, which does not exist in other fish samples and environmental samples. Sebaldella termitidis ATCC 33386, uncultured anaerobic bacterium, Anaerorhabdus furcosa, Desulfovibrio cuneatus, Streptococcus sp. M-S-TSA 29, Bacteroides sp, Cannel-Cat-fish9 and uncultured Verrucomicrobiaceae bacterium were relatively high in grass carp samples. These almost does not exist in other samples.
Analysis of differences between gut microbiota of carp and other samples
We also explored the similarities and differences between carp gut microbiota and other samples. Firstly, at the classification level of phyla (Table 3), Proteobacteria, Firmicutes, Bacteroidetes and Tenericutes were the most abundant phyla in carp. Proteobacteria, Firmicutes and Bacteroidetes were the dominated populations in crucian carp. The high content of grass carp groups were concentrated, which were Proteobacteria, Fusobacteria, Bacteroidetes, Firmicutes and Spirochaetae respectively. Compared with the bacterial groups in the pond sediment, the contents of Cyanobacteria, Chloroflexi, Verrucomicrobia and Acidobacteria in the gut of carp were all small or none. Cyprinus carpio, crucian carp and grass carp contain a variety of high content phyla, such as Proteobacteria, Firmicutes and Bacteroidetes.
Secondly, at the level of genus classification (Fig . 5A), the flora similarities of CC2 and CC3 were high, the dominant genera of the two groups were Aeromonas, Lactococcus, Candidatus Bacilloplasma and Cetobacterium. CI3 and CA2 were clustered with it, and the genus with higher content in CI3 were similar to the flora of CC2 and CC3. But the genus with higher content in CA2 were different from them, in addition to Aeromonas and Cetobacterium, there is Hados.Sed.Eubac.3 norank, CK-1C4-19 norank, and Shewanella species. The flora of CC1 and CA1 were similar, and their higher content species include Lactococcus, Pseudomonas, Arthrobacter and Aeromonas, etc., both clustered together with CA3, and in addition to the highest content of Cetobacterium, Lactococcus, Pseudomonas and Aeromonas were also richer in CA3. CI1 and CI2 clustered together, and the two had high microbiota similarities, with higher levels of species such as Aeromonas, Fusobacterium, Brevinema, Bacteroides, Dysgonomonas, and Sebaldella, not on the same secondary clade as the remaining 7 fish gut samples. Pond sediment and pond water clustered together and were less similar to fish gut samples. Finally, a cluster heat map with higher content OTUs was used, which was similar to the results at the genus classification level (Fig. S1).
Later, in order to more intuitively explore the degree of difference between fish intestinal samples and with the environment, one-way ANOVA analysis was performed on the phyla, genus and species with higher content in carp, crucian carp and grass carp samples. It was shown that there were no significant differences in phylum, genus and species classification (P>0.05) except for the phylum Spirochaetae, Comamonas (P=0.0495) and Comamonas testosteroni (P=0.0476). Secondly, carp, crucian carp and grass carp samples were compared with environmental samples (Table S3-S5) showed that there were significant differences or differences between fish samples and environmental samples with multiple phyla, genera and species with high content.
According to the principal component analysis (PCA) (Fig. 5B), the pond sediment and pond water samples almost overlapped, while the 9 fish samples broke the species boundary and were relatively scattered in the figure, with no samples able to gather together. Furthermore, Anosim analysis was used to compare the differences among gut microbiota of carp, crucian carp and grass carp (Fig. S2A-S2C). Compared with crucian carp, R value was 0.157 and P value was 0.092; compared with carp and grass carp, R value was 0.476 and P value was 0.002; compared with grass carp, crucian carp and grass carp, The R value is 0.626 and the P value is 0.002.
Analysis of core flora
Venn diagrams were constructed to identify the core flora in different groups of intestinal samples, carp, crucian carp and grass carp all had a certain number of core flora. There were 212 OTUs in three-tailed carp, and the number of OTUs unique to each fish was different. Among the total 212 OTUs, the reads number of OTUs was higher in the three-tailed fish, such as Aeromonas (not annotated to species), Lactococcus piscium, Lactococcus (not annotated to species), Arthrobacter species (species not annotated), Pseudomonas species (species not annotated), Pseudomonas reactans, Trichomonas testosterone and Stenotrophomonas maltophilia K279a (Fig. 6A). There were 103 core microbial species in crucian carp and 57 core microbial species in grass carp, which is less than crucian carp and carp (Fig. 6B-6C). There were 22 core bacteria common to carp and crucian carp and grass carp. Among the three-tailed carp, the reads of these 22 OTUs accounted for 54.5%, 67.2% and 60.6% of the total reads, which was higher than that of crucian carp (45.9%, 42.4% and 82.7%) and grass carp (40.7%, 15.8% and 86.1%, respectively). Excluding unclassified and uncultured bacterial species, the reads numbers of these 22 species accounted for 37.8%, 16.8% and 7.3% of the total reads in three-tailed carp, which was similar to the proportion of crucian carp (31.8%, 3.0% and 23.0%, respectively), but much higher than the proportion in grass carp (0.6%, 0.1% and 6.4%, respectively). At the taxonomic levels of family, genera and species, the relative content of OTUs in the 22 common samples was different, but the relative content of these OTUs in pond sediment and pond water was extremely low (Fig S3-S4).
Comparative analysis of intestinal contents and microbiota in intestinal mucus of carp
Next, we continued to explore the differences between fish intestinal tract contents and intestinal mucus flora. At the classification level of phyla, genus and species (Fig. 7,Fig. S5-S6), there were no significant differences in the bacterial species in the intestinal contents and intestinal mucus of common carp, and this trend was similar in crucian carp and grass carp. However, at the classification level of class (Fig. 8), the relative contents of Gammaproteobacteria in intestinal mucus and intestinal contents of carp were 28.82% and 46.74%, and the contents of Mollicutes were 1.42% and 8.53% respectively. The content of Clostridia had little change between the two groups, which was 4.63% in the contents and 8.37% in the mucus. The contents of Bacill are 32.23% and 16.51% respectively in the mucus and the contents, and the contents of Bacteroidia are 7.51% and 1.06% respectively.
It was later seen by comparing the intestinal contents of carp with the Shannon index of intestinal mucus (Fig. 9A), the diversity of contents of carp was lower than that in mucus, and the diversity index in mucus was more average. This trend was also present in crucian carp with the same diet, but in the intestines of grass carp there was a completely opposite trend. Further diversity analysis at the taxonomic level of family, genus and species showed (Fig. 10) (Fig. S7A-S7B) carp mucus and crucian carp mucus were clustered first, with high similarity. and the two were clustered with carp contents and then clustered with crucian carp contents; grass carp mucus and contents clustered together; pond water and pond sediment were clustered together separately, not on the same primary branch as fish intestinal samples. PCA analysis showed (Fig. 9B) carp mucus and crucian carp mucus samples were relatively small distances on the X axis and Y axis, indicating that in carp and crucian carp with the same diet, the composition of intestinal mucus and contents has more distinction. Finally the intestinal contents and intestinal mucus of carp, crucian carp and grass carp were analyzed between the Anosim groups and showed that the intestinal contents of carp and mucus (R=- 0.037, P=0.5), the intra-group differences between intestinal contents and mucus (R=- 0.074, P=0.6) and between intestinal contents and mucus of grass carp (R=- 0.296, P=0.8) were greater than the between-group differences, and the differences were not significant.