Turtle growth and morphology under different habitats
The mortality was negligible both in paddy fields and ponds during the experiment period. However, only a small number of turtles were caught from lake at 60d and no marked turtles were recaptured at 120d, resulting in incomplete statistics on mortality and growth for turtles in lake. We selected a small sample size for turtle resource protection purpose and the difficult in sample collection from nature lake, and the differences on growth, physiology and gut microbiota were distinct among different groups. The body weight of turtles in paddy fields and ponds were obviously higher than those in lake (p<0.05) and the divergence occurred in the early days. The growth rates of turtles were 0.76%/d, 0.68%/d and 0.40%/d for habitats of paddy fields, ponds and lake in first 60d. It was 0.72%/d and 0.62%/d for turtles in paddy fields and ponds during the whole 120d. The hepato-smatic index and clumpy fat index were highest in ponds, secondly in paddy fields and lowest in lake (p<0.05). The gut-smatic index on weight (DSIW) for turtles from pond was significantly higher than lake and paddy field (p<0.05). Inversely, the gut-smatic index on length (DSIL) was higher for turtles from paddy field and lake compared to pond. Measured values are presented as mean ± standard deviation, the different superscript letters in same row indicated significant difference(p<0.05)(Table 1).
There was no obvious trauma for most of turtles from lake except occasional leeches parasitic on calipash. Meanwhile, more bruises or scars were observed for the turtles from ponds than paddy fields. The appearance such as color of carapace and plastron were different among turtles in different habitats. The carapace of turtles cultured in ponds presented bottle green, but the individuals from paddy fields presented bottle green with slight golden yellow, which were similar with turtles from lake. There was no significant difference on main somatotype index (p>0.05), but the calipash lateral width was relative higher for turtles from ponds than paddy fields (Table 1).
Composition and diversity of turtle gut microbiota
The grouping details for samples from different habitats, cultured days and intestinal segment were listed in Table 2. The gut samples A total of 1 723 158 valid bacterial 16S rRNA gene reads were obtined and 4 901 OTUs were identified from all samples. The observed total OTUs varied in 64~822. The total number of OTUs was significantly less in initial groups IF and IL, and more in groups F1F and F1L from paddy fields at 60d. The number was 17~48 on OTUs more than 0.01% of total OTUs (Table S1). Significant differences were found in OTU composition among groups(Fig S1). Guts sampled at 120d had few unique OTUs, both in former and later part. The alpha diversity was calculated according to the composition and relative abundance of OTUs. Generally, the alpha diversity indices of microbes in later gut were higher than those in former gut. Besides, it was obviously lower in initial turtle guts from hothouse (p<0.05), whereas obviously higher in turtle gut sampled from paddy fields than ponds and lake (Fig.1). The species and number of OTUs varied significantly at 60d, different from that relatively harmonious at 120d. The microbial abundance was higher in samples from paddy fields than lake and ponds during the experiment. The microbial community presented relatively high similarity in guts sampled at same time. The PCA (Principal Component Analysis) showed high microbial community similarity in guts from the same individual or group, and significant discrepancy in samples from different habitats and sampling time (Fig.2). Generally, both sampling time and living habitats affected the variation of gut microbial communities.
Dominant microbes
The recognized microbes belonged to 27 phyla, 59 classes, 97 orders, 151 families, 219 genera from all the samples based on GreenGene. The phylum and genus level were emphasized in analysis. Bacteroidetes, Firmicutes, Fusobacteria and Proteobacteria were the most dominant phyla, accounting for more than 95% of the total bacteria in all samples. Firmicutes was the most abundant phylum in turtle gut sampled from hothouse at initial, while Proteobacteria was the most abundant phylum after cultivated in different habitats then followed by Bacteroidetes. Firmicutes and Fusobacteria commonly existed at 60d but rarely present at 120d in all the three habitats (Fig.3a). Additionally, the unidentified bacteria were more in lake compared to paddy fields and ponds.
There was significant difference on dominant genera among initial samples and subsequent samples from different habitats. The dominant genera in initial samples were an unclassified genus belonging to Bacteroidales, Romboutsia, Cetobacterium, Weissella, Lactococcus, Lactobacillus, Clostridium, Edwardsiella, Plesiomonas, and Sarcina. As for samples from the mentioned three habitats, the dominant genera were Cetobacterium, Chryseobacterium,Clostridium, Epulopiscium, Flavobacterium, Helicobacter, Pseudomonas, Stenotrophomonas and another unclassified genus belonging to Xanthomonadaceae. The abundance of dominant genera varied with different habitats, sampling time and gut location. For turtles sampled from paddy fields, the most dominant genus in foregut sampling at 60d was Clostridium and in hindgut was Cetobacterium, while at 120d, the most dominant genus was Stenotrophomonas both in former and later gut. For turtles sampled from pond, the most dominant genus at 60d was Flavobacterium and Cetobacterium in former and later gut, while at 120d, the most dominant genus was also Stenotrophomonas. For turtles sampled from lake, the most dominant genus at 60d was Flavobacterium and Cetobacterium in former and later gut, respectively (Fig.3b).
The dominant species in different gut location was also distinct. In foregut, the dominant species were Weissella cibaria, Enterococcus durans, Lactobacillus sakei, Lactococcus lactis, Lactococcus garvieae, Sarcina sp. and Pseudomonas sp., whereas in hindgut, Clostridium sensu stricto, Romboutsia sp., Weissella cibaria, Escherichia coli, Plesiomonas shigelloides, Edwardsiella tarda, Paeniclostridium sp., Cetobacterium sp., Terrisporobacter sp. and other two unclassified species belonging to Bacteroidales were abundant.
Microbial community in different habitats and sampling time
The microbial community was relatively complex at 60d, especially in former gut. At 60d, the species of microbes were significantly more in field, following by pond and lake. There were 140 common species (8.2%) in former gut from the three different habitats (Fig.4a), thereinto, Flavobacterium sp., Pseudomonas sp., Chryseobacterium sp. and two species belonging to Xanthomonadaceae were relatively abundant. Cetobacterium somerae was more abundant in paddy field than pond and lake. For later gut, there were 205 common species (8.1%) in the three different habitats (Fig.4b). Among these, one specie belonging to Bacteroidaceae was abundant in all habitats. Cetobacterium somerae, Epulopiscium sp., Pseudomonas sp., Stenotrophomonas sp. and Flavobacterium sp. were more abundant in paddy field and lake than in pond, while Clostridium sp. and Epulopiscium sp. were relatively abundant in pond. Meanwhile, Chryseobacterium sp., Parabacteroides sp., Sphingobacterium faecium, Clostridium perfringens, Pseudomonas sp., Bacteroides sp. and Pseudomonas sp. commonly existed in samples from lake and paddy field but did not appear in pond. At 120d, the specific microbes were more in pond (74%) than paddy field (33.4%) for former gut and the common species accounted for 18.6%, the specific microbes were more in paddy field (44%) than pond (34.4%) for later gut and the common species accounted for 26.1% (Fig.S2).
The LEfSe analysis was also conducted to identify representative microbes among various groups. For initial groups, representative genera were Weissella, Cetobacterium, Chryseobacterium, Epulopiscium, Escherichia, Flavobacterium, Lactococcus, Leuconostoc, Plesiomonas, Romboutsia, Sarcina and Stenotrophomonas. For groups cultured in different habitats, it showed F1L contained more species differed from other groups including Cetobacterium, Lactobacillaceae, Bacteroides, Parabacteroides, Plesiomonas and several species belonging to phyla Firmicutes presented higher LDA score. For F1F, representative genera were Sutterella, Bacteroides and Clostridiales. For samples from lake, Xanthomonadaceae and Pseudomonadales were representative genera, especially at 60d. The representative microbes in ponds were numerous and belonging to various phyla, especially the phyla Proteobacteria, also there were some unassigned species were indicated in pond (Fig.S3).
Functional predictions
The nearest sequenced taxon index (NSTI) were developed to quantify the availability of nearby genome representatives for groups (Table S2). Totally 41 predicted functional categories which represented 7 pathway maps in KEGG level 2 were indicated by PICRUSt. Cultural periods had significant effect on metabolism especially the amino acid and carbohydrate metabolism, membrane transport as well as replication and repair. At 60d, the functional microbiota related to amino acid and carbohydrate metabolism was distinct in lake compared to pond and paddy field (Fig.S4).