Mixed sexes healthy Goldfish (Carassius auratus), showing no clinical signs of disease and having a mean body length 10.4 ± 0.45 cm and weight 15.2 ± 0. 53 g were purchased from a local farm at Jiangsu province, China and maintained for 10 days in 60 L tanks containing aerated fresh water at a temperature of 28 ± 2.0°C, about 80% water in the aquariums was exchanged every week. The fish were fed with 3% basal diet per body weight, twice a day. This experiment was approved by the China Pharmaceutical University Animal Care and Use Committee.

The basal pelleted diet comprising 25% soybean meal, 15% fish meal, 20% rice bran, 5% corn starch, 30% rapeseed oil cake, 3.5% soybean oil and 1.5% vitamin and mineral mixture was prepared. The ingredients of the basal feed were analyzed according to the AOAC (Association of Official Analytical Chemists, 1997) method which revealed 2% moisture, 29.8% crude protein, 5.29% crude fat and 13.1% ash. For the preparation the diets with Bacillus sp. QSI-1, QSI-1 were grown in nutrient broth medium with aeration in a shaking incubator at 120 rpm at 37°C. After cultured for up to 80% spore formation, the cells were harvested by centrifuged at 4000 g to obtain pellet. The bacterial pellet was mixed with basal diet @ 5 × 10⁸ CFU/g. To achieve the well-distributed QSI-1 diet, the bacterial suspension was slowly added to feed power with gradual mixing in a laminar airflow bio-clean room. The mixed feed was oven-dried at 37°C for 3 h and stored at −20°C in sealed plastic ziploc™ bags until used. To ensure high QSI-1 level in the supplemented feed, fresh diets were prepared on weekly basis.

Experimental treatments were carried out with randomly taken goldfish in six plastic aquaria (50 L capacity) at a density of 20 fish per aquarium. The experiments were run in triplicate. Fish were fed with basal diet or QSI-1 diet and equal amount were provided at 08:00 and 19:00 h for 15 days. The QSI-1 amount in the experimental feeds was adjusted approximately 4.0 × 10⁸ cells per g fish day. Control group were fed with the dry basal pellet without QSI-1. At the same time, fish health was also recorded by visual observation of behavior.

MS-222 (tricaine methane-sulfonate, Sigma, USA) at the concentration of 100 mg L⁻¹ was used to anesthetize the fish. After dissection, the whole intestines of each fish were removed from the body under sterile conditions. Gut content samples were collected and immediately stored at −20°C. To isolate bacterial cells, the whole intestine from the fishes in the same group were put together and cut into small pieces and homogenized in ~10-fold dilution of sterile 1 × phosphate-buffered saline solution using autoclaved homogenizer, which made it possible to isolate most of the gut wall-attached microbes (Dawood et al., 2016).

Fish gut content were sampled in each group after feeding for 14 d to determinate the total intestinal microbiota, Bacillus spp. lactic acid bacteria (LAB), Escherichia coli and Aeromonas spp. The fish were taken from the rearing tank onto a clean anatomical plate. The fish surface was sterilized with 70% ethanol before opening the abdominal cavity with sterile scissor. Intestinal tract was picked up, weighed and washed with PBS thoroughly three times. Samples were serially diluted with PBS and 100 μl of the different dilutions was spread onto triplicate TSA (Trypto-Soya agar) media to determine total bacterial populations. DeMan, Rogosa, and Sharpe agar media (MRS), Bacillus medium base and Eosin-Methylene blue media (EMB) to detect viable LAB species, Bacillus spp. and E. coli and Aeromonas spp. medium with ampicillin supplement (Ryan) were also used to detect Aeromonas species, all medium were purchased from Qingdao Hope Bio-Technology Co., Ltd, China. The agar plate inoculated with each dilution was incubated for 3–5 days at 28°C. Colony forming units (CFU ml⁻¹) were determined for viable bacterial populations (Nikoskelainen et al., 2003).

The AHL reporter strains Chromobacterium violaceum CV026 and Agrobacterium tumefaciens AT 136 (pZLR4) strains were used to detect the presence of short-chain and medium- to long-chain AHLs. For AT136 bioassay, 20 μl (50 μg/ml) 5-bromo-4-chloro-3-indolyl-b-D-galactopyranoside (X-gal) were spread on the LB plate before bioassay was performed. The AHL reporter strains was streaked on the LB agar plate. Then the gut content sample was placed approximately 0.5 cm apart from the streak. The plates were incubated at 28°C for 24 h. If the bacteria in the sample can produce AHLs, the AHLs can diffuse through the agar and activate the coloration of the biosensor strains.

The whole bacterial DNA was extracted from samples using the QIAamp DNA Stool mini kit (Qiagen, Germany) according to the manufacturer's protocol. In addition, samples were treated in FastPrep Lysing Matrix E tubes (MP Biomedicals, USA) three times for 60 s in a bead-beater (Mini-Beadbeater 8 Bio-Spec Products, USA) with one intervening minute on ice. The DNA concentration was determined by a Nano-Drop 1000 spectrophotometer (Thermo Scientific Inc., Wilmington, DE, USA). Universal primers (341F 5′-CCTAYGGGRBGCASCAG-3′ and 806R 5′-GGACTACNNGGGTATCTAAT-3′) targeting the V3-V4 region of 16S rRNA gene were chosen for the amplification of the PCR products and analyze the taxonomic composition of the bacterial community (Baker et al., 2003). The progress of the PCR amplification consisted of an initial denaturation step at 95°C for 4 min, followed by 25 cycles, where 1 cycle consisted of 94°C for 30 s (denaturation), 56°C for 30 s (annealing), 72°C for 60 s (extension), and a final extension of 72°C for 10 min. The PCR products were visualized on agarose gels (2% in TBE buffer) containing ethidium bromide. The 16S rRNA gene was analyzed to evaluate the bacterial composition by using Illumina Miseq (Majorbio Bio-Pharm Technology Co., Ltd. in Shanghai, China).

Pyrosequencing paired end reads were assigned, truncated and merged using Pandaseq software (V 2.7) to generate raw tags. High quality clean tags from the raw tags were obtained and compared with reference database to detect and remove chimeric sequences using Usearch (vsesion 7.1 http://drive5.com/uparse/) to obtain effective tags. QIIME platform (http://qiime.org/scripts/assign_taxonomy.html) was used to analyze the effective tags, and tags with ≥ 97% similarity were assigned to the same operational taxonomic unit (OTU). The phylogenetic affiliation of each 16S rRNA gene sequence was analyzed using the RDP Classifier (version 2.2 http://sourceforge.net/projects/rdp-classifier/) against the Silva (SSU115) 16S rRNA database, with a confidence threshold of 70% (Amato et al., 2013). Principal component analysis (PCA) and weighted Fast UniFrac principal coordinate analysis (PCoA) based on OTUs were performed to provide an overview of gut microbial dynamics in response to with or without probiotic treatments. Analysis of Community richness (Chao1 and Ace), community diversity (Shannon, Simpson), Sequencing depth (Good's coverage) and Observed species; and beta diversity on both weighted and unweighted unifrac distance metrics of OTUs were calculated with the MOTHUR program (http://www.mothur.org) (Schloss et al., 2009).

Results were expressed as means and standard deviations. At least, three different measurements were accomplished for each mean. Statistical differences were determined using ANOVA, followed by Bonferroni's multiple comparison test. All statistical analyses were performed using Prism 6 (GraphPad Software, San Diego, CA, USA). A t-test was used to identify significant differences in relative abundance of bacterial taxa. P < 0.05 were considered significant.

The effect of QSI-1 on the intestinal bacteria was determined using the traditional culture-based techniques. Total intestinal culturable microbiota, LAB species, Bacillus spp. E. coli and Aeromonas spp. from goldfish intestines were counted on TSA, MRS, Bacillus spp. EMB, and Aeromonas spp. selective media, respectively. Figure 1 shows the total viable cell counts, LAB, Bacillus spp., E. coli and Aeromonas spp. counts with the fish fed with or/without QSI-1 supplementations. No significant differences were observed in total bacteria levels between 0d and 14d with the levels around log 8 CFU g⁻¹ (P > 0.05). LAB and Bacillus spp. levels were all significantly higher in the gut in the fish fed QSI-1supplemented diets (P < 0.05). But E. coli and Aeromonas spp. levels were all significantly lower in the gut after fed with QSI-1 for14 days (P < 0.05).

First, we determined whether QS signals are present in the environment of fish gut. We selected bacterial isolates from Aeromonas selective medium for its ability to produce AHLs by cross-streaking using biosensor strains. Violacein was induced in the C. violaceum CV026 AHL sensor and but did not stimulate the AHL sensor strain A. tumefaciens AT 136 (pZLR4) change to green, which responds to longer chain AHLs (Figure 2), it means the Aeromonas isolates can produce short-chain and medium AHLs but not long-chain AHLs.

Ethyl acetate extracts of AHLs from Aeromonas spp. isolates cultures retained the ability to induce reporter strains to produce pigment. AHLs extracts applied to TLC and visualized with the application of the C. violaceum CV026 sensor revealed the presence of two different AHLs in multiple extractions and chromatographs. On the basis of the mobility of the AHLs detected and comparison to synthetic standards, we tentatively identified these molecules as C4 HSL, and 3- C6 HSL (Figure 3).

To determine whether QSI-1 can influence the production of AHL signal molecules in fish gut, QSI-1 was administered orally to fish for 14 days, and then the AHL signal molecules in the fish gut content were detected by reporter strains. Figure 4 showed that after 14 days orally administered of QSI-1, short-chain AHLs cannot be detected by CV026, and the medium- to long-chain AHLs were also decreased.

High-throughput sequence analysis of bacterial 16S rRNA V3-V4 region was conducted on gut content samples. After filtering the low-quality reads, trimming the longer homopolymer runs, adapters, barcodes and primers, removing all cyanobacteria/chloroplast sequences and rarefying the datasets, the analysis revealed a highly diverse microbiota with a total of 760826 sequences (with an average length of 417.66 ± 6.5 bp), representing 316 OTUs, from 1702746 raw reads. All novel sequence data was deposited at NCBI's Sequence Read Archive under accession number SRP 093472. The relationships among samples was evaluated based on differences in phylogenetic diversity. Using the weighted-UniFrac metric, principal component analysis (PCA) and principal co-ordinates analysis (PCoA) plots were calculated from weighted UniFrac distances for the evaluation of the community composition, the results demonstrates a clear separation between each groups (Figures 5A,B), and suggesting that there is clear dissimilarity between the gut microbiota fed with or without QSI-1 supplemented diet. Figure 5C is the comparative presentation of the bacterial taxonomic analysis of intestinal microbiota composition between the groups with QSI-1 for 14 days and without QSI-1. At the phylum level, the bacterial communities in all treatments were dominated by three phyla: Bacteroidetes, Proteobacteria and Fusobacteria, but the higher relative abundance of Proteobacteria was observed after QSI-1 orally administrated. As shown in Table 1, differences were found in the diversity indices (Shannon and Simpson) and richness estimators (ACE and Chao) of the intestinal microbiota between the groups with QSI-1 for 14 days and without QSI-1.

A. hydrophila can produce two different short-chain AHLs, one is N-butanoyl-L-homoserine lactone (C4-HSL) and the other is N-hexanoyl-L-homoserine lactone (C6-HSL), so the level of A. hydrophila in fish gut was analyzed, we found that after administrated of QSI-1, the percentage of A. hydrophila in fish gut was decreased, and changed significantly at 7 and 14 days from 0.37 to 0.08 and 0.06% (Figure 6).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.