Characterizing the bacterial community in bioocs, culture water and intestines of GIFT tilapia (Oreochromis niloticus) reared in a biooc technology (BFT) system

GIFT Oreochromis niloticus was cultured in a bio-oc technology (BFT) system to identify the oral structure of the biooc water and in the intestine of the sh. High-throughput sequencing technology was used to detect the microbes in the culture water, in the intestines and gills of farmed sh. The effective sequence coverage rate of each sample exceeded 99% and the sequencing results were true and reliable. The results show that according to Chao1 and Simpson indices, the microbial richness and diversity of ocs increase with the increase of culture time. Flocs initially favoured the thick-walled bacterium ( Firmicutes ) as the dominant fungi with abundance of 86%, whiles the rest were all less than 10% in abundance. At the end of the experiment the abundance of Chloroexi was 43% , Proteobacteria 26%, and Bacteroidetes 12%. During the early stages, Bacillus was the most dominant group of about 86% abundance. However, this declined towards the end of the breeding period and the abundance was recorded to reduce to about 3%. The highest abundance was associated with the bacterial described as unclassied species called N orank-f-caldilineaceae representing 40% and the abundance of Flavobacterium found to be 2%. The dominant bacteria in the intestinal tract of sh were Proteobacteria (35%), Bacillus (16%) and Cetobacterium (13%). The main heterotrophic assimilation functions in the BFT system were bacillus and chlorocampylobacter . However, avobacterium was found in the BFT system indicating the presence of on-going denitrication


Introduction
In aquaculture practice, farmers attempt to achieve higher yields by adopting high-density culture of sh and through the utilization of high-protein feeds. In addition, aquaculturist also promote nitrogen utilization in aquatic animal feed which is often maintained in the range of 25-35% in an effort to increase production ( The main methods for controlling ammonia nitrogen in bio oc technology system include; ammonia assimilation by heterotrophic bacteria and nitri cation by autotrophic bacteria (Hargreaves, J A. 2006;Schryver, P D et al. 2008). Bio ocs Technology Aquaculture Systems (BFTs) control ammonia nitrogen through the process of maintaining the appropriate C/N to promote heterotrophic assimilation of ammonia nitrogen. In general, the C/N in BFTs is high, which is bene cial to the growth of heterotrophic bacteria (Mirzakhani Neda et al. 2019) and the water characterized by higher suspended matter. BFT controls ammonia nitrogen through bacteria suspended in water and bacteria in ocs. Therefore, bacteria are an indispensable component of BFTs. Some of these are responsible for water puri cation through the removal of these toxic substances and provide suitable conditions for farmed animals (Burford et al. 1975). In addition, the changes in the intestinal ora structure caused by the type of biological occulation, carbon sources, breeding objects, and the environment can all affect the distribution of microbial communities in biological ocs Anand PSS et al. 2014;Yang Zhangwu et al. 2015).
The oral structure of the cultured water has a signi cant effect on the intestinal ora of cultured animals, which in turn affects their growth and welfare (Panigrahi et al. 2019). Yale et al. (2019) showed that the structure of the intestinal ora of P. vannamei is similar to that of the BFTs in which it is thriving. GIFT tilapia is a common breed for BFTs. Tilapia can directly feed on bio ocs in BFTs. Our previous study investigated the growth, digestive activity, welfare, cost-effectiveness, water quality dynamics (Luo et al., 2017) and nitrogen budget of tilapia in BFTs, has also been described in our previous paper (Cao et al., 2019). This study focused on characterizing the bacterial community in BFT water and bio ocs as well as the bacterial community in sh intestines reared in the BFT environment. The ndings of this study will contribute to increasing the knowledge-base and understanding of bacterial communities in closed aquaculture systems and as well help improve the operational design of zero-exchange aquaculture systems and thus increase their sustainability.

Cultured system
The experiment was conducted in Shanghai Ocean University, China. Most of the time during the experiment, the air temperature was less than 10 °C. The experiment used three large water tanks and two large air pumps for aeration. The operating volume of each water tank was 2 cubic meters. A 100 L bucket was used to settle the excess oc in each bucket and the supernatant from the bucket was returned to the BFT tank.

Experimental design 2.3.1 Start-up of BFT
All the tanks and water treatments were cleaned prior to lling with dechlorinated tap water. Commercial aquatic feed contained ≥ 48% crude protein, ≤ 17% crude ash and ≥ 4% crude lipid. The initial TSS was 200 mg/L. Glucose was added to increase the input C/N to 20:1. When the concentrations of NH 3 -N,NO 2 − -N and NO 3 − -N were stable less than 0.33 ± 0.04, 0.06 ± 0.03 and 40.32 ± 13.03 mg/L, respectively. It was therefore considered that the start-up period of BFT was completed (Hargreaves, 2013). The start up period lasted for 45 days.

Stocking and management
Tilapia ngerling with initial mean weight 3.54 ± 2.82 g (n = 30) were stocked at a density of 5.12 ± 0.04 kg/m³ per tank. In order to give time for the tilapia to adapt to the environment, the water temperature was maintained about 24 ℃ for 20 days. Then the temperature decreased slowly and the water temperature was maintained in the range of 18 to 20 ℃. By the time the air temperature increased up to 20 ℃, the experiment was over. This culture lasted for a period of 64 days. The water was continuously aerated to maintain the DO within the range of 6-7.5 mg/L in each tank. Considering the high TSS contained, more aeration was supplied. The pH in all systems was maintained in the range of 7.0 to 7.5 by adding NaHCO 3 . In total 11.34 kg of NaHCO 3 was added into BFT to supply the consumption of alkalinity and maintain pH. PCL was used as the additional carbonate for BFT tanks as describe by Luo et al. (2018). Whenever the TSS level exceeded 600 mg/L in any BFT tank, some of the water containing ocs was taken out and put into the settling container. After settling for 1 h, the supernatant was returned to the tank and the bio ocs were collected and recorded.
Fish were fed to apparent satiation three times a day (9:00, 14:00 and 19:00). Only dechlorinated tap water was added to each tank every day to compensate for water lost through evaporation and removal of redundant ocs.

Water quality parameters and air temperature
Water temperature (℃), DO and pH were measured daily using a WTW meter (WTW Multi 3430 SETF, German). NH 3 -N, NO 2 − -N, NO 3 − -N, total nitrogen (TN), total phosphate (TP), TSS were analyzed every four days according to standard methods (SEPA 2002). DOC was analyzed every four days using a Total Organic Carbon Analyzer (Multi N/C 2100, Analytikjenamulti, German). Air temperature was measured daily. Seven days were selected randomly to determine water temperature every 1 h from 6:00 to 24:00.

PCR ampli cation and high-throughput sequencing
All samples taken for the determination of bacterial community diversity were sent to Shanghai Meiji Biomedical Technology Co., Ltd. for PCR ampli cation and high-throughput sequencing. All samples in this experiment were PCR ampli ed targetting the V3-V4 region of 16SrRNA as a universal primer. The primer used were 338F (5'-ACTCCTACGGGAGGCAGCAG-3') and 806R (5'-GGACTACHVGGGTWTCTCATAT-3'). The reaction conditions of PCR were: denaturation at 95℃ for 3 min; denaturation at 95℃ for 30 s; annealing at 55℃ for 30 s; extension at 72℃ for 45 s for 27 cycles, and then extension at 72℃ for 10 min. The PCR ampli cation results were checked by 2% agarose gel electrophoresis, and the ampli ed samples were subsequently subjected to high-throughput sequencing.
Sequences used for these analysis were deposited in GenBank accessions. The raw data were submitted to the NCBI Raw DataBFTe under the SRA accession ID PRJNA529947.

Data processing
The i-sanger cloud platform (www.i-sanger.com) was used to analyze and process the obtained data. The repeats and single sequences in the optimized sequence were removed, and the non-repeated sequences obtained were OTU clustered according to 97% similarity. Chimeras were removed to obtain the OTU representative sequence. The RDP (ribosomal dataBFTe project) classi er Bayesian algorithm was used to perform a taxonomic analysis of 97% similar OTU representative sequences and the taxonomic information represented by each OTU obtained. Analysis of Chao1 and simpson indices were done.
Chao1 index represents the richness of the sample microbial community. The larger the value is, the richer it is. On the other hand, Simpson index represents the diversity of the sample microbial community. The larger the value, the lower the diversity. The calculation formula is as follows: In the formula: SChao1 is the Chao1 index, S obs is the actual number of OTUs, n1 has only one sequence of OTUs, and n2 is the number of OTUs with only two sequences.
D Simpson in the formula is the Simpson index, S obs is the actual number of OTUs obtained, n i is the number of sequences contained in the i-th OTU, and N is the number of all sequences.
In order to study the similarity and difference of microbial community composition between different samples, QIIME software was used to calculate the beat diversity distance matrix, and perform hierarchical clustering to construct a sample hierarchical clustering tree.
The water quality data obtained from this study was analysed using SPSS software (IBM Co.LtD version 20). Signi cant differences in the water quality parameters were determined using ANOVA. The analysed were presented in graphs.

Dissolved nitrogen syndrome and DOC/TN in the BFTs
Ammonia nitrogen and nitrite were controlled at lower levels, with average concentrations of 1.56 ± 0.76 mg/L, 1.58 ± 0.56 mg/L, and 0.47 ± 0.29 mg/L, 0.09 ± 0.04 mg/L, and nitrate accumulation reached 336.91 mg/L. Due to the added carbon source, the DOC/TAN of the BFT group was above 20 as shown in Fig. 1. The growth performance of tilapia and the detailed changes in other parameters using a similar design has been demonstrated in our previous study (Cao et al., 2019).

Diversity and richness of bacteria community
Samples of BFT water, oc, and sh intestines in BFT were used to analyze the ora structure. The bacterial community structures of the bio ocs were determined by Illumina high-throughput sequencing for 16S rRNA genes. The indices used to describe the richness and diversity of bacteria communities are given in Fig. 2. Chao 1 richness estimator and Shannon Index were used to describe the bacterial community richness and diversity respectively (Pei et al., 2012). With the increase of culture time, microbial richness in ocs; that is both the degree and diversity increased and the microbial diversity increased signi cantly. Its Simpson index decreased from 0.6449 to 00889.

Microbial community distribution in the BFT system
A total of 27 phylums, 320 genus, and 673 OTUs were found in the oc in the BFT system and the number of OTUs shared in the early and late breeding period was 212, as shown in Fig. 3.
In terms of phylum level, as shown in Fig. 4, the abundance of Firmicutes was 86% at the beginning of the culture, which was an absolute advantage, but was only 3% at the end. The abundance of Chloro exi which became the dominant bacteria at the end of breeding period was 43% but intially was only 3% at the beginning. Proteobacteria increased from the initial 4-26% and Bacteroides increased from the initial 4-12% at the end. The dominance of Actinobacteria did not change signi cantly as it only 3% at the beginning and 2% at the end.
At the genus level, as shown in Fig. 4, Bacillus was the main species in the BFT tank during the initial stage of cultivation reaching 86%, while the abundance of Bacillus spp decreased to 3% at the end of the cultivation. Also, Norank-f-Caldilineaceae was 40% at the end of the study. Initially it was Cetobacterium genus that was detected with 8% dominance at the end of the expeiment. The abundance of Flavobacterium was 2%, the others were Norank-c-SBR2076 which had not been identi ed to the genus level. At the end, Norank-f-Saprospiraceae was also 2% and the similarity in the lter was similar.
Reyranella and Norank-f-Cytophagaceae also made up 2% of the BFT at the end of the experiment.

Microbial community distribution in the sh intestines
In the intestinal tract of tilapia cultured in the BFT system, 15 phylums, 16 genera and 684 OTUs of bacteria were found.
During the entire breeding process, the intestinal bacterial community of the sh changed dramatically and as there were noticeable shifts in the dominance of bacterial phylum. The dominance of the Proteobacteria initially, was only 4%, which increased to 20% at the end. Campylobacter phyllophyta increased from 7-16% and the Actinobacteria group increased from 12-20%. Firmicutes increased from the original 15-20%. It is worth noting that the dominance of the Chlamydia phylum in the intestines of both sh species at the end of the culture was 0.26%, which was later increased to 1%.
At the genus level, Cetobacterium, which initially accounted for 62%, was only 13% at the end of the experiment. Bacillus, which was not detected at the beginning, increased to 16% at the end. The dominance of Microbacterium and Romboutsia was 3%, which was not detected in the intestine. The unclassi ed bacteria norank-o-JG30-KF-CM45 belonging to the phylum Curvularia phylum with an initial dominance of 6% decreased to 0.19% at the end.

Analysis of microbial community and diversity of the BFT system
Flocculation is the process of the combination of particles, algae and fungi in water through the action of ocs and heterotrophic microorganisms (Lu et al. 2010;De Schryver et al. 2008). Studies have shown that C/N greater than 15 is conducive to the growth of heterotrophic bacteria (Goldman et al.1987). In this experiment, C/N in the BFT group was maintained above 15, which provided the su cient conditions for the growth and reproduction of bacteria. In situ BFT system, the breeding objects are put together with the biological oc and the sh can ingest the oc (Azim et al.2008). So the microbial abundance in the BFT system decreased at the end of this experiment. A study by  showed that among the bacterial species found in the biological oc, Proteobacteria constituted a dominant bacteria group. The bacterial phyla in BFT are known to be involved in the beta deformation of ocs ora is important in addition to the nitrogen in the wastewater treatment of bacteria (Kapley et al. 2007). The nitrosation spirillum (Nitrosospira) and nitrosation belongs to bacterium (Nitrosomonas) as the main ammonia oxidizing bacteria (Ballinger lewis et al. 1998) in charge of the breakdown of nitrate nitrogen. Ammonia nitrogen convertion during the experiment was found in the occules undergoing nitrosation by the bacterium. But the abundance was not high and may be due to the over-wintering temperature which was too low. This affected the proliferation of bacteria by inhibiting their growth. However, the ammonia nitrogen was controlled below the level of safe breeding. Somasundaram (2012), showed that the main nitrite oxidizing bacteria, and denitrifying spiral bacteria all function to translate nitrate nitrogen and dinitrogen. Additionally, in this experiment, the occules were also found to contain nitration spiral bacteria genera which further proves that in BFT systems autotrophic nitri cation phenomenon exists in the system.
In addition to autotrophic nitri cation and heterotrophic assimilation, the maximum growth rate of heterotrophic bacteria in the bio occulation system was signi cantly higher than that of nitrifying bacteria and the microbial biomass was 40 times higher than that of autotrophic nitrifying bacteria (Ebeling et al., 2006). . In this experiment, there was the advantage of the early establishment of the bacterium Bacillus genus, which reached an abundance of 86%. Bacillus can secrete amylase and protease to degrade organic pollutants in the water, and can effectively inhibit pathogenic bacteria breeding under certain condition. The Bacillus can both improve the aquaculture water quality (Vijayalakshmi et al, 2003). When it is used as feed additive, it can promote the growth of farmed animals and enhance the body's immune and disease-resistant ability (Shih et al, 2001). In the later water environment, C. lustrum was the dominant ora, which was inconsistent with the dominant ora of Proteobacteria. This was also observed Xia Yun et al. (2012) in the biological oc. C. lustrum is a facultative anaerobe that can carry out photosynthesis but does not produce oxygen This indicated the existence of anaerobic environment in the culture system, which may be related to the high density of the culture in the later stage. In the later period of culture, the phylum abundance of Fusarium increased from 0.5-2%, mainly belonging to the genus Isosphaera. It is an ammonia-oxidizing bacteria, which can directly convert ammonia nitrogen and nitrite nitrogen into nitrogen through anaerobic ammonia oxidation . It is the main bacteria that can remove ammonia nitrogen in the system. At the genus level, the genus Xanthobacter reached 22%. This type of facultative anaerobic bacteria can use nitrate and nitrite as electron acceptor to conduct denitri cation in the anaerobic environment. Nitrate is converted into reduction products and organic matter is oxidized (Djordje et al., 2020).

Effects of culture water on intestinal microorganisms of sh
Gut microbes can adjust the physiological activities of the host, promote metabolism and growth of sh (Clements et al, 2014). Research shows that sh intestinal micro ora depends on the culture water, their genetics, diets and sh aspects of in uence (Bolnick et al, 2014). Therefore, different systems of cultivation, can in uence the structure of intestinal ora and also cause changes in the sh. Xia et al. In order to better study the impact of oc on sh intestinal ora structure, data from another study ( Cetobacterium can produce vitamin B12, which promote the metabolism of sh, protecting the nervous system and inhibiting the growth of harmful bacteria. However, the abundance of Cetobacterium in the BFT group was 13%, while that in the RAS group was only 3%. BFT aquaculture water signi cantly changed. The culture water of the BFT system, signi cantly in uenced the bacterial community structure of the sh intestinal tract, indicating that different breeding environments can signi cantly affect the intestinal microorganisms of sh. Moreover, the abundance of bene cial bacteria such as Bacillus and Cetobacterium in the sh intestinal tract of the BFT breeding mode is higher than that in the RAS breeding mode, which leads to the difference in the diversity and richness of intestinal microorganisms in the sh cultured in the two systems.
The main functions of heterotrophic assimilation in BFT system of the current study was carried out by Bacillus, Campylobacter, and Flavobacterium. This bacterial composition was an indication that denitri cation was ongoing. The BFT breeding model had a higher abundance of bene cial ora, which can inhibit the reproduction of pathogenic bacteria, reduce the incidence during breeding, improve digestive ability and promote the healthy growth of breeding species.
Declarations Acknowledgement This study was funded by the Shanghai Science and Technology Commission Project (19DZ2284300).

Authors' contributions
Conceptualization: YZ and BC; data collection and analysis: YZ, BC and GA; drafting of the manuscript: YZ, GA, and GL; review and editing: GL, YZ and YAL. All authors read and approved the nal manuscript.

Funding
This study was funded by the Shanghai Science and Technology Commission Project (19DZ2284300)

Data availability
All data are fully available on request.

Compliance with ethical standards
Ethics approval and consent to participate This article does not contain any studies with human participants or vertebrate animals performed by any of the authors.

Consent for publication
This manuscript does not contain any individual person's data.