1 INTRODUCTION

The gastrointestinal tract of fish generally coexists with a diverse, large, and dynamic microbiota (Roeselers et al., 2011). Gut microbiota is regarded as an integral component due to long-term co-evolution with the host and provides important services to the host, such as immune defense, digestion of complex carbohydrates, and production of enterocytes (Weimer et al., 1999; Ley et al., 2008). However, opportunistic pathogens hiding in the digestive tract can cause serious diseases for the host when the gut microbiota become imbalanced (Roeselers et al., 2011; Wu et al., 2012). Therefore, a comprehensive investigation of fish gut microbiota is needed, considering their importance in host health management. Although fish are the most diverse vertebrate group, knowledge of their gut microbiota is inadequate and lags compared with that of terrestrial vertebrates (Nayak, 2010).

Silver carp (Hypophthalmichthys molitrix) is cultured intensively in Chinese lakes and reservoirs and has been widely introduced into foreign countries due to its high culture yield and excellent role controlling algal blooms (Liang et al., 1981; Ke et al., 2007). Experiments conducted in Lake Taihu and Lake Donghu demonstrated that silver carp suppress cyanobacterial blooms effectively. In addition, silver carp exhibit a high growth rate (8.14 g/d) with a high ratio of Microcystis spp. in the gut (up to 84.4% of total phytoplankton) (Chen et al., 2006; Ke et al., 2007; Zhang et al., 2008). Cellulose is the most important component in the algal cell wall. It is assumed that bacteria in gut contents (allochthonous) and mucus (autochthonous) contribute to breakdown phytoplankton in the gut because silver carp lacks the complex enzymatic system or acid hydrolysis mechanism for digesting cellulose (Bitterlich, 1985). The gut of silver carp, which is 7-9 times longer than its body (Ke et al., 2008), provides many inhabitable microenvironments for microbes to decompose polysaccharides (e.g., cellulose, hemicelluloses, and xylan) into a usable form for the host. Therefore, understanding gut microbiota is of great importance to better manage productivity and health of silver carp. The silver carp gut microbiota has been surveyed based on conventional culture-dependent and modern molecular methods (Bairagi et al., 2002; Ye et al., 2013), and bacterial strains, such as Bacillus, Acinetobacter, Aeromonas, and Microbacterium, which secrete extracellular digestive enzymes, have been detected (Bairagi et al., 2002). Nevertheless, these groups represent only a small proportion of the gut microbiota, as most microbes cannot be cultured (Bairagi et al., 2002). In recent years, significant progress has been achieved in the field of gut microbiota with the development of cultureindependent methods, such as terminal restriction fragment length polymorphism, fluorescence in situ hybridization, clone libraries, and denaturing gradient gel electrophoresis (DGGE) (Langendijk et al., 1995; Roeselers et al., 2011; Niu et al., 2011; Wu et al., 2012). These methods have been used to identify predominant gut microbiota, and many novel sequences have been discovered. However, little is known about the bacterial community structure in silver carp gut contents and mucus during a cyanobacterial bloom.

Lake Taihu is the third largest freshwater lake in China (total area: 2 338 km², mean depth: 2 m) and is located in the southern Changjiang (Yangtze) River Delta (Jin and Hu, 2003). Due to increases in the human population, industry, and agriculture in the surrounding area, Lake Taihu is undergoing rapid entrophication (Chang, 1996). During the past few decades, cyanobacterial blooms have occurred regularly in certain regions, and planktivorous silver carp (Hypophthalmichthys molitrix) has been widely used to control these blooms (Ke et al., 2007). The purpose of this study was to use polymerase chain reaction (PCR)-DGGE and quantitative polymerase chain reaction (qPCR) methods to evaluate the structure and diversity of the dominant and cellulolytic bacterial communities in the contents and mucus along the gut of silver carp during a cyanobacterial bloom. Moreover, bacteria in the fish-associated environment were also detected.

2 MATERIAL AND METHOD 2.1 Fish breeding and sample collection

Silver carp were cultured in a cage surrounding dense algal blooms in Meiliang Bay on Lake Taihu (31°47′N, 120°22′E) from March to August 2012. Thirty silver carp were collected from the cage with a net on August 12, 2012, transported immediately to the laboratory, an anesthetized with an appropriate dose of MS-222. Twelve fish (mean weight: 175 g; total length: 19.2 cm) were selected and divided randomly into three groups for subsequent experiments. All sampled fish were aseptically dissected with sterile anatomy tools on a clean bench. The gut was removed carefully from the abdominal cavity, and the spleen, gallbladder, liver, and fat deposits surrounding the gut were removed. The gut was divided into the foregut, midgut, and hindgut, based on morphological characteristics (Xie, 1999). Each segment was opened with sterile tools, and the contents and mucus were collected as described by Wu et al. (2010). Samples from four fish in the same group were pooled. Pooling samples is a reasonable way to study the gut bacterial community of fish, and pooled samples generally well represent the gut microbiota of an individual (Romero and Navarrete, 2006). The foregut contents were collected to examine food items using the method described by Xie (1999). In addition, samples to check particle-attached bacteria (PAB) in the water, free-living bacteria (FLB) in the water, and bacteria in sediment (BS) were collected according to the methods of Niu et al. (2011) with minor modifications. Briefly, water was collected in a 2.5 L Schindler sampler at depths of 50, 100, and 150 cm. The water samples were pooled (300 mL) and sieved through a 5 μm MF^TM membrane filter (SMWP04700; Millipore, Billerica, MA, USA) to collect PAB, and all of the 5 μm filtrate was filtered through a 0.22 μm Nuclepore Track-Etched Membrane (111106; Whatman, Dassell, Germany) to collect the FLB fraction. Sediment samples were collected using a Petersen grab, and the unconsolidated surface sediment was gathered. All samples were stored at-80℃ until analysis.

2.2 DNA extraction and PCR amplification

DNA from all samples was extracted using a Bacterial DNA Kit (Omega stool DNA kit, D4015-02; Bio-tek, Winooski, VT, USA) according to the manufacturer's protocol. The QIAamp quick PCR purification kit (28104; Qiagen, Valencia, CA, USA) was used for purification. DNA purity and concentration were checked with a Nanodrop ND-1000 spectrophotometer (NanoDrop Technology, Wilmington, DE, USA). The purified DNA was used to amplify the variable V3 region of the bacterial 16S rRNA with the general 341F and 518R primers published by Muyzer et al. (1993). The PCR reactions were performed in 50 μL mixtures containing 180 μmol/L of each dNTP, 2 mmol/L MgCl₂, 0.2 μmol/L of each primer, 5 U Taq DNA polymerase, 1× PCR buffer, and 50 ng template DNA. Touchdown PCR was carried out with a T100^TM thermal cycler (Bio-Rad, Hercules, CA, USA) : initial incubation at 94℃ for 5 min, 35 cycles at 94℃ for 1 min, 1 min at annealing temperature (10 cycles at 67-58℃ and 25 cycles at 58℃), and 72℃ for 1 min; followed by a 10 min extension at 72℃. PCR products were checked by 1.5% agarose gel electrophoresis at 150 V for 30 min.

2.3 Denaturing gradient gel electrophoresis

PCR products containing approximately equal amounts of DNA in each sample were identified by 8% polyacrylamide gel electrophoresis (acrylamide: biscarylamide 37.5:1) in 1×TAE buffer (20 mmol/L Tris, 10 mmol/L acetic acid, and 5 mmol/L EDTA, pH 8.0) in a 42%-58% denaturing gradient. DGGE was performed with a Decod system at 60℃ for 12 h at 85 V. After electrophoresis, the gel was stained with 1:1 000 GelRed^TM (Biofilm, 41003; Biotium, Hayward, CA, USA) nucleic acid staining solution for 25 min, and the gel was photographed using the Bio Image System (Syngene, Cambridge, UK) under UV light.

2.4 DGGE band sequencing

All visible DGGE bands were excised with a sterile razor and immersed overnight at 4℃ in 40 μL sterile deionized water after mashing with a pipet tip. The supernatant was used as the template for PCR amplification with the 341F (without a GC clamp) and 518R primers following the procedure described above. PCR products were checked by 1.5% agarose gel electrophoresis and purified with a Gel Extraction Kit (Omega, D2500-01; Bio-Tek). The purified DNA was cloned directly into the PMD18-T plasmid vector system (D101A; TaKaRa Bio, Shiga, Japan) according to the manufacturer's instructions, and selected clones were sequenced by Beijing Yingjun Biotechnology (Beijing, China).

2.5 qPCR analysis of nucleic acids from gut contents and mucus

qPCR was used to quantify the abundance of cellulolytic bacteria (Clostridium Ⅰ, Clostridium Ⅲ, Clostridium XIVab, and Fibrobacter) and total bacteria based on the 16S rRNA gene with genusspecific primer pairs (Table 1). Amplification was performed using the Step One Plus^TM Real-time PCRsystem (Applied Biosystems, Foster City, CA, USA). Each system contained 40 ng DNA template, 12.5 μL of 1× SYBR real-time PCR premix (QPS-101B; Toyobo, Tokyo, Japan), 0.4 μL (10 μmol/L) of each primer, and sterile water added to a final volume of 25 μL. The cycling conditions were 95℃ for 10 min, followed by 43 cycles of 95℃ for 10 s, annealing at the temperatures displayed in Table 1 for 15 s, and 72℃ for 35 s. Specificity was evaluated based on melting curves obtained from continuous acquisition of fluorescence at 60-94℃ with 0.5℃ intervals after amplification. The purified plasmid DNAs were used as standards for the qPCR assays. Amplification efficiency was determined from the given slopes in the SDS software package (Applied Biosystems), and all values were 90%-110%. Plasmid copy number was calculated using the method described in Li et al. (2009).

2.6 DGGE profiles and statistical analyses

A DGGE fingerprinting profile was used to construct a binary matrix by assigning values for absence (0) or presence (1) of the bands to evaluate the bacterial communities in the different samples. Pairwise similarities between samples were quantified using the Dice similarity coefficient (S_D). The S_D values were used to construct a dendrogram with the unweighted pair group method with arithmetic average (UPGMA) using the NTSYS ver. 2.10e program (Exeter Software, Setauket, NY, USA) (Niu et al., 2011).

Relative band intensity in the DGGE profile was analyzed using Quantity-One ver. 4.6.2 software (Bio-Rad). Two indices were calculated to estimate changes in the bacterial community: (1) species richness (R) was calculated based on the total number of bands, and (2) the Shannon index (H′) was calculated based on a previous reference (Niu et al., 2011). Means of the diversity and richness values were compared using the independent t-test. Twoway analysis of variance was used to analyze the effects of sample type (contents vs. mucus) and gut region (foregut, midgut, and hindgut) on cellulolytic bacteria using SPSS 13.0 for Windows software (SPSS Inc., Chicago, IL, USA). A P-value < 0.05 was considered significant.

3 RESULT 3.1 Bacterial community structure in the gut and fish-associated environments

The dominant and cellulolytic bacteria communities in the gut of silver carp and fish-associated environments (PAB, FLB, and BS) during a cyanobacterial bloom were investigated. The guts of silver carp were green and full of algae. Moreover, microscopic observations revealed that the gut contents were composed primarily of Microcystis (up to 92.2%), followed by Melosira, detritus, and sporadic zooplankton. The DGGE patterns of the bacterial communities are shown in Fig. 1 and were composed of two independent DGGE profiles (A and B). Twentynine distinct bacterial bands were observed. Based on the DGGE patterns and subsequent statistical analysis, the bacterial communities among gut regions and ample types were significantly different (Table 2). Species richness (R) and diversity (H) increased significantly (P < 0.05) from the foregut to the hindgut, both in the gut contents and mucus samples. Species richness and diversity were higher in the contents than in the mucus within the same gut segment.

The cluster analysis showed that bacteria from the gut samples were clustered into one group, except the FM samples, whereas the adjacent gut regions clustered more closely than the segregated gut regions (Fig. 2). Moreover, the PAB showed higher similarity with gut samples than with the BS and FLB.

3.2 Sequence and phylogenetic analyses of the DGGE bands

Twenty-nine sequences obtained from different gut samples in this study were blasted in GenBank. All samples were collected from Lake Taihu, so the first band from the foregut contents (FC) sample was labeled TH-FC-1, followed by TH-FC-2, and THFC-3; the other bands were named in a similar manner. The results show that the silver carp gut microbiota were closely related to the following eight groups: Alphaproteobacteria (11 bands), Betaproteobacteria (2 bands), Gammaproteobacteria (2 bands), Actinobacteria (5 bands), Firmicutes (5 bands), Cyanobacteria (2 bands), Fusobacteria (1 band), and Bacteroidetes (1 band) (Table 3). Most of the sequences (19 bands) were identified as unable to be cultured. The relative abundance of the phylotype in the DNA template mixture is positively correlated to the corresponding band intensity (Fromin et al., 2002). According to this correlation, Betaproteobacteria (band TH-FC-13) were more abundant in the contents than in the mucus throughout the gut, whereas Clostridium XI (TH-FC-8) dominated the mucus samples. Firmicutes (TH-MC-1), Gammaproteobacteria (TH-FC-3), and Actinobacteria (TH-FC-4) were uniquely detected in the midgut and hindgut samples. Furthermore, novel bands for Alphaproteobacteria (TH-HC-1 and TH-HC-2) and Actinobacteria (TH-HC-3) were detected in hindgut samples. The phylogenetic relationships of the bacteria retrieved from the gut are shown in Fig. 3.

3.3 qPCR analysis of the cellulolytic and total bacteria 16S rRNA genes

The qPCR results revealed that four cellulolytic bacterial genera (Clostridium Ⅰ, Clostridium Ⅲ, Clostridium XIVab, and Fibrobacter) showed the highest abundance in the silver carp gut, and they were significantly different among gut regions and sample types (Table 4). Clostridium XIVab was the most prevalent among the four cellulolytic species and was >0.39% of all gut bacteria at all sites. The percentages of Clostridium Ⅰ, Clostridium Ⅲ, and Fibrobacter were < 0.02% in most samples, except Clostridium Ⅰ in the foregut samples (0.19%). The highest proportions of Clostridium Ⅲ, Clostridium XIVab, and Fibrobacter were observed in FM, whereas the highest Clostridium Ⅰ proportion was in foregut contents.

4 DISCUSSION

Although studies on silver carp gut microbiota have been sporadically reported, the microbial ecology of the silver carp gut during a cyanobacterial bloom is poorly understood. We found that silver carp maintain a dynamic and diverse microbiota throughout the gut and that sample types and gut regions had a pronounced effect on these bacterial communities.

The DGGE banding patterns and statistical analysis demonstrated that the composition of the bacterial community increased significantly in contents and mucus throughout the different gut regions. Additionally, all bacterial species that appeared in the foregut and midgut were found in the hindgut, suggesting that the hindgut might be the most ideal sampling region for further study of silver carp gut microbiota. Regulation of bacterial communities along the gut can be attributed to several factors, such as enzyme activities, nutrient concentrations, flow rate, and pH (Palframan et al., 2002; Nayak, 2010). The silver carp foregut contents were more heterogeneous and included intact phytoplankton and detritus from the diet, whereas the hindgut contents were more homogeneous and included well-digested material (Stevens and Hume, 1996). Moreover, pH in the silver carp gut increased continuously from the foregut to the hindgut, whereas digestive enzyme (e.g., trypsin and amylase) activities decreased (Bitterlich, 1985). Thus, differences in the microenvironment along the gut might lead to variations in gut microbiota. In the present study, samples of the same type (contents or mucus) between adjacent gut regions were more similar and clustered together in the phylogenetic analysis. However, the opposite results are obtained from animals with a complex digestive tract that is divided into a stomach, small intestine, caecum, and large intestine (Kohl et al., 2014; Malmuthuge et al., 2014). This finding suggests that this trait of silver carp is unique for stomachless fish because their digestive tract is simple, and the internal environment changes gradually along the gut (Bitterlich, 1985).

The bacterial community in the gut contents was quite different from that in the mucus, which is consistent with previous reports (Kim et al., 2007; Wu et al., 2010), in which content-associated bacterial communities were distinct from bacteria that adhered to the mucus. The bacterial communities in the contents and mucus samples were consistent at the taxonomic level, but differed in their abundance. For example, the representative sequence affiliated with Betaproteobacteria (band TH-FC-13) was more abundant in content samples, whereas Clostridium XI (band TH-FC-8) was dominant in mucus samples. In addition, bacterial diversity and richness were higher in contents than those in local mucus. The distribution of gut microbiota in the contents and mucus are determined by the host and the biological characteristics of the bacteria, as only bacteria than can resist the extreme conditions (e.g., low pH, fast flow, and digestive enzyme activity) in the gut can successfully colonize the gut's epithelial surface (Nayak, 2010).

In our study, Proteobacteria and Firmicutes were the dominant species in all gut samples; the same conclusion was reported for bighead carp, common carp, and grass carp (Ye et al., 2013; Li et al., 2015). Among the Firmicutes detected, two bands (TH-FC-3 and TH-FC-8) showed high similarity to Clostridium XI, which ferments polysaccharides and proteins (Lubbs et al., 2009). Actinobacteria has been reported to occupy only a small proportion of the fish gut (Li et al., 2015). However, we that more Actinobacteria than Firmicutes in the silver carp gut. PCR biases were excluded as a cause of this discrepancy, as the primers (F341/R518) used here are more sensitive for Firmicutes than for Actinobacteria (Wu et al., 2012). This finding confirms that Actinobacteria are naturally prevalent in the silver carp gut. The UPGMA analysis suggested that the Actinobacteria in the silver carp gut may have mainly originated from PAB.

It is well known that cyanobacteria are an important food source for silver carp (Ke et al., 2007). In our study, two clear bands associated with cyanobacteria were detected in the DGGE profiles and likely support their importance as food sources for silver carp. However, bright bands were also observed in the hindgut, indicating that digestion in the foregut might be incomplete. Kamjunke et al. (2002) reported that growth of silver carp continuing Microcystis was equal to that of starving fish. The photosynthetic activities of cyanobacteria fully recover after passing through the silver carp gut (Zeng et al., 2014). Incomplete digestion in the silver carp gut could be blamed for the lack of a complex enzyme system (Bitterlich, 1985). However, Zhu and Deng (1983) used a radioisotope tracer technique and reported that silver carp absorb some nutrients from Microcystis. Although it is unclear whether silver carp gain nutrition from consuming cyanobacteria, the ability to ingest cyanobacteria as a food resource is a competitive advantage for other species during cyanobacterial blooms.

The qPCR analysis showed that the abundance of Clostridium Ⅰ, Clostridium Ⅲ, Clostridium XIVab, and Fibrobacter varied along the silver carp gut. Empirically, Fibrobacter have been reported to only exist in the mammalian gastrointestinal tract (Kobayashi et al., 2008). However, the relatively high Fibrobacter copies detected in our study indicate that Fibrobacter may occupy a wider ecological distribution and their abundance suggests a potential role degrading cellulosic material in the silver carp gut. The four cellulolytic bacteria detected here not only regulate gut pH (Williams and Coleman, 1992) but also ferment polysaccharides to end-products that provide the host with additional nutrients (McDonald, 2002). Generally, the high ratio of short chain fatty acids (SCFAs) is in line with the high abundance of cellulolytic species (Mountfort et al., 2002). The copy numbers of the four cellulolytic bacteria were consistently higher in foregut samples, suggesting their function in cellulose breakdown and that SCFAs are mainly produced in the silver carp foregut.

The highest proportion of the four cellulolytic species examined here (Clostridium Ⅰ, Clostridium Ⅲ, Clostridium XIVab, and Fibrobacter) accounted for 1.8% of all bacteria in the foregut. This is much lower than that reported in the terrestrial vertebrate gastrointestinal tract. Stevenson and Weimer (2007) discovered that the bovine rumen contains >2% Fibrobacter, whereas other reports showed that members of Clostridium Ⅰ and Clostridium XIVab represent 29.7% of all sequences in the giant panda gut based on pyrosequencing (Zhu et al., 2011). According to Weimer et al. (1999), the abundance of cellulolytic bacteria depends on the concentration of cellulose in the diet. Cellulose is the major component of the terrestrial plant cell wall, representing 30% of the dry weight of a plant, whereas cellulose generally makes up only 1%-8% of marine plants and algae (Choat and Clements, 1998). Thus, the discrepancy in cellulose abundance between the digestive tract of terrestrial and aquatic vertebrates leads to the main difference in cellulolytic bacteria.

5 CONCLUSION

The current study demonstrated that the dominant and cellulolytic bacteria communities were significantly different among the gut regions and between contents and mucus in the silver carp gut. In addition, some characteristics of bacterial structure may be unique for such stomachless fish. For example, the bacteria were more similar between adjacent gut regions and became more homogenous along the digestive tract. Although Microcystis is not an ideal food source for silver carp because of its poor digestibility, the high abundance of cellulolytic bacteria in the digestive tract could help silver carp obtain energy from Microcystis. Therefore, we conclude that the silver carp gut microbiota may be of help during times of low food availability, such as during cyanobacterial blooms.