A comparison of bacterial community structure in seawater pond with shrimp, crab, and shellfish cultures and in non-cultured pond in Ganyu, Eastern China

Abstract

The aim of this study was to gain an understanding of the effects of mariculture on the bacterial communities in a co-culture pond, which is a popular culture model for shrimp, crab, and shellfish along the eastern coast of China. Six seawater samples were collected from a pond with cultures of shrimp, crab, and shellfish, and six other samples were collected from a non-cultured pond (control). The diversity of the bacterial communities in the samples was examined using the MiSeq desktop sequencer (Illumina) to amplify and sequence the V4–V5 region of the 16S ribosomal DNA analysis. The Meta-Stat computer program was used to assess the differences in the two communities. The sequences produced from all 12 samples were categorized into 1533 unique phylotypes, including 30 phyla, 81 classes, 270 families, and 414 genera. The top five dominant communities in the culture pond were Proteobacteria, Chloroflexi, Actinobacteria, Firmicutes, and Acidobacteria, while Proteobacteria, Planctomycetes, Bacteroidetes, Actinobacteria, and Cyanobacteria were the dominant communities in the control pond. Higher abundance was observed in the culture pond at the phylum, class, and genus levels, and the same result was also observed in the rarefaction curves. Elusimicrobia, Fibrobacteres, Fusobacteria, Tenericutes, MVP-21, SM2F11, and WCHB1-60 were present, whereas Deferribacteres and Lentisphaerae, as well as the candidate divisions BRC1, OD1, and OP11, disappeared from the culture pond. Moreover, abundant phylotypes (Aeromonadaceae, Pseudomonadaceae, Enterobacteriaceae, Bradyrhizobiaceae, Clostridiaceae, and Xanthomonadaceae) were present in the culture pond, but only rarely present in the control pond. These results suggest that mariculture contributed to the change in bacterial communities in the culture pond compared with the non-culture pond. Proteobacteria was the most dominant community in both ponds, but the second most dominant community was Planctomycetes in the control pond and Chloroflexi in the culture pond. Seven new phyla were observed, but five phyla disappeared in the culture pond. The disease- or metabolism-related members of the bacterial communities, namely, Aeromonadaceae, Pseudomonadaceae, Enterobacteriaceae, and Xanthomonadaceae, were abundant phylotypes in the culture pond. Our results contribute to an improved understanding of the establishment and maintenance of the bacterial community structure in a complex aquaculture ecosystem and may have practical applications in terms of improving seawater quality and early disease warning in this popular mariculture model in China.

Introduction

Aquaculture is the fastest-growing food production sector in the world, with the industry providing nearly half of all fish for human consumption in 2012. China is the largest aquaculture producer, with an average annual growth rate of 5.5 % in the 2000–2012 period. China’s aquaculture production for 2012 reached 41.11 million tons, which accounts for 61.69 % of the world’s total production. More specifically, mariculture produces 1.03 million tons, which accounts for 2.5 % of the total aquaculture production of China (FAO 2014). A wide variety of mariculture production modes are in use in China, such as open-ocean and harbor mariculture and pond, cage, and factory farming. The pond farming mode, particularly seawater ponds with shrimps, crabs, and shellfish polycultures, is widely used along the east coast of China. Its advantages include the need for only a small farming area (usually < 1 ha), easy management, high stocking density, symbiotic complementarity, and high unit:area yield, and the requirements for successful pond farming are obtained through the use of fertilizers, specialized feed, oxygen, and other technical measures (Hong and Zhang 2001). Although this modern form of aquaculture has long been used, basic information on the microbial populations and types in this ecosystem mode is scarce. Such information is critical for the development of preventive measures to safeguard mariculture production from infectious agents that cause diseases and to prevent severe financial losses. Aquatic microorganisms influence the water quality and are known to be closely associated with the physiological status, disease, and postharvest quality of cultured fish (Al-Harbi and Uddin 2005). Fish are intimately in contact with a complex and dynamic microbial world. A large fraction of microorganisms adhere to and colonize epithelial surfaces. In rare circumstances, microorganisms cause diseases directly by damaging or traversing epithelial layers and indirectly by inducing tissue-damaging inflammatory responses (Gómez and Balcázar 2008).

Studies on the microbiological impact of fish farming have traditionally been performed using cultivation methods designed to yield pure cultures of bacterial isolates for characterization (La Rosa et al. 2001; Maki et al. 2006). However, these cultivation-based approaches introduce a strong bias because only up to 1 % of bacteria from any environment can be cultivated using standard laboratory techniques (Amann et al. 1995). Consequently, methods based on molecular techniques have become popular alternatives. Studies using fingerprinting methods, such as denaturing gradient gel electrophoresis (Sandaa et al. 2003; Payne et al. 2006; Le Nguyen et al. 2008; Zhang et al. 2014) (GTG)₅-PCR fingerprinting (Huys et al. 2007), phospholipid fatty acid analysis (Green and Scow 2000), restriction fragment length polymorphism (Sun et al. 2011; Li et al. 2013; Mengoni et al. 2013), and fluorescence in situ hybridization (Balcázar et al. 2010), have been conducted on the most abundant bacterial groups, but in depth studies which characterize microbes in the aquaculture environment by cloning and sequencing are lacking. Next-generation sequencing (NGS) technologies contribute to the ease of implementing low-cost, high-throughput sequencing (Ye and Zhang 2013). For example, the 454 pyrosequencing platform (Roche Diagnostics, Indianapolis, IN) is one of the most popular high-throughput sequencing systems and has the capacity to create over 400,000 reads with an average accuracy rate of >99.5 % (Glenn 2011). A specified number of DNA samples can be sequenced simultaneously in one run by incorporating barcode sequences on the primers (Tamminen et al. 2011). The MiSeq sequencing platform (Illumina Inc., San Diego, CA) can provide data that are at least as good as those generated by the 454 pyrosequencing platform while providing considerably higher sequencing coverage for a fraction of the cost (Kozich et al. 2013). This latter technology has been used to identify the microbiota in distinct samples, such as river water (Staley et al. 2013), soil (Caporaso et al. 2012), skin (Smeekens et al. 2013; Becker et al. 2014), and microsatellites (Norrell et al. 2014).

To explore the bacterial diversity and community structure of a seawater polyculture pond with shrimps, crabs, and shellfish at different stages of aquaculture, we compared a typical polyculture pond with a non-culture pond (control) in the Ganyu county of Lianyungang, Jiangsu Province, China. The composition of the bacterial community was characterized by 16S ribosomal DNA analysis of the V4–V5 region using theMiSeq sequencing platform. To the best of our knowledge, this study is the first to use the high-throughput MiSeq sequencing platform to characterize the phylogenetic diversity of a seawater polyculture pond. The results of this study contribute to an improved understanding of the bacterial community in a complex aquaculture ecosystem that will have direct applications in the regulation of water quality, reduction of disease incidence, and enhancement of aquaculture production in terms of quantity and quality.

Materials and methods

Sample collection

Samples were collected from seawater ponds located in Ganyu County, Lianyungang city, Jiangsu province, eastern China (34°58′23″W-119°11′36″E). The ponds are two adjacent of ten artificial ponds, each pond is about 200 m long, 50 m wide and 1.0 m water depth. Samples labeled “A” came from a non-culture pond (control group) without any mariculture. A1–A6 represent the sampling dated for Jul. 15, Jul. 25, Agu. 15, Agu. 25, Sep. 15, Sep. 25 respectively. Samples labeled “B” came from an integrated culture pond with shrimps (Penaeus japonicus Bate), crabs (Portunus trituberculatus), and shellfish (Ruditapes philippinarum). B1–B6 represent the sampling dated for Jul. 15, Jul. 25, Agu. 15, Agu. 25, Sep. 15, Sep. 25 respectively. Ten liters of water were passed through neutral filter to remove most metazoans and planktons, after which water was filtered through 0.45 μm filters, Filters were then divided into quartering packs and stored at −70 °C for extracting DNA.

DNA extraction and PCR amplification

Microbial DNA was extracted from the aquaculture water samples using the E.Z.N.A.® soil DNA kit (Omega Bio-tek, Norcross, GA) according to manufacturer’s protocols, except that glass beads were not used during the extraction. The V4–V5 region of the bacterial 16S ribosomal RNA gene was amplified by PCR (amplification regime: 95 °C for 2 min, followed by 25 cycles at 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s, with a final extension at 72 °C for 5 min). The PCR primers used were 515 F (5ʹ-barcode- GTGCCAGCMGCCGCGG-3ʹ) and 907R (5ʹ-CCGTCAATTCMTTTRAGTTT-3ʹ), where barcode was an 8-base sequence unique to each sample. The PCR reactions were performed in multiple (triplicate) independent 20-μL reaction mixtures, with each reaction sample containing 4 μL of 5× FastPfu Buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of each primer (5 μM), 0.4 μL of FastPfu Polymerase, and 10 ng of template DNA, on an ABI GeneAmp® 9700 PCR thermal cycler (Applied Biosystems, Foster City, CA).

Illumina MiSeq sequencing platform

Amplicons were extracted from 2 % agarose gels and purified using the AxyPrep DNA Gel Extraction kit (Axygen Biosciences, Union City, CA) according to the manufacturer’s instructions and then quantified using QuantiFluor™ -ST (Promega, Madison, WI). Purified amplicons were pooled in equimolar mixtures and paired-end sequenced (2× 250) on an Illumina MiSeq platform following standard protocols. The raw reads were submitted to the DDBJ Sequence Read Archive (SRA) database (Accession no. DRR024550-24561).

Processing of sequencing data

The raw FASTQ files were de-multiplexed and quality filtered using the QIIME (ver. 1.17) bioinformatics software program with the following criteria. First, the 250-bp reads were truncated at any of the sites, with an average quality score of <20 over a 10-bp sliding window, whereas the truncated reads that were shorter than 50 bp were discarded. Second, instances of exact barcode matching, two-nucleotide mismatch during primer matching, and reads containing ambiguous characters were removed. Three, only sequences with overlaps >10 bp were assembled according to their overlapping sequences. For further analyses, reads that could not be assembled were discarded.

Operational taxonomic units (OTUs) were clustered with a 97 % similarity cutoff using UPARSE (ver. 7.1; http://drive5.com/uparse/). Chimeric sequences were identified and removed using UCHIME [Electronic Supplementary Material (ESM) Table S1]. The phylogenetic affiliation of each 16S rRNA gene sequence was analyzed by the RDP Classifier (http://rdp.cme.msu.edu/) against the SILVA (SSU115) 16S rRNA database using a confidence threshold of 70 % (Amato et al. 2013).

Results

Phylogenetic analysis and taxonomic richness of seawater ponds

To investigate the bacterial community structure of a seawater mariculture pond with a integrated system consisting of shrimp, crab, and shellfish cultures, we sequenced a total of 158,679 PCR amplicons that spanned the V4–V5 region of 16S rRNA gene in our DNA preparations. Each sample produced 8864 (sample B2 from co-culture pond) to 21,903 (sample A5 from control pond) effective reads (ESM Table S1). Given a 97 % sequence similarity threshold, these clusters were analyzed to determine the OTUs, abundance-based coverage estimator (ACE), Chao estimator, exponential Shannon index, and Simpson’s index (Table 1). The sequences produced from the 12 samples were assigned to 1533 unique OTUs, which included 30 phyla, 81 classes, 270 families, and 414 genera (ESM Table S2). Taken together, the data revealed a complex bacterial community structure and a wide range of diversity in the seawater pond. The number of OTUs during aquaculture ranged in the control pond from 281 (A1) to 558 (A5) and in the co-culture pond from 683 (B2) to 853 (B5) (Table 1). At the phylum level, all OTUs could be classified into 20 formally described bacterial phyla and ten candidate phyla (Tables 2 and 3). Proteobacteria was predominant in all samples and accounted for 38.24 % (B4) to 90.69 % (A1) of all bacterial amplicons. The differences between samples only emerged at levels lower than the phylum level. No distinct changes were detected during the sampling period, suggesting that the bacterial communities are very sensitive to subtle variations in nutritional input and environmental factors, such as rain, baiting, and human activities.

Table 1 Operational taxonomic units and species richness estimates based on 97 % sequence similarity threshold

Table 2 Bacterial community distribution in samples (A1–A6) from the control pool and relative abundances at the phylum level

Table 3 Bacterial community distribution in samples from co-culture pond and relative abundances at phylum level

Comparison of bacterial community structure between the seawater co-culture and control ponds

To identify the effects of mariculture on the bacterial community structure in the seawater ponds, we focused on the differences in bacterial community structure between the co-culture and control ponds. Proteobacteria was the most dominant community in both groups at the phylum level (Tables 2, 3; Fig. 1a). The second-most dominant community was Planctomycetes in the control pond and Chloroflexi in the co-culture pond. In the co-culture and control microbial groups, eight and five phyla, respectively, had >2000 PCR amplicons. The eight most dominant communities that accounted for more than 96 % of bacteria in the culture pond included Proteobacteria, Chloroflexi, Actinobacteria, Firmicutes, Acidobacteria, Planctomycetes, Bacteroidetes, and Cyanobacteria. In particular, Chloroflexi, Actinobacteria, and Firmicutes become dominant communities in the co-culture pond but not in the control pond (Fig. 1a). In comparison, Proteobacteria, Planctomycetes, Bacteroidetes (lost in co-culture pond), Actinobacteria, and Cyanobacteria were the five most dominant communities in the control pond. In addition, the undetermined phylotypes (designated as “others” in Fig. 1a) were more abundant in the co-culture than in the control group. Similar results were observed at the class level (Fig. 1b). This result suggests that co-culture enriched the bacterial community structure in seawater ponds. The rarefaction of curves at 97 % similarity was analyzed in terms of the OTUs, and the results indicated that the bacterial community diversity of the co-culture pond was significantly higher than that of the control pond (Fig. 2). This result provides further support for our hypothesis that mariculture leads to the composition and structure of bacterial community of seawater ponds becoming more complex. In addition, seven new phyla (marked ▲ in Table 3) emerged in the co-culture pond, as follows: Elusimicrobia, Fibrobacteres, Fusobacteria, Tenericutes, MVP-21, SM2F11, and WCHB1-60. Alternatively, five phyla (marked ▼ in Table 2 ) disappeared in the co-culture pond: Deferribacteres and Lentisphaerae, as well as candidate divisions BRC1, OD1, and OP11 (Table 2). These changes are clearly seen in Fig. 1a. To assess the differences between the two pond communities, we performed a meta-analysis using the Meta-Stat software program. The results revealed that at least 37 classes of bacteria from the culture pond were more distinct and more abundant compared with the control pool (p < 0.005); 31 of these classes had distinctly separate clades (p < 0.001; Fig. 3a). Only three classes were beneficial in the control pond: Alphaproteobacteria (p < 0.001), Deltaproteobacteria, and SPOTSOCT00m83 (p < 0.005; Fig. 3b).

Fig. 1

Distribution and relative abundance of bacterial communities in 12 seawater samples (A1–A6 control pool samples; B1–B6 co-culture pool samples) at the phylum level (a) and at the class level (b). Relative abundance is presented as a percentage of the total effective bacterial sequences per sample

Fig. 2

Rarefaction analysis for the 12 seawater samples. The curves were generated based on a 97 % sequence similarity threshold level of the operational taxonomic unit (OTU)

Fig. 3

Meta-Stat analysis of bacterial communities at the class level between the control and co-culture groups. a Classes listed are noticeably abundant in the co-culture group, b classes listed are noticeably abundant in the control group. Error bars indicate the percentage error

Principal component analysis of the bacterial community in the seawater ponds

To determine the bacterial diversity of the seawater pond ecosystem, data obtained from the MiSeq sequencing platform were analyzed with respect to sampling date using principal component analysis (PCA). The similarity of the microbial communities in the 12 samples was monitored with PCA at OTU levels of 0.03 (Fig. 4). These 12 samples were collected from different stages during the culture cycle. The first principal component (PC1) explained 47.80 % of the variation. The cultured pond samples exhibited a noticeable and regular separation from the control pond samples. Samples A1–A6 from the control pond clearly belonged to a single cluster, whereas samples B1–B6 from the cultured pond belonged to a different cluster. The second principal component (PC2) represented 20.72 % of the variation, with the control samples in the range of −2000 to +2000. The abovementioned data indicate that cultured organisms, rather than sampling dates, were the main factors influencing the composition of the microbial community.

Fig. 4

Principal component (PC) analysis of bacterial communities in the 12 seawater samples at the 0.03 OTU level

Abundant and rare microorganisms in the seawater pond

The tremendous diversity of the bacterial communities in the seawater pond was highlighted at different levels. Based on previous reports, the rare phylotypes were assumed to have a frequency of ≤0.01 % per sample and the abundant phylotypes to have a frequency of ≥1 % per sample (Pedrós-Alió 2006; Galand et al. 2009). The distribution of rare phylotypes in each sample was within the range of 11.07–46.26 % based on the OTUs, but <5 % when compared with the abundant OTUs (Table 1, with the exception of sample A5). The distribution of rare phylotypes in each sample was similar, but it ranged from 31.74 to 39.70 % of OTUs in the co-culture pond and from 52.95 to 77.29 % of OTUs in the control pond (Fig. 5).

Fig. 5

Phylogenetic composition of rare bacterial phylotypes (≤0.01 %) and richness of abundant phylotypes (≥1 % frequency) for the 12 seawater samples. Blue Rare species OTUs, green abundant sequences

Overall, 29,421 and 53,660 abundant reads were counted in the co-culture and control groups, covering 36.68 % and 68.38 % of the number of sequences, respectively. We counted 366 rare phylotypes from among 1067 OTUs in the co-culture group and 448 rare phylotypes from among 888 OTUs in the co-culture group. In addition, a total of 52 and 23 abundant phylotypes were identified (Table 1). The abundant and rare genera in the control and co-culture samples are shown in detail in Fig. 6da, b. Seven genera were abundant in the bacterial community of the co-culture pool but identified only rarely in the bacterial community of the control pool (Fig. 6a): Aeromonas, Pseudomonas, Shimwellia, Bradyrhizobium, Clostridium, Dyella and norank (belonging to phylum Cyanobacteria). Six genera were abundant in the bacterial community of the control pool but identified only rarely in the co-culture pond: Planctomyces, Planctomyces (uncultured species), Crocinitomix, uncultured (Saprospiraceae family), Marinicella, and Marinobacter (Fig. 6b). This result suggests that the distribution of the bacteria varies from being rare to being abundant in the biosphere, with obvious differences between the bacterial communities of the co-culture and control pools.

Fig. 6

Heat map and clustering of bacteria at the genus level. a Control, b co-culture

Discussion

Bacteria play an important role in biological chains in a seawater pond ecosystem, and the structural diversity of the bacterial community is relevant to the prevailing conditions in this ecosystem.

In our samples, Proteobacteria was the most important phylum and accounted for 62.5 % of the top 16 OTUs in the co-culture community. Gammaproteobacteria was the major class of Proteobacteria. Interestingly, the main members of Gammaproteobacteria, namely, Aeromonadaceae (Aeromonas hydrophila), Pseudomonadaceae (Pseudomonas), Enterobacteriaceae (Shimwellia uncultured), and Xanthomonadaceae (Dyella), varied from rare to abundant phylotypes in the co-culture. This finding suggests that these organisms are related to the mariculture system.

A. hydrophila is a common species of Aeromonadaceae and is widely distributed in freshwater, seawater and soil. A. hydrophila is also very toxic to several organisms, especially to freshwater fish cultures or feral groups where it causes diseases (Cipriano et al. 1984; Nielsen et al. 2001; Gopalakannan and Arul 2006; Akinbowale et al. 2007; da Silva et al. 2012). When A. hydrophila enters the body of its victim, the pathogen travels through the bloodstream to the first available organ. The cell produces aerolysin cytotoxic enterotoxin (ACT), a toxin that can cause tissue damage, such as hemorrhagic septicemia (Xu et al. 1993; Nielsen et al. 2001; da Silva et al. 2012), hemolytic ascitesosis (Sun et al. 1991), intussusception (Liu et al. 2008), tail or fin rot (Liu et al. 1993; Rahman et al. 2001), stigmatosis (Xu et al. 1980), and epizootic ulcerative syndrome (Austin and Adams 1996; Roberts 1997). The increased quantity of Aeromonas in the co-culture pond suggests that it would also be a potential pathogen of cultured organisms, such as shrimps, crabs, and shellfish. It should also be pointed out that the above inference is only valid for the purpose of a preliminary understanding of the ecosystem. Positive identification of the species will require further work.

Pseudomonas is a physiologically and genetically diverse group named according to the type genus Pseudomonadaceae; these bacteria have great ecological significance, demonstrate a very broad range of metabolic diversity, and consequently colonize a wide range of niches (Madigan 2005; Cornelis 2008; Zago and Chugani 2009). The following species have been previously studied: (1) P. aeruginosa is an opportunistic human pathogen (Lister et al. 2009; Morita et al. 2010; Rello et al. 2014) and is antagonistic to Vibrio parahaemolyticus (Vinoj et al. 2013); (2) the plant pathogen, P. syringae (Kennelly et al. 2007; Carrion et al. 2014); (3) the soil bacterium, P. putida (Dos Santos et al. 2004; Puchałka et al. 2008). Pseudomonas has been applied in bioremediation systems (Yu et al. 2001; Poblete-Castro et al. 2012). The plant growth-promoting effects and the environmental sensing and adaption of P. fluorescens have been studied by Martinez-Granero et al. (2014). There is a sufficient body of evidence showing that Pseudomonas cultures perform significant functions in life, infection, and/or metabolism. However, information on its exact functions in the field is still lacking.

Enterobacteriaceae is another representative of phylum Gammaproteobacteria. Similar to various harmless symbionts, Enterobacteriaceae is represented by pathogens such as Salmonella, Escherichia coli, Yersinia pestis, Klebsiella, Shigella, Proteus, Enterobacter, Serratia, and Citrobacter (Brenner et al. 2005). Among these organisms, Y. pestis, Klebsiella, and Serratia are common pathogens in aquatic environments. This information supports the hypothesis that the presence of Enterobacteriaceae contributed to the alteration of an abundant group in the co-culture pool.

Xanthomonadaceae is an independent family of Gammaproteobacteria, and the members of this family are typically characterized as environmental organisms, occupying diverse ecological niches, such as soil and water, as well as plant tissues. Various members of the Xanthomonadaceae cause plant diseases, especially the species of genera Xanthomonas and Xylella (Mhedbi-Hajri et al. 2011). In our study, Xanthomonadaceae was distributed in six OTUs, and its amplicon percentage increased by 3.534 % in the culture pond, whereas only four OTUs showed a slight increase (0.051 %) in the control, namely, Stenotrophomonas (Denton and Kerr 1998; Looney et al. 2009; Lakatos et al. 2014), Dyella (Xie and Yokota 2005; Jung et al. 2009; Anandham et al. 2011; Son et al. 2013), Rhodanobacter (Kaci et al. 2014; Lee et al. 2014), and Silanimonas (Manucharova et al. 2008; Shi et al. 2014).

Stenotrophomonas is a confirmed opportunistic pathogen (Mhedbi-Hajri et al. 2011), whereas little is known of Dyella, Rhodanobacter, and Silanimonas. Kaci et al. (2014) suggested that Rhodanobacter thiooxydans and Stenotrophomonas were related to metal resistance. Silanimonas lenta reportedly plays a main role in chitin transformation and nitrogen consumption (Manucharova et al. 2008). These data imply that Xanthomonadaceae plays variable roles in the culture pond. This phenomenon is not difficult to understand because Xanthomonadaceae existed in the culture pond where cultured organisms and nutrients were supplied.

In conclusion, our work reveals the special structure diversity of the bacterial community in a seawater culture pond system with shrimp, crab, and shellfish. These changes suggest that early attention must be paid to preventing latent disease. The co-culture pond is a complex ecosystem, and a number of factors, such as introduction of non-native microbial species, input and circulation of nutrients, and human interference, can mediate microbial communities. Consequently, this system requires more detailed study.