Chinese Journal of Oceanology and Limnology   2018, Vol. 36 issue(5): 1643-1654     PDF
Institute of Oceanology, Chinese Academy of Sciences

Article Information

LIU Hongmei(刘红梅), WANG Bin(王斌), HU Xiaoke(胡晓珂)
Sediment bacterial communities are more complex in coastal shallow straits than in oceanic deep straits
Chinese Journal of Oceanology and Limnology, 36(5): 1643-1654

Article History

Received May. 4, 2017
accepted in principle Aug. 7, 2017
accepted for publication Sep. 23, 2018
Sediment bacterial communities are more complex in coastal shallow straits than in oceanic deep straits
LIU Hongmei(刘红梅)1, WANG Bin(王斌)2, HU Xiaoke(胡晓珂)2     
1 School of Ocean, Yantai University, Yantai 264003, China;
2 Key Laboratory of Coastal Biology and Bioresource Utilization, Yantai Institute of Coastal Zone Research, Chinese Academy of Sciences, Yantai 264003, China
Abstract: Straits are ideal models to investigate the bacterial community assembly in complex hydrological environments. However, few studies have focused on bacterial communities in them. Here, comparable bacterial communities in costal shallow Bohai Strait (BS) and oceanic deep Fram Strait (FS) were studied. The Shannon and Chao1 indices were both higher in BS than in FS. The relative abundances of the classes Deltaproteobacteria and Bacilli and the family Halieaceae were higher in BS than in FS, in contrast to the families OM1_clade and JTB255_marine_benthic_group, revealing typical characteristics of bacterial communities in coastal and oceanic regions. Cluster analysis based on the Bray-Curtis index showed that samples were clustered by depth layer in FS and BS, indicating that structures of bacterial communities would differ with increasing water depth in straits. Additionally, the cluster relationships among samples in abundant and rare communities were both similar to those in entire communities. However, the dissimilarities among samples showed a descending order as rare communities, entire communities and abundant communities. Network analysis indicated that the BS network was obviously more complex than the FS network. Filamentous bacteria Desulfobulbaceae exhibited high degree values in BS but not in FS, indicating key roles of Desulfobulbaceae in the BS. Our study provides different and common evidences for understanding microbial ecology in coastal shallow and oceanic deep straits.
Keywords: bacterial communities     Bohai Strait     Fram Strait     currents     microbial ecological network    

The bacterial communities are major decomposers, secondary producers (Li and Gao, 2012) and major drivers of biogeochemical processes in the marine (Fuhrman et al., 2015). It is therefore important to investigate the marine bacterial communities. One fundamental question is the bacterial community assembly laying emphasis on the structures, diversities, abundant and rare bacteria and coexist relationships (Pedrós-Alió, 2012; Nemergut et al., 2013; Lynch and Neufeld, 2015). According to the niche theory (Pontarp et al., 2012), the structures and diversities of bacterial communities are determined by environmental variables. The hydrological environments in the marine exert great effects on the environmental variables and the microbial dispersion (Gilbert et al., 2012), suggesting that the assembly of marine bacterial community should be related to the hydrological environments such as currents. In the water body, for instance, regions with strong currents and large thermal fluctuations selected microbial communities with the greatest capacities for flexibility and evolution (Doblin and van Sebille, 2016), and the taxonomic and functional richness increased with depth, possibly resulting from vertical stratification with changes in physicochemical parameters (Sunagawa et al., 2015). Thus, these findings implied that water movements will exert effects on the microbial communities. Since sediment transportation is carried by currents (Shao et al., 2007), and the structures of bacterial communities in the sediments of Fram Strait with West Spitsbergen Warm Current (WSWC) and East Greenland Cold Current (EGCC) were similar along the direction of currents (Jacob et al., 2013), we inferred that the currents might influence the bacterial communities in the area they flowed along. However, investigations on the bacterial communities in sediments in the regions with strong currents remain sparse.

Additionally, the abundances of bacterial communities are affected not only by environmental variables but also by coexisting bacteria in the habitats. Various relationships among bacteria (e.g., competition and symbiosis) exert distinct influences on the parties (Hibbing et al., 2010). The widespread biofilm also provides excellent evidence (Singh et al., 2006). Bacterial associations thus should be included to contribute better understandings of the bacterial communities assembly (Faust and Raes, 2012; Nemergut et al., 2013; Coutinho et al., 2015). Recently, network analysis, which was used to investigate the associations among different phylotypes, has been introduced into bacterial communities in soil and marine environments (Chaffron et al., 2010; Zhou et al., 2011a; Barberan et al., 2014).

The straits, which separate and connect two large bodies of water, are influenced by strong tidal currents. According to the morphology, oceanic and atmospheric forcing, and flow characteristics of straits, straits have been classified into different types including shallow strait and deep strait (Li et al., 2015). The Bohai Strait (BS) is the foremost channel for the materials exchange between the Bohai Sea and the Yellow Sea. It is a typical shallow strait with depth ranged from 14 m in the south to 80 m in the north. The regional current pattern described as "north-in south-out" of the BS is the important factor controlling the sediment transportation in the BS (Li et al., 2016). Meanwhile, the BS adjacent to well-developed economic zones including Dalian and Yantai in the north and south, respectively, is an important fishery and navigable seaway. The anthropogenic influences exacerbate the complexity of environments in the BS. Thus, the BS is a representative to investigate the bacterial communities in the sediments with currents and anthropogenic influences. Despite some studies have investigated the microbial communities in the deep strait such as Fram strait (Jacob et al., 2013) and Bransfield Strait (Signori et al., 2014), few studies were reported in the coastal shallow straits. The Fram Strait is typical deep strait with depth around 2 500 m. It is the only deep-water connection between the Arctic Ocean and the World Oceans. Considering the little studies on the microbial ecology in straits, BS and FS are pretty much the extremes, at least for which samples probably exist, although there are lots of straits with the depth ranged from 20 m to 2 500 m including Strait of Hormuz, Strait of Gibraltar and Strait of Messina. Comparative studies on bacterial communities between in the BS and FS may provide more understanding on the ecological roles of bacterial communities in the straits.

This study aimed to investigate the structures, biogeographic patterns and bacterial associations of bacterial communities in surface sediments of BS in winter compared to those in FS. Our study will provide more evidences for the bacterial ecology in straits representing hydrologically complex environments.

2 MATERIAL AND METHOD 2.1 Sediment sampling, DNA extraction, PCR amplification, and Illumina-Miseq sequencing

Eleven stations were sampled to cover the BS (Fig. 1). At the time of sampling, composite samples of surface sediments were collected between December 17, 2013, and December 21, 2013. At each station, sediments were sampled in triplicates for the 0–5 cm surface layer using box grab. The sediment samples were transferred to sterile polythene bags, mixed, and stored at -80℃ in the dark until processing. Information regarding the depth of the stations was obtained from a navigational report. Eleven stations in the Fram Strait (FS) were selected for comparative analysis (Table 1). The samples in FS were collected in July 2009. For the environmental variables, only depth was both detected in the BS and FS.

Figure 1 Locations of the eleven survey stations in the Bohai Strait and Fram Strait (plotted with Surfer 11.0TM) Map drawing No. GS(2016)2945 (accessed from the National Administration of Surveying, Mapping and Geoinformation of China).
Table 1 Summary of environments, sequencing, and alpha diversities of survey stations

Bacterial genomic DNA was extracted using MoBio PowerSoilTM kit (MoBio, Carlsbad, CA, USA) according to the manufacturer's instructions. Three replications were extracted and mixed together for each sample. Eleven DNA extracts were obtained. The universal primers 515F (5′-GTGCCAGCMGCCGCGG-3′) and 907R (5′-CCGTCAATTCMTTTRAGTTT-3′) were used to amplify the V4–V5 region of the bacterial 16S rRNA gene using PCR. After extracting and purifying amplicons from 2% agarose gels using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, US), the purified amplicons were quantified using QuantiFluorTM-ST (Promega, Fitchburg, WI, USA). An Illumina-Miseq platform (Illumina Inc., San Diego, CA, USA) was applied for paired-end sequencing (2×250) according to the standard protocols in Majorbio (Shanghai, China).

2.2 Sequence data processing

The FS sequence data were downloaded from the GenBank Sequence Read Archives ( under BioProject ID: PRJNA208712. The sequences were obtained using 454 pyrosequencing of the 16S rRNA gene's V4–V6 region described previously (Jacob et al., 2013). The BS and FS sequence data were then transformed into operational taxonomic unit (OTU)-by-sample tables using Mothur software (version 1.36.1) (Schloss et al., 2009) according to the standard operating procedure (Kozich et al., 2013) and previous description (Jacob et al., 2013). Because of the different amplification region, the sequence data from BS and FS were processed separately. There were three alterations: (ⅰ) the latest release of the Silva reference (downloaded from, Release 123) with more reference sequences was applied for assigning taxonomic information; (ⅱ) singletons containing only one sequence across all samples are more likely to be artefacts, and we preferred to remove them to reduce the impact of spurious errors for both FS and BS. Each OTU was filtered to contain more than two sequences from at least two samples (Zhou et al., 2011b; Flynn et al., 2015); (ⅲ) as the primers we used are supposed to amplify members of the bacteria, thus those OTUs affiliated to archaea, chloroplasts and mitochondria, which might be undesirable and random mistakes or have no functional role in microbial communities (Kozich et al., 2013) were removed.

2.3 Comparative analysis between FS and BS

First of all, the following data processing was executed separately for the BS and FS. Those OTUs with average relative abundance (RA) of less than 0.1% were considered to be rare (Coveley et al., 2015; Lynch and Neufeld, 2015). Shannon and chao1 indices were applied to estimate the α-diversity. Because sequencing depth has great influence on estimated diversity (Magurran, 2003) and the sequencing depths of FS and BS were different, original datasets were subsampled to 9 900 sequences per sample (the lowest number of sequences in samples was 9 966). The Bray-Curtis index was applied to evaluate dissimilarities in community composition within samples. Cluster analysis was applied based on the average distance matrices, which were respectively calculated with entire OTUs, rare OTUs, and abundant OTUs of subsamples. Subsampling and calculating the three indices were executed 103 times to obtain the mean values. For the taxonomic composition, high Good's coverage across all samples was necessary (Magurran, 2003). The original datasets with higher Good's coverage were thus normalized into RA. The RAs of OTUs were summarized at the class, family, and genus levels. T-test was used to analyze the significant differences in the RA at different levels between in FS and BS. OTUs that occurred in fewer than five samples in FS or BS were removed, considering the following reasons: (ⅰ) the 11 samples were considered to be parallel; and (ⅱ) the network with reduced complexity facilitated the determination of core communities. Datasets with 9900 rarefied sequences per sample were used to calculate the Spearman correlation coefficients (P < 0.05, adjusted with "Benjamini Hochberg") to explore the abundance associations among OTUs in FS and BS. Those correlations were visualized into the networks, in which the nodes and connections represented the bacterial populations and the correlations. Based on network topology, the keystone species in the bacterial communities can be discerned (Zhou et al., 2011a). The degree defined as the number of connections connected to one node was used to evaluate the influence of the nodes in the network (Barberan et al., 2014). Significant correlations between two species were often related to their direct or indirect interactions (Faust and Raes, 2012; Barberan et al., 2014; Lupatini et al., 2014; Coutinho et al., 2015). An OTU with a higher degree value, which means more OTUs were correlated with the OTU, was thus considered to perform stronger influence (directly or indirectly) on the abundances of the correlated members in the network. R software (version 3.3.2) was employed to perform calculations and to plot the results.

3 RESULT AND DISCUSSION 3.1 The data were comparable

The survey stations are shown in Fig. 1. Some measures have been taken to make sure that the data of FS and BS were comparable: (ⅰ) the sequence data was processed in the same methods; (ⅱ) considering the different amplification region, the sequence data of FS and BS were processed separately; (ⅲ) considering that the sequencing depth affects estimating alpha- and beta-diversity, subsampled datasets with same sequences were used; (ⅳ) comparative analysis was performed on the taxonomic level with relative abundances, when Good's coverage (BS: 95.34%±1.5%, FS: 94.13%±1.39%) indicated that sequencing captured majority of the bacterial diversity in the samples (Table 1).

3.2 The bacterial community in BS was more complex than that in FS

First, after quality and OTU filtering, 289 043 and 169 268 16S rRNA gene sequences with 5 309 and 4 313 OTUs, identified at a 3% dissimilarity cutoff were obtained across samples in FS and BS, respectively (Table 1). The Shannon and Chao1 values were both higher in BS (Shannon: 5.81±0.46, Chao1: 3 160.49±393.84) than those in FS (Shannon: 5.27±0.35, Chao1: 2 599.45±339.86, P < 0.01, Table 1), indicating that the bacterial communities in BS were more diverse and richer than those in FS.

Second, the taxonomic compositions in BS and FS showed typical characteristics of bacterial communities in coastal and oceanic regions, respectively. At class level, the relative abundance of Deltaproteobacteria was higher in BS (20.04%±5.63%) than in FS (7.62%±1.07%, P < 0.05, Table 2). The major contributors at family level were Desulfobacteraceae and Desulfobulbaceae. At a global scale, the relative abundance of Deltaproteobacteria in sediments tended to decrease from the coastal sea to the open sea (Liu et al., 2015). At family level, the Halieaceae (Haliea genus and unclassified in BS, Halioglobus genus and unclassified in FS) affiliated to Gammaproteobacteria is aerobic anoxygenic phototrophic bacteria (AAPB) and mainly inhabits in coastal sea. Members of this family require dissolved organic carbon (DOM) as carbon source, and are capable of using bacteriochlorophyll a and carotenoids to harvest light and utilize light as additional energy source (Spring et al., 2015). Its relative abundance in BS was 3.78%±2.05% while the number in FS was 0.92%±0.24%. Taking into account the availability of DOM and light limitation for AAPB, the shallow depth and vertically well mixed water are benefit for the colonization of AAPB in BS. The family OM1_ clade affiliated to the class Acidimicrobiia and the family JTB255_marine_benthic_group affiliated to Gammaproteobacteria showed higher relative abundance in FS (4.38%±0.72% and 20.18%±4.63%) than in BS (0.57%±0.18% and 6.89%±4.70%, Table 2). Both of them were more sequence-abundant in polar, cold regions (Bienhold et al., 2016).

Table 2 Summary of the classes with significantly different RA between in the FS and BS groups, and the major contributors at family and genus levels

Moraxella and Psychrobacter were psychrophiles with widespread distributions in deep seas such as in the Japan Trench and cold Antarctic environments (Maruyama et al., 2000). However, the genera Acinetobacter, Moraxella, and Psychrobacter affiliated to Moraxellaceae were not observed in FS but in BS. Meanwhile, our previous study had isolated one crude oil-degrading Acinetobacter sp. strain HC8-3S from sediments of the Bohai Sea (Lin et al., 2014; Liu et al., 2016). Likewise, the genus Pseudomonas in the family Pseudomonadaceae, which is able to colonize a wide range of niches (Madigan et al., 2014), was abundant in BS but not detected in FS, despite that Pseudomonas has been previously found in deep seas (Pan and Hu, 2015; Yoshida et al., 2015). Therefore, further explorations on these unusual phenomena might be helpful to investigate the microbial ecology in straits.

Third, compared to the FS, the BS was more susceptible to the anthropogenic influences. At class level, the relative abundance of Bacilli in BS was higher (8.77%±8.09%) than that in FS (7.62%±1.07%, P < 0.05, Table 2). The genus Lactococcus (7.18%±6.53% in BS; not observed in FS), which contributed greatly to the high abundance of Bacilli in BS, was found to be abundant in fish and shellfish (Itoi et al., 2014). Since the Bohai Sea is an important fishing farm in China, the fishery breeding was supposed to influence the bacterial communities in the surface sediments of BS.

3.3 The structures of bacterial community differed with increasing water depth

The tidal current scour in straits may lead to depth gradients. In the BS, the depth gradients are due to the scour of coastal and tidal currents (Li et al., 2016). Depth may shape the composition and structure of microbial communities in the global ocean (Sunagawa et al., 2015). In BS, cluster an