The eutrophic Bohai Sea receives large amount of suspended material, nutrients and contaminant from terrestrial runoff, and exchanges waters with the northern Yellow Sea through a narrow strait. This coastal region provides an ideal model system to study microbial biogeography. We performed high-throughput sequencing to investigate the distribution of bacterial taxa along spatial and environmental gradients. The results showed bacterial communities presented remarkable horizontal and vertical distribution under coastal gradients of spatial and environmental factors. Fourteen abundant taxa clustered the samples into three distinctive groups, reflecting typical habitats in shallow coastal water (seafloor depth ≤ 20 m), sunlit surface layer (at water surface with seafloor depth >20 m) and bottom water (at 2–3 m above sediment with seafloor depth >20 m). The most significant taxa of each cluster were determined by the least discriminant analysis effect size, and strongly correlated with spatial and environmental variables. Environmental factors (especially turbidity and nitrite) exhibited significant influences on bacterial beta-diversity in surface water (at 0 m sampling depth), while community similarity in bottom water (at 2–3 m above sediment) was mainly determined by depth. In both surface and bottom water, we found bacterial community similarity and the number of OTUs shared between every two sites decreased with increasing geographic distance. Bacterial dispersal was also affected by phosphate, which was possible due to the high ratios of IN/IP in this coastal sea area.
Microbes are recognized as the vital biological engines that drive global biogeochemical cycling (
In addition, water depth is another noteworthy factor that can shape vertical distribution of microbes in coastal seas (
Therefore the driving forces, including physicochemical factors, geographic distance and water depth, should be thoroughly considered when investigating microbial biogeography in coastal waters. Owing to pervasive applications of molecular approaches, especially high-throughput sequencing, plenty of studies have been conducted to reveal microbial biogeography and to compare the relative influences of above major forces in coastal surface water and sediments (
Here we choose the sea area from Bohai Sea to northern Yellow Sea (
The sampling sites were determined along the center axis from the edge of Yellow River Estuary to northern Yellow Sea. Water collection was performed at seven stations (
Environmental parameters, including temperature, salinity and depth, were measured
The total DNA on filters was extracted using the FastDNA SPIN Kit for soil (MP BIO, Santa Ana, CA, USA) according to the manufacturer’s instruction. The quality of the DNA extracts was determined by agarose gel electrophoresis. We used a NanoDrop 2000c spectrophotometer (ThermoFisher, Waltham, MA, USA) to measure the DNA concentration.
The universe prokaryotic primers 515F (5′-GTG CCA GCM GCC GCG GTA A-3′) and 907R (5′-CCG TCA ATT CCT TTG AGT TT-3′) were used to target the V4–V5 region of 16S rRNA genes. Triplicate PCRs were carried out by ABI GeneAmp 9700, using purified DNA as a template. The PCR process started with a pre-denaturation period at 95 °C for 5 min, followed by 35 cycles of denaturation at 94 °C for 30 s, annealing at 55 °C for 30 s and elongation at 72 °C for 40 s, and ended with a final extension at 72 °C for 7 min.
PCR products were purified by agarose gel electrophoresis with an AxyPrep DNA Gel Extraction Kit (Axygen, Corning, NY, USA), and then quantified with a QuantiFluor-ST Fluorometers (Promega, Madison, WI, USA). High-throughput sequencing was performed at Majorbio Co., Ltd. (Shanghai, China) using the 250-bp PE Illumina MiSeq sequencing platform.
Raw sequence profiles were processed using QIIME (
To estimate biodiversity of bacterial communities, we used the core_diversity_analyses.py script to calculate alpha-diversity indexes and beta-diversity distance with an evenly sequencing depth of 10,000 reads per sample. Multiple indices for alpha-diversity of bacterial communities were generated, including observed OTUs, Good’s coverage, phylogenetic diversity, Chao1 and Shannon-Wiener index. Weighted UniFrac distance was calculated to reveal beta-diversity. Principal coordinate analysis (PCoA) was performed to visualize community dissimilarities (beta-diversity) using R v.3.3.1 (
The dominant phyla, which had a relative abundance >1% in at least one sample, were utilized when making heat maps and conducting hierarchical clustering analysis (HCA). The discriminant taxa from each major cluster (suggested by HCA) were determined by the LDA (least discriminant analysis) effective size (LEfSe), using Kruskal–Wallis sum-rank test (
We measured environmental parameters and calculated Pearson’s correlations between physicochemical parameters and spatial factors (longitude, latitude, water depth and collection depth;
Most of clean sequences (98%) were assigned into bacterial taxa. We focused our analyses on fourteen taxonomic groups that are greater than 1% relative abundance in at least one sample. These groups included Alpha-proteobacteria (30.5%), Gamma-proteobacteria (19.2%), Bacteroidetes (16.1%), Cyanobacteria (10.5%), Actinobacteria (5.8%), Planctomycetes (3%), Firmicutes (2.6%), Rhodothermaeota (2.4%), Delta-proteobacteria (2.4%), Beta-proteobacteria (1.8%), Verrucomicrobia (1.4%), SAR406 (1%), Tenericutes (0.4%), Chloroflexi (0.4%).
Hierarchical clustering analysis (HCA) showed that these abundant phyla/classes can roughly classify samples into surface and bottom clusters, except for P1s and B8s (
The abundant taxa are defined with relative abundance >1% in at least one sample. The phylum Proteobacteria is shown at class level. Color legend shows the relative abundance of bacterial taxa.
Color legend shows correlation coefficients. More details are shown in
Alpha-diversity indexes including observed OTUs, phylogenetic diversity, Chao1 and Shannon-Wiener index, were calculated to estimate richness and biodiversity of bacterial community (
Correlation tests were performed separately using all samples, samples from surface water and bottom water. Significant correlations are shown in this table, and highlighted with bold type (
Samples | Variables | Observed OTUs | Phylogenetic Diversity | Chao1 | Shannon–Wiener |
---|---|---|---|---|---|
All samples | Temperature | −0.51* | −0.47 | ||
pH | −0.49 | ||||
Turbidity | |||||
Dissolved oxygen | |||||
NO2-N | |||||
NO3-N | 0.42 | 0.44 | 0.24 | ||
SiO3-Si | |||||
Samples from surface water | Longitude | ||||
Water Depth | |||||
Salinity | −0.63* | −0.63* | −0.63* | ||
pH | −0.68* | −0.68* | −0.64* | −0.75* | |
Turbidity | |||||
Chl |
0.5* | ||||
Samples from bottom water | PO4-P | −0.40 | −0.46 | −0.29 |
Using weighted UniFrac distance, we conducted PCoA for all samples. The PCoA plots suggested separation of bacterial communities from surface and bottom water (
Samples from surface and bottom water are labeled with blue squares and red dots.
Tests were performed separately using all samples, samples from surface water and bottom water. Correlations are shown in this table and are labeled with an asterisk when |
Samples | Variables | Mantel tests | Partial Mantel tests control factors | |||
---|---|---|---|---|---|---|
Geographic distance | Water depth | Collection depth | Environment | |||
All samples | Geographic distance | 0.18 | – | 0.10 | 0.19 | 0.18 |
Water depth | 0.16 | 0.04 | – | 0.12 | 0.13 | |
Collection depth | – | |||||
Environment | 0.08 | – | ||||
Samples from surface water | Geographic distance | – | 0.47 | – | 0.35 | |
Water depth | 0.21 | −0.13 | – | – | 0.26 | |
Collection depth | −0.08 | 0.07 | −0.05 | – | −0.01 | |
Environment | 0.57* | 0.47 | 0.58* | – | – | |
Samples from bottom water | Geographic distance | – | 0.09 | 0.08 | ||
Water depth | – | 0.17 | ||||
Collection depth | −0.02 | – | ||||
Environment | 0.34 | 0.37 | 0.19 | 0.34 | – |
Mantel and partial Mantel tests were performed to further explore the effects of geographic distance, spatial factors (water depth and collection depth) and physicochemical variables (environment) (
Effect evaluation of each physicochemical parameter revealed that temperature, pH, turbidity, dissolved oxygen, nitrite and silicate was the most important environmental factors influencing the beta-diversity of entire 14 bacterial communities (
Scatter plots were generated separately for samples from surface (A–B) and bottom water (C–D). Regression lines, along with regression coefficients (R) and probability (P), were generated using general linear model (GLM).
Numbers of pairwise shared OTUs decreased with increasing geographic distance of every two sites in both surface and bottom water, as well as pairwise dissimilarity (Euclidean distance) of water depth, longitude, salinity and phosphate (
Correlation tests were performed separately using samples from surface water and bottom water. Significant correlations are shown in this table and are labeled with an asterisk when |
Variables | Samples from surface water | Samples from bottom water |
---|---|---|
Geographic distance | ||
Water depth | ||
Longitude | ||
Temperature | −0.21 | |
Salinity | ||
PO4-P |
We observed horizontal salinity and turbidity gradients ranging from the Bohai Sea to northern Yellow Sea (
For instance, genera
Cluster II was dominated by Bacteroidetes (including class Flavobacteriia and genus
Discriminant taxa in Cluster III were Planctomycetes (including genus
In summary, the dominant microbial taxa of each province were greatly affected by spatial control and environmental conditions, which was consistent with their physiological requirements, dispersal limitation and ecological functions. Meanwhile a considerable number of abundant taxa dispersed ubiquitously between surface and bottom layers, and among entire sampling sites (
Our results demonstrated geographic distance, spatial factors and physicochemical variables contributed significantly to the diversity of bacterial communities, and their relative influences varied in the surface and bottom communities. For bacterioplankton in surface water, both alpha- and beta-diversity had significant correlations with turbidity (
For bacteria in bottom water, community similarity was mainly determined by water depth and collection depth, presenting a depth-decay pattern (
Sea water in the study area is subject to annual circles of winter mixing and summer stratification, and also affected three important water bodies, the Yellow River, Bohai Sea and Yellow Sea, which might lead to the large core bacterial species bank as well as the community distinctions existing in surface and bottom samples (
In this study, we observed significant differences of environmental conditions and a distinct separation of bacterial communities in surface and bottom waters. Vertical separation of bacterial communities has been reported in ocean by numerous studies (
Now we may answer the questions driving this study: (i) do the abundant bacteria vary along spatial and physicochemical gradients? Yes. In our study, those discriminant taxa exhibited limited distributions in the study area, while the others presented a ubiquitous distribution. Bacterial distribution along the gradient from the shallow coast to relative open sea shows remarkable horizontal and vertical patterns in bacterial communities. (ii) Do geographic distance and depth contribute more than environmental factors do to spatial variation of microbial communities? The situation differs in bacterial communities form surface and bottom waters. Environmental factors significantly affected the composition and biodiversity of bacterial communities in surface water. In coastal sea, water depth plays a noticeable role in biogeography of bacteria in bottom water, and leads to a depth-decay pattern of community similarity. Geographic distance enhanced community dissimilarity as previously reported (
(A) Environmental parameters determined for each samples.
(B) Pearson’s correlations between environmental variables. Significant correlations are highlighted with bold type for probabilities (P) in lower triangle, and with gray shadow for coefficients (
(C) Pearson’s correlations between environmental variables in surface water. Significant correlations are highlighted with bold type for probabilities (P) in lower triangle, and with gray shadow for coefficients (
(D) Pearson’s correlations between environmental variables of bottom water. Significant correlations are highlighted with bold type for probabilities (P) in lower triangle, and with gray shadow for coefficients (
Significant correlations (
The authors acknowledge the editing of English by Prof. Kevin McCartney (University of Maine at Presque Isle) and Prof. Andrzej Witkowski (University of Szczecin).
The authors declare there are no competing interests.
The following information was supplied regarding data availability:
The raw sequence data of this study have been deposited to the Sequence Read Archive of NCBI (