Distinct Bacterial Communities in Wet and Dry Seasons During a Seasonal Water Level Fluctuation in the Largest Freshwater Lake (Poyang Lake) in China

Abstract

Water level fluctuations (WLFs) are an inherent feature of lake ecosystems and have been regarded as a pervasive pressure on lacustrine ecosystems globally due to anthropogenic activities and climate change. However, the impacts of WLFs on lake microbial communities is one of our knowledge gaps. Here, we used the high-throughput 16S rRNA gene sequencing approach to investigate the taxonomic and functional dynamics of bacterial communities in wet-season and dry-season of Poyang Lake (PYL) in China. The results showed that dry-season was enriched in total nitrogen (TN), nitrate (NO₃ ^-), ammonia (NH₄ ⁺), and soluble reactive phosphorus (SRP), while wet-season was enriched in dissolved organic carbon (DOC) and total phosphorus (TP). Bacterial communities were distinct taxonomically and functionally in dry-season and wet-season and the nutrients especially P variation had a significant contribution to the seasonal variation of bacterial communities. Moreover, bacterial communities responded differently to nutrient dynamics in different seasons. DOC, NO₃ ^-, and SRP had strong influences on bacterial communities in dry-season while only TP in wet-season. Alpha-diversity, functional redundancy, taxonomic dissimilarities, and taxa niche width were higher in dry-season, while functional dissimilarities were higher in wet-season, suggesting that the bacterial communities were more taxonomically sensitive in dry-season while more functionally sensitive in wet-season. Bacterial communities were more efficient in nutrients utilization in wet-season and might have different nitrogen removal mechanisms in different seasons. Bacterial communities in wet-season had significantly higher relative abundance of denitrification genes but lower anammox genes than in dry-season. This study enriched our knowledge of the impacts of WLFs and seasonal dynamics of lake ecosystems. Given the increasingly pervasive pressure of WLFs on lake ecosystems worldwide, our findings have important implications for conservation and management of lake ecosystems.

Keywords: 16S rRNA; Changjiang River; hydrological regime; nutrient; shallow lake.

FIGURE 1

Study area and sample sites. Samples were collected from 10 and 13 sites in wet-season (July 2017) and dry-season (January 2017), respectively. TGD represents Three Gorges Dam, the world’s largest hydroelectric dam. The map was created in ArcGIS 14.0 (http://desktop.arcgis.com/en/arcmap/) using Landsat images download from USGS (https://earthexplorer.usgs.gov/). The images show the change of the lake area. The images were acquired in July and January 2017, respectively, corresponding to the sampling time.

FIGURE 2

Nutrient concentrations in dry and wet seasons of Poyang Lake, China. Statistical significance between dry and wet seasons was assessed by t-test and indicated by asterisks (^∗∗represents P < 0.01).

FIGURE 3

Basic differences of bacterial communities between dry-season and wet-season. (A) Venn diagram showing the unique and shared OTUs in dry-season and wet-season. The volcano plot showing the shared OTUs that significantly (t-test, P < 0.05) enriched in dry-season (red dots) and wet-season (green dots). The volcano plot was constructed using log2 fold change on x-axis and –log10 p-values of t-test on y-axis. (B) Alpha diversity (Chao 1, observed OTUs, Shannon diversity, and phylogenetic diversity) of bacterial communities in dry-season and wet-season. (C) Chord diagram showing the relative abundances of dominant phyla (the phyla with a relative abundance >1%) in dry-season and wet-season. “Others” represents the unsigned OTUs and the phyla with a relative abundance <1%. Statistical significance between dry-season and wet-season was assessed by t-test and indicated by asterisks (^∗represents P < 0.05, ^∗∗represents P < 0.01).

FIGURE 4

(A) Principal coordinates analysis (PCoA) based on Bray–Curtis distance in terms of the relative abundance of OTUs. (B) Z-score normalized heatmap of the top 100 OTUs. (C) Distance-based redundancy analysis (dbRDA) of bacterial community (symbols) and nutrient variables (arrows). All nutrient variables had goodness of fit at the significant level P < 0.05 by envfit function. (D) Variance partition analysis (VPA) determined the relative contributions of C-factor (including DOC), N-factor (including TN, NO₃^-, and NH₄⁺), P-factor (including TP and SRP), and the interactions between two or three of these factors (C × N, C × P, N × P, and C × N × P). The relative variance proportions that the corresponding components could explain are shown in percentages. Blank circles on the end of the triangle show the percentage of variance explained by C, N, and P alone. The percentage of variance explained by interactions between two or three of these factors is depicted as black dots on the edges or black triangle in the middle.

FIGURE 5

Results of Mantel test (A) between taxonomic matrix and environmental matrix, and (B) between functional matrix and environmental matrix. Spearman correlation coefficients (r) and associated p-values were calculated. Taxonomic dissimilarity and functional dissimilarity were represented by Bray–Curtis distances in terms of relative abundance of OTUs and KOs, respectively. Environmental distance is represented by Euclidean distance in terms of environmental variables (DOC, TN, NO₃^-, NH₄⁺, TP, and SRP).

FIGURE 6

Distribution and distinction of functional composition of the bacterial communities between dry-season and wet-season. (A) PCoA based on Bray–Curtis distance in terms of relative abundances of KOs associated to overall function, C-metabolism, N-metabolism, and P-cycle. (B) Non-parametric statistical tests of ADONIS (analysis of variance using distance matrices), ANOSIM (analysis of similarity), and MRPP (multi-response permutation procedure analysis).

FIGURE 7

Relative abundances of the major KEGG pathways associated with (A) carbon metabolism, (B) nitrogen metabolism, and (C) functional categories of phosphorus cycle. Statistical significance between dry-season and wet-season was assessed by t-test and indicated by asterisks (^∗represents P < 0.05, ^∗∗represents P < 0.01).

FIGURE 8

Relationship between the functional composition and nutrient variables. (A) Distance-based redundancy analysis (dbRDA) of the functional composition of the bacterial communities (symbols) and nutrient variables (arrows), in terms of overall function, carbon metabolism, nitrogen metabolism, and phosphorus cycle. (B) Corresponding variance partition analysis (VPA) determined the relative contributions of carbon (C, including DOC), nitrogen (N, including TN, NO₃^-, and NH₄⁺), phosphorus (P, including TP and SRP), and the interactions between two or three of these factors (C × N, C × P, N × P, and C × N × P). The relative variance proportions that the corresponding components could explain are shown in percentages on the right. Blank circles on the end of the triangle show the percentage of variance explained by C, N, and P alone. The percentage of variance explained by interactions between two or three of these factors is depicted as black dots on the edges or black triangle in the middle.