Linking Shifts in Bacterial Community Composition and Function with Changes in the Dissolved Organic Matter Pool in Ice-Covered Baiyangdian Lake, Northern China

Abstract

The relationship between CDOM (Chromophoric dissolved organic matter) and the bacterial community was investigated in ice-covered Baiyangdian Lake. The results showed that environmental parameters significantly differed in Baiyangdian Lake, whereas a-diversity was not significantly different. Moreover, the microbial and functional communities exhibited significant differences, and T (Temperature), pH, ORP (Oxidation-reduction potential), DO (Dissolved oxygen), NO₃^--N, NH₄⁺-N, and Mn (Manganese) were the main environmental factors of these differences, based on redundancy analysis and the Mantel test. Biomarkers of the microbial and functional communities were investigated through linear discriminant analysis effect size and STAMP analysis. The number of biomarkers in the natural area was highest among the typical zones, and most top functions were related to carbohydrate metabolism. Two protein-like components (C1 and C2) and one humic-like component (C3) were identified by parallel factor analysis, and C1 was positively related to C2 (R = 0.99, p < 0.001), indicating the same sources. Moreover, CDOM significantly differed among the typical zones (p < 0.001). The high biological index, fluorescence index, β:α, and low humification index indicated a strong autochthonous component and aquatic bacterial origin, which was consistent with the results of UV-vis absorption spectroscopy. Network analysis revealed non-random co-occurrence patterns. The bacterial and functional communities interacted closely with CDOM. The dominant genera were CL500-29_marine_group, Flavobacterium, Limnohabitans, and Candidatus_Aquirestis. Random forest analysis showed that C1, C2, and C3 are important predictors of α- and β-diversity in the water bacterial community and its functional composition. This study provides insight into the interaction between bacterial communities and DOM (Dissolved organic matter) in ice-covered Baiyangdian Lake.

Keywords: Bacterial community; Baiyangdian Lake; Excitation-emission matrix-parallel factor analysis (EEM-PARAFAC); Functional composition; Network analysis; chromophoric dissolved organic matter (CDOM).

Figure 1

Location of sampling sites in Baiyangdian Lake, Xiongan New Area of China.

Figure 2

Alpha diversity and environment factors in Baiyangdian Lake. (A) environment factors (different letters are significantly different (p < 0.05)); (B), Alpha diversity (different letters are significantly different (p < 0.05)); (C), PCoA of alpha diversity; (D), PCoA of environment factors; (E) correlation between environment factors and alpha diversity (*, **, *** indicate the significance of the correlation at p < 0.05, p < 0.01, and p < 0.001).

Figure 3

LEfSe analysis of microbial community in Baiyangdian Lake. (A) Cladogram; (B) LEfSe Bar.

Figure 4

The functional composition of significant changes based on PICRUSt2, Tax4Fun, and FARPROTAX using the response ratio method at a 95% confidence interval (CI).

Figure 5

RDA, and Mantel test of taxonomic community and functional community in Baiyangdian Lake. (A) RDA of taxonomic community; (B) RDA of functional community based on PICRUSt2; (C) RDA of functional community based on Tax4Fun; (D) RDA of functional community based on FARPROTAX; (E) Mantel test between taxonomic community and environment factors; (F) Mantel test between functional community and environment factors.

Figure 6

Fluorescence components and spatial distributions of fluorescence intensities and relative abundance in Baiyangdian Lake. (A) Fluorescence components; (B) Fluorescence intensities of fluorescence components; (C), Relative abundance of fluorescence components; (D) PCoA of CDOM based on UV and EEM; (E) Fluorescence indices; (F) correlation analysis (*, *** indicate the significance of the correlation at p < 0.05, and p < 0.001).

Figure 7

Network visualizes the OTU-DOM and functional community-DOM interactions in Baiyangdian Lake. Positive correlations were displayed in red and negative correlations were displayed in green. The nodes were coloured according to different types of modularity classes. The size of each node is proportional to the degree. (Spearman′s |r|> 0.6, p <0.05). (A) microbial network analysis based on OTU vs DOM (module level); (B) microbial network analysis based on OTU vs DOM (phylum level); (C) microbial network analysis based on function vs DOM (PICRUSt2); (D) microbial network analysis based on function vs DOM (Tax4Fun); (E) microbial network analysis based on function vs DOM (FARPROTAX).

Figure 8

Random Forest analysis is an effective predictor of environment factors as drivers of αdiversity and β-diversity microbial community in Baiyangdian Lake. (A) for microbial community; (B) for functional community based on PICRUSt2; (C) for functional community based on Tax4Fun; (D) for functional community based on FARPROTAX.