Vertical distribution and seasonal dynamics of planktonic cyanobacteria communities in a water column of deep mesotrophic Lake Geneva

Abstract

Background: Temperate subalpine lakes recovering from eutrophication in central Europe are experiencing harmful blooms due to the proliferation of Planktothrix rubescens, a potentially toxic cyanobacteria. To optimize the management of cyanobacteria blooms there is the need to better comprehend the combination of factors influencing the diversity and dominance of cyanobacteria and their impact on the lake's ecology. The goal of this study was to characterize the diversity and seasonal dynamics of cyanobacteria communities found in a water column of Lake Geneva, as well as the associated changes on bacterioplankton abundance and composition.

Methods: We used 16S rRNA amplicon high throughput sequencing on more than 200 water samples collected from surface to 100 meters deep monthly over 18 months. Bacterioplankton abundance was determined by quantitative PCR and PICRUSt predictions were used to explore the functional pathways present in the community and to calculate functional diversity indices.

Results: The obtained results confirmed that the most dominant cyanobacteria in Lake Geneva during autumn and winter was Planktothrix (corresponding to P. rubescens). Our data also showed an unexpectedly high relative abundance of picocyanobacterial genus Cyanobium, particularly during summertime. Multidimensional scaling of Bray Curtis dissimilarity revealed that the dominance of P. rubescens was coincident with a shift in the bacterioplankton community composition and a significant decline in bacterioplankton abundance, as well as a temporary reduction in the taxonomic and PICRUSt2 predicted functional diversity.

Conclusion: Overall, this study expands our fundamental understanding of the seasonal dynamics of cyanobacteria communities along a vertical column in Lake Geneva and the ecology of P. rubescens, ultimately contributing to improve our preparedness against the potential occurrence of toxic blooms in the largest lake of western Europe.

Keywords: Planktothrix; bacterioplankton; column; cyanobacteria; lake.

Figure 1

Bacterioplankton community composition in Lake Geneva. Main bacteria genus identified in the (A) top (0–30 m depth) and (B) bottom (30–100 m depth) layers of the studied water column. Compositional data was obtained by averaging the relative abundances determined in 155 and 58 water samples collected at different depths within the top and bottom layer, respectively.

Figure 2

Non-metric multidimensional scaling (MDS) plot showing Bray–Curtis dissimilarity among samples. dbRDA ordinations of the bacteria communities obtained using Bray Curtis distance estimates and colored based on the sampling date.

Figure 3

Cyanobacteria community composition in Lake Geneva. (A) Main cyanobacteria genus identified in the studied water column. The compositional data was obtained by averaging the relative abundance of cyanobacteria in 218 water samples. (B) The panel below shows the spatiotemporal abundance of genera Planktothrix and Flavobacterium as relative abundance of reads.

Figure 4

Environmental variables associated with the increase of Planktothrix rubescens in the current study. Temperature is presented in Celsius degrees, conductivity is in μS/cm, global solar irradiation in Wm⁻², turbidity in NTU, oxygen saturation in %, DO in mg/L and depth is in meters. The abundance of Planktothrix rubescens reads was extracted for each sample of the study (n = 218).

Figure 5

Relationship between the dominance of Planktothrix rubescens reads, 16srRNA gene abundance and diversity estimations. (A) Abundance of the 16S rRNA gene determined by qPCR in lake samples collected across different depths and dates from August 2019 to December 2020. Results are represented as Log10 genomic copies per liter of lake water. (B) Estimates of taxonomic diversity according to Chao1 indices. (C) Estimates of functional diversity according to Chao1 indices, calculated from the predicted KEGG pathways obtained using PICRUST2.

Figure 6

PICRUSt-predicted functional markers for different dates. The determination of the markers was done using Lefse algorithm based on KEGG pathway predictions obtained with PICRUSt2. LDA stands for linear discriminant analysis.