Multiple stressors affect function rather than taxonomic structure of freshwater microbial communities

Abstract

Microbial community responses to environmental stressors are often characterised by assessing changes in taxonomic structure, but such changes, or lack thereof, may not reflect functional changes that are critical to ecosystem processes. We investigated the individual and combined effects of nutrient enrichment ( + 10 mg/L N, + 1 mg/L P) and salinisation ( + 15 g/L NaCl)-key stressors in freshwater systems-on the taxonomic structure and metabolic function of benthic microbial communities using 1000 L open freshwater ponds established >10 years ago in the field. Combined stressors drove strong decreases in maximum and mean total carbon metabolic rates and shifted carbon metabolic profiles compared to either stressor individually and compared to ambient conditions. These metabolic functional changes did not recover through time and occurred without significant alterations in bacterial community taxonomic structure. These results imply that critical functions, including organic carbon release, are likely to be impaired under multiple stressors, even when taxonomic structure remains stable.

Fig. 1. Bacterial community structure and diversity of benthic freshwater communities on tiles under ambient (A; green symbols), elevated nutrients (N; yellow symbols), elevated salinity (S; blue symbols) and combined elevated salinity and elevated nutrients (SN; purple symbols) treatments 1, 7, 30 and 90 days after tiles with established microbial communities were deployed in the treated ponds.

A nMDS ordination on square root transformed Bray Curtis distances calculated on absolute abundances of ASVs (stress = 0.192). B Mean Shannon’s diversity calculated on absolute ASV abundances ( ± SE) (n = 3, except SN treatment on Day 7 and Day 30 where n = 4). An nMDS combining points from samples at each time into one panel is found in Supplementary Fig. 4.

Fig. 2. Concentration of groups of phototrophic organisms in benthic freshwater microbial communities on tiles under ambient (A; green bars), elevated nutrients (N; yellow bars), elevated salinity (S; blue bars) and combined elevated salinity and elevated nutrients (SN; purple bars) treatments 1, 7, 30 and 90 days after tiles with established microbial communities were deployed in the treated ponds (data are means ± SE, n = 4).

A Concentration of green algae. B Concentration of diatoms. C Concentration of cyanobacteria. * = pairwise contrast between treatments p <0.05, ** = pairwise contrast p <0.01.

Fig. 3. Multivariate metabolic profiles of freshwater benthic bacterial communities on tiles under ambient (A; green symbols), elevated nutrients (N; yellow symbols), elevated salinity (S; blue symbols) and combined elevated salinity and elevated nutrients (SN; purple symbols) treatments, 1 day (circles), 7 days (triangles) and 30 days (squares) after tiles with established microbial communities were deployed in the treated ponds.

PCA ordination on Euclidean distances calculated on maximum slopes of OD590 values for each plate. PCA 1 explains 45% of variance and PCA 2 explains 11% of variance. Ellipses represent a confidence interval of 0.8.

Fig. 4. Metabolic rates of freshwater benthic bacterial communities across 31 carbon sources under ambient (A; green bars), elevated nutrients (N; yellow bars), elevated salinity (S; blue bars) and combined elevated salinity and elevated nutrients (SN; purple bars) treatments, 1, 7 and 30 days after tiles with established microbial communities were deployed in the treated ponds (data are means ± SE, n = 4).

A Maximum slope of OD590 values across all carbon sources; B mean slope of OD590 values across all carbon sources. Letters indicate significant pairwise comparisons (alpha = 0.05) within Day groups.

Fig. 5. Relative maximum metabolic rates of freshwater benthic bacterial communities under elevated nutrients (N), elevated salinity (S) and combined elevated salinity and elevated nutrients (SN) treatments, 1, 7 and 30 days after tiles with established microbial communities were deployed in treatment ponds.

Values are maximum slopes of average well colour development absorbance values divided by ambient controls. Relative maximum slope values were square-root transformed and scaled by centring them around the ambient mean slope which was set at 0. Relative mean metabolic rates are seen in Supplementary Fig. 8.

Fig. 6. Experimental and sampling design. Circles represent replicate ponds.

Square represent trays containing 16 ceramic tiles with biofilms.