Microbial community composition in sediments resists perturbation by nutrient enrichment

Abstract

Functional redundancy in bacterial communities is expected to allow microbial assemblages to survive perturbation by allowing continuity in function despite compositional changes in communities. Recent evidence suggests, however, that microbial communities change both composition and function as a result of disturbance. We present evidence for a third response: resistance. We examined microbial community response to perturbation caused by nutrient enrichment in salt marsh sediments using deep pyrosequencing of 16S rRNA and functional gene microarrays targeting the nirS gene. Composition of the microbial community, as demonstrated by both genes, was unaffected by significant variations in external nutrient supply in our sampling locations, despite demonstrable and diverse nutrient-induced changes in many aspects of marsh ecology. The lack of response to external forcing demonstrates a remarkable uncoupling between microbial composition and ecosystem-level biogeochemical processes and suggests that sediment microbial communities are able to resist some forms of perturbation.

Figure 1

Photograph of a very highly fertilized plot at GSM taken autumn 2007. The fertilized plot is clearly visible in the center of the photograph and is surrounded by unfertilized marsh that has begun fall senescence. Note the difference in macrophyte community composition, above ground biomass and elevation between the fertilized and unfertilized portions of the marsh. Photo by I Valiela.

Figure 2

Number of sequences from different microbial taxa in control and fertilized marshes at PIE. Top: marsh pair one compares Sweeney (fertilized) with West (control). Bottom: marsh pair two compares Clubhead (fertilized) and Nelson (control).

Figure 3

Number of sequences from different microbial taxa in C compared with LF (top), HF (middle) and XF (bottom) from GSM.

Figure 4

Stacked bar plot of the magnitude of the hybridization signal of the 15 most abundant nirS targets from GSM and PIE. The RFR indicates the relative signal strength of that archetype in each sample. Analysis of variance results indicate no significant differences between control and fertilized marshes in either experiment (GSM: F=0.846, P=0.471; PIE: F=0.115, P=0.734).

Figure 5

Quantitative PCR results comparing the absolute abundance of nirS genes in DNA from each of the GSM plots with the absolute abundance of the 16S rRNA genes in the same plots. Inset: the % of 16S rRNA that is nirS. There were no significant differences among treatments with respect to gene copy number for either 16S (F=1.11, P=0.37) or nirS (F=0.08, P=0.96), although the difference between the abundance of nirS and 16S were clearly different from each other (F=91.07, P<0.01).