Ecological inference on bacterial succession using curve-based community fingerprint data analysis, demonstrated with rhizoremediation experiment

Abstract

Nucleic acid-based community fingerprinting methods are valuable tools in microbial ecology, as they offer rapid and robust means to compare large series of replicates and references. To avoid the time-consuming and potentially subjective procedures of peak-based examination, we assessed the possibility to apply direct curve-based data analysis on community fingerprints produced with bacterial length heterogeneity PCR (LH-PCR). The dataset comprised 180 profiles from a 21-week rhizoremediation greenhouse experiment with three treatments and 10 sampling times. Curve-based analysis quantified the progressive effect of the plant (Galega orientalis) and the reversible effect of the contaminant (fuel oil) on bacterial succession. The major observed community shifts were assigned to changes in plant biomass and contamination level by canonical correlation analysis. A novel method to extract relative abundance data from the fingerprint curves for Shannon diversity index revealed contamination to reversibly decrease community complexity. By cloning and sequencing the fragment lengths, recognized to change in time in the averaged LH-PCR profiles, we identified Aquabacterium (Betaproteobacteria) as the putative r-strategic fuel oil degrader, and K-strategic Alphaproteobacteria growing in abundance later in succession. Curve-based community fingerprint analysis can be used for rapid data prescreening or as a robust alternative for the more heavily parameterized peak-based analysis.