Discriminating between effects of metals and natural variables in terrestrial bacterial communities

This study aimed to assess effects of metals on bacterial communities in grassland soil and to discriminate these effects from natural variability in soil properties. Changes in gross parameters of bacterial communities were investigated by the determination of 14C-leucine and 3H-thymidine incorporation rates, CO2 evolution, biomass indicators and N mineralization rates. 14C-leucine incorporation rate and CO2 evolution showed correlations with all metals tested as well as with organic matter content. 3H-thymidine incorporation rates, biomass indicators and N mineralization rates did not strongly correlate with any physical or chemical parameters. Further, community-level physiological profiles (CLPP) and polymerase chain reaction (PCR)-amplified denaturing gradient gel electrophoresis (DGGE) of 16S rDNA were performed. Monte-Carlo permutation testing was performed to analyse CLPP and DGGE data to allow for a stringent discrimination between sources of effects. CLPP changes correlated with the Pb concentration and pH in the soil. DGGE changes correlated with Pb and Cu concentrations and organic matter content. Pollution-induced community tolerance (PICT) was not found for any of the metals assessed. A causal relation between the effects observed on bacterial communities and the presence of metals was not established with PICT. The negative outcome of PICT can probably be attributed to indirect effects or to methodological problems. We concluded that the observed shifts in CLPP and DGGE patterns are a strong indication of metals effects on bacterial communities in this grassland. These effects could only be filtered from the total variation by multivariate analyses.