Microbial changes linked to the accelerated degradation of the herbicide atrazine in a range of temperate soils

Abstract

Accelerated degradation is the increased breakdown of a pesticide upon its repeated application, which has consequences for the environmental fate of pesticides. The herbicide atrazine was repeatedly applied to soils previously untreated with s-triazines for >5 years. A single application of atrazine, at an agriculturally relevant concentration, was sufficient to induce its rapid dissipation. Soils, with a range of physico-chemical properties and agricultural histories, showed similar degradation kinetics, with the half-life of atrazine decreasing from an average of 25 days after the first application to <2 days after the second. A mathematical model was developed to fit the atrazine-degrading kinetics, which incorporated the exponential growth of atrazine-degrading organisms. Despite the similar rates of degradation, the repertoire of atrazine-degrading genes varied between soils. Only a small portion of the bacterial community had the capacity for atrazine degradation. Overall, the microbial community was not significantly affected by atrazine treatment. One soil, characterised by low pH, did not exhibit accelerated degradation, and atrazine-degrading genes were not detected. Neutralisation of this soil restored accelerated degradation and the atrazine-degrading genes became detectable. This illustrates the potential for accelerated degradation to manifest when conditions become favourable. Additionally, the occurrence of accelerated degradation under agriculturally relevant concentrations supports the consideration of the phenomena in environmental risk assessments.

Keywords: Adaptation; Atrazine; Fate modelling; Microbial communities; Risk assessment; Soil pH.

Fig. 1

Dissipation of atrazine over three applications to nine temperate soils. Atrazine concentration in soil sub-samples was monitored at regular intervals by HPLC-UV. Error bars show the standard error between replicates, n = 12 for GA_2012 and GS_2012 applications 1 and 2, n = 6 for application 3, and n = 4 for all other soils. Soil identifier: CA Cotril agricultural, CS Cotril set aside, GA Ganthorpe agricultural, MA Mount agricultural, MS Mount set aside, GRA Grange agricultural, GRS Grange set aside, GA_2012 Ganthorpe agricultural collected in 2012, GS_2012 Ganthorpe set aside collected in 2012

Fig. 2

Comparison of modelling approaches for the dissipation of atrazine in GRA over three applications. Using the regulatory single first-order (SFO) approach (a), with each application modelled separately and the ‘growth-linked model’ described in this study (b). The growth-linked model enabled all applications to be modelled simultaneously. In both modelling approaches, the model fit of % atrazine remaining is shown as a solid black line and individual soil sub-samples as diamonds (n = 4). For the growth-linked model (b), the grey dashed line represents the number of atrazine degraders

Fig. 3

Percentage of the bacterial community that contain the atrazine-degrading gene trzN in the GA_2012 and GS_2012 soils. The trzN gene was monitored in the Ganthorpe agricultural soil, GA_2012 (A), and Ganthorpe set-aside soil, GS_2012 (S), 14 days after the second (2) or third application (3) of atrazine to each soil. TrzN was measured in atrazine-treated and control sub-samples. The proportions of the community carrying trzN was normalised against the 16S rRNA gene for each sample. Error bars show the standard error between experimental replicates, n = 6. The significant differences between the proportion of the community containing trzN between treated and control soils are indicated by asterisk (p < 0.05)

Fig. 4

Non-metric multidimensional scaling plot of the association of bacterial communities with atrazine treatment in the GA_2012 and GS_2012 soils. Each bacterial community is represented by a triangle, originating from the Ganthorpe agricultural soil, GA_2012 (A), and Ganthorpe set-aside soil, GS_2012 (S). The bacterial communities are based on OTU clustering of the pyrosequencing of 16S rRNA genes. The variables included in the analysis were soil history: set aside (S) or agricultural (A), duration in days under incubation conditions (0 or 120 days) and atrazine treatment, treated (T) or control (C). The similarity ellipses are based on hierarchical clustering shown in the Online Resource 26

Fig. 5

Principal component analysis (PCA) of the association of nine temperate soils with various physico-chemical properties. Measured soil properties were normalised, and the corresponding data matrix was subject to PCA. Each triangle represents an individual soil. The association between different soils is plotted along the first two principal components, which represent 68 and 20% of the variation between the soils. Soil properties: MWHC maximum water holding capacity, MC moisture content, C/N ratio carbon/nitrogen ratio, and total N total nitrogen. Soil identifiers: CA Cotril agricultural, CS Cotril set aside, GA Ganthorpe agricultural MA Mount agricultural, MS Mount set aside, GRA Grange agricultural, GRS Grange set aside, GA_2012 Ganthorpe agricultural collected in 2012, and GS_2012 Ganthorpe set aside collected in 2012

Fig. 6

Effect of soil pH on atrazine dissipation over two applications in the GRA_pH and GRS_pH soils. Atrazine was applied to the Grange set-aside soil (GRS_pH) and Grange agricultural soil (GRA_pH) which were collected in 2014 and their pH was amended (-a) to approximately pH 7 (GRSa) and pH 4 (GRAa), respectively. Error bars represent the standard error between replicates (n = 4). Parameters used for the SFO model fits are provided in the Online Resource 10; asterisk shows that the pH of the amended soils (GRSa and GRAa) are only approximate as there was minor variation in their soil pH throughout the experiment (Online Resource 4)