Bacterial community structure and biogeochemical activity in an aquifer contaminated with pesticides
Résumé
Bacterial communities play a pivotal role in biogeochemical cycle,
however there is still no consensus on the effect of pesticide
contamination on bacterial community function, especially on
their ability to reduce nitrate, which is an issue in several pesticides
impacted sites. Atrazine is an herbicide which have been widely
used for weed control in corn, soja and sorgho cultures, until
2003 when it has been withdrawn in France. Desethylated
atrazine (DEA) is among its metabolite the one most observed
in soil and groundwater and it has been reported with higher
effect to aquatic life, than ATZ. Ten years after its withdrawn, ATZ
and DEA concentrations exceeding the legal EU thresholds for
groundwater and drinking waters (0.1 μg/L) are still reported.
Our objective was to assess the effect of pesticide mixtures on
groundwater microbial abundance, community structure and
their function in the nitrate reduction at the catchment level.
This is, to our best knowledge, the first study on pesticide impact
on groundwater microbial community diversity structure and
function; it has the potential to provide sound-based arguments
to be considered when improving the current strategy to manage
water quality, as well as when proposing end points to monitor
the microbial community in the biodiversity objective under the
European water directive framework.
Two-year monitoring. The Ariège alluvial plain (France) is
contaminated with a large panel of pesticides with concentations
up to the ppm level. Water (1 L) was sampled from 17 selected
springs on a monthly basis during 2 years (March 2012- March
2014, n = 50).
Microcosm. Water was sampled in July 2014 in two wells having
different contamination profils. Water (700 mL) was placed in 1 L
microcosm and ATZ, DEA or ATZ+DEA was spiked at 0, 1 and 10 μg/L. Units were sacrificed at the start and following 15-day and
30-day incubation (n = 58).
Bacterial analyses. Water was filtered through 0.22 μm filters and
microbial DNA was extracted. Abundance of the universal marker
(16S rRNA gene) and of nitrate-reducing bacteria (narG and
napA genes) were assessed by quantitative PCR (qPCR). Diversity
was assessed using fingerprinting technic, CE-SSCP (Capillary
Electrophoresis-Single Strand Conformational Polymorphism).
Chemical analyses. Water samples for pesticides were analyzed
by LC-MS/MS following an on line-solid phase extraction,
and samples for anions and cations were analyzed by ion
chromatography.
Statistical analyses. Divergence between diversity profiles were
analyzed with StatFingerprints software. Diversity indexes were
also calculated (richness, Shannon-weaver index, eveness).
Statistical differences were analyzed using two-way ANOVA with
treatments, chemical or incubation time as factors (p < 0.05).
Principal component analyses (PCA) were performed using
XLSTAT Version 2011.2.02.
Biodiversity was higher thorough the experiment in the water
C+, historically contaminated with various pesticides than in
the water C- where none of the 51 pesticides monitored was
observed during 2 years (Figure 1). Pesticide concentrations in
the water historically contaminated often exceeded the legal EU
threshold for groundwater and drinking waters (e.g. 2-year mean:
0.09 ± 0.01 μg ATZ /L, 0.43 ± 0.06 μg DEA /L (n = 23)). On the other
side, during microcosm incubation, biodiversity decreased when
spiked-chemical concentration increased from 1 to 10 μg/L. In the
water C+, no effect of the incubation duration was noticed. In the
water C-, biodiversity decreased with the chemical concentration
and increased with the incubation duration when exposed to 1
μg/L, while no increase was observed when the community was
exposed to 10 μg/L, suggesting that adaptation to pesticides
might occur and this process depends on the chemical
concentration. In both waters, there was no difference in the ATZ,
DEA or ATZ+DEA effect to the microbial diversity. Abundance of
bacteria reducing nitrate among the total community drastically
decreased during the experiment, but this decrease was also
observed in the control unit (p < 0.05). Total biomass was similar
in both waters and during the whole experiment (p > 0.05).
No biodegradation of ATZ and DEA was observed during the
1-month exposure.
Analyses of the two-year monitoring at an agricultural catchment
level are undergoing; preliminary results suggests that biomass
is similar in all samples while biodiversity and nitrate-reducing
bacteria show important differences between samples.
The undergoing analyses on natural waters monitored during two
years at a catchment level will hopefully show boundaries within
the complex relationship between biodiversity and chemical
concentrations observed in the microcosm, which is positive at
the environmental level (between water C- and C+) and negative
at the spiked-concentratoin level (between 1 and 10 μg/L).
In the microcosms, ATZ, DEA or ATZ+DEA exhibited similar effect
on the microbial community. Comparison at the catchment level
of the effect of the 51 monitored pesticides taken individually or
summed, on the microbial biodiversity will enable to assess the
mixture effect on the microbial community.
Biomass was similar in all conditions (p > 0.05), suggesting that
this is not a sensitive endpoint to assess water quality.
Relationship between pesticide contamination and bacteria nitrate-reducing community will be assessed further to consider
the risk of nitrate acumulation or of inhibition effect of nitrate on
pesticide biodegradation.
Acknowledgement: The authors thank the Water Agency Adour-Garonne
(France), the FEDER grants (Europe) and the BRGM for their financial
support.