Effect of a karst system (France) on extended spectrum beta-lactamase (ESBL)-producing Escherichia coli.

Karst aquifers are an important water resource worldwide particularly exposed to anthropogenic pollution, including antibiotic-resistance. The release of antibiotic-resistant bacterial pathogens in the environment is a major public health challenge worldwide. In this One Health study, we aimed to determine the effect of karst on antibiotic-resistant bacteria. For this purpose, we determined the concentrations of extended-spectrum ß-lactamases-producing Escherichia coli (ESBL-Ec) for 92 weeks in a rural karst hydrosystem providing drinking water. ESBL-Ec isolates (n = 130) were sequenced by whole genome sequencing. We analysed the isolates at different levels of granularity, i.e., phylogroup, sequence type, presence of antibiotic-resistance genes, mutations conferring antibiotic-resistance, and virulence genes. The ESBL-Ec concentrations were spatially and temporally heterogeneous in the studied karst hydrosystem. ESBL-Ec isolates survived in the karst and their concentrations were mostly explained by the hydrodynamic of the hydrosystem. We demonstrate that the studied karst has no filtration effect on ESBL-Ec, either quantitatively (i.e., in the ESBL-Ec concentrations) or qualitatively (i.e., in the genetic characteristics of ESBL-Ec isolates).

Calibration comparison between two passive samplers -o-DGT and POCIS- for 109 hydrophilic emerging and priority organic compounds.

The Polar Organic Chemical Integrative Samplers (POCIS) is the most widely used passive sampler for hydrophilic compounds, but unsuitable for certain ionic organic contaminants. The Diffusive Gradient in Thin-Film technique (o-DGT) has shown positive results for both ionic and hydrophilic compounds. However, a calibration step is now needed to evaluate kinetic constant of accumulation for a wide range of molecules. In this study, o-DGT and POCIS were compared for the sampling of three families of micropollutants of potential risk to aquatic environments: 53 pesticides, 36 pharmaceuticals and 20 hormones. A calibration experiment was conducted to compare the kinetic models and constants from a scientific and practical perspective. The results are discussed in a single table that summarizes the performance of both passive samplers for the 109 compounds of interest. The advantage of o-DGT is that it allows linear accumulation for 72 compounds versus only 33 with POCIS. The mean times to equilibrium obtained with o-DGT are higher than those obtained with POCIS. These results confirm that the presence of a diffusion gel delays the achievement of equilibrium during compound accumulation. Therefore, o-DGT can be considered for situations where POCIS cannot be used due to non-linear accumulation over a typical 14-day deployment period. However, overall sampling rates and mass transfer coefficients also appear reduced with o-DGT, which is explained by the smaller exchange surface area, as well as the consideration of an additional diffusive layer in this device. This paper also showed that the most appropriate membrane to sample polar compounds with o-DGT was a polyethersulfone polymer with a pore size of 5 µm.

Determination of diffusion coefficients in agarose and polyacrylamide gels for 112 organic chemicals for passive sampling by organic Diffusive Gradients in Thin films (o-DGT).

The diffusive gradient in thin film technique was recently adapted to organic compounds. The diffusional coefficient (D) is a key parameter needed to calculate the time-weighted average concentration. In this study, two methods are used for D measurement in two gels (agarose and polyacrylamide): the diffusion cell method (Dcell) and the slice stacking method (Dstack). Thus, D were discussed and compared for 112 organic compounds, including pesticides, hormones, and pharmaceuticals. Dstack tends to be higher than Dcell. It could be explained by the presence of a non-negligible diffusive boundary layer thickness in diffusion cell. Consequently, the use of sampling rates (RS) should be more adequate to determine water concentration, for a given bulk flow velocity. Dstack also corresponds to the diffusion in gel only, allowing the determination of the maximal RS, and would be considered as a reference value that can be adjusted to in situ conditions, by applying the appropriate DBL thickness. The range and variability of D values found in the literature and obtained in this work were discussed. Relationships between D and compound physicochemical properties (molecular mass, log Dow, polar surface area, van der Waals volume) were investigated. We did not find clear and robust correlation between D and any single physicochemical property, for the set of compounds tested.