A biomimetic approach to the detection and identification of estrogen receptor agonists in surface waters using semipermeable membrane devices (SPMDs) and bioassay-directed chemical analysis.

Goal, Scope And Background: Some anthropogenic pollutants posses the capacity to disrupt endogenous control of developmental and reproductive processes in aquatic biota by activating estrogen receptors. Many anthropogenic estrogen receptor agonists (ERAs) are hydrophobic and will therefore readily partition into the abiotic organic carbon phases present in natural waters. This partitioning process effectively reduces the proportion of ERAs readily available for bioconcentration by aquatic biota. Results from some studies have suggested that for many aquatic species, bioconcentration of the freely-dissolved fraction may be the principal route of uptake for hydrophobic pollutants with logarithm n-octanol/water partition coefficient (log Kow) values less than approximately 6.0, which includes the majority of known anthropogenic ERAs. The detection and identification of freely-dissolved readily bioconcentratable ERAs is therefore an important aspect of exposure and risk assessment. However, most studies use conventional techniques to sample total ERA concentrations and in doing so frequently fail to account for bioconcentration of the freely-dissolved fraction. The aim of the current study was to couple the biomimetic sampling properties of semipermeable membrane devices (SPMDs) to a bioassay-directed chemical analysis (BDCA) scheme for the detection and identification of readily bioconcentratable ERAs in surface waters.

Methods: SPMDs were constructed and deployed at a number of sites in Germany and the UK. Following the dialytic recovery of target compounds and size exclusion chromatographic cleanup, SPMD samples were fractionated using a reverse-phase HPLC method calibrated to provide an estimation of target analyte log Kow. A portion of each HPLC fraction was then subjected to the yeast estrogen screen (YES) to determine estrogenic potential. Results were plotted in the form of 'estrograms' which displayed profiles of estrogenic potential as a function of HPLC retention time (i.e. hydrophobicity) for each of the samples. Where significant activity was elicited in the YES, the remaining portion of the respective active fraction was subjected to GC-MS analysis in an attempt to identify the ERAs present.

Results And Discussion: Estrograms from each of the field samples showed that readily bioconcentratable ERAs were present at each of the sampling sites. Estimated log Kow values for the various active fractions ranged from 1.92 to 8.63. For some samples, estrogenic potential was associated with a relatively narrow range of log Kow values whilst in others estrogenic potential was more widely distributed across the respective estrograms. ERAs identified in active fractions included some benzophenones, various nonylphenol isomers, benzyl butyl phthalate, dehydroabietic acid, sitosterol, 3-(4-methylbenzylidine)camphor (4-MBC) and 6-acetyl-1,1,2,4,4,7-hexamethyltetralin (AHTN). Other tentatively identified compounds which may have contributed to the observed YES activity included various polycyclic aromatic hydrocarbons (PAHs) and their alkylated derivatives, methylated benzylphenols, various alkyl-phenols and dialkylphenols. However, potential ERAs present in some active fractions remain unidentified.

Conclusions And Outlook: Our results show that SPMD-YES-based BDCA can be used to detect and identify readily bioconcentratable ERAs in surface waters. As such, this biomimetic approach can be employed as an alternative to conventional methodologies to provide investigators with a more environmentally relevant insight into the distribution and identity of ERAs in surface waters. The use of alternative bioassays also has the potential to expand SPMD-based BDCA to include a wide range of toxicological endpoints. Improvements to the analytical methodology used to identify ERAs or other target compounds in active fractions in the current study could greatly enhance the applicability of the methodology to risk assessment and monitoring programmes.