Identification of toxic compounds in the aquatic environment
My name is Willem Jonker and I am a third-year PhD student. After I obtained my MSc degree in Analytical Chemistry at the VU University Amsterdam I started my PhD project in the AIMMS Division of Biomolecular Analysis. I work within the STW project “High Throughput Effect-Directed Analysis: a novel platform for rapid and sensitive identification of toxic compounds in the aquatic environment”.
A common problem for laboratories monitoring water quality is the ever increasing list of toxic pollutants that have to be monitored according to governmental regulations. Unfortunately, by screening for the presence of only preselected compounds, unknown toxicants will be missed. Therefore, water laboratories are searching for new ways to identify newly emerging compounds in water involving bioactivity as an essential parameter.
The project aims to develop a platform that efficiently combines bioassay testing with chemical analysis for the detection and identification of newly emerging toxicants in the aquatic environment. In environmental chemistry, this is called effect-directed analysis (EDA). Unfortunately, conventional EDA studies are time consuming and in many cases do not successfully elucidate the toxicant identity. This has to be improved in order to make EDA a useful approach for water laboratories. We developed two EDA platforms which are illustrated in the figure. The LC and GC platforms combine bioassay and chemical analysis with compound bioactivity and identity as simultaneous output.
The “LC platform” makes use of liquid chromatography (LC) as a separation technique to reduce the sample complexity for bioassay testing. After the column, the eluate is split towards a high-resolution mass spectrometer and a fraction collector. We are able to collect small fractions in microtiter plates maintaining chromatographic separation. After fraction collection, each fraction is tested with a bioassay of choice. The biological endpoints of interest for this project are estrogenicity, androgenicity, glucocorticoid and thyroid activity. We selected different mammalian gene reporter assays because of their selectivity and sensitivity required for trace level analysis. The readout of each well correlates to a retention time and a bioassay chromatogram can be reconstructed. This bioassay chromatogram can directly be correlated with the MS chromatogram and bioactives can be easily pinpointed. For subsequent toxicant identification the exact mass can be extracted from the MS data and is used for molecular formula calculation and database searching. We currently apply the platform for the detection of estrogenic and anti-estrogenic compounds in environmental water samples.We also developed a “GC platform” that allows us to analyze samples by gas chromatography (GC) and fractionate these into components. In contrast to previously developed setups where maximum 12 fractions can be collected, we are able to collect up to 384 fractions. The platform consists of a standard gas chromatograph with flame ionization detector (FID) and multipurpose autosampler. In step one, the sample is injected after which it enters the column. The end of the column is connected to a flow splitter that directs a small part towards an FID detector. The major part goes to a second split. At this split, a preheated volatile solvent (heated by a modified FID detector) is added that mixes with the column eluate and is guided via a capillary outside the GC oven. Once outside, the volatile solvent containing the separated analytes condenses and is collected in microtiter plates. The capillary is attached to a fraction collection tip that was picked up from its standby position (2) and moves with small time intervals from one well to another (3) and collects the whole separation. The platform has been characterized by analysis of an alkane mix (C7-C30). Different parameters such as injection volume, fractionation stability, recovery and the influence of different infusion solvents have been studied. After fractionation each fraction was reanalyzed by GCFID and the measured response was correlated to each fraction and plotted. Clearly octane was the most suitable infusion solvent and is capable to maintain the separation and in addition showed good recoveries. We envision applications for different research fields in analytical chemistry and are currently hyphenating the platform with a gene reporter assay specific for androgenic compounds.