Application of preparative capillary gas chromatography (pcGC), automated structure generation and mutagenicity prediction to improve effect-directed analysis of genotoxicants in a contaminated groundwater.

BACKGROUND, AIM AND SCOPE: The importance of groundwater for human life cannot be overemphasised. Besides fulfilling essential ecological functions, it is a major source of drinking water. However, in the industrial area of Bitterfeld, it is contaminated with a multitude of harmful chemicals, including genotoxicants. Therefore, recently developed methodologies including preparative capillary gas chromatography (pcGC), MOLGEN-MS structure generation and mutagenicity prediction were applied within effect-directed analysis (EDA) to reduce sample complexity and to identify candidate mutagens in the samples. A major focus was put on the added value of these tools compared to conventional EDA combining reversed-phase liquid chromatography (RP-LC) followed by GC/MS analysis and MS library search.

MATERIALS AND METHODS

We combined genotoxicity testing with umuC and RP-LC with pcGC fractionation to isolate genotoxic compounds from a contaminated groundwater sample. Spectral library information from the NIST05 database was combined with a computer-based structure generation tool called MOLGEN-MS for structure elucidation of unknowns. Finally, we applied a computer model for mutagenicity prediction (ChemProp) to identify candidate mutagens and genotoxicants.

RESULTS AND DISCUSSION

A total of 62 components were tentatively identified in genotoxic fractions. Ten of these components were predicted to be potentially mutagenic, whilst 2,4,6-trichlorophenol, 2,4-dichloro-6-methylphenol and 4-chlorobenzoic acid were confirmed as genotoxicants.

CONCLUSIONS AND PERSPECTIVES

The results suggest pcGC as a high-resolution fractionation tool and MOLGEN-MS to improve structure elucidation, whilst mutagenicity prediction failed in our study to predict identified genotoxicants. Genotoxicity, mutagenicity and carcinogenicity caused by chemicals are complex processes, and prediction from chemical structure still appears to be quite difficult. Progress in this field would significantly support EDA and risk assessment of environmental mixtures.