Identification of potentially toxic compounds in complex extracts of environmental samples using gas chromatography-mass spectrometry and multivariate data analysis

• Published in: Environmental toxicology and chemistry Vol. 26; no. 2; pp. 208 - 217
• Main Authors: Bergknut, Magnus, Kucera, Adam, Frech, Kristina, Andersson, Erika, Engwall, Magnus, Rannug, Ulf, Koci, Vladimir, Andersson, Patrik L., Haglund, Peter, Tysklind, Mats
• Format: Journal Article
• Language: English
• Published: Hoboken Wiley Periodicals, Inc 01.02.2007; Blackwell Publishing Ltd
• Subjects: Algal growth; Ash; Bioassay; Biological effects; Carbon; Chemicals; Chemistry; Data analysis; Desmodesmus subspicatus; Environmental chemistry; Environmental Pollutants - toxicity; Environmental toxicology; Folsomia candida; Gas chromatography; Gas Chromatography-Mass Spectrometry - methods; Hordeum vulgare; Incinerators; Kemi; Mass spectrometry; Miljökemi; Miljötoxikologi; Multivariate Analysis; Multivariate techniques; Municipal solid waste; NATURAL SCIENCES; NATURVETENSKAP; Plant growth; Polycyclic aromatic compounds; Polycyclic aromatic hydrocarbons; Preservatives; Salmonella; Sinapis alba; Soil; Soils; Vibrio fischeri; Wood preservatives
• ISSN: 0730-7268; 1552-8618; 1552-8618
• DOI: 10.1897/06-204R.1

—In this study, we examined 31 samples of varying chemical composition, including samples of soils from gasworks, coke production sites, and sites where wood preservatives were heavily used; ash and soot from municipal solid waste incinerators; antiskid sand; and dust from areas with heavy road traffic. The samples were comprehensively chemically characterized, especially their polycyclic aromatic compound contents, using gas chromatography‐time‐of‐flight mass spectrometry, whereas their biological effects were assessed using dehydrogenase activity, root growth (Hordeum vulgare), reproduction of springtails (Folsomia candida), algal growth (Desmodesmus subspicatus), germinability (Sinapis alba), Vibrio fischeri, DR‐CALUX, and Ames Salmonella assays. The number of compounds detected in the samples ranged from 123 to 527. Using the multivariate regression technique of partial‐least‐squares projections to latent structures, it was possible to find individual compounds that exhibited strong correlations with the different biological responses. Some of the results, however, indicate that a broader chemical characterization may be needed to identify all the compounds that may cause the measured biological responses.