QUANTITATIVE STRUCTURE-ACTIVITY-RELATIONSHIPS FOR CHEMICAL TOXICITY TO ENVIRONMENTAL BACTERIA

Quantitative structure-activity relationships (QSARs) were developed for nonreactive chemical toxicity to each of four groups of bacteria of importance in environmental engineering: aerobic heterotrophs, methanogens, Nitrosomonas, and Microtox. The QSARs were based on chemicals covering a range of structures and including important environmental pollutants (i.e., chlorinated and other substituted benzenes, phenols, and aliphatic hydrocarbons). QSARs were developed for each chemical class and for combinations of chemical classes. Three QSAR methods (groups of chemical describing parameters) were evaluated for their accuracy and ease of use: log P, linear solvation energy relationships (LSER), and molecular connectivity. Successful QSARs were found for each group of bacteria and by each method, with correlation coefficients (adjusted r2) between 0.79 and 0.95. LSER QSARs incorporated the widest range of chemicals with the greatest accuracy. Log P and molecular connectivity QSARs are easier to use because their parameters are readily available. Outliers from the QSARs likely due to reactive toxicity included acryls, low pKa compounds, and aldehydes. Nitro compounds and chlorinated aliphatic hydrocarbons and alcohols showed enhanced toxicity to the methanogens only. Chemicals with low IC50 concentrations (log IC50 μmol/liter < 1.5) were often outliers for Nitrosomonas. QSARs were validated statistically and with literature data. A suggested method is provided for use of the QSARs. © 1991.