Structural alerts for the excess toxicity of acrylates, methacrylates, and propiolates derived from their short-term and long-term bacterial toxicity.

For eight acrylates, three methacrylates, and three propiolates as three subclasses of α,β-unsaturated esters, short-term and long-term bacterial toxicity with Vibrio fischeri was determined in terms of EC(50) (effective concentration 50%) values for the 30-min bioluminescence and 24-h growth inhibition. To this end, experimental exposure concentrations were corrected for volatilization through a thermodynamic model based on Henry's law constant of the compounds. Moreover, toxicity enhancements T(e) as the ratio of narcosis-predicted over actual EC(50) were determined and discussed in terms of underlying mechanisms of reaction of the electrophiles with endogenous nucleophiles such as glutathione (GSH) and proteins. Overall, log EC(50) [M] ranges from -2.28 to -3.70 (30 min) and from -2.80 to -5.28 (24 h), respectively, indicating a significantly larger sensitivity of the growth inhibition bioassay for the reactive toxicity of these Michael acceptors. The latter is also reflected in the observed toxicity enhancements, where log T(e) > 1 was obtained for only 5 of 14 30-min EC(50) values but for 11 of 13 24-h EC(50) values. Moreover, the average long-term to short-term difference in log T(e) is 1 unit for the acrylates and 0.7 units for both methacrylates and propiolates. Methacrylates exert narcosis-level toxicity except for the methyl derivative in the long-term assay. The log EC(50) (24 h) of a subset of 10 mostly excess-toxic acrylates and a propiolate correlates with their logarithmic rate constants of reaction with GSH, log k(GSH), significantly more than with log K(ow) (r(2) 0.76 vs 0.47), yielding a respective regression rms of 0.34 log units. For allyl and propargyl acrylate as well as propargyl methacrylate, the observed excess toxicity is likely caused by initial enzymatic hydrolysis and subsequent oxidation of the α,β-unsaturated alcohols to the respective carbonyls. The latter shows that in the context of nonanimal testing schemes such as for REACH, the metabolic capacity of in vitro screens requires attention.