Characterization of toxic compound in thermally oxidized oil.

Abstract

AbstractIn preliminary experiments, soybean oil was heated at 275 C for 12 hr in the presence of N2 or air. Feeding studies with rats showed that the oil heated in the presence of air, oxidatively polymerized oil, retarded weight gain more than that heated with N2, thermally polymerized oil. Oxidatively polymerized oil was fractionated with silicic acid column chromatography, and the fraction eluted with ether (fraction III) proved to be most toxic to mice. For chemical studies combined with bioassays, additional oxidatively polymerized oil was prepared by aeration at 185 C for 90 hr, and the product was fractionated by silicic acid column chromatography. The material originally eluted with ether (fraction III) was eluted in stages, and fraction IIIc proved to be most toxic in a mouse bioassay. IR, UV and NMR analyses did not indicate the presence of C−O−O−C or C−O−C linkages or aldehyde groups or aromatic compounds. Fraction IIIc was converted to its methyl esters and molecularly distilled to yield two fractions: one containing unpolymerized fatty acid esters and one giving evidence of more functional groups per molecule and of dimeric material. This “dimeric” fraction was more toxic to mice. IR analysis of this fraction revealed carbonyl groups, and NMR showed several functional groups on alkyl chains. Treatment with sodium borohydride and with hydroiodic acid revealed no C−O−O−C and no C−O−C linkages. Mass spectrography showed peaks at mass 586 and 293; in the reduced fraction a peak also occurred at mass 143, suggestive of cleavage of some of the chains. It is tentatively concluded that a highly toxic material formed in oil heated in air is a dimer of triglyceride molecules.