GLUTATHIONE TRANSFERASE ACTIVITIES TOWARD HERBICIDES USED SELECTIVELYIN SOYBEAN

Using extracts from suspension-cultured cells of soybean (Glycine max cv. Mandarin) as a source of active enzymes, the activities of glutath ione transferases (GSTs) catalysing the conjugation of 1-chloro-2,4-di nitrobenzene (CDNB) and selective herbicides were determined to be in the order CDNB much greater than fomesafen > metolachlor = acifluorfen > chlorimuron-ethyl. GST activities showed a thiol dependence in a su bstrate-specific manner. Thus, GST activities toward acifluorfen and f omesafen were greater when homoglutathione (hGSH), the endogenously oc curring thiol in soybean, was used as the co-substrate rather than glu tathione (GSH). Compared with GSH, hGSH addition either reduced or had no effect on GST activities toward other substrates. In the absence o f enzyme, the rates of hGSH conjugation with acifluorfen, chlorimuron- ethyl and fomesafen were negligible, suggesting that rapid hGSH conjug ation in soybean must be catalysed by GSTs. GST activities were subseq uently determined in 14-day-old plants of soybean and a number of annu al grass and broadleaf weeds. GST activities of the plants were then r elated to observed sensitivities to postemergence applications of the four herbicides. When enzyme activity was expressed on a mg-l protein basis, all grass weeds and Abutilon theophrasti contained considerably higher GST activity toward CDNB than soybean. With fomesafen as the s ubstrate, GST activities were determined to be in the order soybean mu ch greater than Echinochloa crus-galli > Digitaria sanguinalis > Solgh um halepense = Setaria faberi with none of the broadleaf weeds showing any activity. This order related well to the observed selectivity of fomesafen, with the exception of A. theophrasti, which was partially t olerant to the herbicide. Using metolachlor as the substrate the order of the GST activities was soybean > A. theophrasti much greater than S. halepense > Amaranthus retroflexus > Ipomoea hederacea, with the re maining species showing no activity. GST activities toward metolachlor correlated well with the selectivity of the herbicide toward the broa dleaf weeds but not toward the grass weeds. Acifluorfen and chlorimuro n-ethyl were selectively active on these species, but GST activities t oward these herbicides could not be detected in crude extracts from wh ole plants.