PREDICTING ACTIVITY OF PROTOPORPHYRINOGEN OXIDASE-INHIBITORS BY COMPUTER-AIDED MOLECULAR MODELING

Protoporphyrinogen oxidase (Protox) is the primary site of action of herbicides belonging to at least ten chemical classes. Structure-activity relationships studies of diphenyl ether and phenopylate herbicides have shown the bicyclic structure of Protox inhibitors to approximate one half of the enzyme substrate, protoporphyrinogen IX. We determined the effect of a member of each of ten different chemical classes of Protox inhibitors on Protox activity and ability to induce protoporphyrin IX accumulation and to cause Light-induced cellular leakage. Molecular properties determined with MOPAC of these compounds were then correlated with these biological activities by regression analysis. The regression equations based on the bulk, electronic, energy, and lipophilicity properties accounted for most of the variation in the three biological activities, and appear to be specific to the particular biological activity. Predictive equations generated by this method were then tested on enantiomers with chirality in the meta-substitution of the phenyl ring. The equations were somewhat predictive for active compounds, but were not predictive for the inactive compounds.