Submitted to: Pesticide Science
Publication Type: Peer reviewed journal
Publication Acceptance Date: 3/20/2000
Publication Date: N/A
Citation: Interpretive Summary: Predicting the activity of natural products is not easily achieved because there is often not enough compounds to obtain a reliable result. Cyperine, a natural product from several plant fungal pathogens has a chemical structure similar to many commercial herbicides. We used a modern 3-dimensional computer modeling technique called Comparative Molecular Field Analysis (CoMFA) to predict the biological activity of this natural plant toxin using models developed with 31 synthetic analogues. We were able to accurately predict the activity of cyperine and the analyses suggest that the binding of the natural herbicide may be specific for one side of the natural herbicide.
Technical Abstract: Protoporphyrinogen oxidase (Protox) is the target site of a large number of commercial herbicides. Previous QSARs performed at a 2-dimensional (2-D) level reproduced the activity of individual data sets relatively well, but these models could not be used to predict the activity of structurally related derivatives. We developed a more reliable model by applying 3-dimensional (3-D) molecular techniques to a set of 31 phenyl ether (PE) analogues. Inhibitory activity at the molecular site of action was chosen because it circumvents the effects of uptake, translocation, and metabolism of the compounds occurring in whole plant studies. Increased predictability was achieved by aligning the diphenyl ether (DPE) analogues along the trifluoromethyl phenyl (q2=0.70) ring rather than along the nitrophenyl (q2=0.65) ring or along the centroids (q2=0.69). This new analysis differentiated between R and S enantiomers and allowed the prediction of the inhibitory activity of the natural diphenyl ether cyperine. The prediction model suggests that the binding of cyperine on the active site of Protox is stereospecific.