Crop Protection 19 (2000) 583}589 Natural products as sources for new mechanisms of herbicidal action Stephen O. Duke*, Joanne G. Romagni, Franck E. Dayan United States Department of Agriculture, Agricultural Research Service, Natural Products Utilization Research Unit, P.O. Box 8048, University, MS 38677, USA Abstract New mechanisms of action for herbicides are highly desirable to "ght evolution of resistance in weeds, to create or exploit unique market niches, and to cope with new regulatory legislation. Comparison of the known molecular target sites of synthetic herbicides and natural phytotoxins reveals that there is little redundancy. Comparatively little e!ort has been expended on determination of the sites of action of phytotoxins from natural sources, suggesting that intensive study of these molecules will reveal many more novel mechanisms of action. Examples of natural products that inhibit unexploited steps in the amino acid, nucleic acid, and other biosynthetic pathways are given. AAL-toxin, hydantocidin, and various plant-derived terpenoids are discussed. Strategies and potential pitfalls of using natural products as leads for new herbicide classes are summarized. Published by Elsevier Science Ltd. Keywords: Natural products; Herbicides; Mode of action 1. Introduction In the 60 years since its introduction, the area of chemical weed control has been a dynamic research "eld requiring constant innovation. The need for better tools, possessing broader weed control spectra and proper en- vironmental behavior, intertwined with an increasing e!ort to cope with evolved weed resistance to herbicides, has primarily been addressed by massive synthetic chem- istry programs generating numerous compounds. These molecules are submitted to biological and toxicological screens to identify potential lead compounds that are subsequently taken through rigorous structure optimiza- tion processes to develop commercial products. This approach has generated a tremendously large number of biologically active molecules that a!ect a rela- tively small number of target sites in plants. The advent of genetic engineering is currently revolutionizing this paradigm by enabling the use of non-selective herbicides on crops that have been genetically altered to be resistant to the compounds (Duke, 1996). Resistance can be achieved either by introducing an enhanced mechanism of degradation of the herbicide and/or modifying the * Corresponding author. Tel.: #1-662-915-1036; fax: #1-662-915- 1035. E-mail address: sduke@olemiss.edu (S.O. Duke). target site. While the idea is appealing because of obvi- ous bene"ts, such as ease of use and decreased envir- onmental impact of the herbicides, problems are emerging. Agricultural practices relying on the use of herbicide-resistant crops are leading to shifts in weed populations to naturally resistant species (Owen, 1997). Another problem involves the lack of acceptance of transgenic crops by some countries. Therefore, the need for new chemical weed control tools continues to be pressing. As mentioned above, commercial herbicides have a limited number of target sites. The use of natural products as herbicides or as lead structures for herbicide discovery programs is an alternative that has not been exploited as fully as it has been for insecticides and fungicides. Phytotoxic natural products are in general structurally more complex than synthetic herbicides and would not have been obtained by traditional synthetic approaches that tend to be limited by the cost of the synthesis of the "nal molecule. Henkel et al. (1999) dem- onstrated in a comparison between biologically active synthetic and natural molecules that natural products generally have higher molecular weights and more struc- tural complexity than synthetic compounds, and `heavya atoms such as halogens are seldom present. On the other hand, natural products have a greater proportion of oxygen and nitrogen, than most synthetic compounds. The diversity found in natural products is slowly being 0261-2194/00/$ - see front matter. Published by Elsevier Science Ltd. PII: S 0 2 6 1 - 2 1 9 4 ( 0 0 ) 0 0 0 7 6 - 4