Acetolactate Synthase and Ketol-Acid Reductoisomerase: Targets for Herbicides Obtained by Screening and de novo Design

Several major classes of herbicides, discovered by conventional screening techniques, have been found to inhibit the first common enzyme of branched-chain amino acid biosynthesis, acetolactate synthase, as their mode of action. These herbicides seem to bind to an evolutionary vestige of a quinone-binding site, extraneous to the active site, that is present due to the evolutionary history of this enzyme. Besides their herbicidal effect on sensitive plants, these compounds can effect stasis in the growth of bacteria and yeast. Recently is has been reported that an experimental herbicide from Hoechst. Hoe 704. that was discovered by conventional screening techniques, inhibits the second common enzyme of branched-chain amino acid bio- synthesis [Schultz etal., FEBS Lett. 238, 375-378 (1988)]. We have also recently designed novel reaction-intermediate analogs (e.g. N-isopropyl oxalylhydroxamate) that arc exceptionally potent (Ki = 22 pM: half-time for release approximately six days) and selective inhibitors of the second common enzyme, ketol-acid reductoisomerase. Both of these selective inhibitors of the second common enzyme will kill sensitive plants, but will only inhibit the growth (without killing) of bacteria. The effects in bacteria parallel those obtained by mutations in the relevant genes, where loss of either the first or second common enzyme in the pathway gives an organ- ism that is auxotrophic for branched-chain amino acids, but does not result in a conditionally lethal phenotype. Higher plant mutants have only been obtained to date that arc deficient in functional leucine-specific gene products (as yet uncharacterized), threonine deaminase (isoleucine specific), and dihydroxyacid dehydratase (common). The phenotypes of these mutants. at least at the level of cell culture, are similar to those of their bacterial counterparts, in that auxotrophy, but not conditional lethality, is obtained. These results highlight the potential non-equality of the enzymes of branched-chain amino acid biosynthesis as targets in herbicide design.