Biosynthesis of rotenoids by Amorpha fruticosa: sequence, specificity, and stereochemistry in the development of the hemiterpenoid segment

Abstract

By isolation and radiochemical methods rot-2′-enonic acid, dalpanol, rotenone, and amorphigenin have been identified in Amorpha fruticosa seedlings, and ordered in biosynthetic sequence. Its specific activity shows that 12aβ-hydroxyamorphigenin has origins other than direct 12aβ-hydroxylation of amorphigenin: its occurrence and labelling establish it as a true natural product.

Except for isopentenyl alcohol, the potential hemiterpene precursors mevalonic acid, 3-hydroxy-3-methylglutaric acid, and dimethylallyl alcohol are poor precursors for amorphigenin. By employing the already prenylated (E)-[4′-¹⁴C]rot-2′-enonic acid, it is shown that the 4′-C of this compound becomes 7′-C of rotenone. By assuming normal rear-side attack on an intermediate epoxide, and utilising the known absolute configurations at position 5′ of dalpanol and rotenone, a stereochemical sequence can be written. (E)-4′-Labelled rot-2′-enonic acid leads to a (2′S,3′S)-epoxide, which on intramolecular attack by phenolate anion would give (5′R,6′S)-dalpanol, dehydration to rotenone then involving the labelled (pro-S)-7′-methyl group of dalpanol.

Neither (6′R)- nor (6′S)-amorphigenol is a precursor of amorphigenin. Administration of [7′-¹⁴C]rotenone to A. fruticosa seedlings has led, in three experiments designed to avoid inadvertent chemical scrambling of the allylic label, to amorphigenin having even label distribution between C-7′ and C-8′. Possible interpretations are considered.