The polyketide nonactin, a polyketide possessing antitumor and antibacterial activity, is produced by an unusual biosynthesis pathway in Streptomyces griseus that uses both enantiomers of the nonactin precursor, nonactic acid. Despite many studies with labeled precursors, much of the biosynthesis pathway remains unconfirmed, particularly the identity of the last achiral intermediate in the pathway, which is believed to be 4,6-diketoheptanoyl-CoA. We set out to confirm the latter hypothesis with feeding studies employing [4,5-(13)C(2)]-, [5,6-(13)C(2)]-, and [6,7-(13)C(2)]-4,6-diketoheptanoate thioester derivatives. In each case the isotopic label was incorporated efficiently into nonactin; however, at positions inconsistent with the currently accepted biosynthesis pathway. To resolve the discrepancy, we conducted additional feeding studies with a [3,4-(13)C(2)]levulinate thioester derivative and again observed efficient label incorporation. The latter result was intriguing, as levulinate is not an obvious precursor to nonactin. Levulinate, however, is known to be efficiently degraded into propionate even though the pathway for the conversion is not known. On the basis of both our levulinate and diketoheptanoate isotope incorporation data we can now postulate a pathway from levulinate to propionate that can also account for the conversion of 4,6-diketoheptanoate into levulinate in S. griseus.