Curacin A is a mixed polyketide/nonribosomal peptide from the marine cyanobacterium Lyngbya majuscula. The natural product has three unusual chemical groups: a cyclopropane ring, a cis double bond and a terminal alkene. Our studies of the Cur biosynthetic gene cluster focus on discovery of the enzymatic tools that synthesize these unusual groups. Cyclopropane formation is catalyzed by a halogenase, by four enzymes encoded within an “HCS cassette” in the Cur gene cluster, and by a cyclopropanase/ enoylreductase (Gu et al. Nature, in press). The enzyme for cis double bond formation is unknown due to a mismatch between catalytic domains predicted by sequence analysis of the Cur modular PKS and those needed in the predicted biosynthetic pathway (Chang et al. (2004) J. Nat. Prod. 67, 1356). Four dehydratase domains are available to catalyze five dehydration reactions. A dehydratase is missing from two modules, including the module where cis double bond formation is expected. A dehydratase exists in yet another module where no dehydration reaction is needed. Structural data are beginning to solve this puzzle. The terminal alkene is formed in a novel set of reactions catalyzed by the C-terminal domains of the CurM polypeptide. Structure-function studies of catalytic domains have revealed a novel strategy for polyketide chain termination and terminal alkene formation.