The biosynthesis of the pyrimidine moiety and the uronic acid moiety of the polyoxins and the formation of unnatural polyoxins has been studied in Streptomyces cacaoi. Experimental evidence is provided for the biosynthesis of thymine via a pathway that is independent of thymidylate synthetase. This new thymine pathway is based on two experimental approaches. First, two known inhibitors of DNA synthesis (1-formylisoquinoline thiosemicarbazide and 5-fluoro-2′-deoxyuridine), when added to polyoxin-producing cultures of S. cacaoi, inhibit the synthesis of TMP from exogenously supplied uracil but do not inhibit the synthesis of the thymine or hydroxymethyluracil in the polyoxin complex. Second, exogenously supplied thymine and hydroxymethyluracil are taken up by S. cacaoi but are not incorporated into the thymine or hydroxymethyluracil of the polyoxin complex. The thymine is incorporated into the DNA. The uracil in polyoxin L could be the parent pyrimidine chromophore with C-1 additions occurring at carbon-5 to form thymine and hydroxymethyluracil. Carbon-3 of serine but not the methyl group of methionine is a one-carbon source for the formation of the thymine and hydroxymethyluracil in the polyoxin complex. S. cacaoi can synthesize unnatural polyoxins, as evidenced by the incorporation of 5-fluoro, 5-bromo, and 6-azauracil into the polyoxins; 5-iodo-, 2-thio-, or 4-thiouracil is not a substrate. Two new polyoxin analogs synthesized and characterized when 5-fluorouracil is added to the cultures are 5-fluoropolyoxin L and 5-fluoropolyoxin M. There is a marked change in the molar ratio of the uracil:thymine:hydroxymethyluracil chromophores in the polyoxin complex following the incorporation of 5-fluoro-, 5-bromo-, or 6-azauracil. Apparently, the unnatural polyoxins inhibit the addition of the C-1 unit to carbon-5 of uracil in the polyoxin complex. Polyoxin L and polyoxin C do not inhibit Escherichia coli and Streptococcus faecalis, but 5-fluoropolyoxin L and 5-fluoropolyoxin C inhibit both these organisms. There is little or no difference in the inhibition of the fluorinated and natural polyoxins against leukemia L-1210 cells. The fluoro group on carbon-5 of the uracil ring does not affect the enzyme-inhibition complex with chitin synthetase since the inhibition constant of fluoropolyoxins L is the same as has been reported for polyoxins A, D, and L. The 14C-labeling pattern in the 5′-amino-5′-deoxy- d-allofuranosyluronic acid moiety of the polyoxins from 14C-labeled glucose, allose, and glycerol suggests that the formation of this unique C-6 uronic acid in the polyoxins does not proceed via the direct oxidation of either d-glucose or d-allose to the -onic or -uronic acids. Glucose is converted to two three-carbon trioses, followed by either (i) the oxidation of one of the trioses to a threecarbon acid and subsequent condensation with another three-carbon sugar to form the C-6 uronic or (ii) an 80:20 equilibrium of the two trioses followed by condensation to a hexose which is then oxidized to the C-6 uronic acid.