Many important antibiotics such as tetracyclines, erythromycin, adriamycin, monensin, rifamycin and avermectins are polyketides. In their biosynthesis, multifunctional synthases catalyse iterated condensation of thio-esters derived from acetate, propionate or butyrate to yield aliphatic chains of varying length and carrying different alkyl substituents. Subsequent modifications, including aromatic or macrolide ring closure or specific methylations or glycosylations, generate further chemical diversity. It has been suggested that, if different polyketide synthases had a common evolutionary origin, cloned DNA coding for one synthase might be used as a hybridization probe for the isolation of others. We show here that this is indeed possible. Study of a range of such synthase genes and their products should help to elucidate what determines the choice and order of condensation of different residues in polyketide assembly, and might yield, by in vitro recombination or mutagenesis, synthase genes capable of producing novel antibiotics. Moreover, because genes for entire antibiotic pathways are usually clustered in Streptomyces, cloned polyketide synthase genes are valuable in giving access to groups of linked biosynthetic genes.

Antibiotic non-producing mutants were isolated from an oxytetracycline (OTC) producing strain of Streptomyces rimosus. Cosynthesis tests and feeding known intermediates of the OTC pathway allowed the classification of the mutants into several groups. The biosynthetic lesion was determined for several of the mutants. Some of the mutants deficient in the reduction of 5a,l la-dehydrooxytetracycline, the final step in the pathway, were unable to synthesize a cofactor (CSF 1) essential for this reaction. Mutations apparently in genes for enzymes of the main OTC pathway were found for three of a possible four steps between 4-aminodedimethylaminoanhydrotetracycline (4-amino-ATC) and OTC. Mutations affecting three other steps of a possible five before 4-amino-ATC were found, but unambiguous identification of these was not possible. The genes for OTC production, including those for CSFl synthesis, were located in two diametrically opposite clusters on the S. rimosus chromosomal linkage map. No evidence for plasmid-borne genes was obtained.

The isolation of mutants of Streptomyces rimosus which were blocked in oxytetracycline (OTC) production was described previously. The genes for the early steps of antibiotic biosynthesis mapped together. Genomic DNA fragments of S. rimosus which conferred resistance to OTC and complemented all of these non-producing mutants have been cloned. The cloned DNA is physically linked within approximately 30 kb of the genome of S. rimosus. The gene cluster is flanked at each end by a resistance gene each of which, independently, can confer resistance to the antibiotic. In OTC-sensitive strains of S. rimosus, the entire gene cluster including both resistance genes has been deleted. Complementation of blocked mutants by cloned DNA fragments in multi-copy vectors was often masked by a secondary effect of switching off antibiotic production in strains otherwise competent to produce OTC. This adverse effect on OTC production was not observed with recombinants using low copy-number vectors.

The anhydrotetracycline (ATC) oxygenase enzyme which carries out the conversion of ATC to dehydrotetracycline was purified and the N-terminal amino acid sequence was determined. The sequence displays a significant similarity to that of the p-hydroxybenzoate hydroxylase from Pseudomonas fluorescens. This is consistent with the activity of the oxygenase, i.e., addition of a hydroxyl moiety to an aromatic ring structure.Oligonucleotide probes were designed and used to clone the corresponding fragment of chromosomal DNA from Streptomyces rimosus. This DNA fragment was used to screen a cosmid library, allowing the isolation of flanking DNA sequences. Surprisingly, the gene was located within the previously cloned cluster of genes involved in the synthesis of the biosynthetic intermediate ATC and not as had been expected (P. M. Rhodes, N. Winskill, E. J. Friend, and M. Warren, J. Gen. Microbiol. 124:329-338, 1981) at a separate locus on the other side of the chromosome. Subcloning of an appropriate DNA fragment from one of the cosmid clones onto pIJ916 produced Streptomyces lividans transformants which synthesized oxytetracycline.Oxytetracyline (OTC) is a commercially important broadspectrum antibiotic produced by Streptomyces rimosus. OTC is a member of the polyketide class of antibiotics which includes, among others, avermectin, monensin, erythromycin, and other tetracyclines such as 7-chlorotetracycline.The OTC biosynthetic pathway of S. rimosus has been the subject of both biochemical and genetic analysis (15). Figure 1 indicates the proposed biosynthetic pathway from malonyl coenzyme A to the final product. The isolation and characterization of mutants blocked at various points in the pathway indicated that the genes involved in OTC biosynthesis are clustered into two distinct positions mapping at diametrically opposite locations on the circular map of S. rimosius.The loci thought to be responsible for all pathway steps leading to the production of anhydrotetracycline (ATC) were assigned to the "4 o'clock" locus, whereas the genes encoding enzymes for the final part of the pathway, including those responsible for the production of a flavinlike cofactor (cosynthetic factor 1 [CSF1]) were assigned to a 10 o'clock location.The isolation (3; M. J. Butler, E. J. Friend, I. S. Hunter, F. S. Kaczmarek, D. A. Sugden, and M. Warren, Mol. Gen. Genet., in press) of two genes, otrA and otrB, involved in host resistance to OTC and the analysis of flanking DNA indicated that the 4 o'clock cluster was organized as a single group of biosynthetic genes flanked by the two resistance loci (Fig. 2). However, the genes responsible for the conversion of ATC to OTC remained unidentified. This paper describes the isolation of the gene (otcC) encoding the ATC oxygenase, the enzyme catalyzing the hydroxylation of ATC to dehydrotetracycline. otcC (7) except that S. lividans mycelium was grown for 3 days instead of 2 days in YEME medium containing glycine and MgCl2 at 0.5% and 5 mM, respectively. S. rimosus protoplast formation, regener...

No abstract