Strategic Strain Improvement of Antibiotic Producer.

Abstract

We have applied knowledges obtained during studies of antibiotic biosynthesis to improve antibiotic producers as follows; (1) The increase in antibiotic productivity and the production of a new component by gene dosage effect: (2) The production of new analogs of antibiotics by the introduction of heterogeneous biosynthetic gene(s): (3) Selective production of useful components by inactivation or activation of genes involving in antibiotic production.Mutants that do not produce antibacterial, antifungal and antimycoplasma antibiotic kalafungin were isolated from kalafungin-producing Streptomyces tanashiensis and classified into seven distinct biosynthesis phenotypes. By colony hybridization with the genes for polyketide synthase (actI and III) as probes, a clone carrying a large DNA fragment was picked up from a genomic library of this strain. In this DNA fragment, the gene cluster for kalafungin biosynthesis containing at least eight genes (kalI to VII, and a regulatory gene) were identified. Although kalafungin is an intermediate of actinorhodin biosynthesis in S. coelicolor A3(2), this gene cluster for kalafungin biosynthesis in S. tanashiensis was not the same as that in S. coelicolor. When the wild type strain and kal mutants of S. tanashiensis were transformed with the cloned DNA fragment, not only the increase in the production of kalafungin and dihydrokalafungin but also production of a new antimicrobial product named tetrahydrokalafungin were caused.The anthelmintic antibiotic avermectin complex produced by Streptomyces avermitilis is a family of eight closely related components, A1a, A1b, A2a, A2b, B1a, B1b, B2a and B2b. The component B1a is the most highly effective. However, since it is not easy to separate the “a” and “b” components in the industrial processes, the “B1” components B1a and B1b and their hydrogenated products are used as an important anthelmintic agent. After a mutagenesis, two kinds of mutants, K2021 and K2034, which produced only the specific components, were obtained. Strain K2034 lacked avermectin B2 5-O-methyltransferase activity and produced the “B” components, B1a, B1b, B2a and B2b. In the other mutant, K2021, the incorporation of L-isoleucine and its keto acid was efficient, but very little L-valine and its keto acid were incorporated into avermectins. This mutant produced the “a” components, A1a, A2a, B1a and B2a. The recombinant strains, which have both phenotypes of the above mutants, were designed by in vivo recombination. The strains produced the components B1a and B2a only. Genetic mapping showed that the locus of the mutation affecting the selectivity of the incorporation of branched-chain keto acids into the avermectin skeleton apparently is sufficiently distant from the gene cluster for avermectin biosynthesis.