Abstract:
The aminocoumarin antibiotics novobiocin and clorobiocin are potent inhibitors of bacterial DNA gyrase produced by different Streptomyces strains. Since they are closely related in structure, and their biosynthetic gene clusters have been cloned and sequenced, they represent interesting starting compounds for the development of new antibacterial agents by genetic engineering. However, the functional analysis of biosynthetic genes and the availability of genetic tools for manipulation of producer strains are pre-requisites for such approaches.
Clorobiocin is more potent than novobiocin. It differs from the latter, clinically approved drug in the substitution pattern at C-8´ of the aminocoumarin moiety, carrying a chlorine atom instead of a methyl group, and in the presence of a 5-methylpyrrole-2-carboxyl moiety instead of a carbamoyl group at 3´´-OH of the deoxysugar unit. One aim of this thesis was the production of hybrid antibiotics, combining structural features of these two compounds. By gene inactivation, clo-hal was identified as the gene of the halogenase responsible for the introduction of the chlorine atom of clorobiocin. Expression of the methyltransferase gene novO in the clo-hal- mutant of the clorobiocin producer S. roseochromogenes led to the very efficient formation of a clorobiocin analog with -CH3 instead of -Cl at C-8´ (= novclobiocin 102). However, earlier attempts to generate a hybrid antibiotic with a chlorine atom at C-8´ and a carbamoyl group at 3´´-OH (novclobiocin 114) by conventional gene inactivation/gene expression experiments in the novobiocin producer S. spheroides had remained unsuccessful (Pojer and Heide, unpublished results), due to difficulties in the genetic manipulation of this natural producer. In order to circumvent this problem, the entire biosynthetic gene cluster of novobiocin was now expressed in S. coelicolor and S. lividans which are well characterized and easy to manipulate. S. coelicolor M512 derivatives produced at least five times more of the antibiotic than S. lividans TK24 and further investigations were carried out using the former strain as host.
Inactivation of the methyltransferase gene novO could be rapidly achieved in the cosmid by lambda-Red-mediated recombination in Escherichia coli. Subsequent integration of the novO- cosmid into the S. coelicolor genome, followed by co-expression of clo-hal, readily allowed the production of the desired hybrid antibiotic novclobiocin 114. BhaA, responsible for 3-chlorination of the beta-hydroxy-tyrosyl moieties of the glycopeptide antibiotic balhimycin, was unable to replace Clo-hal, suggesting that the two halogenases have different substrate specificities.
Comparison of the antibacterial activity of clorobiocin and novobiocin analogues with –Cl, –H or -CH3 at C-8´ showed that substitution with either -CH3 or -Cl was important for biological activity. -CH3 or -Cl led to comparable activity, whereas -H led to a strong loss in activity.
Moreover, functional proof for the role of NovG as a positive regulator of novobiocin biosynthesis was provided by genetic and biochemical approaches. Using the method described above, novG- mutants were created. These novG- mutants produced only 2% of the novobiocin amount formed by the S. coelicolor strains carrying the intact novobiocin cluster. The production could be restored by introducing an intact copy of novG into the mutant. The introduction of novG on a multicopy plasmid into the strain carrying the intact novobiocin cluster led to almost three-fold overproduction of the antibiotic, suggesting that novobiocin biosynthesis is limited by the availability of NovG protein. Furthermore, purified N-terminal His6 tagged NovG showed specific DNA-binding activity for the novG-novH and for the cloG-cloY intergenic regions of the novobiocin and clorobiocin biosynthetic gene cluster, respectively. By comparing the DNA sequences of the fragments binding NovG, conserved inverted repeats were found in both fragments, similar to those identified as the binding sites for StrR, the well-studied pathway-specific transcriptional activator of streptomycin biosynthesis. The consensus sequence for the StrR and the putative NovG binding sites is GTTCRACTG(N)11CRGTYGAAC. Therefore, NovG and StrR apparently belong to the same family of DNA-binding regulatory proteins.
Furthermore, inactivation of novE by gene replacement using PCR targeting proved that novE does not have an essential catalytic role in novobiocin biosynthesis, since some novobiocin was still produced by the mutants. This gene is likely to have a regulatory function, as a 25-fold drop in productivity was observed in the mutants.