1. Analysis of genes coding for a variety of structurally different fatty acid synthases (FASs) and polyketide synthases (PKSs) from selected Gram positive bacteria of industrial importance;
2. Analysis of temporal (and where relevant, spatial) patterns of transcription for these gene sets, and of the effects of targetted gene disruption on cell physiology and secondary metabolite production;
3. Characterisation and manipulation of novel activator genes within or outside polyketide synthase biosynthetic gene clusters, which appear to exert specific effects on polyketide production;
4. Analysis of the molecular mechanisms by which these cluster-specific activator genes interact with their target to increase transcription, and by which they respond to normal cellular signals during development.
Objective 1: Complete gene sequences were obtained for a second (newly-discovered) type I FAS from Brevibacterium ammoniagenes (Erlangen) and for the type II FAS of Streptomyces coelicolor (Norwich). The existence of two homologous type I FAS genes in B. ammoniagenes was a major surprise, while characterisation of the authentic FAS genes from S. coelicolor represented the successful culmination of a long term search. A substantial portion of a gene cluster from Streptomyces argillaceus (mtm), encoding the type II PKS for the important antitumour compound mithramycin, was cloned and sequenced (Oviedo). Also, part of a gene cluster from Streptomyces longisporoflavus (ten) which governs synthesis of the polyether tetronasin, was identified and shown to encode a type I PKS (Cambridge). The latter two findings significantly extend the range of polyketides for which we have information about the PKSs.
Objective 2: Insight into the function of these gene sets encoding FAS and PKS has been gained by targetted disruption. The S. coelicolor fas genes were found to be essential for growth. Of the two FAS multienzymes in B. ammoniagenes, FasA is the predominant form,and fasA disruptants require fatty acids for growth, and particularly oleic acid; while FasB appears dispensable and fasB disruptants have no fatty acid requirement. The identity of the mtm gene cluster was demonstrated by insertional inactivation. The identity of the ten gene cluster still rests on its propinquity on the S. longisporoflavus chromosome to a gene (tenC) (vide infra) which confers resistance to tetronasin on S. lividans.
The potential for 'cross-talk' in the same organism between the products of fas and pks genes is illustrated by the work on S. coelicolor (Norwich). The FAS ACP gene acpP was cloned in place of the ACT ACP gene actI-ORF3. Despite this co-expression, pigmented polyketide production was drastically reduced. This demonstrated for the first time that FAS and PKS co-existing in the same cell can be prevented from unwanted crosstalk by a simple biochemical incompatibility (e.g. loss of proper protein:protein interactions). In contrast, in vitro experiments using cloned genes expressed in E. coli indicated that the FAS malonyl acyltransferase, and possibly also the FAS ketoacyl-ACP synthase III, may be common to all three pathways, FAS, ACT and WhiE, which has important implications for their co-ordinated control.
Objective 3: An actII-ORF4 analogue from the mtm cluster, mtmR, was shown to trigger actinorhodin production in S. lividans, behaviour also seen for actII-ORF4, cross-talk not seen with redD. The tenC of S. longisporoflavus was found to resemble afsR of S. coelicolor over most of its length, and actII-ORF4 and redD in the N-terminal part. Jointly between Norwich and Cambridge, it was shown that a deletion mutant of S. coelicolor lacking both redD and actII-ORF4 was not complemented by tenC. When tenC was introduced into an afsR deletion mutant of S. coelicolor, which is known to be conditionally defective in ACT and RED production, there was on certain media a stimulation of RED production. TenC represents a wholly new mechanism of antibiotic resistance, mediated by an afsR-like pleiotropic activator. Remarkably, in-frame deletion of the (actII-ORF4-like) N-terminal portion of tenC does not affect the ability to confer resistance to tetronasin on S. lividans TK64.
Objective 4: The pathway-specific activators of the ActII-orf4 family do not show a recognisable DNA-binding motif. However, it was successfully shown for the first time that, at least for ActII-ORF4 itself, the activator protein appears to bind to two intergenic regions in the act cluster, consonant with a direct effect on transcription (Madrid). The analogous RedD in the RED pathway of S. coelicolor does not bind tightly to the red DNA under these conditions (Norwich), hinting at subtle differences in the mode of action of these two activators.
Another important link was forged between the pathway-specific activators and the response to normal cellular signals in development: independently, in both Norwich and Madrid, the gene from S. coelicolor relA encoding the regulatory enzyme (p)ppGpp synthetase was cloned and sequenced. Under some conditions, (p)ppGpp synthetase appears to be essential for triggering antibiotic production via its effect on redD and actII-ORF4.
Funding SchemeCSC - Cost-sharing contracts