Molecular genetic analysis reveals a putative bifunctional polyketide cyclase/dehydrase gene from Streptomycea coelicolor and Streptomyces violoceoruber, and a cyclase/O-methyltransferase from Streptomyces glaucescens

Genes for polyketide secondary metabolic pathways in microorganisms and plants

Recent advances in molecular genetics have led to the isolation, sequencing and functional analysis of genes encoding synthases that catalyse the formation of several classes of polyketides. The structures of the genes and their protein products differ strikingly in the various examples. For Streptomyces aromatic polyketides, exemplified by granaticin and tetracenomycin, the synthases correspond to Type II (bacterial and plant) fatty acid synthases in consisting of distinct proteins for such processes as condensation, acyl carrier function and ketoreduction. In contrast, for actinomycete macrolides such as erythromycin, similar catalytic functions are performed by a set of multifunctional proteins resembling Type I (animal) fatty acid synthases, but with every step in chain-building being catalysed by a different enzymic domain. Penicillium patulum has a simple Type I synthase for 6-methylsalicylic acid. For plant chalcones and stilbenes, a single small polypeptide acts as a condensing enzyme for carbon chain-building and may be unrelated to any of the other polyketide and fatty acid synthases. Thus, although these systems share a common general mechanism of chain assembly, they must differ in the ways that synthase 'programming' has evolved to determine chain length, choice of chain starter and extender units, and handling of successive keto groups during chain assembly, and so control the great diversity of possible chemical products.

A polyketide biosynthetic gene cluster from Streptomyces antibioticus includes a LysR-type transcriptional regulator

In the search for Type II polyketide synthases (PKSs) a DNA fragment was isolated from Streptomyces antibioticus ATCC 11891 (a producer of oleandomycin). DNA sequencing of the cloned fragment revealed six complete ORFs whose deduced products showed similarities to those of other genes known to be involved in polyketide biosynthesis. Several S. coelicolor strains mutated in different steps of actinorhodin biosynthesis (actI, actIII, actV(A) and actVII) were complemented by the cloned genes, suggesting that the isolated genes encode an aromatic polyketide of unknown structure and function. The cluster also contains a putative LysR-type transcriptional regulator (ORF0), which controls PKS gene expression in a heterologous host. DNA binding assays and transcriptional analysis suggest that the pathway-specific regulator for actinorhodin biosynthesis (actII-ORF4) is also involved in the expression of the cloned PKS in the host strain.

Domain analysis of the molecular recognition features of aromatic polyketide synthase subunits

Bacterial aromatic polyketide synthases (PKSs) are a family of homologous multienzyme assemblies that catalyze the biosynthesis of numerous polyfunctional aromatic natural products. In the absence of direct insights into their structures, the use of gene fusions can be a powerful tool for understanding the structural basis for their properties. A series of truncated and hybrid proteins were constructed and analyzed within a family of PKS subunits, designated aromatases/cyclases (ARO/CYCs). When expressed alone, neither the N-terminal nor the C-terminal domain of the actinorhodin (act) or the griseusin (gris) ARO/CYC exhibited substantial aromatase activity. However, in the presence of each other, the half proteins were active. Furthermore, analysis of a set of hybrid proteins derived from the act and gris ARO/CYCs allowed us to localize the chain length dependence of this aromatase activity to their N-terminal domains. Unexpectedly, however, when the C-terminal domain of the gris ARO/CYC was expressed in a context where aromatase activity was absent, it could modulate the chain length specificity of the tetracenomycin (tcm) minimal PKS, leading to the formation of a novel 18-carbon product in addition to the expected 20-carbon one. It was also found that monodomain ARO/CYCs such as tcmN cannot substitute for the the N-terminal domain of didomain ARO/CYCs, even though they exhibit high sequence similarity with the N-terminal domain. Together, these results illustrate the utility of protein engineering approaches for dissecting the structure-function relationships of PKS subunits and for the generation of mutant alleles with novel biosynthetic properties.

Heterologous expression of an engineered biosynthetic pathway: functional dissection of type II polyketide synthase components in Streptomyces species

Polyketides are an extensive class of secondary metabolites with diverse molecular structures and biological activities. A plasmid-based multicomponent polyketide synthase expression cassette was constructed using a subset of actinorhodin (act) biosynthetic genes (actI-orf1, actI-orf2, actI-orf3, actIII, actVII, and actIV) from Streptomyces coelicolor which specify the construction of the anthraquinone product aloesaponarin II, a molecule derived from acetyl coenzyme A and 7 malonyl coenzyme A extender units. This system was designed as an indicator pathway in Streptomyces parvulus to quantify polyketide product formation and to examine the functional significance of specific polyketide synthase components, including the act beta-ketoacyl synthase (beta-KS; encoded by actI-orf1 and actI-orf2) and the act cyclase/dehydrase (encoded by actVII-orf4). Site-directed mutagenesis of the putative active site Cys (to a Gln) in the actI-orf1 beta-KS product completely abrogated aloesaponarin II production. Changing the putative acyltransferase active-site Ser (to a Leu) located in the actI-orf1 beta-KS product led to significantly reduced but continued production of aloesaponarin II. Replacement of the expression cassette with one containing a mutant form of actI-orf2 gave no production of aloesaponarin II or any other detectable polyketide products. However, an expression cassette containing a mutant form of actVII-orf4 gave primarily mutactin with low-level production of aloesaponarin II.

Cloning, sequencing and deduced functions of a cluster of Streptomyces genes probably encoding biosynthesis of the polyketide antibiotic frenolicin

A 10.2-kb fragment of DNA from Streptomyces roseofulvus, which contains polyketide synthase (PKS)-encoding genes (fren) presumed to determine production of the antibiotics frenolicin and the nanaomycins, was cloned. A 5530-bp continuous segment of this DNA was sequenced. Analysis of the sequence revealed five complete open reading frames (ORFs) transcribed in one direction (ORFs 1, 2, 3, 5, 4) and one (ORFX), located between ORF3 and ORF5, transcribed in the opposite direction. The deduced amino-acid sequences of ORFs 1, 2, 3, 4 and 5 closely resemble the sequences of known components of the type-II PKS from other Streptomyces species: putative heterodimeric (ORF1 + 2) ketosynthase, acyl carrier protein, cyclase and ketoreductase, respectively. A resemblance between the N-terminal and C-terminal halves of the ORF4 product--also discovered in the corresponding genes from other isochromanequinone antibiotic producers--suggests a possible origin of the cyclase-encoding gene by duplication. ORFX appears to represent a novel class of genes of unknown function present not only in the fren cluster, but also in other clusters of aromatic antibiotic biosynthetic genes in Streptomyces species. The fren-ORF1-5 genes, encoding a PKS that constructs a nascent polyketide of either 16 or 18 carbons, compared with fixed lengths of 16 and 20 for other available examples, are proving to be valuable for understanding the mechanisms controlling polyketide chain length and patterns of reduction and cyclisation.

Cloning, sequencing, and analysis of the griseusin polyketide synthase gene cluster from Streptomyces griseus

A fragment of DNA was cloned from the Streptomyces griseus K-63 genome by using genes (act) for the actinorhodin polyketide synthase (PKS) of Streptomyces coelicolor as a probe. Sequencing of a 5.4-kb segment of the cloned DNA revealed a set of five gris open reading frames (ORFs), corresponding to the act PKS genes, in the following order: ORF1 for a ketosynthase, ORF2 for a chain length-determining factor, ORF3 for an acyl carrier protein, ORF5 for a ketoreductase, and ORF4 for a cyclase-dehydrase. Replacement of the gris genes with a marker gene in the S. griseus genome by using a single-stranded suicide vector propagated in Escherichia coli resulted in loss of the ability to produce griseusins A and B, showing that the five gris genes do indeed encode the type II griseusin PKS. These genes, encoding a PKS that is programmed differently from those for other aromatic PKSs so far available, will provide further valuable material for analysis of the programming mechanism by the construction and analysis of strains carrying hybrid PKS.

Cloning and nucleotide sequence of a region of the Kibdelosporangium aridum genome homologous to polyketide biosynthetic genes

The actinomycete Kibdelosporangium aridum naturally produces ardacin, a new glycopeptide antibiotic, the biosynthetic pathway of which should involve the participation of a polyketide synthase (PKS). A K. aridum 2.9 kb BamHI genomic fragment homologous to actI (a locus of the PKS cluster catalyzing polyketide chain assembly for actinorhodin biosynthesis in Streptomyces coelicolor) was isolated by shotgun cloning. This DNA fragment, called ardI, was sequenced and the deduced protein products were compared with those of other polyketide synthase genes, revealing similarities ranging from 50 to 80%. ardI was further used to probe a cosmid library of the K. aridum genome. Three hybridizing cosmids were obtained which contain overlapping inserts, together covering a 50 kb region, and including, 15 kb away from ardI, a fragment homologous to actIII, which codes for the ketoreductase of the actinorhodin PKS of S. coelicolor. All these findings indicate that at least part of a polyketide biosynthetic gene cluster has been isolated from the genome of the ardacin producer K. aridum.

Engineered biosynthesis of novel polyketides

Polyketide synthases (PKSs) are multifunctional enzymes that catalyze the biosynthesis of a huge variety of carbon chains differing in their length and patterns of functionality and cyclization. Many polyketides are valuable therapeutic agents. A Streptomyces host-vector system has been developed for efficient construction and expression of recombinant PKSs. Using this expression system, several novel compounds have been synthesized in vivo in significant quantities. Characterization of these metabolites has provided new insights into key features of actinomycete aromatic PKS specificity. Thus, carbon chain length is dictated, at least in part, by a protein that appears to be distinctive to this family of PKSs, whereas the acyl carrier proteins of different PKSs can be interchanged without affecting product structure. A given ketoreductase can recognize and reduce polyketide chains of different length; this ketoreduction always occurs at the C-9 position. The regiospecificity of the first cyclization of the nascent polyketide chain is either determined by the ketoreductase, or the chain-extending enzymes themselves. However, the regiospecificity of the second cyclization is determined by a distinct cyclase, which can discriminate between substrates of different chain lengths.

The tcmVI region of the tetracenomycin C biosynthetic gene cluster of Streptomyces glaucescens encodes the tetracenomycin F1 monooxygenase, tetracenomycin F2 cyclase, and, most likely, a second cyclase

Certain mutations in the tcmVI region of the Streptomyces glaucescens chromosome affect formation of the D ring of the polyketide antibiotic tetracenomycin C (TCM C). This region lies immediately upstream from the TCM C polyketide synthase genes (tcmKLM), and the nucleotide sequence reveals the presence of three small genes, tcmH, tcmI, and tcmJ. On the basis of the phenotypes of mutants and the effects of these genes, when coupled on a plasmid with the tcmKLMN177 genes (tcmN177 is a 3'-truncated version of tcmN), on the production of TCM intermediates in a TCM- mutant, the tcmH gene encodes the C-5 monooxygenase that converts TCM F1 to TCM D3, the tcmI gene encodes the D-ring cyclase that converts TCM F2 to TCM F1 (mutations in this gene are responsible for the type VI phenotype), and the tcmJ gene most likely encodes the B-ring cyclase that acts in the biosynthesis of TCM F2. Furthermore, it appears that the N-terminal domain of the tcmN gene product (encoded by the tcmN177 gene) acts later in the biosynthesis of TCM F2 than the product of tcmJ, suggesting that the N-terminal domain of the TcmN protein is the C-ring cyclase.

Tetracenomycin F2 cyclase: intramolecular aldol condensation in the biosynthesis of tetracenomycin C in Streptomyces glaucescens

Tetracenomycin (Tcm) F2 cyclase, which catalyzes the cyclization of the anthrone Tcm F2 to the naphthacenone Tcm F1 in the biosynthesis of the anthracycline antibiotic Tcm C in Streptomyces glaucescens, has been purified to homogeneity and characterized. The N-terminal sequence of the enzyme establishes that it is encoded by the tcmI gene, whose deduced product has a molecular weight of 12,728. SDS-PAGE analysis gave a single band with a molecular weight of 12,500, whereas gel-filtration chromatography yielded a molecular weight of 37,500, indicating that the Tcm F2 cyclase is a homotrimer in solution. Under pH > or = 8.0, the enzyme catalyzes the cyclization of Tcm F2 to Tcm F1 and has a Km of 121 +/- 18.2 microM and Vmax of 704 +/- 62.3 nmol.min-1.mg-1. In contrast, under pH < or = 6.5, it catalyzes the cyclization of Tcm F2 to 9-decarboxy Tcm F1, a known shunt metabolite of the Tcm C biosynthetic pathway. Tcm F2 cyclase represents the first discrete enzyme for carbon-carbon bond formation via an intramolecular aldol condensation-dehydration mechanism, a key biochemical operation proposed in the early steps of the biosynthesis of all aromatic polyketides.

Hybridization and DNA sequence analyses suggest an early evolutionary divergence of related biosynthetic gene sets encoding polyketide antibiotics and spore pigments in Streptomyces spp

The whiE gene cluster of Streptomyces coelicolor, which is related to gene sets encoding the biosynthesis of polycyclic aromatic polyketide antibiotics, determines a spore pigment. Southern blotting using probes from three different parts of the whiE cluster revealed related gene sets in about half of a collection of diverse Streptomyces strains. A 5.2-kb segment of one such cluster, sch, previously shown to determine spore pigmentation in Streptomyces halstedii, was sequenced. Seven open reading frames (ORFs), two of them incomplete, were found. Six of the ORFs resemble the known part of the whiE cluster closely. The derived gene products include a ketosynthase (= condensing enzyme) pair, acyl carrier protein and cyclase, as well as two of unidentified function. The seventh ORF diverges from the main cluster and encodes a protein that resembles a dichlorophenol hydroxylase. Comparison with sequences of related gene sets for the biosynthesis of antibiotics suggests that gene clusters destined to specify pigment production diverged from those destined to specify antibiotics early in the evolution of the Streptomyces genus.

Functional replacement of genes for individual polyketide synthase components in Streptomyces coelicolor A3(2) by heterologous genes from a different polyketide pathway

Streptomyces coelicolor A3(2) and Streptomyces violaceoruber Tü22 produce the antibiotics actinorhodin and granaticin, respectively. Both the aglycone of granaticin and the half-molecule of actinorhodin are derived from one acetyl coenzyme A starter unit and seven malonyl coenzyme A extender units via the polyketide pathway to produce benzoisochromane quinone moieties with identical structures (except for the stereochemistry at two chiral centers). In S. coelicolor and S. violaceoruber, the type II polyketide synthase (PKS) is encoded by clusters of five and six genes, respectively. We complemented a series of S. coelicolor mutants (act) defective in different components of the PKS (actI for carbon chain assembly, actIII for ketoreduction, and actVII for cyclization-dehydration) by the corresponding genes (gra) from S. violaceoruber introduced in trans on low-copy-number plasmids. This procedure showed that four of the act PKS components could be replaced by a heterologous gra protein to give a functional PKS. The analysis also served to identify which of three candidate open reading frames (ORFs) in the actI region had been altered in each of a set of 13 actI mutants. It also proved that actI-ORF2 (whose putative protein product shows overall similarity to the beta-ketoacyl synthase encoded by actI-ORF1 but whose function is unclear) is essential for PKS function. Mutations in each of the four complemented act genes (actI-ORF1, actI-ORF2, actIII, and actVII) were cloned and sequenced, revealing a nonsense or frameshift mutation in each mutant.

Characterisation of actI-homologous DNA encoding polyketide synthase genes from the monensin producer Streptomyces cinnamonensis

Cloned DNA encoding polyketide synthase (PKS) genes from one Streptomyces species was previously shown to serve as a useful hybridisation probe for the isolation of other PKS gene clusters from the same or different species. In this work, the actI and actIII genes, encoding components of the actinorhodin PKS of Streptomyces coelicolor, were used to identify and clone a region of homologous DNA from the monensin-producing organism S. cinnamonensis. A 4799 bp fragment containing the S. cinnamonensis act-homologous DNA was sequenced. Five open reading frames (ORFs 1-5) were identified on one strand of this DNA. The five ORFs show high sequence similarities to ORFs that were previously identified in the granaticin, actinorhodin, tetracenomycin and whiE PKS gene clusters. This allowed the assignment of the following putative functions to these five ORFS: a heterodimeric beta-ketoacyl synthase (ORF1 and ORF2), an acyl carrier protein (ORF3), a beta-ketoacyl reductase (ORF5), and a bifunctional cyclase/dehydrase (ORF4). The ORFs are encoded in the order ORF1-ORF2-ORF3-ORF5-ORF4, and ORFs-1 and -2 show evidence for translational coupling. This act-homologous region therefore appears to encode a PKS gene cluster. A gene disruption experiment using the vector pGM160, and other evidence, suggests that this cluster is not essential for monensin biosynthesis but rather is involved in the biosynthesis of a cryptic aromatic polyketide in S. cinnamonensis. An efficient plasmid transformation system for S. cinnamonensis has been established, using the multicopy plasmids pWOR120 and pWOR125.

Nucleotide sequence of the tcmII-tcmIV region of the tetracenomycin C biosynthetic gene cluster of Streptomyces glaucescens and evidence that the tcmN gene encodes a multifunctional cyclase-dehydratase-O-methyl transferase

Mutations in the tcmII-tcmIV region of the Streptomyces glaucescens chromosome block the C-3 and C-8 O-methylations of the polyketide antibiotic tetracenomycin C (Tcm C). The nucleotide sequence of this region reveals the presence of two genes, tcmN and tcmO, whose deduced protein products display similarity to the hydroxyindole O-methyl transferase of the bovine pineal gland, an enzyme that catalyzes a phenolic O-methylation analogous to those required for the biosynthesis of Tcm C. The deduced product of the tcmN gene also has an N-terminal domain that shows similarity to the putative ActVII and WhiE ORFVI proteins of Streptomyces coelicolor. The tcmN N-terminal domain can be separated from the remainder of the tcmN gene product, and when coupled on a plasmid with the Tcm C polyketide synthase genes (tcmKLM), this domain enables high-level production of an early, partially cyclized intermediate of Tcm C in a Tcm C- null mutant or in a heterologous host (Streptomyces lividans). By analogy to fatty acid biosynthesis, the tcmKLM polyketide synthase gene products are probably sufficient to produce the linear decaketide precursor of Tcm C; thus, the tcmN N-terminal domain is most likely responsible for one or more of the early cyclizations and, perhaps, the attendant dehydrations that lead to the partially cyclized intermediate. The tcmN gene therefore appears to encode a multifunctional cyclase-dehydratase-3-O-methyl transferase. The tcmO gene encodes the 8-O-methyl transferase.