Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2).

Streptomyces coelicolor is a representative of the group of soil-dwelling, filamentous bacteria responsible for producing most natural antibiotics used in human and veterinary medicine. Here we report the 8,667,507 base pair linear chromosome of this organism, containing the largest number of genes so far discovered in a bacterium. The 7,825 predicted genes include more than 20 clusters coding for known or predicted secondary metabolites. The genome contains an unprecedented proportion of regulatory genes, predominantly those likely to be involved in responses to external stimuli and stresses, and many duplicated gene sets that may represent 'tissue-specific' isoforms operating in different phases of colonial development, a unique situation for a bacterium. An ancient synteny was revealed between the central 'core' of the chromosome and the whole chromosome of pathogens Mycobacterium tuberculosis and Corynebacterium diphtheriae. The genome sequence will greatly increase our understanding of microbial life in the soil as well as aiding the generation of new drug candidates by genetic engineering.

A new mode of stereochemical control revealed by analysis of the biosynthesis of dihydrogranaticin in Streptomyces violaceoruber Tü22.

A class of Streptomyces aromatic polyketide antibiotics, the benzoisochromanequinones, all shows trans stereochemistry at C-3 and C-15 in the pyran ring. The opposite stereochemical control found in actinorhodin (3S, 15R, ACT) from S. coelicolor A3(2) and dihydrogranaticin (3R, 15S, DHGRA) from S. violaceoruber Tü22 was studied by functional expression of the potentially relevant ketoreductase genes, actIII, actVI-ORF1, gra-ORF5, and gra-ORF6. A common bicyclic intermediate was postulated to undergo stereospecific reduction to provide either the 3-(S) or the 3-(R) configuration of an advanced intermediate, 4-dihydro-9-hydroxy-1-methyl-10-oxo-3-H-naphtho[2,3-c]pyran-3-acetic acid (DNPA). Combinations of the four ketoreductase genes were coexpressed with the early biosynthetic genes encoding a type II minimal polyketide synthase, aromatase, and cyclase. gra-ORF6 was essential to produce (R)-DNPA in DHGRA biosynthesis. Out of the various recombinants carrying the relevant ketoreductases, the set of gra-ORF5 and -ORF6 under translational coupling (on pIK191) led to the most efficient production of (R)-DNPA as a single product, implying a possible unique cooperative function whereby gra-ORF6 might encode a "guiding" protein to control the regio- and stereochemical course of reduction at C-3 catalyzed by the gra-ORF5 protein. Updated BLAST-based database analysis suggested that the gra-ORF6 product, a putative short-chain dehydrogenase, has virtually no sequence homology with the actVI-ORF1 protein, which was previously shown to determine the 3-(S) configuration of DNPA in ACT biosynthesis. This demonstrates an example of opposite stereochemical control in antibiotic biosynthesis, providing a key branch point to afford diverse chiral metabolic pools.

Beta-ketoacyl acyl carrier protein synthase III (FabH) is essential for fatty acid biosynthesis in Streptomyces coelicolor A3(2).

The Streptomyces coelicolor fab (fatty acid biosynthesis) gene cluster (fabD-fabH-acpP-fabF) is cotranscribed to produce a leaderless mRNA transcript. One of these genes, fabH, encodes a ketoacyl synthase III that is essential to and is proposed to be responsible for initiation of fatty acid biosynthesis in S. coelicolor.

Functional complementation of pyran ring formation in actinorhodin biosynthesis in Streptomyces coelicolor A3(2) by ketoreductase genes for granaticin biosynthesis.

A mutation in actVI-ORF1, which controls C-3 reduction in actinorhodin biosynthesis by Streptomyces coelicolor, was complemented by gra-ORF5 and -ORF6 from the granaticin biosynthetic gene cluster of Streptomyces violaceoruber Tü22. It is hypothesized that, while gra-ORF5 alone is a ketoreductase for C-9, gra-ORF6 gives the enzyme regiospecificity also for C-3.

Directed transfer of large DNA fragments between Streptomyces species.

The biosynthesis of complex natural products in bacteria is invariably encoded within large gene clusters. Although this facilitates the cloning of such gene clusters, their heterologous expression in genetically amenable hosts remains a challenging problem, principally due to the difficulties associated with manipulating large DNA fragments. Here we describe a new method for the directed transfer of a gene cluster from one Streptomyces species to another. The method takes advantage of tra gene-mediated conjugal transfer of chromosomal DNA between actinomycetes. As proof of principle, we demonstrate transfer of the entire approximately 22-kb actinorhodin gene cluster, and also the high-frequency cotransfer of two loci that are 150 to 200 kb apart, from Streptomyces coelicolor to an engineered derivative of Streptomyces lividans.

Chemical characterisation of disruptants of the Streptomyces coelicolor A3(2) actVI genes involved in actinorhodin biosynthesis.

The actVI genetic region of Streptomyces coelicolor A3(2) is part of the biosynthetic gene cluster of actinorhodin (ACT), the act cluster, consisting of six ORFs: ORFB, ORFA, ORF1, ORF2, ORF3, ORF4. A newly devised method of ACT detection with a combination of HPLC and LC/MS was applied to the analysis of the disruptants of each ORF. ACT was produced by those of ORFB, ORFA, ORF3, and ORF4. Instead of ACT, the ORF1 disruptant produced 3,8-dihydroxy-1-methylanthraquinone-2-carboxylic acid (DMAC) and aloesaponarin II as shunt products. The ORF2 disruptant gave 4-dihydro-9-hydroxy-1-methyl-10-oxo-3-H-naphtho-[2,3-c]-pyran-3-(S)-acet ic acid, (S)-DNPA. These results support our previous proposal for stereospecific pyran ring formation in the biosynthesis of ACT, most importantly suggesting that the actVI-ORF2 product would recognize (S)-DNPA as a substrate for stereospecific reduction at C-15. The disruptant of ORFA produced (S)-DNPA together with ACT, suggesting that actVI-ORFA might play a role such as stabilising the multicomponent, type II PKS complex.

Proof that the ACTVI genetic region of Streptomyces coelicolor A3(2) is involved in stereospecific pyran ring formation in the biosynthesis of actinorhodin.

Pyran ring formation in the biosynthesis of actinorhodin in Streptomyces coelicolor A3(2) was studied using the act cluster deficient strain, CH999, carrying pRM5-based plasmids harbouring combinations of the actVI genes. The strain, CH999/pIJ5660 (pRM5 + actVI-ORF1), produced a chiral intermediate, (S)-DNPA, suggesting that the actVI-ORF1 product is a reductase determining the C-3 stereochemical centre.

Cloning and characterization of a gene cluster from Streptomyces cyanogenus S136 probably involved in landomycin biosynthesis.

From a cosmid library of Streptomyces cyanogenus S136, DNA fragments encompassing approximately 35 kb of the presumed landomycin biosynthetic gene cluster were identified and sequenced, revealing 32 open reading frames most of which could be assigned through data base comparison.

[Over-expression of a polyketide synthase (PKS) module of a giant polyene antibiotic gene cluster in E. coli by double induction].

A 2,671 bp DNA carrying a type I PKS module with KS and AT domains from Streptomyces sp. FR-008 was cloned in-frame into the BamHI site immediately downstream of the PT7 promoter of the E. coli expression vector pET-15b, no considerable expression under IPTG induction was detected. The same PKS gene cloned downstream of the tandem PRPL promoters of pBV220 also yielded no over-expression under 42 degrees C induction. This gene was, however, over-expressed when it was cloned downstream of the tandem PRPT7 or PRPLPT7 promoters. In the case of the tandem PRPLPT7 promoters, the over-expression was dependent on the 42 degrees C plus IPTG double induction. While in the case of the tandem PRPT7 promoters, over-expression could be achieved when the gene was induced by IPTG or 42 degrees C individually or by IPTG and 42 degrees C double induction. Based on these experiences an expression vector pHZ330 containing the tandem PRPT7 promoters was constructed. In addition, the PKS protein expressed in E. coli was injected into rabbits to generate PKS-specific antibodies. Western blotting experiment indicated that these antibodies were PKS-specific which could be used either for the study of the PKS gene cluster or for the detection of the heterologous expression of Streptomyces sp. FR-008 PKS genes.

The granaticin biosynthetic gene cluster of Streptomyces violaceoruber Tü22: sequence analysis and expression in a heterologous host.

BACKGROUND: The granaticins are members of the benzoisochromanequinone class of aromatic polyketides, the best known member of which is actinorhodin made by Streptomyces coelicolor A3(2). Genetic analysis of this class of compounds has played a major role in the development of hypotheses about the way in which aromatic polyketide synthases (PKSs) control product structure. Although the granaticin nascent polyketide is identical to that of actinorhodin, post-PKS steps involve different pyran-ring stereochemistry and glycosylation. Comparison of the complete gene clusters for the two metabolites is therefore of great interest. RESULTS: The entire granaticin gene cluster (the gra cluster) from Streptomyces violaceoruber T-22 was cloned on either of two overlapping cosmids and expressed in the heterologous host, Streptomyces coelicolor A3(2), strain CH999. Chemical analysis of the recombinant strains demonstrated production of granaticin, granaticin B, dihydrogranaticin and dihydrogranaticin B, which are the four known metabolites of S. violaceoruber. Analysis of the complete 39,250 base pair sequence of the insert of one of the cosmids, pOJ466-22-24, revealed 37 complete open reading frames (ORFs), 15 of which resemble ORFs from the act (actinorhodin) gene cluster of S. coelicolor A3(2). Among the rest, nine resemble ORFs potentially involved in deoxysugar metabolism from Streptomyces spp. and other bacteria, and six resemble regulatory ORFs. CONCLUSIONS: On the basis of these resemblances, putative functional assignments of the products of most of the newly discovered ORFs were made, including those of genes involved in the PKS and tailoring steps in the biosynthesis of the granaticin aglycone, steps in the deoxy sugar pathway, and putative regulatory and export functions.

Heterologously expressed acyl carrier protein domain of rat fatty acid synthase functions in Escherichia coli fatty acid synthase and Streptomyces coelicolor polyketide synthase systems.

INTRODUCTION: Fatty acid synthases (FASs) catalyze the de novo biosynthesis of long-chain saturated fatty acids by a process common to eubacteria and eukaryotes, using either a set of monofunctional proteins (Type II FAS) or a polypeptide containing several catalytic functions (Type I FAS). To compare the features of a Type I domain with its Type II counterpart we expressed and characterized an acyl carrier protein (ACP) domain of the Type I rat FAS. RESULTS: An ACP domain of rat FAS was defined that allows expression of a small percentage of active holo-ACP both in Escherichia coli, increasing fivefold upon co-expression with an E. coli holo-ACP synthase, and in Streptomyces coelicolor. The rat ACP domain functions with some components of the E. coli FAS, and can replace the actinorhodin polyketide synthase (PKS) ACP in S. coelicolorA3(2). Purification of the rat ACP domain from E. coli resulted in loss of its functionality. Purified apo-ACP could be converted to its holo-form upon incubation with purified E. coli holo-ACP synthase in vitro, however, suggesting that the loss of functionality was not due to a conformational change. CONCLUSIONS: Functionality of the recombinant rat ACP was shown in distantly related and diverse enzyme systems, suggesting that Type I and Type II ACPs have a similar conformation. A procedure was described that might permit the production of rat FAS holo-ACP for structural and further biochemical characterization.

Identification of a monooxygenase from Streptomyces coelicolor A3(2) involved in biosynthesis of actinorhodin: purification and characterization of the recombinant enzyme.

The oxidation of phenols to quinones is an important reaction in the oxidative tailoring of many aromatic polyketides from bacterial and fungal systems. Sequence similarity between ActVA-Orf6 protein from the actinorhodin biosynthetic cluster and the previously characterized TcmH protein that is involved in tetracenomycin biosynthesis suggested that ActVA-Orf6 might catalyze this transformation as a step in actinorhodin biosynthesis. To investigate the role of ActVA-Orf6 in this oxidation, we have expressed the actVA-Orf6 gene in Escherichia coli and purified and characterized the recombinant protein. ActVA-Orf6 was shown to catalyze the monooxygenation of the tetracenomycin intermediate TcmF1 to TcmD3, strongly suggesting that it catalyzes oxidation of a similar intermediate in actinorhodin biosynthesis. The monooxygenase obeys simple reaction kinetics and has a Km of 4.8 +/- 0.9 microM, close to the figure reported for the homologous enzyme TcmH. The enzyme contains no prosthetic groups and requires only molecular oxygen to catalyze the oxidation. Site-directed mutagenesis was used to investigate the role of histidine residues thought to be important in the reaction; mutants lacking His-52 displayed much-reduced activity, consistent with the proposed mechanistic hypothesis that this histidine acts as a general base during catalysis.

Solution structure of the actinorhodin polyketide synthase acyl carrier protein from Streptomyces coelicolor A3(2).

The solution structure of the actinorhodin acyl carrier protein (act apo-ACP) from the polyketide synthase (PKS) of Streptomyces coelicolor A3(2) has been determined using 1H NMR spectroscopy, representing the first polyketide synthase component for which detailed structural information has been obtained. Twenty-four structures were generated by simulated annealing, employing 699 distance restraints and 94 dihedral angle restraints. The structure is composed, principally, of three major helices (1, 2, and 4), a shorter helix (3) and a large loop region separating helices 1 and 2. The structure is well-defined, except for a portion of the loop region (residues 18-29), the N-terminus (1-4), and a short stretch (57-61) in the loop connecting helices 2 and 3. The RMS distribution of the 24 structures about the average structure is 1.47 A for backbone atoms, 1.84 A for all heavy atoms (residues 5-86), and 1.01 A for backbone atoms over the helical regions (5-18, 41-86). The tertiary fold of act apo-ACP shows a strong structural homology with Escherichia coli fatty acid synthase (FAS) ACP, though some structural differences exist. First, there is no evidence that act apo-ACP is conformationally averaged between two or more states as observed in E. coli FAS ACP. Second, act apo-ACP shows a disordered N-terminus (residues 1-4) and a longer flexible loop (19-41 with 19-29 disordered) as opposed to E. coli FAS ACP where the N-terminal helix starts at residue 3 and the loop region is three amino acids shorter (16-35). Most importantly, however, although the act apo-ACP structure contains a hydrophobic core, there are in addition a number of buried hydrophilic groups, principally Arg72 and Asn79, both of which are 100% conserved in the PKS ACPs and not the FAS ACPs and may therefore play a role in stabilizing the growing polyketide chain. The structure-function relationship of act ACP is discussed in the light of these structural data and recent genetic advances in the field.

Relationships between fatty acid and polyketide synthases from Streptomyces coelicolor A3(2): characterization of the fatty acid synthase acyl carrier protein.

We have characterized an acyl carrier protein (ACP) presumed to be involved in the synthesis of fatty acids in Streptomyces coelicolor A3(2). This is the third ACP to have been identified in S. coelicolor; the two previously characterized ACPs are involved in the synthesis of two aromatic polyketides: the blue-pigmented antibiotic actinorhodin and a grey pigment associated with the spore walls. The three ACPs are clearly related. The presumed fatty acid synthase (FAS) ACP was partially purified, and the N-terminal amino acid sequence was obtained. The corresponding gene (acpP) was cloned and sequenced and found to lie within 1 kb of a previously characterized gene (fabD) encoding another subunit of the S. coelicolor FAS, malonyl coenzyme A:ACP acyl-transferase. Expression of S. coelicolor acpP in Escherichia coli yielded several different forms, whose masses corresponded to the active (holo) form of the protein carrying various acyl substituents. To test the mechanisms that normally prevent the FAS ACP from substituting for the actinorhodin ACP, acpP was cloned in place of actI-open reading frame 3 (encoding the actinorhodin ACP) to allow coexpression of acpP with the act polyketide synthase (PKS) genes. Pigmented polyketide production was observed, but only at a small fraction of its former level. This suggests that the FAS and PKS ACPs may be biochemically incompatible and that this could prevent functional complementation between the FAS and PKSs that potentially coexist within the same cells.

Conserved secondary structure in the actinorhodin polyketide synthase acyl carrier protein from Streptomyces coelicolor A3(2) and the fatty acid synthase acyl carrier protein from Escherichia coli.

The acyl carrier protein (ACP) of Streptomyces coelicolor A3(2) functions as a molecular chaperone during the biosynthesis of the polyketide actinorhodin (act). Here we compare structural features of the polyketide synthase (PKS) ACP, determined by two-dimensional 1H-NMR, with the Escherichia coli fatty acid synthase (FAS) ACP. The PKS ACP contains four helices (residues 7-16 [A], 42-53 [B], 62-67 [C], 72-86 [D]), and a large loop (residues 17-41) having no defined secondary structure with the exception of a turn between residues 21 and 24. The act ACP shows 47% sequence similarity with the E. coli FAS ACP and the results demonstrate that the sequence homology is extended to the secondary structure of the proteins.

A set of ordered cosmids and a detailed genetic and physical map for the 8 Mb Streptomyces coelicolor A3(2) chromosome.

A Supercos-1 library carrying chromosomal DNA of a plasmid-free derivative of Streptomyces coelicolor A3(2) was organized into an ordered encyclopaedia of overlapping clones by hybridization. The minimum set of overlapping clones representing the entire chromosome (with three short gaps) consists of 319 cosmids. The average insert size is 37.5 kb and the set of clones therefore divides the chromosome into 637 alternating unique and overlapping segments which have an average length of approx. 12.5 kb. More than 170 genes, gene clusters and other genetic markers were mapped to their specific segment by hybridization to the encyclopaedia. Genes could be cloned by direct transformation and complementation of S. coelicolor mutants with cosmids isolated from Escherichia coli, selecting for insertion into the chromosome by homologous recombination. As in other streptomycetes, the ends of the chromosome have long terminal inverted repeats.

Production of actinorhodin-related "blue pigments" by Streptomyces coelicolor A3(2).

The genetically well-known strain Streptomyces coelicolor A3(2) produces the pH indicator (red/blue) antibiotic actinorhodin, but not all the "blue pigment" produced by this strain is actinorhodin. When the organism was subjected to various nutrient limitations (ammonium, nitrate, phosphate, or trace elements), and also during growth cessation caused by a relatively low medium pH, blue pigment production was initiated but the pigment and its location varied. At pH 4.5 to 5.5, significant formation of actinorhodin occurred and was located exclusively intracellularly. At pH 6.0 to 7.5 a different blue pigment was produced intracellularly as well as extracellularly. It was purified and identified as gamma-actinorhodin (the lactone form of actinorhodin). Analysis of act mutants of S. coelicolor A3(2) confirmed that both pigments are derived from the act biosynthetic pathway. Mutants with lesions in actII-ORF2, actII-ORF3, or actVA-ORF1, previously implicated or suggested to be involved in actinorhodin export, were impaired in production of gamma-actinorhodin, suggesting that synthesis of gamma-actinorhodin from actinorhodin is coupled to its export from the cell. However, effects on the level of actinorhodin production were also found in some mutants.

Engineered biosynthesis of novel polyketides: properties of the whiE aromatase/cyclase.

The ORFVI from the cluster of genes, which is responsible for the biosynthesis of the Streptomyces coelicolor spore pigment, the whiE cluster, has been described as a bifunctional aromatase/cyclase. In order to evaluate its potential use for generating novel polyketides, combinations of this gene with those encoding minimal polyketide synthase enzymes with or without a ketoreductase from S. coelicolor A3(2) were constructed and analyzed in vivo. Analysis of the polyketide products generated from these constructs indicates that the whiE-ORFVI enzyme has properties similar to those of TcmN, although the whiE aromatase/cyclase normally acts on a polyketide intermediate that is four carbons longer than the TcmN substrate. The whiE aromatase/cyclase can influence the regiospecificity of the first cyclization of unreduced, but not reduced, backbones and is also responsible for the second ring aromatization. An unusual new polyketide, EM18, was identified which is not seen in equivalent strains expressing the tcmN aromatase/cyclase or the act aromatase genes. The structure of EM18 suggests that the WhiE-ORFVI product might have some unique properties within this family of polyketide synthase subunits, and may therefore be useful in the design of combinatorial biosynthetic strategies.

Ectopic expression of the Streptomyces coelicolor whiE genes for polyketide spore pigment synthesis and their interaction with the act genes for actinorhodin biosynthesis.

The whiE gene cluster of Streptomyces coelicolor is normally expressed shortly before sporulation in the aerial mycelium, leading to production of the grey polyketide spore pigment. By placing the whiE genes under the control of the thiostrepton-inducible tipA promoter, they were artificially expressed on plasmids or in the chromosome during vegetative growth in a strain deleted for the act genes, which control biosynthesis of the polyketide antibiotic actinorhodin. Certain combinations of whiE-ORFI-VII led to production of mycelial pigments; these were exported into the medium when whiE-ORFI was absent, but poorly in its presence. Combined with comparative sequence data, the results allowed deductions to be made, or confirmed, about the normal roles of the eight known genes, whiE-ORFI-VIII, as follows: whiE-ORFIII, IV, V encode the three components (ketosynthase, chain length factor and acyl carrier protein) of the whiE 'minimal' polyketide synthase (PKS) needed for assembly of the carbon chain of the spore pigment precursor; whiE-ORFII, VI, VII are likely to be involved in cyclizations of the nascent carbon chain; whiE-ORFVIII controls a late step in the spore pigment biosynthetic pathway, probably a hydroxylation; and whiE-ORFI may encode a protein needed for correct targeting or retention of spore pigment at an appropriate cellular location. In other experiments, genes encoding components of the act-PKS and whiE-PKS were artificially co-expressed. Each of the three whiE minimal PKS subunit genes could complement lesions in the corresponding act-PKS genes to produce actinorhodin or related mycelial pigments, and each of the three act minimal PKS genes could complement lesions in the whiE minimal PKS genes to cause spore pigmentation. Thus the two sets of PKS subunits, which are encoded by genes that have presumably diverged from a common ancestor, are still capable of biochemical 'cross-talk', but this is normally prevented because the gene sets are expressed in different 'tissues' of the differentiated Streptomyces colony. Ectopic expression of sets of whiE-PKS genes presumed to be sufficient to assemble a carbon chain caused inhibition of early growth of the strains, perhaps by causing interference with fatty acid biosynthesis; this yielded circumstantial evidence that the whiE-PKS gene products can also interact with those of the fatty acid synthase(s) of the organism.