| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Graduate School of Pharmaceutical Sciences, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
Correspondence
Koji Ichinose
ichinose{at}mol.f.u-tokyo.ac.jp
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This paper is dedicated to Professor Sir David Hopwood, a pioneer in the field of Streptomyces research, on the occasion of his 70th birthday.
The GenBank accession number for the sequence reported in this paper is AB103463.
1H and 13C-NMR spectra of medermycin isolated from S. coelicolor CH999/pIK340 are available as supplementary data with the online version of this paper at http://mic.sgmjournals.org.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Its structure was shown (Tatsuta et al., 1990) to be identical with lactoquinomycin A, independently characterized as an anticancer compound from Streptomyces tanashiensis (Tanaka et al., 1985; Okabe et al., 1985). Medermycin featured in the first production of hybrid antibiotics by genetic engineering (Hopwood et al., 1985), where a medermycin producer, Streptomyces sp. AM-7161, was transformed to produce mederhodins A and B (
mura et al., 1986). This was achieved by introduction of a segment of the biosynthetic gene cluster of a structurally related antibiotic, actinorhodin, from Streptomyces coelicolor A3(2).
Medermycin and actinorhodin are among a class of Streptomyces aromatic antibiotics known as benzoisochromanequinones (BIQs; Ichinose et al., 1998a). BIQs present a number of interesting biosynthetic problems concerning the polyketide synthase (PKS) (Hopwood, 1997), post-PKS modification (Rix et al., 2002) (‘tailoring’), and for the glycosylated BIQs, such as medermycin and granaticin, deoxysugar biosynthesis (Trefzer et al., 1999). Of particular interest are the common chemical modifications at C-10 via a C–C bond either by glycosylation (medermcyin and granaticin) or by dimerization (in actinorhodin) (Fig. 1). The biosynthetic gene clusters for the BIQs that have been cloned and sequenced to date are for actinorhodin (the act cluster: Hallam et al., 1988; Fernández-Moreno et al., 1991, 1992, 1994; Caballero et al., 1991), frenolicin B (the fren [frn] cluster: Bibb et al., 1994; GenBank accession no. AF058302), griseusin B (the gris cluster: Yu et al., 1994), and granaticin (the gra cluster: Sherman et al., 1989; Bechthold et al., 1995; Ichinose et al., 1998b). Sequence analysis of ORFs in these clusters revealed rather diverse genetic organizations, especially for genes encoding tailoring enzymes (Ichinose et al., 1998b). Some common ‘unknown’ genes are also present, whose deduced functions are unavailable from database searches, but are apparently important for BIQ biosynthesis (Cole et al., 1987; Taguchi et al., 2000a, b). The present paper describes the complete DNA sequence of the medermycin biosynthetic gene cluster of Streptomyces sp. AM-7161. This is the third example (after actinorhodin and granaticin) of the cloning of an entire cluster for BIQ biosynthesis, allowing further insight into aromatic polyketide biosynthesis in Streptomyces spp. on the basis of comparative genomics.
|
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
mura of the Kitasato Institute, and maintained as described by mura et al. (1986). S. coelicolor CH999 (proA1 argA1 redE60 
act : : ermE SCP- SCP2-) was described by McDaniel et al. (1993). Escherichia coli DH5
was used for standard cloning experiments. E. coli strains (Stratagene) for cosmid manipulations were XL-1 Blue-MR and XL-1 Blue-MRF'. Plasmids were passed through E. coli ET12567 (dam dcm hsdS) to generate unmethylated DNA before their use to transform S. coelicolor CH999 (MacNeil et al., 1992). pBluescript SK-/+ (pBS-SK-/+) and pT7Blue(R) T-Vector were from Stratagene and Novagen, respectively. Cosmid pKU402 (Pang et al., 1994) was kindly provided by Professor Haruo Ikeda, Kitasato University. Plasmid pTST59.1 was a generous gift from Dr Josef Altenbuchner, University of Stuttgart, Germany.
General genetic manipulation.
Genomic DNA was isolated as described by Kieser et al. (2000). Restriction mapping and general molecular biology methods were performed as described by Sambrook et al. (1989). PCRs were performed with AmpliTaq Gold (Perkin Elmer) and Ex-Taq (Takara). Synthetic oligonucleotides for PCR primers were obtained from Nihon Bioservice (Saitama, Japan). The following primers were used in this study: PS-A-S, 5'-GGVTTCGGSGGSTTCCAGAGCGC-3'; PS-A-A, 5'-CCAGGCGAASGACTSGTASRCGCTCAC-3'; PS-B-S, 5'-CSGGSGSSGCSGGSTTCATCGG-3'; PS-B-A, 5'-GGGWRCTGGYRSGGSCCGTAGTTG-3'; Neo-F, 5'-AGACAATCGGCTGCTCTGATG-3'; Neo-R, 5'-TAAAGCACGAGGAAGCGGTCAGCCC-3'.
PCR amplification and hybridization.
Genomic (approx. 0·5 µg template) PCR with the PS-A-S and PS-A-A primers was performed in a final volume of 50 µl for 40 cycles of amplification using a step programme (0·5 min at 94 °C, 0·5 min at 64 °C and 1 min at 72 °C) in the Robocycler Gradient 40 (Stratagene) according to the manufacturer's protocol except for the addition of 5 % DMSO. The PCR product of expected size (approx. 0·45 kb) was subcloned into pT7Blue T-vector and propagated in E. coli DH5. DNA sequencing revealed a unique 467 bp gene. Amplification with PS-B-S and PS-B-A was carried out in an essentially similar manner except for the annealing temperature of 68 °C to afford an approximately 0·55 kb product. Sequence analysis of the product identified a unique 541 bp gene. Both products were DIG-labelled (PKS and DOH probes) following the manufacturer's protocol (Roche Diagnostics). Southern hybridization was performed with Nytran N membranes (Schleicher & Schuell) at 68 °C. Membranes were washed once with 1x SSC in 0·1 % SDS at room temperature for 5 min and twice with 0·1x SSC in 0·1 % SDS at 68 °C for 15 min. Subsequent colorimetric detection was carried out with 4-nitroblue tetrazolium chloride (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (X-phosphate) following the standard protocol (Roche Diagnostics).
Library construction and screening.
A genomic library of Streptomyces sp. AM-7161 was prepared in cosmid pKU402 as described by Pang et al. (1994). A total of 1152 independent clones were screened with the PKS and DOH probes to identify seven positive clones, two of which (pIK130 and pIK133) were positive for the both probes. pIK130 was further characterized by restriction mapping and the overlapping subclones were subsequently generated based on pBS-SK-/+ for sequence analysis.
DNA sequencing and computer-assisted sequence analysis.
DNA was sequenced on an automated DNA sequencer model 4000L or 4200L (LI-COR). Sequencing reactions were carried out on double-stranded plasmid DNA (pBS-SK+/-) with the Thermo Sequenase Cycle Sequencing kit (Amersham-Pharmacia). DNA sequences were analysed using the DNASIS software package (version 3.7; Hitachi Software Engineering). A version of the FRAME program (Bibb et al., 1984), MacFRAME (version 1.2, developed by K. Kendal, Tulane University, LA, USA) was used to identify potential protein-coding regions. Deduced gene products were analysed using a version of BLASTP provided by the DNA Database of Japan (DDBJ, http://www.ddbj.nig.ac.jp). Identity/similarity scores were obtained by the BLAST2 SEQUENCES program available from National Center for Biotechnology Information, National Library of Medicine, NIH (http://www.ncbi.nlm.nih.gov/). Phylogenetic analysis was carried out with a version (http://spiral.genes.nig.ac.jp/homology/clustalw.shtml) of the CLUSTAL W program, provided by DDBJ, based on a neighbour-joining method. A phylogenetic tree was created with the TreeView program (version 1.6.2, freely available from the Taxonomy and Systematics server at the University of Glasgow, UK). The number of bootstrap replications was 1000.
Construction of expression plasmid.
pTST59.1 is a derivative of the SUPERCOS1 cosmid (Stratagene) carrying the attP site and integrase derived from 
C31 (Lomovskaya et al., 1980). The vector was digested with XbaI followed by calf intestine alkaline phosphatase treatment and BamHI digestion. The entire insert of pIK130 was released by DraI digestion, and ligated with a BamHI–SmaI adaptor and pSmaI linker (Takara). The resulting fragment was ligated with the foregoing vector, and the mixture was packaged followed by transfection into E. coli XL-1 Blue MRF'. The completed expression plasmid was designated as pIK340.
Heterologous expression of the med cluster and chemical characterization.
Protoplast formation, transformation and regeneration of protoplasts from S. coelicolor CH999 were carried out by standard procedures (Kieser et al., 2000). Genome integration was confirmed by colony PCR with the primer set Neo-F and Neo-R for kanamycin-resistant clones as described by Ishikawa et al. (2000). Transformants were liquid-cultured as described previously (Taguchi et al., 2000a). The medium separated from the broth by centrifugation was directly subjected to reversed-phase HPLC analysis on a TOSOH 8020 system under the following conditions: column, TSK gel ODS-80TM (4·6 mm i.d. x 150 mm, TOSOH); column temperature, 40 °C; gradient elution, solvent A (0·5 % acetic acid in CH3CN) and solvent B (0·5 % acetic acid in distilled H2O), gradient profile (0–5 min, 20 % A; 5–25 min, 20–70 % A; 25–28 min, 70–95 % A; 28–32 min, 95 % A; 32–35 min, 95–20 % A); flow rate, 1·0 ml min-1; photo-diode array detector (PD-8020, TOSOH), 250–600 nm. LC/MS spectra were recorded on a Thermoquest LCQ equipped with a Hewlett Packard HP1100 series LC system under the same chromatographic conditions as used for the HPLC analysis. Atmospheric pressure chemical ionization (APCI) was applied to detect positive and negative ions of samples. Medermycin was also isolated from CH999/pIK340 as follows. Culture broth (1·2 l) was neutralized with 1 M NaOH, and extracted with ethyl acetate. The combined extracts were washed with brine and water, followed by drying over Na2SO4. The residue, after evaporation of the solvent, was subjected to chromatography on silica gel (Wako-gel C200) with chloroform/methanol (15 : 1) to give medermycin (5 mg). The purified compound gave NMR data (available as supplementary data with the online version of this paper at http://mic.sgmjournals.org) consistent with the published values for natural medermycin/lactoquinomycin A (Okabe et al., 1985; Tatsuta et al., 1990; Williamson et al., 2002).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
), and those for many other Streptomyces antibiotics, the resultant genes were to be used as screening probes of a genomic library of the medermycin producer, Streptomyces sp. AM-7161. Genetic organization is conserved among the minimal PKS (type II) components, ketosynthase (KS), ketosynthase
(KS) and acyl carrier protein (ACP). In particular, the universal tandem pair of KS and KS, which are normally translationally coupled, led us to design degenerate primers from the C-terminus of KS (PS-A-S) to the N-terminus of KS (PS-A-A) based on the known examples of BIQs, actinorhodin (Fernández-Moreno et al., 1992), frenolicin B (Bibb et al., 1994), griseusin B (Yu et al., 1994) and granaticin (Sherman et al., 1989). A DOH-specific enzyme, NDP-glucose-4,6-dehydratase, was successfully used for designing degenerate primers to clone the genes involved in DOH biosynthesis directed to a number of Streptomyces antibiotics. The suggested primers, PS-B-S and PS-B-A (Decker et al., 1996), were applied to this study. PCR using genomic DNA of Streptomyces sp. AM-7161 as a template afforded PCR products of the expected sizes (approx. 450 bp for PKS; approx. 550 bp for DOH). Subcloning and sequence analysis of the PCR products afforded 467 bp PKS and 541 bp DOH fragments, likely to encode PKS (KS-KS) and a DOH enzyme (NDP-glucose-4,6-dehydratase), respectively. Both proteins closely resembled (>70 % similarity) the proteins used for the PCR primer design. Genomic hybridization analysis (data not shown) with the PCR products as probes revealed a single positive restriction fragment on electrophoresis: PKS (7·7 kb BamHI, 3·3 kb PstI); and DOH (10 kb BamHI, 4·2 kb PstI). A cosmid library was constructed using approximately 40 kb insert DNA prepared by partial Sau3AI digestion of the genomic DNA. Library screening with the PCR products as hybridization probes identified seven positive clones out of 1152, with two clones (pIK130, 133) positive to both probes. We chose pIK130 for further characterization. Southern blot analysis suggested that the clone contained the same BamHI fragments as revealed in the genomic blot analysis.
DNA sequence analysis
Restriction mapping was performed for the pIK130 insert, and its overlapping subclones were sequenced to cover the complete insert (36 202 bp, overall G+C content 72·8 mol%). Probable ORFs were identified by the FRAME program (Bibb et al., 1984), and by potential ribosome-binding sites (Strohl, 1992). The deduced ORFs were functionally designated based on database searches as shown in Fig. 2. We identified 34 complete ORFs (Table 1), with an incomplete ORF at either end, and several inverted repeats in intergenic regions. The cluster was designated as the med cluster. Following is the description of the med-ORFs with the corresponding act genes, if any, as a reference.
|
|
). Nearly all the known examples of the minimal PKS (Hopwood, 1997) include a gene encoding the ACP immediately downstream of a set of KS and KS. Surprisingly, the ACP gene (med-ORF23) was discovered 20 kb upstream of the med-ORF2 (KS). Such an unusual organization is known for the biosynthetic gene clusters for daunorubicin (Grimm et al., 1994), R1128 (Marti et al., 2000) and griseorhodin A (Li et al., 2002). Downstream of med-ORF23 are a pair of genes, med-ORF24 and 22, possibly related to polyketide starter formation and chain extension. The jadomycin B biosynthetic gene cluster from Streptomyces venezuelae contains a similar organization for jadM and jadN (Wang et al., 2001). The med-ORF24 protein is similar to the product of jadM, which was proposed to encode a phosphopantetheinyl transferase, functioning as a holo-ACP synthase for jadomycin B biosynthesis. Other med-ORF24 homologues (similarity/identity at amino acid level) include pgaX from Streptomyces sp. PGA64 (65/53 %, GenBank accession no. AY034378) and mtaA from Stigmatella aurantiaca (48/35 %, GenBank accession no. AF188287). gra-ORF32, previously assigned as ‘unknown’ (Ichinose et al., 1998b), now belongs to this family, encoding a protein with 38 % similarity to the med-ORF24 product. Very high homology was found for med-ORF22 with jadN, and also with lanP from Streptomyces cyanogenus, probably involved in landomycin biosynthesis (Westrich et al., 1999). The homologues both resemble acyl-CoA decarboxylase genes. The med-ORF6 protein is an apparent homologue of the KR encoded by actIII, controlling C-9 keto-reduction. The ARO gene, med-ORF19, was found over 15 kb apart from the KS gene, med-ORF1, rather closer to the ACP gene, med-ORF23. med-ORF3, encoding CYC, is downstream of med-ORF2, with putative translational coupling. Thus, the genes required for the formation of the bicyclic intermediate were identified as expected.
|
; Taguchi et al., 2001). In actinorhodin biosynthesis, actVI-ORF1 was proved to encode the reductase responsible (Ichinose et al., 1999). Since medermycin possesses the same stereochemistry concerning the two chiral centres (3S, 15R) as actinorhodin, an actVI-ORF1 homologue was expected. A homologue, med-ORF12, was indeed found. The product (‘chiral alcohol’) of the C-3 reduction would cyclize to the hemiketal, followed by dehydration to afford (S)-DNPA (Cole et al., 1987) (Fig. 3
). Our earlier experiments (Taguchi et al., 2000a) suggested the actVI-ORF2 protein, an enoyl reductase homologue, to function as the second reductase at C-15 to complete the (3S, 15R) configuration. The presence of another homologue, actVI-ORF4, prompted us to propose that the product might assist an unusual mode of reduction of (S)-DNPA (Taguchi et al., 2000a). The reduction was proposed to proceed in two coupled steps, with stereospecific reduction (by ActVI-ORF2) of the double bond between C-14 and C-15 followed by isomerization (by ActVI-ORF4) to afford the ,-unsaturated carbonyl structure, thus accomplishing the 1,4-reduction concerning C-15 and C-6. Here, we encountered the same situation as for the act cluster: two actVI-ORF2 homologues, med-ORF9 and med-ORF29 (37 % similar to each other), are present, in agreement with our assumption.
Oxygenation and hydroxylation.
The oxygenase/hydroxylase would be involved in the introduction of oxygen functionalities at C-6 and C-8 of the BIQs. In actinorhodin biosynthesis, the actVA-ORF6 protein was proposed (Fernández-Moreno et al., 1994) to oxidize C-6, and later biochemical proof was given for its catalytic activity (Kendrew et al., 1997). This enzyme performs oxygenation without any prosthetic group, metal ion, or cofactor: its novel catalytic mechanism was revealed by a recent X-ray crystallographic study (Sciara et al., 2003). The med cluster lacks an obvious homologue, and instead includes a gene, med-ORF7, similar to actVA-ORF5, deduced to encode a hydroxylase (Caballero et al., 1991). actVA-ORF5 and its close homologue, gra-ORF21, were reasonably assigned to encode a hydroxylase at C-8 based on their significant similarity (>70 %) to pheA, encoding a phenol hydroxylase from Bacillus stearothermophilus (Kim et al., 1995), correlating with the presence of the hydroxyl group at C-8 of actinorhodin and granaticin. The absence of an oxygen at C-8 of medermycin might allow us to assume the med-ORF7 protein to be an oxygenase at C-6. med-ORF13 is a clear homologue of actVB, encoding a flavin : NADH oxidoreductase (Kendrew et al., 1995), and gra-ORF34 (Ichinose et al., 1998b). This finding supports our proposal (Ichinose et al., 1998b) of their possible roles in establishing the redox change at C-10 required for the substitution steps, dimerization for actinorhodin and C-glycosylation for granaticin, since medermycin is C-glycosylated at C-10.
actVI-ORFA homologue.
actVI-ORFA homologues constitute a family of genes widely found in the gene clusters for Streptomyces aromatic polyketides, the members being fren-ORFX (Bibb et al., 1994), gra-ORF31 (Ichinose et al., 1998b), mtmX from the mithramycin cluster of Streptomyces argillaceus (GenBank accession no. X89899), and dpsH from the daunorubicin clusters of Streptomyces peucetius (Gerlitz et al., 1997) and Streptomyces sp. strain C5 (GenBank accession no. U43704). The actVI-ORFA disruptant produced reduced amounts of actinorhodin, together with biosynthetic intermediates and shunt products (Taguchi et al., 2000a, b). A possible involvement of dpsH in polyketide chain cyclization was also postulated (Gerlitz et al., 1997). Although the gene function is still speculative, it apparently controls the efficiency of biosynthetic enzymes in a given cluster. med-ORF10 belongs to this family.
actVI-ORF3 homologue.
The phenotypic and chemical characterization of an actVI-ORF3 disruptant (Fernández-Moreno et al., 1992; Taguchi et al., 2000a) suggested that the product might catalyse chemically spontaneous processes such as the hemiketal formation and dehydration to afford (S)-DNPA (Fig. 3). Recent proteomic analysis (Hesketh et al., 2002) of the act enzymes in S. coelicolor A3(2) suggested an interesting possibility for a role in exporting actinorhodin (see Discussion). Unlike the actVI-ORFA family, the actVI-ORF3 homologues are restricted to the clusters for BIQs. gra-ORF18 and the gene herein discovered, med-ORF5, are the only other family members.
DOH pathway.
The biosynthetic genes for the angolosamine moiety were found to be contiguous and transcribed in the same direction. Examples of such clustering for DOH pathways in Streptomyces glycoside antibiotics are the D-olivose/L-rhodinose genes for urdamycin A formation in Streptomyces fradiae (Hoffmeister et al., 2000), L-mycarose for tylosin in S. fradiae (GenBank accession no. AF147704), L-digitoxose for jadomycin B in S. venezuelae (GenBank accession no. AY026363), and L-desosamine for pikromycin/methymycin (GenBank accession no. AF079138). The first three steps are common with those for the DOH groups (4-keto-2,6-dideoxyglucose and L-rhodinose) of granaticin (Bechthold et al., 1995; Ichinose et al., 1998b): NDP-glucose synthase (encoded by gra-ORF16), NDP-glucose-4,6-dehydratase (encoded by gra-ORF17) and NDP-glucose-2,3-dehydratase (encoded by gra-ORF27), all of which were reasonably assigned to the respective gene products of med-ORF18, med-ORF17 and med-ORF16 based on significant similarity scores (>70 %). The med-ORF20 protein resembles a class of aminotransferases involved in the biosynthesis of Streptomyces antibiotics containing aminodeoxysugars. Representative proteins are DnrJ for daunosamine formation in daunorubicin biosynthesis in S. peucetius (Madduri et al., 1995) and DesV (with 54 % sequence similarity) for desosamine production related to macrolide antibiotics in S. venezuelae (GenBank accession no. AF079138). The med-ORF20 protein was reasonably assigned as an aminotransferase at C-3'. All of the foregoing genes appear to be translationally coupled. The remaining steps, which may be interchangeable, are catalysed by 4'-keto-reductase and N-methyltransferase. The genes responsible were found downstream of med-ORF16 as a translationally coupled pair: med-ORF15 and med-ORF14. 4'-Keto reduction is required for granaticin B (L-rhodinose) biosynthesis: the assigned gene, gra-ORF22 (Tornus et al., 2001), encodes a protein 45 % similar to the med-ORF15 product. Other homologues include urdZ3 for urdamycin A (47 % similarity, Hoffmeister et al., 2000), lanZ3 for landomycin (50 % similarity, Westrich et al., 1999) and snoG for nogalamycin biosynthesis in Streptomyces nogalater (45 % similarity, GenBank accession no. AF187532). Strong similarity was found for med-ORF15 to a family of N-methyltransferase genes, snoX for nogalamycin (Ylihonko et al., 1996) and rdmD for rhodomycin biosynthesis in Streptomyces purpurascens (56 %, GenBank accession no. U10405). Functional assignment of the DOH genes was thus completed (Fig. 4).
|
). The gra-ORF14 protein resembles the product of med-ORF8. Extensive sequence comparison of med-ORF8 with other Streptomyces GT genes indicated its probable role in C-glycosylation, which is required for angolosamine transfer in medermycin biosynthesis (see Discussion).
Regulatory and antibiotic export genes
The pathway-specific activator genes, now known as Streptomyces antibiotic regulatory proteins, SARPs (Wietzorrek et al., 1997), are a growing family of genes found mainly in the gene clusters for Streptomyces aromatic polyketides. actII-ORF4 (Fernández-Moreno et al., 1991) is the best-characterized member. med-ORF11 is a clear homologue. actII-ORF4 contains a rare triplet codon, TTA, for leucine, involved in a key regulatory role in actinorhodin production (Fernández-Moreno et al., 1991), but the med protein contains no TTA codon. In this it resembles redD in the undecylprodigiosin pathway of S. coelicolor, which is controlled by a second regulator, redZ, which contains a TTA codon (Guthrie et al., 1998).
The product of med-ORF26 resembles those of gra-ORF19 (Ichinose et al., 1998b) and frnE (GenBank accession no. AF058302), becoming ‘common’ proteins in the BIQ clusters. We proposed the function to be a redox-sensor by forming a disulphide bridge (Ichinose et al., 1998b). All of the three homologues contain a key motif of CXXC (11CPWC16 for Med-ORF26) found in the glutaredoxin/thioredoxin superfamily. The med-ORF30 protein is homologous with the product of jadR1 in the jadomycin B cluster, which is deduced to encode an essential transcriptional regulator (GenBank accession no. U24659).
The only gene found for a possible antibiotic transporter is med-ORF25, whose product is homologous with a probable transmembrane-transport protein, SMa0185, from Sinorhizobium meliloti (See Table 1). Other regulatory genes in the med cluster are med-ORF21 (a possible kinase gene) and med-ORF28 (a putative Tet-R family transcriptional regulator gene) (see Table 1).
‘Unknown’ and unassigned genes
Downstream of the essential PKS gene, med-ORF3, is med-ORF4. The product resembles a family of chitosanases, which catalyse the hydrolysis of chitosan. Typical protein length is about 300 aa, whereas Med-ORF4 is 100 aa longer (409 aa). Its region of homology to a putative chitosanase gene, csn, from S. coelicolor (Bentley et al., 2002) was localized towards the C-terminus. Since chitosanases are catabolic enzymes, med-ORF4 appears to be unrelated to medermycin biosynthesis. A lower G+C (approx. 65 mol%) region was present upstream of med-ORF23. FRAME analysis suggested a possible protein-coding region, designated as med-ORF27. The deduced product showed partial similarity (approx. 50 %) to the products of ‘unknown’ genes found in Streptomyces antibiotic gene clusters, including pgaK from the rubromycin B-producing Streptomyces sp. PGA64 (GenBank accession no. AY034378), ORF12 from the tylosin producer S. fradiae (Bate et al., 1999) and an incomplete ORF, aur1O, from the auricin-producing Streptomyces aureofaciens (Novakova et al., 2002). All other ORFs (P, Q, R, X, Y and Z) identified were highly homologous with hypothetical proteins revealed by the genome sequencing project (Bentley et al., 2002) of S. coelicolor (Table 1).
Heterologous expression of the med cluster
Deduction of all essential biosynthetic steps to medermycin from the deduced med protein functions induced us to attempt heterologous expression of the entire med cluster. An integrative expression vector (pTST59.1) was chosen for stable replication of the large DNA fragment derived from a cosmid clone. The whole insert of pIK130 was used to construct pIK340, which was then introduced into S. coelicolor CH999 (McDaniel et al., 1993). Transformants on plates gave brownish pigmentation, and they were subjected to liquid culture followed by HPLC analysis. The recombinant, CH999/pIK340, gave a peak corresponding to standard medermycin, which was not observed in the profile derived from the control recombinant, CH999/pTST59.1 (Fig. 5A). LC/MS analysis suggested its molecular mass to be 457 Da, which is consistent with the formula of medermycin (C24H27O8N). MS/MS spectra (the precursor ion at m/z 458) for the sample from CH999/pIK340 and authentic medermycin were identical (Fig. 5B). Final confirmation using NMR spectra (available as supplementary data with the online version of this paper at http://mic.sgmjournals.org) indicated clearly that the sample was medermycin. This result provides definite proof that the med cluster described above is indeed sufficient for medermycin production.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Mechanistically, medermycin shares the early biosynthetic pathway with actinorhodin and granaticin. The genes encoding the PKS and cyclizing enzymes (ARO and CYC) were found, two of which (med-ORF23 for ACP and med-ORF19 for ARO) have unusual locations upstream of those for the core PKS components, KS and KS. Post-PKS tailoring steps for aglycone formation were expected to follow the actinorhodin pathway. The med cluster provides the genes for the key reductases for stereochemical control, med-ORFs 12, 9 and 29. A structural difference of the BIQ chromophore between medermycin and actinorhodin concerns the hydroxyl group at C-8. Earlier functional assignment of ActVA-ORF5 (Caballero et al., 1991) as a hydroxylase at C-8 is open to discussion based on the present finding of med-ORF7. On the other hand, the med cluster lacks an actVA-ORF6 homologue, encoding an essential monooxygenase to introduce oxygen at C-6 of actinorhodin (Kendrew et al., 1997; Sciara et al., 2003). This situation is reasonably explained by our tentative assignment of Med-ORF7 as an oxygenase at C-6.
An increasing number of genes for DOH biosynthesis have been cloned (Trefzer et al., 1999). The angolosamine biosynthetic genes were found to be contiguous in the med cluster, possibly providing a useful gene cassette of DOH production for metabolic engineering (Tang et al., 2001). One of our initial interests was C-glycosylation in the BIQs. The gra cluster for granaticin biosynthesis contains a single GT gene, gra-ORF14, although the producer Streptomyces violaceoruber produces an additional granaticin derivative, granaticin B, which is the O-rhodinoside of granaticin. To reconcile two glycosylation steps with the presence of a single GT, two alternatives were proposed: (a) Gra-ORF14 catalyses the rhodinosyl transfer, and the C-glycosylation is catalysed by an ‘unknown’ enzyme; (b) Gra-ORF14 catalyses both glycosylations. Since the reaction centres of an acceptor molecule for O- and C-glycosylations are quite different, it would be hard to imagine that Gra-ORF14 could have such dual activities. We then proposed (Ichinose et al., 1998b) that the first sugar attachment might proceed in rather unusual fashion, with nucleophilic substitution of the C-10 of an aglycone with the keto group at C'-4, giving a plausible explanation to support possibility (a). In the meantime, a C-GT gene, urdGT2 for urdamycin biosynthesis (Faust et al., 2000), was characterized. Its C-glycosylating activity for D-olivose was demonstrated by gene inactivation to result in the accumulation of aglycone derivatives, urdamycins I, J and K (Künzel et al., 1999). The UrdGT2 protein shows strong similarities to Gra-ORF14 (53 %) and Med-ORF8 (58 %). Phylogenetic analysis of Streptomyces GTs indicated that the three proteins, along with two others (LanGT2 and SnoT), occupy a distinct position from those of O-GTs (Fig. 6A). Recent crystallographic studies shed light on the catalytic mechanism of GTs (Ünligil et al., 2000). Particularly noteworthy is the structural study of the UDP-glucosyltransferase, GtfB, involved in the biosynthesis of vancomycin-group antibiotics (Mulichak et al., 2001), where the highly conserved Gly-rich sequence HHGGAGT, the strictly conserved Pro, and the highly conserved Asp (the potential catalytic base) were highlighted. Sequence alignment of the entries used for the foregoing phylogenetic analysis showed that the Gly-rich motif and the Pro residue were universally aligned, whereas conservation of the possible base residue, Asp, was mainly restricted to O-GTs (Fig. 6B). Since the general base is supposed to abstract a proton from an acceptor hydroxyl group, this residue might determine the specificity for C-glycosylation. Taking these results together, we favour the idea that Gra-ORF14 and Med-ORF8 both function as C-GTs.
|
). The recent completion (Bentley et al., 2002) of the genome sequencing project for the actinorhodin producer, S. coelicolor A3(2), provided a new direction of post-genome research, such as proteomics. Proteomic analysis (Hesketh et al., 2002) of mycelial extracts suggested that ActVI-ORF3 is exported to the cell wall matrix, with its N-terminal signal peptide truncated. This provides an intriguing possibility for dual activities of ActVI-ORF3, extracellular as well as intracellular, leading to the efficient biosynthesis of actinorhodin. Since no definite roles of the ‘unknowns' such as ActVI-ORF3 are available from database analysis, combined proteomics and comparative genomics would be the most efficient approach to highlight the critically important genes in a given metabolic pathway. The finding of an actVI-ORF3 homologue, med-ORF5, clearly contributes to this approach. What other genes are common among the three complete BIQ clusters? Table 2 shows the sequence-based comparison of the cluster genes. Apart from the functionally defined genes, the actVI-A, actVA-ORF5 and actVB families each contain members in all three clusters. Priority should be given to these genes for further characterization in the quest for comprehensive understanding of the biosynthesis of BIQ antibiotics.
|
|
ACKNOWLEDGEMENTS |
|---|
mura, The Kitasato Institute, and Professor Haruo Ikeda, Kitasato University, for the generous gifts of strain and materials. We thank Dr Josef Altenbuchner, University of Stuttgart, for valuable technical advice. This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas ‘Genome Biology’ from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (14014210), to K. I. We thank David Hopwood for critical reading of the manuscript. |
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bechthold, A., Sohng, J. K., Smith, T. M., Chu, X. & Floss, H. G. (1995). Identification of Streptomyces violaceoruber Tü22 genes involved in the biosynthesis of granaticin. Mol Gen Genet 248, 610–620.[CrossRef][Medline]
Bentley, S. D., Chater, K. F. Cerdeno-Tarraga, A. M. & 40 other authors (2002). Complete genome sequence of the model actinomycete Streptmyces coelicolor A3(2). Nature 417, 141–147.[CrossRef][Medline]
Bibb, M. J., Findlay, P. R. & Johnson, M. W. (1984). The relationship between base composition and codon usage in bacterial genes and its use for the simple and reliable identification of protein-coding sequences. Gene 30, 157–166.[CrossRef][Medline]
Bibb, M. J., Sherman, D. H., mura, S. & Hopwood, D. A. (1994). Cloning, sequencing and deduced functions of a cluster of Streptomyces genes probably encoding biosynthesis of the polyketide antibiotic frenolicin. Gene 142, 31–39.[CrossRef][Medline]
Caballero, J. L., Martinez, E., Malpartida, F. & Hopwood, D. A. (1991). Organisation and functions of the actVA region of the actinorhodin biosynthetic gene cluster of Streptomyces coelicolor. Mol Gen Genet 230, 401–412.[CrossRef][Medline]
Cole, S. P., Rudd, B. A. M., Hopwood, D. A., Chang, C.-J. & Floss, H. G. (1987). Biosynthesis of the antibiotic actinorhodin. Analysis of blocked mutants of Streptomyces coelicolor. J Antibiot 40, 340–347.[Medline]
Decker, H., Gaiser, S., Pelzer, S., Scheider, P., Westrich, L., Wohlleben, W. & Bechthold, A. (1996). A general approach for cloning and characterizing dNDP-glucose dehydratase genes from actinomycetes. FEMS Microbiol Lett 141, 195–201.[CrossRef][Medline]
Faust, B., Hoffmeister, D., Weitnauer, G. & 7 other authors (2000). Two new tailoring enzymes, a glycosyltransferase and an oxygenase, involved in biosynthesis of the angucycline antibiotic urdamycin A in Streptomyces fradiae Tü22. Microbiology 146, 147–154.
Fernández-Moreno, M. A., Caballero, J. L., Hopwood, D. A. & Malpartida, F. (1991). The act cluster contains regulatory and antibiotic export genes, direct targets for transcriptional control by the bldA tRNA gene of Streptomyces. Cell 66, 769–780.[CrossRef][Medline]
Fernández-Moreno, M. A., Martinez, E., Boto, L., Hopwood, D. A. & Malpartida, F. (1992). Nucleotide sequence and deduced functions of a set of co-transcribed genes of Streptomyces coelicolor A3(2) including the polyketide synthase for the antibiotic actinorhodin. J Biol Chem 267, 19278–19290.
Fernández-Moreno, M. A., Martinez, E., Caballero, J. L., Ichinose, K., Hopwood, D. A. & Malpartida, F. (1994). DNA sequence and functions of the actVI region of the actinorhodin biosynthetic gene cluster of Streptomyces coelicolor A3(2). J Biol Chem 269, 24854–24863.
Gerlitz, M., Meurer, G., Wedt-Pienkowski, E., Madduri, K. & Hutchinson, C. R. (1997). Effects of the daunorubicin dpsH gene on the choice of starter unit and cyclization pattern reveals that type II polyketide synthases can be unfaithful yet intriguing. J Am Chem Soc 119, 7392–7393.[CrossRef]
Grimm, A., Madduri, K., Ali, A. & Hutchinson, C. R. (1994). Characterization of the Streptomyces peucetius ATCC 29050 genes encoding doxorubicin polyketide synthase. Gene 151, 1–10.[CrossRef][Medline]
Guthrie, E. P., Flaxman, C. S., White, J., Hodgson, D. A., Bibb, M. J. & Chater, K. F. (1998). A response-regulator-like activator of antibiotic synthesis from Streptomyces coelicolor A3(2) with an amino-terminal domain that lacks a phosphorylation pocket. Microbiology 144, 727–738.[Abstract]
Hallam, S. E., Malpartida, F. & Hopwood, D. A. (1988). Nucleotide sequence, transcription and deduced function of a gene involved in polyketide antibiotic synthesis in Streptomyces coelicolor. Gene 74, 305–320.[CrossRef][Medline]
Hesketh, A., Chandra, G., Shaw, A. D., Rowland, J. J., Kell, M. J., Bibb, M. J. & Chater, K. F. (2002). Primary and secondary metabolism, and post-translational protein modifications, as portrayed by proteomic analysis of Streptomyces coelicolor. Mol Microbiol 46, 917–932.[CrossRef][Medline]
Hoffmeister, D., Ichinose, K., Domann, S. & 9 other authors (2000). The NDP-sugar co-substrate concentration and the enzyme expression level influence the substrate specificity of glycosyltransferases: cloning and characterization of deoxysugar biosynthetic genes of the urdamycin biosynthetic gene cluster. Chem Biol 7, 821–831.[CrossRef][Medline]
Hopwood, D. A. (1997). Genetic contributions to understanding polyketide synthases. Chem Rev 97, 2465–2497.[CrossRef][Medline]
Hopwood, D. A., Malpartida, F., Kieser, H. M., Ikeda, H., Duncan, J., Fujii, I., Rudd, B. A. M., Floss, H. G. & mura, S. (1985). Production of ‘hybrid’ antibiotics by genetic engineering. Nature 314, 642–644.[CrossRef][Medline]
Ichinose, K., Taguchi, T., Ebizuka, Y. & Hopwood, D. A. (1998a). Biosynthetic gene clusters of benzoisochromanequinone antibiotics in Streptomyces spp. – identification of genes involved in post-PKS tailoring steps. Actinomycetologica 12, 99–109.
Ichinose, K., Bedford, D. J., Tornus, D., Bechthold, A., Bibb, M., Revill, W. P., Floss, H. G. & Hopwood, D. A. (1998b). The granaticin biosynthetic gene cluster of Streptomyces violaceoruber Tü22: sequence analysis and expression in a heterologous host. Chem Biol 5, 647–659.[CrossRef][Medline]
Ichinose, K., Surti, C., Taguchi, T., Malpartida, F., Booker-Milburn, K. I., Stephenson, G. R., Ebizuka, Y. & Hopwood, D. A. (1999). Proof that the actVI genetic region of Streptomyces coelicolor A3(2) is involved in stereospecific pyran ring formation in the biosynthesis of actinorhodin. Bioorg Med Chem Lett 9, 395–400.[CrossRef][Medline]
Ichinose, K., Taguchi, T., Bedford, D. J., Ebizuka, Y. & Hopwood, D. A. (2001). Functional complementation of pyran ring formation in actinorhodin biosynthesis in Streptomyces coelicolor A3(2) by ketoreductase genes for granaticin biosynthesis. J Bacteriol 183, 3247–3250.
Ishikawa, J., Tsuchizaki, N., Yoshida, M., Ishiyama, D. & Hotta, K. (2000). Colony PCR for detection of specific DNA sequences in actinomycetes. Actinomycetologica 14, 1–5.
Kendrew, S. G., Harding, S. E., Hopwood, D. A. & Marsh, E. N. G. (1995). Identification of a flavin : NADH oxidoreductase involved in the biosynthesis of actinorhodin. J Biol Chem 270, 17339–17343.
Kendrew, S. G., Hopwood, D. A. & Marsh, E. N. G. (1997). Identification of a monoxygenase from Streptomyces coelicolor A3(2) involved in biosynthesis of actinorhodin: purification and characterization of the recombinant enzyme. J Bacteriol 179, 4305–4310.
Kieser, T., Bibb, M. J., Buttner, M. J., Chater, K. F. & Hopwood, D. A. (2000). Practical Streptomyces Genetics. Norwich: John Innes Foundation.
Kim, I. C. & Oriel, P. J. (1995). Characterization of the Bacillus stearothermophilus BR219 phenol hydroxylase gene. Appl Environ Microbiol 61, 1252–1256.[Abstract]
Künzel, E., Faust, B., Oelers, C., Weißbach, U., Bearden, D., Weitnauer, G., Westrich, L. & Bechthold, A. (1999). The inactivation of the urdGT2 gene, which encodes a glycosyltransferase responsible for the C-glycosyltransfer of activated D-olivose, leads to the formation of the three novel urdamycins I, J and K. J Am Chem Soc 121, 11058–11062.[CrossRef]
Li, A. & Piel, J. (2002). A gene cluster from a marine Streptomyces encoding the biosynthesis of the aromatic spiroketal polyketide griseorhodin A. Chem Biol 9, 1017–1026.[CrossRef][Medline]
Lomovskaya, N. D., Chater, K. F. & Mkrtumian, N. M. (1980). Genetics and molecular biology of Streptomyces bacteriophages. Microbiol Rev 44, 206–229.
MacNeil, D. J., Gewain, K. M., Ruby, C. L., Dezeny, G., Gibbons, P. H. & MacNeil, T. (1992). Analysis of Streptomyces avermitilis genes required for avermection biosynthesis utilizing a novel integration vector. Gene 111, 61–68.[CrossRef][Medline]
Madduri, K. & Hutchinson, C. R. (1995). Functional characterization and transcriptional analysis of the dnrR1 locus, which controls daunorubicin biosynthesis in Streptomyces peucetius. J Bacteriol 177, 1208–1215.
Marti, T., Hu, Z., Phol, N. L., Shah, A. N. & Khosla, C. (2000). Cloning, nucleotide sequence, and heterologous expression of the biosynthetic gene cluster for R1128, a non-steroidal estrogen receptor antagonist. J Biol Chem 275, 33443–33448.
McDaniel, R., Ebert-Khosla, S., Hopwood, D. A. & Khosla, C. (1993). Engineered biosynthesis of novel polyketides. Science 262, 1546–1550.
Mulichak, A. M., Losey, H. C., Walsh, C. T. & Garavito, R. M. (2001). Structure of the UDP-glucosyltransferase GtfB that modifies the heptapeptide aglycone in the biosynthesis of vancomycin group antibiotics. Structure 9, 547–557.[Medline]
Novakova, R., Bisrakova, J., Homerova, D., Rezuchova, B. & Kormanec, J. (2002). Cloning and characterization of a polyketide synthase gene cluster involved in biosynthesis of a proposed angucycline-like poleketide auricin in Streptomyces aureofaciens CCM 3239. Gene 297, 197–208.[CrossRef][Medline]
Okabe, T., Nomoto, K., Funabashi, H., Okuda, S., Suzuki, H. & Tanaka, N. (1985). Lactoquinomycin A, a novel anticancer antibiotic. II. Physico-chemical properties and structure assignment. J Antibiot 38, 1333–1336.[Medline]
mura, S., Ikeda, H., Malpartida, F., Kieser, H. M. & Hopwood, D. A. (1986). Production of new hybrid antibiotics, mederhodins A and B, by a genetically engineered strain. Antimicrob Agents Chemother 29, 13–19.
Pang, C.-H., Shiiyama, M., Ikeda, H., Tanaka, H. & mura, S. (1994). Cosmid vector for cloning and analysis of Streptomyces DNA. Actinomycetologica 8, 21–25.
Rix, U., Fischer, C., Remsing, L. L. & Rohr, J. (2002). Modification of post-PKS tailoring steps through combinatorial biosynthesis. Nat Prod Rep 19, 542–580.[CrossRef][Medline]
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.
Sciara, G., Kendrew, S. G., Miele, A. E., Marsh, N. G., Federici, L., Malatesta, F., Schimperna, G., Savino, C. & Vallone, B. (2003). The structure of ActVA-Orf6, a novel type of monooxygenase involved in actinorhodin biosynthesis. EMBO J 22, 205–215.[CrossRef][Medline]
Sherman, D. H., Malpartida, F., Bibb, M. J., Kieser, H. M., Bibb, M. J. & Hopwood, D. A. (1989). Structure and deduced function of the granaticin-producing polyketide synthase gene cluster of Streptomyces violaceoruber Tü22. EMBO J 8, 2717–2725.[Medline]
Strohl, W. R. (1992). Compilation and analysis of DNA sequences associated with apparent streptomycete promoters. Nucleic Acids Res 20, 961–974.
Taguchi, T., Itou, K., Ebizuka, Y., Malpartida, F., Hopwood, D. A., Surti, C. M., Booker-Milburn, K. I., Stephenson, G. R. & Ichinose, K. (2000a). Chemical characterization of disruptants of the Streptomyces coelicolor A3(2) actVI genes involved in actinorhodin biosynthesis. J Antibiot 53, 144–152.[Medline]
Taguchi, T., Ebizuka, Y., Hopwood, D. A. & Ichinose, K. (2000b). Identification of a novel shunt product produced by a disruptant of the actVI-ORFA gene involved in the biosynthesis of actinorhodin in Streptomyces coelicolor A3(2). Tetrahedron Lett 41, 5253–5356.[CrossRef]
Taguchi, T., Ebizuka, Y., Hopwood, D. A. & Ichinose, K. (2001). A new mode of stereochemical control revealed by analysis of the biosynthesis of dihydrogranaticin in Streptomyces violaceoruber Tü22. J Am Chem Soc 123, 11376–11380.[CrossRef][Medline]
Takano, S., Hasuda, K., Ito, A., Koide, Y., Ishii, F., Haneda, I., Chihara, S. & Koyama, Y. (1976). A new antibiotic, medermycin. J Antibiot 29, 765–768.[Medline]
Tanaka, N., Okabe, T., Isono, F., Kashiwagi, M., Nomoto, K., Takahashi, M., Shimazu, A. & Nishimura, T. (1985). Lactoquinomycin, a novel anticancer antibiotic. I. Taxonomy, isolation, and biological activity. J Antibiot 38, 1327–1332.[Medline]
Tang, L. & McDaniel, R. (2001). Construction of desosamine containing polyketide libraries using a glycosyltransferase with broad substrate specificity. Chem Biol 8, 547–555.[CrossRef][Medline]
Tatsuta, K., Ozeki, H., Yamaguchi, M., Tanaka, M. & Okui, T. (1990). Enantioselective total synthesis of medermycin (lactoquinomycin). Tetrahedron Lett 31, 5495–5498.[CrossRef]
Tornus, D. & Floss, H. G. (2001). Identification of four genes from the granaticin biosynthetic gene cluster of Streptomyces violaceoruber Tü22 involved in the biosynthesis of L-rhodinose. J Antibiot 54, 91–101.[Medline]
