Heterologous production of the antifungal polyketide antibiotic soraphen A of Sorangium cellulosum So ce26 in Streptomyces lividans

Ross Zirkle1,2,{dagger}, James M. Ligon1 and István Molnár1

1 Syngenta Biotechnology Inc., 3054 Cornwallis Road, Research Triangle Park, NC 27709, USA
2 Department of Microbiology, North Carolina State University, 4527 South Gardner Hall, Raleigh, NC 27695, USA

Correspondence
István Molnár
istvan.molnar{at}syngenta.com


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The antifungal polyketide soraphen A is produced by the myxobacterium Sorangium cellulosum So ce26. The slow growth, swarming motility and general intransigence of the strain for genetic manipulations make industrial strain development, large-scale fermentation and combinatorial biosynthetic manipulation of the soraphen producer very challenging. To provide a better host for soraphen A production and molecular engineering, the biosynthetic gene cluster for this secondary metabolite was integrated into the chromosome of Streptomyces lividans ZX7. The upstream border of the gene cluster in Sor. cellulosum was defined by disrupting sorC, which is proposed to take part in the biosynthesis of methoxymalonyl-coenzyme A, to yield a Sor. cellulosum strain with abolished soraphen A production. Insertional inactivation of orf2 further upstream of sorC had no effect on soraphen A production. The genes sorR, C, D, F and E thus implicated in soraphen biosynthesis were then introduced into an engineered Str. lividans strain that carried the polyketide synthase genes sorA and sorB, and the methyltransferase gene sorM integrated into its chromosome. A benzoate-coenzyme A ligase from Rhodopseudomonas palustris was also included in some constructs. Fermentations with the engineered Str. lividans strains in the presence of benzoate and/or cinnamate yielded soraphen A. Further feeding experiments were used to delineate the biosynthesis of the benzoyl-coenzyme A starter unit of soraphen A in the heterologous host.


Abbreviations: ACP, acyl carrier protein; AT, acyl transferase; 6-dEB, 6-deoxyerythronolide B; PAL, phenylalanine ammonia lyase; PKS, polyketide synthase

{dagger}Present address: Martek Biosciences, 4909 Nautilus Court North, Boulder, CO 80301, USA.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Sorangium cellulosum strains are slow-growing, saprophytic and cellulolytic Gram-negative myxobacteria that are also prolific producers of many different bioactive compounds, including the eukaryotic transcription inhibitor sorangicin, the cytotoxic and antitumour compound epothilone, or fungicides like jerangolid A and soraphen A (Reichenbach & Höfle, 1999). Myxobacteria also display a complex social behaviour. When starved for nutrients, myxobacteria aggregate to form a multicellular fruiting body, in which cells differentiate to form myxospores as a culmination of a complex series of developmental events. Swarming motility plays a key role in the coordinated cell movements required for the formation of fruiting bodies, and also leads to the formation of thin, spreading and merging swarms instead of colonies on agar medium in vegetative cultures.

Fermentation and industrial strain development of Sor. cellulosum strains is challenging. Sor. cellulosum strains grow slowly, with doubling times approaching 16 h (Reichenbach, 1999). Sorangium cultures do not grow readily on agar medium when plated under a certain cell density. This behaviour results from a quorum-sensing mechanism that prevents the futile growth of individual cells, reflecting the ecological need for Sorangium strains to form communities that can effectively degrade non-soluble polymeric substrates in the soil (Reichenbach, 1999). Plated above the threshold density, the primary micro-colonies spread and aggregate to form macroscopic swarms. As a result of this process, distinct colonies derived from single cells do not form, making the isolation and propagation of discrete genetic events difficult.

The molecular genetic manipulation of Sor. cellulosum strains is also extremely challenging. Introduction of DNA into these strains relies on low-efficiency intergeneric conjugation from Escherichia coli. Phleomycin resistance has been the only effective selection marker described in Sor. cellulosum until recent reports on the use of a hygromycin resistance marker in strains So ce56 and So ce12 (Pradella et al., 2002; Kopp et al., 2004). No plasmids have been found to replicate in Sor. cellulosum strains, thus the stable maintenance of transferred DNA requires integration into the chromosome via homologous recombination between a cloned DNA fragment and a homologous chromosomal locus (Jaoua et al., 1992). Resolution of the cointegrate formed as a result of homologous recombination is extremely rare, preventing the isolation of gene replacement events.

Soraphen A is an 18-membered macrolide polyketide with an unsubstituted phenyl side ring, produced by Sor. cellulosum So ce26. Soraphen A has a unique mode of action in the inhibition of fungal acetyl-CoA carboxylase (Gerth et al., 1994; Vahlensieck et al., 1994). Due to its potent activity against plant-pathogenic fungi, soraphen A was of considerable commercial interest until it was discovered that it is a weak teratogen. The gene cluster responsible for the biosynthesis of soraphen A in Sor. cellulosum So ce26 has been cloned and sequenced (Ligon et al., 2002; Schupp et al., 1995). The cluster encodes two type I polyketide synthases (PKSs) that together contain a starter module and seven extension modules for the biosynthesis of the soraphen polyketide core, several enzymes that take part in the biosynthesis of the putative polyketide chain extender unit methoxymalonyl-CoA, and at least two enzymes that tailor the nascent polyketide to produce soraphen A (Ligon et al., 2002). Type I PKSs are multifunctional enzymes with domains for substrate recognition and loading (acyl transferase, AT), substrate anchoring (acyl carrier protein, ACP), condensation (ketoacyl synthase, KS) and {beta}-keto processing (ketoreductase, KR; dehydratase, DH; enoyl reductase, ER). These domains are organized into modules and each module is responsible for one round of polyketide chain extension using acyl-CoA substrates (Hopwood, 1997). Apart from the common substrates malonyl- and methylmalonyl-CoA, the soraphen PKS also uses benzoyl-CoA for chain initiation and a ‘glycolate’ unit, probably methoxymalonyl-CoA, during chain extension.

Heterologous expression of secondary metabolic gene clusters in surrogate hosts is emerging as a viable alternative to both classical strain and fermentation process development, and molecular biological manipulation of the native producer strain (Pfeifer & Khosla, 2001). Heterologous expression of these clusters can provide strains with better fermentation characteristics that are more amenable to further optimization, or present the only fermentation-compatible alternative as with sponge- or symbiont-derived metabolites. Surrogate hosts with well developed genetic systems and higher production levels also open up the possibility of combinatorial genetic derivatization of secondary metabolites originating from genetically intractable strains for structure–activity relationship studies. Heterologous production of type I polyketides is, however, extremely challenging because of the large size (up to 10 kDa) of the synthases, the requirement for post-translational modification of the synthases by phosphopantheteinylation, the need for the availability of the activated substrates and our limited knowledge of the regulation of the interaction of primary and secondary metabolism. Despite these difficulties, the last few years have witnessed a growing number of reports on the successful production of these metabolites in surrogate hosts. Thus, Streptomyces coelicolor and its close relative Streptomyces lividans were used to express the polyketide cores for erythromycin, oleandomycin, picromycin, megalomicin, epothilone and their genetically engineered variants (Kao et al., 1994; Shah et al., 2000; Tang et al., 1999, 2000; Volchegursky et al., 2000; Xue et al., 1999). 6-Methylsalicylate and 6-deoxyerythronolide B (6-dEB, the macrolide core of erythromycin) were produced in yeast (Kealey et al., 1998), 6-dEB and yersiniabactin were manufactured in E. coli (Pfeifer et al., 2001, 2003) and epothilone was biosynthesized in a Myxococcus xanthus heterologous host (Julien & Shah, 2002). Although the initial titres with these engineered strains were usually low (0·2–50 mg l–1), further improvements in the production processes by classical microbiological or metabolic engineering methodologies are becoming increasingly feasible (Desai et al., 2002; Lau et al., 2002; Murli et al., 2003).

The formidable difficulties with the fermentation, classical strain improvement and molecular genetic manipulation of soraphen production in Sor. cellulosum So ce26 made it desirable, while the characterization of the gene cluster for soraphen A production made it feasible, to investigate a heterologous expression strategy of this metabolite in a more amenable host. The downstream border of the soraphen A biosynthetic gene cluster was reported previously by our group (Schupp et al., 1995). In this study, we have established the upstream border of the soraphen biosynthetic gene cluster by insertional inactivation of sorC and orf2 in Sor. cellulosum So ce26. Next, the soraphen PKS genes sorA and sorB, together with two genes downstream of sorB, the soraphen C methyltransferase sorM and orf4 of unknown function (Ligon et al., 2002), were cloned into Streptomyces expression vectors and integrated into the Str. lividans ZX7 chromosome. The genes upstream of sorA that were implicated in soraphen A production (sorR, sorC, sorD, sorF and sorE), were also introduced into the engineered Str. lividans strain together with a gene from Rhodopseudomonas palustris that encodes a benzoate-coenzyme A ligase (Egland et al., 1995). The resulting strains of Str. lividans produced soraphen A when sodium benzoate or trans-cinnamic acid was supplied to the fermentations. Further feeding experiments were used to shed light on the biosynthesis of the soraphen starter unit benzoyl-CoA in the engineered Str. lividans strain.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Strains and plasmids.
The streptomycin-resistant mutant SJ3 (Jaoua et al., 1992) of Sor. cellulosum So ce26 (Gerth et al., 1994) was used for these studies. E. coli DH10B (Invitrogen) and SURE (Stratagene) were used for routine cloning, and E. coli ET 12567 with plasmid pUZ8002 (MacNeil et al., 1992) was utilized for conjugative plasmid transfer. Str. lividans ZX7 (John Innes Centre, Norwich, UK) is a derivative of Str. lividans 66. Candida albicans (type strain NRRL Y-12983; Agricultural Research Service Culture Collection, Peoria, IL, USA) is sensitive to soraphen A and was used as an indicator strain for soraphen production assays. E. coli cloning vectors pBluescript II SK (Stratagene), pCR-Blunt II-Topo (Invitrogen) and pNEB193 (New England Biolabs) were used for routine cloning. The mobilizable plasmid pCIB132 was used for conjugative plasmid transfer (Schupp et al., 1995). The expression vector pTBK (I. Molnár, unpublished) contains the thiostrepton-inducible promoter PtipA (Murakami et al., 1989) upstream of the cloning sites PacI and PmeI, the kanamycin resistance marker aphII (Kieser et al., 2000) and the {phi}C31 integrase-attP site (Kuhstoss & Rao, 1991) for site-specific recombination into the chromosome of streptomycetes. The expression vector pTBBH (I. Molnár, unpublished) contains the PtipA promoter upstream of the cloning sites PacI and PmeI, a hygromycin resistance marker (Kieser et al., 2000) and the IS117 transposase and attachment site attM (Henderson et al., 1990) for site-specific integration into the chromosome of streptomycetes. Plasmid pTUE (I. Molnár, unpublished) contains the Streptomyces origin of replication and replication protein gene from pIJ101 (Kieser et al., 2000), the PtipA promoter upstream of the cloning sites PacI and PmeI and an erythromycin resistance marker (Kieser et al., 2000). All three expression vectors also contain the pUC18 origin of replication and an ampicillin resistance marker for propagation in E. coli.

Media and growth conditions.
E. coli strains were grown at 37 °C in Luria broth or on Luria agar with the appropriate antibiotics. Sor. cellulosum So ce26 SJ3 was grown at 30 °C on S42 agar or in G51t broth (Jaoua et al., 1992). Str. lividans ZX7 and its derivatives were grown on ISP-2 agar for sporulation, R5 agar for protoplast regeneration and YEME broth for soraphen production (Kieser et al., 2000), supplemented with antibiotics if necessary. C. albicans was grown at 30 °C in Bacto potato glucose broth (PDB) or agar (PDA; Becton Dickinson).

DNA manipulations.
Routine cloning and transformation procedures for E. coli were as described by Sambrook & Russell (2001). PCR for cloning was performed using Pfu polymerase and, for the analysis of strains, with Herculase polymerase (Stratagene) with 10 % (final concn) dimethyl sulfoxide supplementing each reaction. Oligonucleotides were obtained from Integrated DNA Technologies (Coralville, IA, USA). Genomic DNA was isolated from Sor. cellulosum So ce26 SJ3 with the Puregene Kit (Gentra) and from Str. lividans ZX7 with DNAzol (Invitrogen). The isolation of plasmid DNA from Str. lividans ZX7 was done with Tip-100 columns (Qiagen) as described by Kieser et al. (2000).

Gene disruptions.
Genomic DNA of Sor. cellulosum So ce26 SJ3 was used as template for PCR reactions to amplify internal fragments of the sorC and orf2 coding regions. The SC1 primers (Table 1) that contain NotI sites were used to amplify a 1506 bp fragment internal to the coding region of sorC and cloned into pCR-Blunt II-Topo to create pTsorCko. Plasmid pTsorCko was digested with NotI and the sorC fragment was cloned into NotI-digested pCIB132 to generate pCIB132sorCko. Similarly, the O2 primers (Table 1) were used to generate a 1995 bp fragment internal to orf2, and the amplicon was cloned into pCR-Blunt-Topo to generate pTorf2ko. The orf2 fragment was then cloned into pCIB132 to create pCIB132orf2ko as described above for pCIB132sorCko. Introduction of plasmids pCIB132sorCko and pCIB132orf2ko into Sor. cellulosum So ce26 SJ3 by conjugation and selection of transconjugants were carried out as described by Jaoua et al. (1992).


View this table:
[in this window]
[in a new window]
 
Table 1. Oligonucleotide primers used in PCR reactions

Restriction sites are in bold and the ribosome-binding site is in italic.

 
Construction of soraphen expression vectors.
The primers used in the construction of the following vectors are listed in Table 1. To construct an integrative expression plasmid with the sorA gene, a PCR fragment containing the start codon to a SmaI site within the sorA gene was amplified from cosmid pM15-5 (Ligon et al., 2002). The 370 bp PCR product was cloned as a KpnI–SpeI fragment by sites incorporated from the primers (PS1) into KpnI/SpeI-digested pBluescript II SK(+) to yield pSorAs. A second PCR reaction was used to amplify the 3' end of sorA from cosmid p98/1 (Ligon et al., 2002) from a HindIII site to the stop codon. The PCR product was cloned as an SpeI–XbaI fragment by sites incorporated from the amplification primers (PS2) into SpeI/XbaI-digested pBluescript II SK(+) to yield pSorAe that was digested with XbaI/SpeI. The 1·6 kb fragment was ligated to SpeI/XbaI-digested pSorAs to yield pSorAlink that was further digested with SmaI and HindIII and ligated to a 6·3 kb SmaI–HindIII fragment isolated from cosmid pM15-5. The resulting plasmid, pSorA3m15, was then digested with HindIII and ligated with a 10·6 kb HindIII fragment from cosmid p98/1. The resulting plasmid, pBSsorA, was digested with PacI/PmeI and the 19 kb fragment was ligated into the PacI/PmeI-digested expression vector, pTBK (I. Molnár, unpublished), yielding pSorA.

An integrative expression plasmid that contains the genes sorB, orf4 and sorM was constructed as follows. A 546 bp fragment containing the sorB start codon to a Bst1107I site within sorB was PCR-amplified from cosmid p98/1 and cloned into PacI/PmeI-digested pNEB193 as a PacI–PmeI fragment using restriction sites from the amplification primers (PS3) to yield pSorBs. This plasmid was digested with Bst1107I and SanDI (present in PS3) and ligated to a 29·8 kb Bst1107I–SanDI fragment from cosmid p98/1 containing the rest of sorB, the flanking genes orf4 and sorM, and the C-terminal part of orf5. The resulting plasmid, pNEBsorB was digested with PacI/PmeI and the 30·3 kb fragment was ligated into the PacI/PmeI-digested expression vector pTBBH (I. Molnár, unpublished), yielding pSorB4M.

Two expression plasmids containing the genes upstream of sorA were constructed. sorR was PCR-amplified from cosmid pM15-5 and cloned into XbaI/NdeI-digested pNEB193 as an NdeI–XbaI fragment using sites introduced by the amplification primers (PS4) to yield pSorR. A DNA fragment containing the genes sorC, sorD, sorF and sorE was amplified by PCR from cosmid pM15-5 with primers (PS5) that introduced a ribosome-binding site (AGGAGG) upstream of sorC, and was cloned into the pCR-Blunt II-Topo vector, creating pTsorCDFE. This plasmid was digested with EcoRI and SpeI (sites introduced by PS5) and the 6 kb fragment was ligated into EcoRI/SpeI-digested pSorR to yield pNEBsorRCDFE. The badA gene was PCR-amplified (PS6) from plasmid pPE202 (Egland et al., 1995) with SpeI and XbaI restriction sites as well as the ribosome-binding site (AGGAGG) incorporated into the PCR product, and was cloned into pCR-Blunt II-Topo vector, creating pTbadA. This plasmid was digested with SpeI/XbaI and the fragment was cloned into SpeI/XbaI-digested pNEBsorRCDFE to create pNEBsorRCDFE+BL. The PacI–PmeI fragments of plasmids pNEBsorRCDFE and pNEBsorRCDFE+BL were cloned into the Streptomyces expression vector pTUE (I. Molnár, unpublished), yielding pSorRCDFE and pSorRCDFE+BL, respectively.

Construction of Str. lividans ZX7 strains producing soraphen.
Plasmids pSorA and pSorB4M were co-transformed into protoplasts of Str. lividans ZX7 by polyethylene-glycol-mediated transformation (Kieser et al., 2000). The transformants were selected on R5 medium with hygromycin (100 µg ml–1) and kanamycin (25 µg ml–1) for the two integration events. Resistant strains were analysed by Southern hybridizations (data not shown) and strains with the expected integration events were labelled SorAB. Plasmids pSorRCDFE or pSorRCDFE+BL were transformed into protoplasts of SorAB and selected with hygromycin and kanamycin for the integration events and erythromycin (200 µg ml–1) for the presence of the plasmid. These strains were analysed for the structural integrity of the two integrated expression cassettes by a series of PCR reactions covering the sorA (primer pairs SA1–4; Table 1) and the sorB, orf4 and sorM genes (primer pairs SB1–6, Table 1), and the ends of these overlapping 3–6 kb amplicons were sequenced and compared to the published sequence of the soraphen cluster (accession no. U24241). The plasmid pTUE was also transformed into Str. lividans ZX7 to create strain TUE. The presence of the plasmids pTUE, pSorRCDFE or pSorRCDFE+BL was ascertained by plasmid isolation and restriction analysis. Three isolates were selected from each transformed strain and named TUE, SorABRCDFE and SorABRCDFE+BL.

Fermentation of Str. lividans strains.
To test for the production of soraphen A, spores and vegetative mycelia from about 2 cm2 patches of Str. lividans strains TUE, SorAB, SorABRCDFE and SorABRCDFE+BL were inoculated into 25 ml YEME, supplemented with 20 µg erythromycin ml–1 for the plasmid-containing strains, and incubated at 30 °C with shaking at 225 r.p.m. for 2 days. Six millilitres of these starter cultures was used to inoculate 600 ml YEME supplemented with 5 µg thiostrepton ml–1, or 5 µg thiostrepton ml–1 and 20 µg erythromycin ml–1 as applicable. The cultures were grown at 30 °C with shaking at 225 r.p.m. Some of the cultures were also supplemented with one of the following (final concn): 2·5 mM trans-cinnamic acid, 5 mM L-phenylalanine, 5 mM sodium benzoate, 5 mM phenylpyruvic acid, 2·5 mM phenylacetate or 2·5 mM benzaldehyde (all from Sigma), where each stock solution had been adjusted to pH 7·5 with sodium hydroxide. After 3 days of growth, the cultures were freeze-dried and extracted with 500 ml methanol with shaking at 100 r.p.m. overnight at room temperature. The mixtures were centrifuged at 9000 g for 1 h, the methanol phases were concentrated in a SpeedVac and the extracts were redissolved in a final volume of 10 ml methanol. Fermentations were carried out with three independent isolates of each strain and repeated twice.

Sor. cellulosum fermentations.
To analyse strains of Sor. cellulosum for the production of soraphen, 250 ml G51t media were inoculated with 25 ml 5-day-old starter cultures and incubated at 30 °C with shaking at 225 r.p.m. After 12 days, the cultures were freeze-dried and extracted with 100 ml methanol by slow mixing overnight. The mixtures were centrifuged at 9000 g for 30 min, then the methanol phases were concentrated in a SpeedVac and redissolved in 5 ml methanol. Fermentations were carried out with four independent isolates of each strain and repeated twice.

Bioassay for soraphen A production.
Ten to 50 µl from the methanol extracts was spotted onto PDA plates and allowed to dry. An overnight culture of C. albicans was used to seed soft PDA (0·7 % agar) and used to overlay the PDA plates containing the dried extracts. Plates were incubated at 30 °C for a maximum of 3 days and observed every day for the inhibition of growth of C. albicans.

HPLC and LC-MS analysis.
Extracts from strains were analysed by reverse phase HPLC and LC-MS. HPLC was performed on an HP1100 series (Agilent Technologies) fitted with a Synergy 4 micron MAX-PR 80 Å 150x4·6 mm column (Phenomenex). Mobile phase A was 0·01 % trifluoroacetic acid (TFA) in water and mobile phase B was 0·0075 % TFA in acetonitrile. Elution was performed with a gradient of 25–85 % B over 15 min at 40 °C with a flow rate of 0·8 ml min–1 and monitored by a diode array detector. Soraphen A was detected by its absorption at 210 nm at a retention time of 14 min.

LC analysis of extracts was carried out on a HP1100 (Agilent Technologies) with the column described above. The gradient was 25–85 % methanol in water in the presence of 0·1 % sodium formate and 0·1 % formic acid with a flow rate of 0·4 ml min–1. Separation was monitored by a Finnigan LCQ Classic Mass Spectrometer (ThermoFinnigan) using positive mode ESI (capillary temp, 275 °C; capillary voltage, 3·22 V; spray needle voltage, 4·54 kV).


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The upstream border of the soraphen cluster
The soraphen biosynthetic gene cluster has been cloned, sequenced and characterized (Ligon et al., 2002). Genes downstream of the soraphen C methyltransferase gene, sorM (Fig. 1), showed homologies to primary metabolic genes in sugar metabolism (orf6 and orf7) and to hypothetical proteins of unknown function (orf5). Gene disruptions in this region had no effect on the production of any soraphen congener in Sor. cellulosum So ce26 SJ3 (Schupp et al., 1995), and consequently these genes were considered not to be part of the cluster. Genes immediately upstream of the sorA PKS (Fig. 1) have been proposed to take part in soraphen B oxidation (sorR) and in the formation of the putative polyketide extender unit methoxymalonyl-CoA (sorC,D,E). The putative products of genes further downstream of sorE had tentatively been proposed to form the border of the soraphen cluster (Ligon et al., 2002), based on their lack of homology to other proteins in the databases (orf1 and orf2) or their similarity to only conserved hypothetical proteins (orf3). However, there was no direct evidence that suggested the involvement of the sorC,D,E gene products in soraphen biosynthesis or, conversely, the absence of such involvement of the orf13 gene products.



View larger version (17K):
[in this window]
[in a new window]
 
Fig. 1. Heterologous expression strategy for the soraphen cluster. The arrows are not drawn to scale and their shading is coordinated between the cluster and the expression cassettes. Arrows with stripes represent ORFs that are not proposed to take part in soraphen biosynthesis. (a) Genes in the soraphen biosynthetic locus. (b) Expression cassettes in plasmids used to reconstruct the soraphen biosynthetic cluster in Str. lividans. The triangles represent the PtipA promoter. Only the 5' half of orf5 is included in pSorB4M. BL, badA gene from Rhodopseudomonas palustris (Egland et al., 1995).

 
Gene disruption in Sorangium cells is a technically challenging task that can be accomplished by insertional inactivation only, necessitated by the prohibitively low frequency by which genetic replacements are recovered. Insertional inactivation is in itself an inefficient process and has been described to require at least 1 kb of homologous DNA to direct the recombination process (Pradella et al., 2002). Since the genes sorD, E and orf3 are relatively short and thus do not offer practical targets for gene disruption, the genes orf2 and sorC were targeted to establish the upstream border of the soraphen biosynthetic gene cluster. Internal fragments of sorC and orf2 were cloned into the mobilizable plasmid pCIB132 and conjugated into Sor. cellulosum (Jaoua et al., 1992) as described in Methods. Phleomycin-resistant transconjugants were isolated and the integration events at sorC or orf2, respectively, were confirmed by Southern hybridization (results not shown). Bioassays with C. albicans as an indicator strain (see Methods) showed the presence of large zones of growth inhibition for the wild-type SJ3 control strain and for the strains where orf2 was insertionally inactivated, while the strains with the sorC knockout displayed no zones of inhibition (Fig. 2). Further analysis of the extracts by HPLC and LC-MS confirmed that the production of soraphen A remained undisturbed in the orf2 knockout strain, but was completely abolished in the sorC mutant (Fig. 2). Thus, the upstream border of the cloned soraphen locus was localized between orf2 and sorC.



View larger version (37K):
[in this window]
[in a new window]
 
Fig. 2. Analysis of soraphen A production in Sor. cellulosum strains. (a) Bioassay of soraphen A production in orf2 knockout (KO) strains (1–3) and wild-type Sor. cellulosum So ce26 SJ3 (4). (b) Bioassay of soraphen A production in sorC knockout (KO) strains (1–3) and wild-type Sor. cellulosum So ce26 SJ3 (4). (c–f) HPLC analysis of soraphen A (retention time 14·0 min) production. (c) Authentic soraphen A standard; (d) extract from a wild-type Sor. cellulosum So ce26 SJ3 fermentation; (e) extract from a fermentation of an orf2 knockout strain; (f) extract from a fermentation of a sorC knockout strain. (g) Mass spectrum of the peak corresponding to soraphen A (M+Na ion, m/z 543) in the extract from a fermentation of an orf2 knockout strain in LC-MS. Similar mass spectra can be obtained with extracts from all orf2 knockout isolates, extracts of the wild-type strain and from the authentic soraphen A standard. (h) Mass spectrum of the fraction from the extract of a fermentation of a sorC knockout strain corresponding to the retention time of soraphen A in LC-MS (34·8 min).

 
Construction of Str. lividans ZX7 strains carrying the soraphen A gene cluster
To transfer the soraphen A gene cluster to Str. lividans ZX7, the soraphen A PKS genes, sorA and sorB, the methyltransferase gene, sorM, and orf4 were introduced into the chromosome of Str. lividans ZX7 using two compatible integrative expression plasmids. Both sorA and the sorB-orf4-sorM putative operon were placed under the control of the thiostrepton-inducible PtipA promoter (Murakami et al., 1989) that is widely used in Streptomyces expression vectors. The expression plasmid containing the sorA gene (Fig. 1) was integrated into the chromosome of Str. lividans ZX7 by site-specific integration mediated by the integrase and attP site from the Streptomyces phage {phi}C31 (Kuhstoss & Rao, 1991). The expression plasmid (Fig. 1) containing the sorB-orf4-sorM putative operon was integrated into the chromosome of Str. lividans ZX7 by site-specific integration mediated by the transposase and attM site from IS117, a transposable element from Str. coelicolor A3(2) (Henderson et al., 1990). The chromosomal integration of both expression plasmids containing the sorA and the sorB-orf4-sorM cassettes were confirmed by Southern hybridization analysis (data not shown) and the corresponding strain was named Str. lividans SorAB.

The genes upstream of sorA were supplied to Str. lividans SorAB on a replicative expression vector that is compatible with the pSorA and the pSorB4M integrative vectors (Fig. 1). During the course of this work, it was brought to our attention that a small ORF encoding an ACP, not reported in our previous analysis, is located between the sorD and the sorE genes (Tin-Wein Yu, personal communication). This ORF, that we named sorF, extends from bp 8930 to 9181 (on the complementary strand) in the published sequence of the soraphen cluster (U24241) and would link the transcriptionally coupled sorCD genes to the sorE gene in an apparent operon.

Thus, the sorR gene and the sorCDFE operon were placed under the control of the thiostrepton-inducible PtipA promoter in an artificial operon setting. Since benzoate is the starter unit for soraphen biosynthesis (Hill et al., 2003), but no genes for benzoate biosynthesis or activation were found clustered with the soraphen biosynthetic genes, the benzoate-coenzyme A ligase gene badA from Rhodopseudomonas palustris (Egland et al., 1995) was also fused to this artificial operon in construct pSorRCDFE+BL. This gene has previously been shown to be necessary to supply benzoyl-CoA for the biosynthesis of phenyl-substituted lactones in a Saccharopolyspora erythraea expression system utilizing the starter module of the soraphen PKS fused to a triketide lactone synthase derived from the erythromycin PKS (Wilkinson et al., 2001). We have inserted copies of a Streptomyces ribosome-binding site (AGGAGG) into sorRCDFE, the sorRCDFE-badA artificial operons 12 bases upstream of the start codon of sorC and also upstream of the badA gene, as described in Methods. This ribosome-binding site occurs in the PtipA promoter and is widely distributed in genes that are highly expressed in Streptomyces (Strohl, 1992). The expression plasmids pSorRCDFE and pSorRCDFE+BL were transformed into Str. lividans SorAB to produce Str. lividans SorABRCDFE and Str. lividans SorABRCDFE+BL. To ensure that no spurious recombination events disrupted the integrated PKS genes on the Str. lividans chromosome, the sorA and sorB-orf4-sorM expression cassettes were analysed in a series of overlapping PCR reactions as described in Methods. Due to the high level of homology amongst the repeated PKS domains, Southern hybridizations using probes derived from the soraphen PKS were of limited use for this purpose.

Soraphen production with the engineered Str. lividans strains fed with benzoate
Strains of Str. lividans ZX7 carrying the expression vector pTUE, and Str. lividans SorAB, SorABRCDFE and SorABRCDFE+BL were fermented in the presence of thiostrepton to induce the expression of the sorA, and the sorB4M, sorRCDFE or sorRCDFE-badA operons. The fermentations were analysed by Candida bioassays, HPLC and LC-MS, but no trace of soraphen A or soraphen-related compounds were detected in any of the cultures (Table 2). L-Phenylalanine was shown previously to be the precursor of benzoyl-CoA used for soraphen biosynthesis by Sor. cellulosum So ce26 (Gerth et al., 2003; Hill & Thompson, 2003). The Str. lividans fermentations were repeated, feeding L-phenylalanine to the cultures, but again no production of soraphen could be detected. Since benzoate feeding and the benzoate-coenzyme A ligase encoded by badA were both found to be necessary for the production of phenyl-substituted lactones in engineered Sac. erythraea systems (Wilkinson et al., 2001), we have also tried feeding benzoate to the cultures. Str. lividans SorABRCDFE+BL, but none of the other strains, produced a compound that showed antifungal activity against C. albicans (Fig. 3). HPLC analysis revealed a small peak co-migrating with authentic soraphen A standard in extracts of SorABRCDFE+BL only, and LC-MS assays showed the presence of the appropriate molecular ion (m/z 543 for the sodium adduct) in the peak co-eluting with soraphen A at 34·8 min. While the production of soraphen A under these initial conditions was less than 0·3 mg l–1, these experiments indicated that the Str. lividans host is able to produce fully processed soraphen A. The results suggested that the host does not provide an adequate supply of benzoyl-CoA for the pathway, and also that externally supplied benzoate could be channelled into soraphen biosynthesis only in the presence of a suitable coenzyme A ligase.


View this table:
[in this window]
[in a new window]
 
Table 2. Soraphen A production by Str. lividans strains

NF, No feeding; Phe, phenylalanine; B, benzoate; Cin, cinnamate; PP, phenylpyruvate; PA, phenylacetate; BA, benzaldehyde; ND, not done.

 


View larger version (32K):
[in this window]
[in a new window]
 
Fig. 3. Analysis of soraphen A production in Str. lividans strains. (a) Bioassay of soraphen A production in fermentations with Str. lividans SorAB fed with benzoate (1) and three independent isolates of Str. lividans SorABRCDFE+BL fed with benzoate (2–4). (b) HPLC trace of an extract from a fermentation with Str. lividans SorABRCDFE+BL fed with benzoate (soraphen A retention time 14·0 min). (c) LC-MS selected ion chromatogram (m/z 541–545) of an extract from a benzoate-supplemented fermentation of Str. lividans SorABRCDFE+BL (soraphen A retention time 34·8 min). (d) LC-MS of the peak corresponding to soraphen A in the extract from a benzoate-supplemented fermentation of Str. lividans SorABRCDFE+BL (soraphen A M+Na ion, m/z 543).

 
Further feeding experiments
To establish a suitable biosynthetic route for benzoyl-CoA in the engineered Str. lividans strains, various aromatic compounds that were implicated in benzoyl-CoA biosynthesis in different organisms were fed to Str. lividans SorAB, SorABRCDFE and SorABRCDFE+BL (Table 2). Cinnamic acid has successfully been used to label the phenyl ring of soraphen in Sor. cellulosum (Hill & Thompson, 2003) and forms part of the benzoyl-CoA biosynthetic route from phenylalanine in ‘Streptomyces maritimus’, producer of enterocin (Hertweck et al., 2001). Feeding trans-cinnamate led to the production of soraphen in Str. lividans SorABRCDFE+BL and SorABRCDFE (but not in TUE or SorAB), in an amount similar to that obtained with benzoate supplementation (Table 2). Phenylpyruvate and phenylacetate are intermediates in the anaerobic catabolism of phenylalanine to benzoyl-CoA in Thauera aromatica (Schneider et al., 1997), and benzaldehyde has been suggested as a possible intermediate in benzoate biosynthesis in plants (Moore & Hertweck, 2002, and ibid.). Feeding these compounds to the engineered Str. lividans strains, however, did not lead to a detectable production of soraphen A (Table 2).


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The biosynthetic gene cluster for soraphen A production in Sor. cellulosum So ce26 (Schupp et al., 1995) contains genes for a Type I PKS involved in the biosynthesis of the soraphen core, substrate supply genes encoding enzymes proposed to be involved in methoxymalonyl-CoA biosynthesis and genes encoding polyketide tailoring activities that process the nascent product of the PKS (Ligon et al., 2002). Gene disruption has been used to demonstrate the involvement of the PKS genes sorA and sorB in soraphen biosynthesis and localized the downstream border of the gene cluster 5' to the sorM soraphen C O-methyltransferase (Schupp et al., 1995). In this study, we have used insertional inactivation of the sorC and the orf2 genes (Fig. 1) to establish the upstream border of the cluster. Disruption of orf2 had no effect, while insertional inactivation of sorC completely abolished soraphen A production in Sor. cellulosum So ce26 SJ3 (Fig. 2), indicating that the upstream border of the cluster lies somewhere between these two genes. While orf2 shows no homology to known genes, sorC forms an operon with the sorD and sorE oxidases, and the newly described sorF ACP (Tin-Wein Yu, University of Washington, personal communication), which together have been proposed to biosynthesize the unusual polyketide extender unit methoxymalonyl-CoA (Ligon et al., 2002). PKSs most commonly utilize malonyl- or methylmalonyl-CoA substrates, but unorthodox positioning of hydroxyl- or methoxy groups in some polyketides in combination with biosynthetic labelling experiments has suggested the use of extender units derived from glucose or glycerol. Such ‘glycolate’ units are seen in ansamitocin, geldanamycin, leucomycin, FK520 and FK506, as well as in soraphen A (Haber et al., 1977; Hill et al., 1998; Omura et al., 1983). Methoxymalonate biosynthetic subclusters have been characterized in the ansamitocin and FK520 biosynthetic gene clusters (asm13-17 and fkbGHIJK, respectively; Wu et al., 2000; Yu et al., 2002). Disruption of asm13-17 abrogated ansamitocin production in Actinosynnema pretiosum (Yu et al., 2002). Furthermore, co-expression of asm13-17 in Str. lividans with a cassette of the erythromycin PKS in which one of the AT domains had been replaced with an AT that apparently specifies methoxymalonate (Kato et al., 2002) led to the production of a 6-dEB derivative with a methoxymalonate incorporated in the expected position. The products of sorD and sorE, which are homologous to asm13 and 15, and to fbkK and I, respectively, are proposed to be involved in the oxidation of glyceryl-ACP to hydroxymalonyl-ACP. sorF, asm14 and fkbJ all encode type II ACP that might tether the substrate during these oxidations. The loading of the unknown glycerate metabolite onto the ACP and O-methylation of hydroxymalonate to methoxymalonate requires the asm16 and 17, or the fkbH and G products, respectively. The soraphen cluster contains no homologues of asm16 or fkbH, but SorC that features AT, ACP and O-methyltransferase domains residing on a single enzyme represents a unique alternative to the functions of these gene products. The presence of two ACPs (sorF and the ACP domain within sorC) within the proposed soraphen methoxymalonate synthase is unprecedented and would require further investigation. The failure of the sorC knockout strain to produce soraphen A implies that either a functional SorC, or the products of the downstream genes sorD, F and E that might have been affected by the polar effects of the insertion, are necessary for the biosynthesis of soraphen A. In either case, the putative operon formed by the sorC, D, F and E genes is indispensable for soraphen A production and thus was included in our design of a heterologous production strategy (Fig. 1) that would also allow a further functional dissection of this operon.

Heterologous production of polyketide natural products or their derivatives may circumvent the need for strain optimization and fermentation development for every native production strain, some of which cannot (sponges, symbionts, unculturable strains) or have not (environmental DNA libraries) been grown in standard laboratory conditions. Further, horizontal transfer of natural product biosynthetic gene clusters into host strains with well-developed genetics should facilitate the creation of ‘unnatural’ natural products by combinatorial genetics (Pfeifer & Khosla, 2001), a task that has proven especially elusive with the soraphen A producer strain Sor. cellulosum So ce26.

Biosynthetic gene clusters with type I modular PKS have been expressed in several host strains. Thus, the synthase for 6-dEB has been expressed in E. coli together with the Sfp phosphopantheteinyl transferase to yield 10 mg of the macrolide product l–1. E. coli does not produce methylmalonyl-CoA, so the biosynthesis of this substrate of the PKS had to be engineered into the host, requiring a substantial number of genetic modifications (Dayem et al., 2002; Pfeifer et al., 2001). The Sorangium-derived natural product epothilone (Molnár et al., 2000) has been expressed in the myxobacterium host Myxococcus xanthus with an initial yield of 0·2 mg l–1 (Julien & Shah, 2002). Industrial antibiotic-producer Actinomyces strains, optimized for high production levels during fermentation, seem to be especially well suited for heterologous production of polyketides (Martin et al., 2003; Rodriguez et al., 2003), but these strains are often proprietary and their molecular genetic manipulation is far from trivial. Str. coelicolor, and its close relative Str. lividans, have been used much more extensively for the production of complex polyketides, including the macrocyclic core of erythromycin (Kao et al., 1994; Ziermann & Betlach, 1999), megalomicin (Volchegursky et al., 2000), picromycin (Tang et al., 1999), oleandomycin (Shah et al., 2000) or epothilone (Tang et al., 2000). Engineered variants of these metabolites (Tang et al., 2000; Xue et al., 1999), or glycosylated derivatives (Tang & McDaniel, 2001), were also produced in Str. lividans, with the initial titre of the heterologously produced metabolites ranging from 0·2 to 50 mg l–1 in the two host strains (Pfeifer & Khosla, 2001). Both Str. lividans and Str. coelicolor are able to synthesize malonyl- and methylmalonyl-CoA that serve as substrates for these PKS. Both strains also activate type I PKS by phosphopantheteinylation, with the Str. coelicolor PPTase reported to be promiscuous in recognizing diverse synthases as its substrate (Cox et al., 2002). Str. coelicolor also has a wide range of transporters (Bentley et al., 2002) that allowed the secretion of heterologously produced aromatic polyketides (Pfeifer & Khosla, 2001). Based on its close taxonomic position, Str. lividans would also be likely to offer a broad-spectrum PPTase and diverse antibiotic transport systems, and has been described to support similar levels of production of type I polyketides as an expression host to those of Str. coelicolor (Ziermann & Betlach., 1999). The absence of a requirement for methylation-free DNA for genetic transformation of Str. lividans, however, made this strain our preferred choice for the heterologous production of soraphen. Although single-plasmid expression systems have proved useful for the production of complex polyketides, the discovery that PKS complexes readily assemble from subunits expressed from multiple, chromosomally integrated or plasmid-based constructs (Kuhstoss et al., 1996; Ziermann & Betlach, 2000) has simplified the assembly, transfer and later modification of these large gene clusters that often cover 30–100 kb. In the present work, we have used a three-construct approach, integrating the PKS genes (together with a post-PKS methyltransferase and an ORF of unknown function) into two loci in the Str. lividans chromosome for added stability, and supplying the sorR oxidase and the methoxymalonate biosynthetic operon sorCDFE on an autonomously replicating plasmid for added convenience. All these genes were placed under the control of the thiostrepton-inducible promoter PtipA (Murakami et al., 1989) to provide for regulated expression of the cluster during fermentation (Ali et al., 2002; aspects under investigation in our laboratory) that is also decoupled from the developmental regulation of secondary metabolic promoters (Kyung et al., 2001).

The yield of heterologously produced soraphen A was less than 0·3 mg l–1 in these experiments, six to ten times lower than the initial reported titres with the wild-type Sor. cellulosum So ce26 strain (Gerth et al., 1994), but within the range of other type I polyketides produced in Str. lividans or Str. coelicolor (Pfeifer & Khosla, 2001). Although several years of industrial fermentation optimization and strain development had increased the titre of So ce26 fermentations to 150 mg l–1 (Gerth et al., 1994) and later to 1 g l–1 (Gerth et al., 2003), the strain is still largely refractory to molecular genetic manipulations, and challenging for classical genetic manipulations, while its slow growth (generation time 16 h; Reichenbach, 1999) makes industrial-scale fermentations contamination-prone and the volumetric productivity is low. Despite the low initial titres, the Str. lividans soraphen production system might offer a more attractive starting point for fermentation optimization, and microbiological and genetic strain development. Indeed, both classical fermentation optimization and metabolic engineering has been used to increase low initial production levels in heterologous polyketide production systems. Introduction of an additional methylmalonyl-CoA supply pathway increased 6-dEB production in Str. coelicolor from 40 to 200 mg l–1 (Lombo et al., 2001), while media and fermentation regimen optimization boosted the production levels to 1·5 g l–1 (Desai et al., 2002). Fermentation optimization and utilization of an absorber resin to sequester the produced polyketide allowed epothilone titres to be increased from 0·2 to 23 mg l–1 in a Myxococcus xanthus heterologous production system (Julien & Shah, 2002; Lau et al., 2002). A concerted approach of metabolic engineering, expression system optimization and fermentation development allowed the production of g l–1 quantities of 6-dEB even in E. coli (Murli et al., 2003; Pfeifer et al., 2002). The availability of the Str. coelicolor genome sequence (Bentley et al., 2002), together with our increasing understanding of its transcriptome and metabolome (Avignone-Rossa et al., 2002; Huang et al., 2001) also raises the possibility of a rational and rapid metabolic engineering of this host and, by extension, its close relative Str. lividans. Random mutagenesis coupled with whole-genome shuffling could provide an evolutionary approach aimed at rapid optimization of strains for polyketide metabolite production (Zhang et al., 2002).

The biosynthesis of soraphen A is initiated by AT1a in the soraphen starter module loading a benzoate unit from benzoyl-CoA onto ACP1a (Ligon et al., 2002; Wilkinson et al., 2001). The biosynthetic origin of the benzoyl-CoA substrate in Sor. cellulosum is not clear, however. Precursor feeding studies showed the incorporation of label into soraphen from cinnamate and phenylalanine, the latter often quoted as the ultimate source for the phenyl moiety of soraphen A (Gerth et al., 2003; Hill & Thompson, 2003). Phenylalanine might be converted to cinnamate via a phenylalanine ammonia lyase (PAL); although these enzymes are rare in bacteria in general, a PAL is present in the myxalamid biosynthetic gene cluster of the myxobacterium Stigmatella (Silakowski et al., 2001) and a PAL was also detected in the Sor. cellulosum So ce56 genome project (Gerth et al., 2003), indicating that this pathway might be more prevalent in myxobacteria. Interestingly, benzoate fed to So ce26 during fermentation does not seem to be utilized for soraphen biosynthesis (Gerth et al., 2003; Hill et al., 2003), probably due to the lack of a suitable coenzyme A ligase.

The substrate specificity of AT1a, at least in its original module context, seems to be rather strict for benzoyl-CoA, as was shown with an engineered polyketide synthetase in Sac. erythraea. This strain produced exclusively the expected phenyl-substituted lactone upon benzoate supplementation and co-expression of an additional benzoyl CoA ligase, but failed to produce any lactones without benzoate feeding and the additional ligase (Wilkinson et al., 2001). As with Sac. erythraea, there is no evidence that Str. lividans would biosynthesize benzoyl-CoA as part of its normal metabolism. Str. lividans was shown not to degrade benzoate (although it catabolizes p-hydroxybenzoate; An et al., 2000; Grund & Kutzner, 1998) and neither a benzoyl-CoA ligase, nor a PAL can be detected in the Str. coelicolor genome sequence (Bentley et al., 2002). It was thus not surprising that Str. lividans SorABRCDFE did not produce soraphen, even in fermentations where benzoate was supplied to the strain (Table 2). Co-expressing a benzoyl-CoA ligase, the badA gene from Rhodopseudomonas palustris with the soraphen gene cluster, however, allowed the Str. lividans strain SorABRCDFE+BL to produce soraphen A when benzoate was supplied to the medium. The strict requirement for both an external source of benzoate and for an activating enzyme highlights the importance of engineering an appropriate metabolic route for benzoyl-CoA into Str. lividans if this strain is to be utilized for soraphen production.

Benzoyl-CoA was shown to be biosynthesized via several routes in different organisms (Fig. 4). None of these pathways could be shown to supply benzoyl-CoA from phenylalanine for soraphen biosynthesis in Str. lividans, as exemplified by the absence of soraphen production with all the engineered strains in fermentations without feeding or with phenylalanine supplementation (Table 2). Still, we were interested to map out whether parts of these pathways exist in our host and could be utilized to provide the soraphen starter unit by feeding a suitable intermediate. Anaerobic {alpha}-oxidation converts phenylalanine to benzoyl-CoA via phenylpyruvate and phenylacetate in Thauera aromatica (Schneider & Fuchs, 1998; Schneider et al., 1997; Route 1 on Fig. 4). This catabolic pathway seems to be lacking in Str. lividans under the tested conditions, as shown by the absence of soraphen production in all the engineered strains fed with phenylpyruvate or phenylacetate. In plants and the marine actinomycete ‘Str. maritimus’, PAL converts phenylalanine to cinnamate that is activated to cinnamoyl-CoA by a CoA ligase, and this intermediate is converted to benzoyl-CoA via {beta}-oxidation (Route 2; Moore et al., 2002). While the Str. lividans host does not harbour a PAL, cinnamate could be fed to both SorABRCDFE and SorABRCDFE+BL to effect soraphen biosynthesis. A cinnamate/4-coumarate CoA ligase has recently been characterized in Str. coelicolor (Kaneko et al., 2003), so its putative Str. lividans homologue should be able to activate cinnamate and commit it to {beta}-oxidation. Generic {beta}-oxidation enzymes were previously shown to complement for the conversion of cinnamoyl-CoA to benzoyl-CoA in ‘Str. maritimus mutants inactivated in dedicated benzoyl-CoA biosynthetic genes located within the enterocin cluster (Xiang & Moore, 2003). Since {beta}-oxidation proceeds via CoA-activated intermediates, and its last step, the thiolase-catalysed conversion of {beta}-ketophenylpropionyl-CoA yields benzoyl-CoA directly without the involvement of a benzoate-CoA ligase (Moore et al., 2002), the badA gene incorporated into SorABRCDFE+BL was superfluous for soraphen biosynthesis upon cinnamate feeding as shown by the production of soraphen by SorABRCDFE. A non-oxidative, retro-aldol path from cinnamoyl-CoA through benzaldehyde was also reported in plants (Route 3; Moore & Hertweck, 2002; and ibid.). Feeding benzaldehyde to the engineered Str. lividans strains did not provide for soraphen production, so the conversion of cinnamate to benzoyl-CoA in Str. lividans appears to proceed predominantly through the {beta}-oxidative route. As the main focus of these experiments was to establish an effective and convenient supply route for the soraphen starter unit, the stability and bioavailability of the fed precursors during fermentation were not addressed separately from their accessibility to soraphen biosynthesis.



View larger version (23K):
[in this window]
[in a new window]
 
Fig. 4. Biosynthetic routes for benzoyl-CoA used in soraphen A biosynthesis. Route 1, {alpha}-oxidation; Route 2, {beta}-oxidation; Route 3, retro-aldol path; Route 4, benzoate uptake and activation. CL, coenzyme A ligase. See Discussion for a detailed explanation.

 
In conclusion, two biosynthetic routes could be utilized for the provision of the soraphen starter unit in Str. lividans. The first route required benzoate feeding and the expression of a benzoyl-CoA ligase together with the soraphen biosynthetic genes (Fig. 4, Route 4). The second route required cinnamate feeding and seems to rely on an intrinsic cinnamoyl-CoA ligase and {beta}-oxidation enzymes of the host strain (Route 2). While both benzoate and cinnamate are toxic at higher concentrations, and even at the concentrations used in these experiments (2·5 mM for cinnamate and 5 mM for benzoate) would inhibit the growth of the Str. lividans strains to some extent, continuous feeding during a controlled fermentation, or feeding less toxic precursors like phenyl-substituted fatty acids with odd-numbered chain length (a suggestion that was also made by Leadlay and coworkers; Wilkinson et al., 2001) could lead to an increased production of soraphen. Introduction of a PAL into the engineered Str. lividans strains might be able to couple the existing cinnamate degradation pathway to the endogenous phenylalanine pool; significantly, the PAL from ‘Str. maritimus’ has been shown to provide cinnamate when expressed in Str. coelicolor (Moore et al., 2002).

The heterologous production of soraphen thus required the augmentation of the Str. lividans host with both a methoxymalonate and a benzoyl-CoA biosynthetic pathway in addition to the expression of the soraphen polyketide synthetase. The success of this approach serves as a further example for the biosynthetic utility of this genetically well characterized strain and confirms the feasibility of engineering for the lateral transfer of secondary metabolic pathways from industrially and genetically intransigent strains into fermentation- and genetic-manipulation-friendly hosts. Increasing the titre of soraphen production by the engineered Str. lividans strains via fermentation optimization, strain improvement and metabolic engineering will allow us to utilize the flexibility offered by this expression system to decipher the exact role of each gene product encoded in the soraphen gene cluster and to shed more light on the more arcane processes of soraphen A biosynthesis like the introduction of the C9-C10 double bond. Such approaches would also provide strains with better prospects for industrial-scale fermentations and open up new possibilities for the production of ‘unnatural’ natural products by combinatorial biosynthesis.


   ACKNOWLEDGEMENTS
 
We would like to thank Caroline S. Harwood (University of Iowa) for the kind gift of plasmid pPE202; Tin-Wein Yu (University of Washington) for insightful discussions and for bringing the existence of sorF to our attention; Eric Miller, Robert Upchurch and Hosni Hassan (North Carolina State University) for helpful discussions and guidance to R. Z.; Makoto Ono, An Hu and Racella McNair (Syngenta Biotechnology) for their assistance with sequence analysis; James Pomes and Belle Abrera (Syngenta Biotechnology) for soraphen analysis; and Alyssa Gulledge for her assistance with fermentations.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Ali, N., Herron, P. R., Evans, M. C. & Dyson, P. J. (2002). Osmotic regulation of the Streptomyces lividans thiostrepton-inducible promoter, PtipA. Microbiology 148, 381–390.[Medline]

An, H.-R., Park, H.-J. & Kim, E.-S. (2000). Characterization of benzoate degradation via ortho-cleavage by Streptomyces setonii. J Microbiol Biotechnol 10, 111–114.[CrossRef]

Avignone-Rossa, C., White, J., Kuiper, A., Postma, P. W., Bibb, M. & Teixeira de Mattos, M. J. (2002). Carbon flux distribution in antibiotic-producing chemostat cultures of Streptomyces lividans. Metab Eng 4, 138–150.[CrossRef][Medline]

Bentley, S. D., Chater, K. F., Cerdeno-Tarraga, A. M. & 40 other authors (2002). Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2). Nature 417, 141–147.[CrossRef][Medline]

Cox, R. J., Crosby, J., Daltrop, O. & 9 other authors (2002). Streptomyces coelicolor phosphopantetheinyl transferase: a promiscuous activator of polyketide and fatty acid synthase acyl carrier proteins. J Chem Soc Perkin Trans I, 1644–1649.

Dayem, L. C., Carney, J. R., Santi, D. V., Pfeifer, B. A., Khosla, C. & Kealey, J. T. (2002). Metabolic engineering of a methylmalonyl-CoA mutase-epimerase pathway for complex polyketide biosynthesis in Escherichia coli. Biochemistry 41, 5193–5201.[CrossRef][Medline]

Desai, R. P., Leaf, T., Woo, E. & Licari, P. (2002). Enhanced production of heterologous macrolide aglycones by fed-batch cultivation of Streptomyces coelicolor. J Ind Microbiol Biotechnol 28, 297–301.[CrossRef][Medline]

Egland, P. G., Gibson, J. & Harwood, C. S. (1995). Benzoate-coenzyme A ligase, encoded by badA, is one of three ligases able to catalyze benzoyl-coenzyme A formation during anaerobic growth of Rhodopseudomonas palustris on benzoate. J Bacteriol 177, 6545–6551.[Abstract]

Gerth, K., Bedorf, N., Irschik, H., Höfle, G. & Reichenbach, H. (1994). The soraphens: a family of novel antifungal compounds from Sorangium cellulosum (Myxobacteria). I. Soraphen A1 alpha: fermentation, isolation, biological properties. J Antibiot (Tokyo) 47, 23–31.[Medline]

Gerth, K., Pradella, S., Perlova, O., Beyer, S. & Müller, R. (2003). Myxobacteria: proficient producers of novel natural products with various biological activities – past and future biotechnological aspects with the focus on the genus Sorangium. J Biotechnol 106, 233–253.[CrossRef][Medline]

Grund, E. & Kutzner, H. J. (1998). Utilization of quinate and p-hydroxybenzoate by actinomycetes. Key enzymes and taxonomic relevance. J Basic Microbiol 38, 241–255.[CrossRef][Medline]

Haber, A., Johnson, R. D. & Rinehart, K. L., Jr (1977). Biosynthetic origin of the C2 units of geldanamycin and distribution of label from D-[6-13C]glucose. J Am Chem Soc 99, 3541–3544.[Medline]

Henderson, D. J., Brolle, D. F., Kieser, T., Melton, R. E. & Hopwood, D. A. (1990). Transposition of IS117 (the Streptomyces coelicolor A 3 (2) mini-circle) to and from a cloned target site and into secondary chromosomal sites. Mol Gen Genet 224, 65–71.[Medline]

Hertweck, C., Jarvis, A. P., Xiang, L., Moore, B. S. & Oldham, N. J. (2001). A mechanism of benzoic acid biosynthesis in plants and bacteria that mirrors fatty acid {beta}-oxidation. ChemBioChem 2, 784–786.[CrossRef][Medline]

Hill, A. M. & Thompson, B. L. (2003). Novel soraphens from precursor directed biosynthesis. Chem Commun (Camb) 2003 (12), 1360–1361.

Hill, A. M., Harris, J. P. & Siskos, A. P. (1998). Investigation of glycerol incorporation into soraphen A. Chem Commun (Camb) 1998 (12), 2361–2362.

Hill, A. M., Thompson, B. L., Harris, J. P. & Segret, R. (2003). Investigation of the early stages in soraphen A biosynthesis. Chem Commun (Camb) 2003 (12), 1358–1359.

Hopwood, D. A. (1997). Genetic contributions to understanding polyketide synthases. Chem Rev 97, 2465–2497.[CrossRef][Medline]

Huang, J., Lih, C.-J., Pan, K.-H. & Cohen, S. N. (2001). Global analysis of growth phase responsive gene expression and regulation of antibiotic biosynthetic pathways in Streptomyces coelicolor using DNA microarrays. Genes Dev 15, 3183–3192.[Abstract/Free Full Text]

Jaoua, S., Neff, S. & Schupp, T. (1992). Transfer of mobilizable plasmids to Sorangium cellulosum and evidence for their integration into the chromosome. Plasmid 28, 157–165.[Medline]

Julien, B. & Shah, S. (2002). Heterologous expression of epothilone biosynthetic genes in Myxococcus xanthus. Antimicrob Agents Chemother 46, 2772–2778.[Abstract/Free Full Text]

Kaneko, M., Ohnishi, Y. & Horinouchi, S. (2003). Cinnamate : coenzyme A ligase from the filamentous bacterium Streptomyces coelicolor A3(2). J Bacteriol 185, 20–27.[Abstract/Free Full Text]

Kao, C. M., Katz, L. & Khosla, C. (1994). Engineered biosynthesis of a complete macrolactone in a heterologous host. Science 265, 509–512.[Medline]

Kato, Y., Bai, L., Xue, Q., Revill, W. P., Yu, T.-W. & Floss, H. G. (2002). Functional expression of genes involved in the biosynthesis of the novel polyketide chain extension unit, methoxymalonyl-acyl carrier protein, and engineered biosynthesis of 2-desmethyl-2-methoxy-6-deoxyerythronolide B. J Am Chem Soc 124, 5268–5269.[CrossRef][Medline]

Kealey, J. T., Liu, L., Santi, D. V., Betlach, M. C. & Barr, P. J. (1998). Production of a polyketide natural product in nonpolyketide-producing prokaryotic and eukaryotic hosts. Proc Natl Acad Sci U S A 95, 505–509.[Abstract/Free Full Text]

Kieser, T., Bibb, M. J., Buttner, M. J., Chater, K. F. & Hopwood, D. A. (2000). Practical Streptomyces Genetics. Norwich, UK: the John Innes Foundation.

Kopp, M., Irschik, H., Gross, F., Perlova, O., Sandmann, A., Gerth, K. & Müller, R. (2004). Critical variations of conjugational DNA transfer into secondary metabolite multiproducing Sorangium cellulosum strains So ce12 and So ce56: development of a mariner-based transposon mutagenesis system. J Biotechnol 107, 29–40.[CrossRef][Medline]

Kuhstoss, S. & Rao, R. N. (1991). Analysis of the integration function of the streptomycete bacteriophage {phi}C31. J Mol Biol 222, 897–908.[Medline]

Kuhstoss, S., Huber, M., Turner, J. R., Paschal, J. W. & Rao, R. N. (1996). Production of a novel polyketide through the construction of a hybrid polyketide synthase. Gene 183, 231–236.[CrossRef][Medline]

Kyung, Y. S., Sherman, D. H. & Hu, W.-S. (2001). Simultaneous analysis of spatio-temporal gene expression for cephamycin biosynthesis in Streptomyces clavuligerus. Biotechnol Prog 17, 1000–1007.[CrossRef][Medline]

Lau, J., Frykman, S., Regentin, R., Ou, S., Tsuruta, H. & Licari, P. (2002). Optimizing the heterologous production of epothilone D in Myxococcus xanthus. Biotechnol Bioeng 78, 280–288.[CrossRef][Medline]

Ligon, J., Hill, S., Beck, J., Zirkle, R., Molnár, I., Zawodny, J., Money, S. & Schupp, T. (2002). Characterization of the biosynthetic gene cluster for the antifungal polyketide soraphen A from Sorangium cellulosum So ce26. Gene 285, 257–267.[CrossRef][Medline]

Lombo, F., Pfeifer, B., Leaf, T., Ou, S., Kim, Y. S., Cane, D. E., Licari, P. & Khosla, C. (2001). Enhancing the atom economy of polyketide biosynthetic processes through metabolic engineering. Biotechnol Prog 17, 612–617.[CrossRef][Medline]

MacNeil, D. J., Occi, J. L., Gewain, K. M., MacNeil, T., Gibbons, P. H., Ruby, C. L. & Danis, S. J. (1992). Complex organization of the Streptomyces avermitilis genes encoding the avermectin polyketide synthase. Gene 115, 119–125.[CrossRef][Medline]

Martin, C. J., Timoney, M. C., Sheridan, R. M., Kendrew, S. G., Wilkinson, B., Staunton, J. & Leadlay, P. F. (2003). Heterologous expression in Saccharopolyspora erythraea of a pentaketide synthase derived from the spinosyn polyketide synthase. Org Biomol Chem 1, 4144–4147.[CrossRef][Medline]

Molnár, I., Schupp, T., Ono, M. & 13 other authors (2000). The biosynthetic gene cluster for the microtubule-stabilizing agents epothilones A and B from Sorangium cellulosum So ce90. Chem Biol 7, 97–109.[CrossRef][Medline]

Moore, B. S. & Hertweck, C. (2002). Biosynthesis and attachment of novel bacterial polyketide synthase starter units. Nat Prot Rep 19, 70–99.[CrossRef]

Moore, B. S., Hertweck, C., Hopke, J. N. & 9 other authors (2002). Plant-like biosynthetic pathways in bacteria: from benzoic acid to chalcone. J Nat Prod 65, 1956–1962.[CrossRef][Medline]

Murakami, T., Holt, T. G. & Thompson, C. J. (1989). Thiostrepton-induced gene expression in Streptomyces lividans. J Bacteriol 171, 1459–1466.[Medline]

Murli, S., Kennedy, J., Dayem, L. C., Carney, J. R. & Kealey, J. T. (2003). Metabolic engineering of Escherichia coli for improved 6-deoxyerythronolide B production. J Ind Microbiol Biotechnol 30, 500–509.[CrossRef][Medline]

Omura, S., Tsuzuki, K., Nakagawa, A. & Lukacs, G. (1983). Biosynthetic origin of carbons 3 and 4 of leucomycin aglycone. J Antibiot (Tokyo) 36, 611–613.[Medline]

Pfeifer, B. A. & Khosla, C. (2001). Biosynthesis of polyketides in heterologous hosts. Microbiol Mol Biol Rev 65, 106–118.[Abstract/Free Full Text]

Pfeifer, B. A., Admiraal, S. J., Gramajo, H., Cane, D. E. & Khosla, C. (2001). Biosynthesis of complex polyketides in a metabolically engineered strain of E. coli. Science 291, 1790–1792.[Abstract/Free Full Text]

Pfeifer, B., Hu, Z., Licari, P. & Khosla, C. (2002). Process and metabolic strategies for improved production of Escherichia coli-derived 6-deoxyerythronolide B. Appl Environ Microbiol 68, 3287–3292.[Abstract/Free Full Text]

Pfeifer, B. A., Wang, C. C. C., Walsh, C. T. & Khosla, C. (2003). Biosynthesis of yersiniabactin, a complex polyketide-nonribosomal peptide, using Escherichia coli as a heterologous host. Appl Environ Microbiol 69, 6698–6702.[Abstract/Free Full Text]

Pradella, S., Hans, A., Sproer, C., Reichenbach, H., Gerth, K. & Beyer, S. (2002). Characterisation, genome size and genetic manipulation of the myxobacterium Sorangium cellulosum So ce56. Arch Microbiol 178, 484–492.[CrossRef][Medline]

Reichenbach, H. (1999). The ecology of the myxobacteria. Environ Microbiol 1, 15–21.[CrossRef][Medline]

Reichenbach, H. & Höfle, G. (1999). Myxobacteria as producers of secondary metabolites. In Drug Discovery from Nature, pp. 149–179. Edited by S. Grabley & R. Thiericke. Berlin, Heidelberg: Springer.

Rodriguez, E., Hu, Z., Ou, S., Volchegursky, Y., Hutchinson, C. R. & McDaniel, R. (2003). Rapid engineering of polyketide overproduction by gene transfer to industrially optimized strains. J Ind Microbiol Biotechnol 30, 480–488.[CrossRef][Medline]

Sambrook, J. & Russell, D. W. (2001). Molecular Cloning: a Laboratory Manual, 3rd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Schneider, S. & Fuchs, G. (1998). Phenylacetyl-CoA : acceptor oxidoreductase, a new alpha-oxidizing enzyme that produces phenylglyoxylate. Assay, membrane localization, and differential production in Thauera aromatica. Arch Microbiol 169, 509–516.[CrossRef][Medline]

Schneider, S., Mohamed, M. E. & Fuchs, G. (1997). Anaerobic metabolism of L-phenylalanine via benzoyl-CoA in the denitrifying bacterium Thauera aromatica. Arch Microbiol 168, 310–320.[CrossRef][Medline]

Schupp, T., Toupet, C., Cluzel, B., Neff, S., Hill, S., Beck, J. J. & Ligon, J. M. (1995). A Sorangium cellulosum (Myxobacterium) gene cluster for the biosynthesis of the macrolide antibiotic soraphen A: cloning, characterization, and homology to polyketide synthase genes from actinomycetes. J Bacteriol 177, 3673–3679.[Abstract]

Shah, S., Xue, Q., Tang, L., Carney, J. R., Betlach, M. & McDaniel, R. (2000). Cloning, characterization and heterologous expression of a polyketide synthase and P-450 oxidase involved in the biosynthesis of the antibiotic oleandomycin. J Antibiot (Tokyo) 53, 502–508.[Medline]

Silakowski, B., Nordsiek, G., Kunze, B., Blöcker, H. & Müller, R. (2001). Novel features in a combined polyketide synthase/non-ribosomal peptide synthetase: the myxalamid biosynthetic gene cluster of the myxobacterium Stigmatella aurantiaca Sga15(1). Chem Biol 8, 59–69.[CrossRef][Medline]

Strohl, W. R. (1992). Compilation and analysis of DNA sequences associated with apparent streptomycete promoters. Nucleic Acids Res 20, 961–974.[Abstract]

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]

Tang, L., Fu, H., Betlach, M. C. & McDaniel, R. (1999). Elucidating the mechanism of chain termination switching in the picromycin/methymycin polyketide synthase. Chem Biol 6, 553–558.[CrossRef][Medline]

Tang, L., Fu, H. & McDaniel, R. (2000). Formation of functional heterologous complexes using subunits from the picromycin, erythromycin and oleandomycin polyketide synthases. Chem Biol 7, 77–84.[Medline]

Tang, L., Shah, S., Chung, L., Carney, J., Katz, L., Khosla, C. & Julien, B. (2000). Cloning and heterologous expression of the epothilone gene cluster. Science 287, 640–642.[Abstract/Free Full Text]

Vahlensieck, H. F., Pridzun, L., Reichenbach, H. & Hinnen, A. (1994). Identification of the yeast ACC1 gene product (acetyl-CoA carboxylase) as the target of the polyketide fungicide soraphen A. Curr Genet 25, 95–100.[Medline]

Volchegursky, Y., Hu, Z., Katz, L. & McDaniel, R. (2000). Biosynthesis of the anti-parasitic agent megalomicin: transformation of erythromycin to megalomicin in Saccharopolyspora erythraea. Mol Microbiol 37, 752–762.[CrossRef][Medline]

Wilkinson, C. J., Frost, E. J., Staunton, J. & Leadlay, P. F. (2001). Chain initiation on the soraphen-producing modular polyketide synthase from Sorangium cellulosum. Chem Biol 8, 1197–1208.[CrossRef][Medline]

Wu, K., Chung, L., Revill, W. P., Katz, L. & Reeves, C. D. (2000). The FK520 gene cluster of Streptomyces hygroscopicus var. ascomyceticus (ATCC 14891) contains genes for biosynthesis of unusual polyketide extender units. Gene 251, 81–90.[CrossRef][Medline]

Xiang, L. & Moore, B. S. (2003). Characterization of benzoyl coenzyme A biosynthesis genes in the enterocin-producing bacterium ‘Streptomyces maritimus’. J Bacteriol 185, 399–404.[Abstract/Free Full Text]

Xue, Q., Ashley, G., Hutchinson, C. R. & Santi, D. V. (1999). A multiplasmid approach to preparing large libraries of polyketides. Proc Natl Acad Sci U S A 96, 11740–11745.[Abstract/Free Full Text]

Yu, T. W., Bai, L., Clade, D. & 7 other authors (2002). The biosynthetic gene cluster of the maytansinoid antitumor agent ansamitocin from Actinosynnema pretiosum. Proc Natl Acad Sci U S A 99, 7968–7973.[Abstract/Free Full Text]

Zhang, Y. X., Perry, K., Vinci, V. A., Powell, K., Stemmer, W. P. & del Cardayre, S. B. (2002). Genome shuffling leads to rapid phenotypic improvement in bacteria. Nature 415, 644–646.[CrossRef][Medline]

Ziermann, R. & Betlach, M. C. (1999). Recombinant polyketide synthesis in Streptomyces: engineering of improved host strains. BioTechniques 26, 106–110.[Medline]

Ziermann, R. & Betlach, M. C. (2000). A two vector system for the production of recombinant polyketides in Streptomyces. J Ind Microbiol Biotechnol 24, 46–50.[CrossRef]

Received 4 March 2004; revised 20 April 2004; accepted 18 May 2004.



This Article
Abstract
Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Zirkle, R.
Articles by Molnár, I.
Articles citing this Article
PubMed
PubMed Citation
Articles by Zirkle, R.
Articles by Molnár, I.
Agricola
Articles by Zirkle, R.
Articles by Molnár, I.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL MICROBIOLOGY J GEN VIROL
J MED MICROBIOL ALL SGM JOURNALS
Copyright © 2004 Society for General Microbiology.