non-ribosomal peptide synthetase gene clusters in the myxobacterium Stigmatella aurantiaca

Gene 275 (2001) 233–240 www.elsevier.com/locate/gene

Multiple hybrid polyketide synthase/non-ribosomal peptide synthetase gene clusters in the myxobacterium Stigmatella aurantiaca Barbara Silakowski a,b, Brigitte Kunze a, Rolf Mu¨ller a,b,* a

b

Gesellschaft fu¨r Biotechnologische Forschung mbH, Mascheroderweg 1, 38124 Braunschweig, Germany Institut fu¨r Pharmazeutische Biologie, Mendelssohnstrasse 1, Technische Universita¨t, 38106 Braunschweig, Germany Received 14 February 2001; received in revised form 11 June 2001; accepted 21 August 2001 Received by B. Dujon

Abstract Many bacterial and fungal secondary metabolites are produced by polyketide synthases (PKS) and non-ribosomal peptide synthetases (NRPS). Recently, it has been discovered that these modular enzymatic systems can also closely cooperate to form natural products. The analysis of the corresponding biosynthetic machineries, in the form of hybrid systems, is of special interest for combinatorial biosynthesis, because the combination of PKS and NRPS can lead to an immense variety of structures that might be produced. During our screening for hybrid PKS/NRPS systems from myxobacteria, we scanned the genome of Stigmatella aurantiaca DW4/3-1 for the presence of gene loci that encode both the PKS and NRPS genes. In addition to the previously characterized myxothiazol system, we identified three further hybrid loci, three additional PKS and one further NRPS gene locus. These were analyzed by hybridization, physical mapping, PCR with degenerate oligonucleotides and sequencing of fragments of the gene clusters. The function of these genes was not known but it had already been speculated that one compound produced by the strain and detected via HPLC was a secondary metabolite. This was based on the observation that its production is dependent on an active copy of the phosphopantetheinyl transferase gene mtaA. We show here that one of the identified hybrid gene loci is responsible for the formation of this secondary metabolite. In agreement with the genetic data, the chemical structure resembles a cyclic polypeptide with a PKS sidechain. Our data show that S. aurantiaca has a broader genetic capacity to produce natural products than the number of compounds isolated from the strain so far suggests. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Secondary metabolism; Combinatorial biosynthesis; Actinomycetes; Polyketide; Polypeptide; Antibiotics

1. Introduction Bacteria are very potent producers of natural products with biological activity, which are extensively used as pharmaceuticals and agrochemicals. Approximately two-thirds of the known microbial secondary metabolites are produced by members of the order Actinomycetales. Nevertheless, there is an increasing demand for alternative sources of natural products with biological activity, because this class of microorganisms has already been mined for decades. In this respect, only very few groups have focused their research on the Myxococcales, which do have the ability to create a vast variety of metabolites with interesting biological activity (Reichenbach and Ho¨fle, 1999). Currently the epothilons are the most promising myxobacAbbreviations: KS, b-ketoacyl synthase; NRPS, non-ribosomal peptide synthetase; PKS, polyketide synthase * Corresponding author. Gesellschaft fu¨r Biotechnologische Forschung mbH, Mascheroderweg 1, 38124 Braunschweig, Germany. Tel.: 149531-6181420; fax: 149-531-6181284. E-mail address: [email protected] (R. Mu¨ller).

terial compounds due to their potent anticancer activity (Bollag et al., 1995; Gerth et al., 1996; Ho¨fle et al., 1996; Service, 1996; Nicolaou et al., 1998). More than 90 new basic structures have been isolated from myxobacteria; the structure of approximately 400 derivatives has been elucidated (Reichenbach and Ho¨fle, 1999). The work presented here aimed to identify typical biosynthetic genes responsible for secondary metabolite formation in myxobacteria. Interestingly, many of the compounds isolated from myxobacteria have polyketide backbones with incorporated amino acids (e.g. the myxovirescins, epothilons, myxothiazols, myxalamids; see Fig. 1); this led us to assume that hybrid systems of PKS and NRPS genes are responsible for the formation of these secondary metabolites (Beyer et al., 1999). Since we started our genetic screening programme, the biosynthetic gene clusters of the epothilons from Sorangium cellulosum (Julien et al., 2000; Molnar et al., 2000), the myxothiazols from Stigmatella aurantiaca DW4/3-1 (Silakowski et al., 1999), the myxalamids (Beyer et al., 1999; Silakowski et al., 2001) and the myxochelins (Silakowski et al., 2000) from S.

0378-1119/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0378-111 9(01)00680-1

234

B. Silakowski et al. / Gene 275 (2001) 233–240

Fig. 1. Structures of natural products isolated from myxobacteria. Amino acids incorporated into the polyketide backbones are shaded in grey (cysteines into myxothiazol and epothilon, alanine into myxalamid and glycine into myxovirescin).

aurantiaca Sg a15 and myxovirescin from Myxococcus xanthus (Paitan et al., 1999) have been cloned and analyzed; most of them do in fact resemble hybrid systems. Such natural systems, in which both PKS and NRPS enzymes cooperate to form natural products, are promising candidates to study, because the analysis could extend the currently limited knowledge on the interaction of PKS and NRPS, which can be used for combinatorial biosynthesis (Cane and Walsh, 1999). Many PKS systems have been described from actinomycetes and recently Donadio and co-workers, in a genetic screening approach, found that multiple NRPS gene clusters are also present in this class of microorganisms (Sosio et al., 2000). Nevertheless, only a few hybrid systems have been reported from actinomycetes to date (Molnar et al., 1996; Chen et al., 1999). Using our model strain S. aurantiaca, we show here that myxobacteria do not only contain multiple systems of PKS and NRPS genes, but even multiple hybrid systems. A natural product corresponding to one hybrid gene locus was identified from the production spectrum of S. aurantiaca DW4/3-1 by comparison to a mutant of this gene cluster and subsequently the structure was elucidated. This indicates that single strains of S. aurantiaca and most probably other myxobacteria do have the ability to produce more natural products than have been isolated from them so far.

2. Materials and methods 2.1. Bacterial strains and culture conditions Escherichia coli strains were cultured in Luria Broth at 378C. Stigmatella aurantiaca strain DW4/3-1 (Qualls et al., 1978) and its descendants were grown in Tryptone (1% Tryptone, 0.2% MgSO4, pH 7.2) at 308C. Antibiotics were added where necessary in the following final concentrations: kanamycin sulphate 50 mg/ml, tetracyclin 10 mg/ml, and streptomycin sulphate 120 mg/ml. 2.2. DNA manipulations, analysis, Southern blotting, sequencing and PCR Chromosomal DNA was prepared as described (Neumann et al., 1992). Hybridizations using homologous probes, in the case of Southern analysis of chromosomal DNA, were performed using the standard protocol for homologous probes of the DIG High Prime DNA Labeling and Detection Starter Kit II (Roche Molecular Biochemicals). Before screening the cosmid library of S. aurantiaca DW4/3-1, the clones were transferred onto nylon membranes (Roche Molecular Biochemicals) as described in the DIG systems user’s guide for filter hybridization (Roche Molecular Biochemicals). For the hybridization of

B. Silakowski et al. / Gene 275 (2001) 233–240

the gene library using homologous and heterologous probes, the protocol for formamide buffer was modified as follows: (pre-)hybridization was carried out at 378C instead of at 428C and the membranes were washed at 588C instead of at 688C. All other hybridizations experiments were performed under high stringency conditions. For the sequencing of plasmid and cosmid DNA, prepared by the High Pure Plasmid Purification Kit (Roche Molecular Biochemicals), the DYEnamice ET terminator cycle sequencing premix kit (Amersham Pharmacia Biotech) was used. Gels were run on ABI-377 sequencers. PCR was carried out using HotStarTaq Polymerase (Qiagen) according to the manufacturer’s protocol. Dimethylsulfoxide was added to a final concentration of 5%. The conditions used with the Eppendorf Gradient Mastercycler were as follows: 15 min at 958C for activation of the polymerase, denaturation for 30 s at 958C, and annealing for 30 s at 528C for degenerate primers and at 608C for homologous primers. The extension was run for 45 s at 728C. Thirty cycles were run with a final extension for 10 min at 728C. PCR products were purified with the High Pure PCR Product Purification Kit (Roche Molecular Biochemicals) and afterwards ligated into pCR2.1-TOPO using the TOPO TA Cloning Kit (Invitrogen). All other in vitro DNA manipulations were performed according to standard protocols (Sambrook et al., 1989). 2.3. Construction and analysis of S. aurantiaca DW4/3-1 mutants To obtain S. aurantiaca DW4/3-1 mutants, each plasmid harbouring a cloned KS-PCR fragment (see Table 1) was introduced by electroporation as previously described (Beyer et al., 1999). The introduced plasmids integrate by homologous recombination into the genome, leading to gene disruption mutants. No mutants were obtained after the electroporation of plasmids pE119, pE139 and pE159. The integration of all of the other plasmids was verified by genomic Southern blot using the introduced KS fragment as a probe (see Fig. 2). 2.4. Screening and mapping of cosmids For the screening of the cosmid library of S. aurantiaca DW4/3-1 (Silakowski et al., 1999) a hybridization was performed with homologous and heterologous ketosynthase (KS) fragments as probes under low stringency conditions. The KS fragments were derived from three sources: S. cellulosum So ce90, S. aurantiaca Sg a15 (Beyer et al., 1999) and from the mtaB gene, which came from the myxothiazol cluster (Silakowski et al., 1999). From approximately 2000 cosmid clones 90 gave signals. All 90 cosmid clones were analyzed by PCR using degenerate primers based on highly conserved KS sequences (Beyer et al., 1999). Thirty cosmids, including the cosmid E28 representing another mta-cosmid, gave PCR products of the expected size (ca. 700 bp) and were analyzed by restriction with EcoRI,

235

BamHI and BglII. Two cosmids were too small and some showed a very similar restriction pattern, which enabled a first grouping to take place. Twenty different cosmids remained and their corresponding KS-PCR fragments were cloned into pCR2.1-TOPO (Invitrogen) and sequenced. Some of the inserts were identical and could, therefore, be grouped (Table 1). The inserts of cosmids E122, E126, E139, E159, E164, E193, E195 and E196 were end-sequenced using T3 and T7 universal primers. Based on the sequences obtained from cosmids E122, E139, E159, E193 and E195, PCR primer pairs were created and used for PCRs with the other seven cosmids as templates, respectively. The primer pair binding to the T7 end of E193 gave a product with E126 as the template. The primer pair derived from the T3 end of E195 gave a product with E196 as the template. All other PCRs were negative. 2.5. Production and analysis of secondary metabolites For the production of secondary metabolites, the strains were cultivated in Probion liquid medium composed of 0.3% Probion (single cell protein prepared from Methylomonas clarae; Hoechst A.G.), 0.3% soluble starch, 0.2% MgSO4 £ 7H2O, 0.05% CaCl2 £ 2H2O and 50 mM HEPES (pH 7.0). A total of 1% of the adsorber resin XAD-16 (Rohm & Haas) and 1 ml of standard vitamin and trace element solution were added per litre of the medium. The cultivation and isolation of secondary metabolites has been described (Silakowski et al., 1999). The spectrum of secondary metabolites produced by different mutants compared with those of the wild-type strain was determined in aliquots of concentrated acetone extracts by diode-array-detected HPLC analysis using a Hewlett Packard series 1100 instrument. Chromatographic conditions were as follows: Column ET 125 £ 2 mm and precolumn, Nucleosil 120-5-C18. The solvent was methanol/water 45:55 (A)/methanol (B). Isocratic conditions were 0% B for the first 7 min, then a gradient up to 45% B at 20 min, followed by isocratic conditions of 45% B for 6 min (the time point was 26 min in the HPLC run). This was then followed by a gradient up to 81.8% B at 40 min; the flow rate was 0.5 ml/min and detection was at 200–400 nm. 3. Results 3.1. Screening of the S. aurantiaca DW4/3-1 cosmid library A cosmid library of S. aurantiaca DW4/3-1 (Silakowski et al., 1999) was screened for the presence of PKS genes, resulting in the isolation of 90 cosmids hybridizing with the probes used. All cosmids were subjected to PCR analysis using degenerate oligonucleotides (Beyer et al., 1999) designed to bind to conserved regions in KSs. Cosmids which gave rise to amplification products of the expected size were further analyzed by restriction enzyme profiling, which subsequently led to the identification of 20 cosmids,

236

Table 1 PKS and NRPS gene loci in the chromosome of S. aurantiaca DW4/3-1 Cosmids

Name of plasmid harbouring cloned PCR fragment

Homology of aa sequence deduced from DNA sequence generated via PCR to

Homology of aa sequence deduced from DNA sequence from T3 or T7 end of cosmid to

Type of cluster

Inactivation leads to

mta

E25, E28, E133, E138, E158, E160, E190, E201 E122, E178, E192 E159 E119, E139 E114, E115, E118, E126, E137, E140, E156, E193, E194 E164, E167, E168 E130, E131, E195, E196

pE133, pE138, pE160, pE190

PKS and NRPS

(cluster completely sequenced)

Hybrid

myxothiazol negative mutants

pE122, pE178, pE192 pE159 pE119, pE139 pE114, pE126, pE137, pE156, pE193, pE194

PKS PKS PKS PKS

NRPS PKS NRPS PKS

Hybrid PKS Hybrid PKS

? ? ? ?

pE164 pE131, pE195, pE196

PKS PKS

– NRPS

PKS Hybrid

substance 1 negative mutants substances 2 1 3 negative mutants

122 159 139 193

164 196

B. Silakowski et al. / Gene 275 (2001) 233–240

Cosmid group

B. Silakowski et al. / Gene 275 (2001) 233–240

237

DW4/3-1 contains a variety of chromosomal regions harbouring PKS, NRPS and hybrid PKS/NRPS gene clusters. From this approach, it cannot be concluded that gene loci carrying both types of genes are those which are responsible for the production of hybrid structures, because most of the corresponding products are not known. It might well be that the PKS and NRPS genes are just located adjacent to each other by chance. Therefore, we focused next on the generation of gene disruption mutants of each cluster in order to get more information about their biological function.

3.3. Generation of specific mutants of the gene clusters

Fig. 2. Southern analysis of mutants. Genomic DNA of mutants DWPE122 (lane 2) and DWPE131 (lane 4) analyzed via Southern blot in comparison to wild-type DNA (lanes 1 and 3). Lanes 1 and 2, restriction with NotI; lanes 3 and 4, restriction with BamHI. The single wild-type signal disappears in the mutants. Two new fragments hybridize with the probe, because there is one BamHI and one NotI recognition site located in the cloning vector. The sizes of the DNA fragments are calculated according to the DNA molecular weight marker III (digoxigenin labelled).

which could not be grouped together. PCR products from these cosmids were cloned and sequenced. All the sequences are very similar to several known PKS genes (data not shown). Several identical sequences were found, which showed that the corresponding cosmids overlap. Finally, one representative of each group was sequenced from the T3 and the T7 end of the SuperCos vector. Oligonucleotide pairs based on the sequences obtained were designed and used in order to screen the rest of the cosmids for the presence of overlaps. This procedure resulted in the grouping of cosmids shown in Table 1. 3.2. S. aurantiaca DW4/3-1 contains a variety of secondary metabolic gene clusters The results of the hybridization, PCR and sequencing experiments clearly demonstrate that S. aurantiaca

Cloned PCR products carrying specific fragments of each gene cluster (all plasmids are given in Table 1) were used for gene inactivation experiments, as described for S. aurantiaca Sg a15 (Beyer et al., 1999). Each mutant strain was analyzed for the correctness of the integration of the plasmid by Southern analysis, an example of which is shown in Fig. 2. Subsequently, the phenotype of the mutant strains was analyzed by diode-array coupled HPLC. To date we have identified three natural products corresponding to two of the described gene clusters. Substance 1 (possibly related to two further substances represented by peaks around a retention time of 25 min in Fig. 3), the structure of which is currently being elucidated, was detected in the wild-type after comparison of the production spectrum with mutant E164 (see Fig. 3). A PKS is involved in its biosynthesis. Substances 2 and 3 are described below.

3.4. Identification of a compound related to a hybrid cluster as a hybrid structure Two compounds of the same class (substances 2 1 3) produced by S. aurantiaca DW4/3-1 came to our attention after we inactivated the phosphopantetheinyl transferase gene mtaA (mutant BS57; see Fig. 3), which is located in front of the myxothiazol biosynthetic gene cluster (Silakowski et al., 1999). The production of the secondary metabolites was shown to be strictly dependent on an active copy of that gene. After integration of pE196 into the chromosome, production of two substance peaks with the same retention time and UV spectra was abolished (see Fig. 3), which led us to assume that cosmid E196 carries at least fragments of the corresponding biosynthetic gene cluster responsible for the formation of both substances. The T7 end-sequencing of the same cosmid revealed the presence of an NRPS gene, which indicated that the compounds represent hybrid structures. The structural elucidation of the major peak (substance 2) is currently under way. In total agreement with the genetic data presented here, it is already clear that it resembles a cyclic polypeptide with a PKS sidechain (G. Ho¨ fle, unpublished data).

238

B. Silakowski et al. / Gene 275 (2001) 233–240

4. Discussion 4.1. Many secondary metabolic gene clusters in ‘talented’ producers of natural products This communication demonstrates the presence of an enormous amount of typical secondary metabolic gene clusters in S. aurantiaca DW4/3-1. If one assumes that there are a total of eight (see below) gene clusters of an approximate size of 40–60 kbp present, then this would add up to a minimum of 320 kbp up to 480 kbp of total DNA sequence (or 3.2–4.8% of the genome size of 10 MB; compare with Neumann et al., 1993) devoted to ‘secondary metabolism’. The analysis of different bacterial species that have been sequenced thus far shows that this is not a common feature among microorganisms, as most of them carry far fewer genes of this type. The current effort to sequence the genome of the model actinomycete Streptomyces coelicolor as representative of ‘talented’ producers also reveals the presence of several biosynthetic gene clusters representing PKS and NRPS (www.sanger.ac.uk/Projects/S_coelicolor). Further data from actinomycetes are available, indicating that they carry a substantial amount of PKS and NRPS in their genomes (Sosio et al., 2000). Taken together, these data imply that the biodiversity of natural products made by ‘talented’ producers such as actinomycetes and myxobacteria mirrors their genetic diversity. The genes found indicate that these producers have a much broader potential to generate new (to date unknown) structures than can be estimated from the amount of compounds found to date. 4.2. New natural products can be predicted and found via genetic screening approaches

Fig. 3. HPLC analysis of mutants. HPLC traces of production spectra of S. aurantiaca DW4/3-1 wild-type and mutants described in this work. Numbers correspond to substances 1, 2 and 3 (see text). DAD, diode array detection.

The occurrence of multiple secondary metabolic gene clusters, even those of the hybrid type, seems to be a common feature in myxobacteria. We have performed a series of hybridizations and cloning experiments with other myxobacterial strains revealing similar results to the ones presented in this communication; for S. cellulosum So ce90 at least two hybrid gene clusters have been described, one of currently unknown function (Beyer et al., 1999; Mu¨ ller et al., 2000) and one responsible for epothilon biosynthesis (Molnar et al., 2000). The genome of S. aurantiaca Sg a15 revealed the presence of another hybrid gene locus, which has been recently characterized as the myxalamid biosynthesis gene cluster (Silakowski et al., 2001); it also revealed several further PKS gene loci (Beyer et al., 1999). Additionally, it has been demonstrated in this strain that the structure of a natural product can be predicted on the basis of cloned genes with an unknown function. A NRPS was found and shown to be involved in the biosynthesis of the myxochelin type iron chelator, which was not known from S. aurantiaca Sg a15 at the time (Silakowski et al., 2000). S. aurantiaca DW4/3-1 needs to carry a similar set of genes, because we can show that it also produces

B. Silakowski et al. / Gene 275 (2001) 233–240

myxochelins (G. Ho¨ fle, B.S., B.K. and R.M., unpublished data). These data taken together illustrate the strength of the constantly improving technique described here in the identification of new natural products based on genetic screening. New opportunities arise as more microorganisms are sequenced.

4.3. Possible biological functions of deduced PKS and NRPS proteins and their activation PKS and NRPS are generally considered to be typical secondary metabolic enzymes. Nevertheless, they are also involved in the formation of siderophores (see above), which are essential under iron limiting conditions. It has even been speculated that they are involved in the formation of signalling peptides in S. coelicolor (Willey et al., 1991) and they might well have a function as signalling substances during the processes of development and social gliding in S. aurantiaca and other myxobacteria. Such pheromones would presumably not be produced on a scale that enables their detection in simple HPLC screens as described herein. Furthermore, one would have to evaluate the exact conditions under which these substances are produced. In fact, a compound named stigmolone has been isolated from S. aurantiaca DW4/3-1, which induces fruiting body formation (Plaga et al., 1998) and may well be biosynthesized similarly to secondary metabolites. It is tempting to speculate that the gene clusters that could not be inactivated thus far (groups 159,139) may be responsible for the formation of compounds that are essential under the conditions used. It is interesting to note that the production of all known substances from S. aurantiaca DW4/3-1 is dependent on an active copy of the mtaA gene (compare with Fig. 3). After inactivation of this gene, substances 1, 2 and 3 plus myxothiazol cannot be produced any more. This shows the broad substrate specificity of MtaA, which must at least be capable of activating all the megasynthetases of the PKS and the NRPS type involved in the biogenesis of the compounds mentioned. Thus, MtaA is a very promising protein candidate, which can be used for the post-translational activation of PKS/NRPS systems after heterologous expressions (compare with Gaitatzis et al., 2001). The immense genetic capacity described in this communication may also be a result of cryptic gene clusters or of others, for which the conditions for the production of the corresponding compound have not yet been identified. Nevertheless, S. aurantiaca has already been established as a multiproducer of secondary metabolites; strain Sg a15 (myxalamids, stigmatellins, aurachins and myxochelins) as well as strain DW4/3-1 (myxothiazols, substance 1, substances 2 1 3 and myxochelins) both produce at least four secondary metabolites. Based on the findings described here it can be expected that further new compounds will be isolated in the future.

239

Acknowledgements The authors thank Professor Dr H. Reichenbach and Professor Dr G. Ho¨ fle for their support. We would like to acknowledge the skilful technical assistance of A. Hans, M. Weilharter and M. Scharfe. J. Scriven and S. Beyer have contributed to this work by careful proof-reading of the manuscript. We would like to thank Professor Dr H.U. Schairer for sending several ‘care packages’ to Braunschweig. This research was supported by grant Mu 1254/3-1 of the Deutsche Forschungsgemeinschaft to R.M.

References Beyer, S., Kunze, B., Silakowski, B., Mu¨ ller, R., 1999. Metabolic diversity in myxobacteria – identification of the myxalamid and the stigmatellin biosynthetic gene cluster of Stigmatella aurantiaca Sg a15 and a combined polyketide- (poly)peptide gene cluster from the epothilon producing strain Sorangium cellulosum So ce90. Biochim. Biophys. Acta 1445, 185–195. Bollag, D.M., McQueney, P.A., Zhu, J., Hensens, O., Koupal, L., Liesch, J., Goetz, M., Lazarides, E., Woods, M., 1995. Epothilones, a new class of microtubule-stabilizing agents with Taxol-like mechanism of action. Cancer Res. 55, 2325–2333. Cane, D., Walsh, C., 1999. The parallel and convergent universes of polyketide synthases and nonribosomal peptide synthetases. Chem. Biol. 6, R319–R325. Chen, S., von Bamberg, D., Hale, V., Brauer, M., Hardt, B., Mu¨ ller, R., Floss, H.G., Reynolds, K.A., Leistner, E., 1999. Biosynthesis of ansatrienin (mycotrienin) and naphthomycin: identification and analysis of two separate biosynthetic gene clusters in Streptomyces collinus Tu¨ 1892. Eur. J. Biochem. 261, 98–107. Gaitatzis, N., Hans, A., Mu¨ ller, R., Beyer, S., 2001. The mtaA gene of the myxothiazol biosynthetic gene cluster from Stigmatella aurantiaca DW4/3-1 encodes a phosphopantetheinyl transferase that activates polyketide synthases and polypeptide synthetases. J. Biochem. 129, 119–124. Gerth, K., Bedorf, B., Ho¨ fle, G., Irschik, H., Reichenbach, H., 1996. Epothilons A and B: antifungal and cytotoxic compounds from Sorangium cellulosum (myxobacteria) production, physico-chemical and biological properties. J. Antibiot. 49, 560–563. Ho¨ fle, G., Bedorf, N., Steinmetz, H., Schomburg, D., Gerth, K., Reichenbach, H., 1996. Epothilone A and B – novel 16-membered macrolides with cytotoxic activity: isolation, crystal structure, and conformation in solution. Angew. Chem. Int. Ed. Engl. 35, 1567–1569. Julien, B., Shah, S., Ziermann, R., Goldman, R., Katz, L., Khosla, C., 2000. Isolation and characterization of the epothilone biosynthetic gene cluster from Sorangium cellulosum. Gene 249, 153–160. Molnar, I., Aparicio, J.F., Haydock, S.F., Khaw, L.E., Schwecke, T., Ko¨ nig, A., Staunton, J., Leadlay, P.F., 1996. Organisation of the biosynthetic gene cluster for rapamycin in Streptomyces hygroscopicus: analysis of genes flanking the polyketide synthase. Gene 169, 1–7. Molnar, I., Schupp, T., Ono, M., Zirkle, R.E., Milnamow, M., NowakThompson, B., Engel, N., Toupet, C., Stratman, A., Cyr, D.D., Gorlach, J., Mayo, J.M., Hu, A., Goff, S., Schmid, J., Ligon, J.M., 2000. The biosynthetic gene cluster for the microtubule-stabilizing agents epothilones A and B from Sorangium cellulosum So ce90. Chem. Biol. 7, 97– 109. Mu¨ ller, R., Gerth, K., Brandt, P., Blo¨ cker, H., Beyer, S., 2000. Identification of an L-dopa decarboxylase gene from Sorangium cellulosum So ce90. Arch. Microbiol. 173, 303–306. Neumann, B., Pospiech, A., Schairer, H.U., 1992. Rapid isolation of genomic DNA from Gram negative bacteria. Trends Genet. 8, 332–333.

240

B. Silakowski et al. / Gene 275 (2001) 233–240

Neumann, B., Pospiech, A., Schairer, H.U., 1993. A physical and genetic map of the Stigmatella aurantiaca DW4/3.1 chromosome. Mol. Microbiol. 10, 1087–1099. Nicolaou, K., Roschangar, F., Vourloumis, D., 1998. Chemie und Biologie der Epothilone. Angew. Chem. 110, 2120–2153. Paitan, Y., Alon, G., Orr, E., Ron, E.Z., Rosenberg, E., 1999. The first gene in the biosynthesis of the polyketide antibiotic TA of Myxococcus xanthus codes for a unique PKS module coupled to a peptide synthetase. J. Mol. Biol. 286, 465–474. Plaga, W., Stamm, I., Schairer, H.U., 1998. Intercellular signaling in Stigmatella aurantiaca: purification and characterization of stimolone, a myxobacterial pheromone. Proc. Natl. Acad. Sci. USA 95, 11263– 11267. Qualls, G.T., Stephens, K., White, D., 1978. Morphogenetic movements and multicellular development in the fruiting myxobacterium, Stigmatella aurantiaca. Dev. Biol. 66, 270–274. Reichenbach, H., Ho¨ fle, G., 1999. Myxobacteria as producers of secondary metabolites. In: Grabley, S., Thieriecke, R. (Eds.), Drug Discovery from Nature. Springer-Verlag, Berlin, pp. 149–179. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

Service, R.F., 1996. Tumor-killer made; how does it work? Science 274, 2009. Silakowski, B., Schairer, H.U., Ehret, H., Kunze, B., Weinig, S., Nordsiek, G., Brandt, P., Blo¨ cker, H., Ho¨ fle, G., Beyer, S., Mu¨ ller, R., 1999. New lessons for combinatorial biosynthesis from myxobacteria: the myxothiazol biosynthetic gene cluster of Stigmatella aurantiaca DW4/3-1. J. Biol. Chem. 274, 37391–37399. Silakowski, B., Kunze, B., Nordsiek, G., Blo¨ cker, H., Ho¨ fle, G., Mu¨ ller, R., 2000. The myxochelin iron transport regulon of the myxobacterium Stigmatella aurantiaca Sg a15. Eur. J. Biochem. 267, 6476–6485. Silakowski, B., Nordsiek, G., Kunze, B., Blo¨ cker, H., Mu¨ ller, R., 2001. Novel features in a combined polyketide synthase/non-ribosomal peptide synthetase: the myxalamid biosynthetic gene cluster of the myxobacterium Stigmatella aurantiaca Sg a15. Chem. Biol. 8(1), 59– 69. Sosio, M., Bossi, E., Bianchi, A., Donadio, S., 2000. Multiple peptide synthetase gene clusters in Actinomycetes. Mol. Gen. Genet. 264, 213–221. Willey, J., Santamaria, R., Guijarro, J., Geistlich, M., Losik, R., 1991. Extracellular complementation of a developmental mutation implicates a small sporulation protein in aerial mycelium formation by S. coelicolor. Cell 65, 641–650.