Heterologous expression of intact biosynthetic gene clusters in Fusarium graminearum
Mikkel Rank NielsenRasmus Dam WollenbergKlaus Ringsborg WestphalTeis Esben SøndergaardReinhard WimmerDonald M. GardinerJens Laurids Sørensen
21
Citation
90
Reference
10
Related Paper
Citation Trend
Keywords:
Polyketide synthase
Heterologous expression
Gene cluster
Secondary metabolism
Gene cluster
Polyketide synthase
Thioesterase
Cite
Citations (67)
Coronafacic acid (CFA) is the polyketide component of the phytotoxin coronatine, a virulence factor of the plant pathogen Pseudomonas syringae . Our current knowledge of polyketide biosynthesis largely is based on the analysis of polyketide synthases (PKSs) in actinomycetes and other Gram-positive bacteria. Consequently, the cloning and characterization of the CFA biosynthetic gene cluster will contribute significantly to our knowledge of polyketide synthesis in Pseudomonas . In this report, we describe two genes in the CFA biosynthetic gene cluster that encode PKSs that are structurally and functionally similar to the multifunctional modular PKSs, which catalyze the synthesis of macrolide antibiotics. The CFA PKS genes were overproduced in Escherichia coli and shown to cross-react with antisera made to a modular PKS involved in erythromycin synthesis. A scheme for CFA biosynthesis is presented that incorporates the activities of all proteins in the CFA PKS. In this report a gene cluster encoding a pseudomonad polyketide has been completely sequenced and the deduced gene functions have been used to develop a biosynthetic scheme.
Gene cluster
Polyketide synthase
Pseudomonas syringae
Cite
Citations (62)
Salinipyrones and pacificanones are structurally related polyketides from Salinispora pacifica CNS-237 that are proposed to arise from the same modular polyketide synthase (PKS) assembly line. Genome sequencing revealed a large macrolide PKS gene cluster that codes for the biosynthesis of rosamicin A and a series of new macrolide antibiotics. Mutagenesis experiments unexpectedly correlated salinipyrone and pacificanone biosynthesis to the rosamicin octamodule Spr PKS. Remarkably, this bifurcated polyketide pathway illuminates a series of enzymatic domain- and module-skipping reactions that give rise to natural polyketide product diversity. Our findings enlarge the growing knowledge of polyketide biochemistry and illuminate potential challenges in PKS bioengineering.
Polyketide synthase
Gene cluster
Natural product
Cite
Citations (20)
Polyketides are a group of natural products that have gained much interest due to their use as antibiotics, cholesterol lowering agents, immunosuppressors, and as other drugs. Many organisms that naturally produce polyketides are difficult to cultivate and only produce these metabolites in small amounts. It is therefore of general interest to transfer polyketide synthase (PKS) genes from their natural sources into heterologous hosts that can over-produce the corresponding polyketides. In this study we demonstrate the heterologous expression of 6-methylsalicylic acid synthase (6-MSAS), naturally produced by Penicillium patulum, in the yeast Saccharomyces cerevisiae. In order to activate the PKS a 4'-phosphopantetheinyl transferase (PPTase) is required. We therefore co-expressed PPTases encoded by either sfp from Bacillus subtilis or by npgA from Aspergillus nidulans. The different strains were grown in batch cultures. Growth and product concentration were measured and kinetic parameters were calculated. It was shown that both PPTases could be efficiently used for activation of PKS's in yeast as good yields of 6-MSA were obtained with both enzymes.
Polyketide synthase
Heterologous expression
Heterologous
Aspergillus nidulans
Cite
Citations (55)
Gene cluster
Polyketide synthase
Natural product
Cite
Citations (24)
The polyketides are a diverse group of natural products with great significance as human and veterinary pharmaceuticals. A significant barrier to the production of novel genetically engineered polyketides has been the lack of available heterologous expression systems for functional polyketide synthases (PKSs). Herein, we report the expression of an intact functional PKS in Escherichia coli and Saccharomyces cerevisiae . The fungal gene encoding 6-methylsalicylic acid synthase from Penicillium patulum was expressed in E. coli and S. cerevisiae and the polyketide 6-methylsalicylic acid (6-MSA) was produced. In both bacterial and yeast hosts, polyketide production required coexpression of 6-methylsalicylic acid synthase and a heterologous phosphopantetheinyl transferase that was required to convert the expressed apo-PKS to its holo form. Production of 6-MSA in E. coli was both temperature- and glycerol-dependent and levels of production were lower than those of P. patulum , the native host. In yeast, however, 6-MSA levels greater than 2-fold higher than the native host were observed. The heterologous expression systems described will facilitate the manipulation of PKS genes and consequent production of novel engineered polyketides and polyketide libraries.
Polyketide synthase
Heterologous expression
Heterologous
Acyl carrier protein
Cite
Citations (216)
Polyketide synthase
Gene cluster
Secondary metabolism
Gibberella fujikuroi
Cite
Citations (49)
The biosynthetic gene cluster for the 26-membered ring of the polyene macrolide pimaricin extends for about 110 kilobase pairs of contiguous DNA in the genome of Streptomyces natalensis. Two sets of polyketide synthase (PKS) genes are separated by a group of small polyketide-functionalizing genes. Two of the polyketide synthase genes, pimS0 and pimS1, have been fully sequenced and disrupted proving the involvement of each of these genes in pimaricin biosynthesis. The pimS0 gene encodes a relatively small acetate-activating PKS (∼193 kDa) that appears to work as a loading protein which "presents" the starter unit to the second PKS subunit. The pimS1 gene encodes a giant multienzyme (∼710 kDa) harboring 15 activities responsible for the first four cycles of chain elongation in pimaricin biosynthesis, resulting in formation of the polyene chromophore. The biosynthetic gene cluster for the 26-membered ring of the polyene macrolide pimaricin extends for about 110 kilobase pairs of contiguous DNA in the genome of Streptomyces natalensis. Two sets of polyketide synthase (PKS) genes are separated by a group of small polyketide-functionalizing genes. Two of the polyketide synthase genes, pimS0 and pimS1, have been fully sequenced and disrupted proving the involvement of each of these genes in pimaricin biosynthesis. The pimS0 gene encodes a relatively small acetate-activating PKS (∼193 kDa) that appears to work as a loading protein which "presents" the starter unit to the second PKS subunit. The pimS1 gene encodes a giant multienzyme (∼710 kDa) harboring 15 activities responsible for the first four cycles of chain elongation in pimaricin biosynthesis, resulting in formation of the polyene chromophore. polyketide synthase acyl carrier protein acyltransferase β-ketoacyl ACP synthase kilobase(s) Polyketides are a large and highly diverse group of natural products (1O′Hagan D. The Polyketide Metabolites. Ellis Horwood Ltd., Chichester, UK1991Google Scholar) that includes antibacterial, antifungal, anticancer, antiparasitic, and immunosuppressant activities, among others. Despite their structural diversity, these metabolites are synthesized by a common pathway in which units derived from acetate, propionate, or butyrate are condensed onto the growing chain by a polyketide synthase (PKS)1 in a process resembling fatty acid biosynthesis (2Hopwood D.A. Sherman D.H. Annu. Rev. Genet. 1990; 24: 37-66Crossref PubMed Google Scholar, 3Staunton J. Angew. Chem. Int. Ed. Engl. 1991; 30: 1302-1306Crossref Scopus (44) Google Scholar, 4Katz L. Donadio S. Annu. Rev. Microbiol. 1993; 47: 875-912Crossref PubMed Scopus (321) Google Scholar), except that the β-keto function introduced at each elongation step may undergo all, part, or none of a reductive cycle comprising β-ketoreduction, dehydration, and enoyl reduction (5Schwecke T. Aparicio J.F. Molnár Y. König A. Khaw L.E. Haydock S.F. Oliynyk M. Caffrey P. Cortés J. Lester J.B. Böhm G.A. Staunton J. Leadlay P.F. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7839-7843Crossref PubMed Scopus (399) Google Scholar). The structural variety in this class of products arises from choice of monomers, the extent of β-ketoreduction, and dehydration and the stereochemistry of each chiral center. Yet further diversity is produced by functionalization of the polyketide chain by the action of glycosylases, methyltransferases, and oxidative enzymes. Macrocyclic polyketides are produced mainly by Streptomycesand related filamentous bacteria through the action of so-called type I modular PKSs (5Schwecke T. Aparicio J.F. Molnár Y. König A. Khaw L.E. Haydock S.F. Oliynyk M. Caffrey P. Cortés J. Lester J.B. Böhm G.A. Staunton J. Leadlay P.F. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7839-7843Crossref PubMed Scopus (399) Google Scholar, 6Donadio S. Staver M.J. McAlpine J.B. Swanson S.J. Katz L. Science. 1991; 252: 675-679Crossref PubMed Scopus (752) Google Scholar, 7Cortés J. Haydock S.F. Roberts G.A. Bevitt D.J. Leadlay P.F. Nature. 1990; 348: 176-178Crossref PubMed Scopus (515) Google Scholar, 8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar). These enzymes usually consist of several extremely large polypeptides in which different modules (sets) of enzymic activities catalyze each successive round of elongation (6Donadio S. Staver M.J. McAlpine J.B. Swanson S.J. Katz L. Science. 1991; 252: 675-679Crossref PubMed Scopus (752) Google Scholar, 7Cortés J. Haydock S.F. Roberts G.A. Bevitt D.J. Leadlay P.F. Nature. 1990; 348: 176-178Crossref PubMed Scopus (515) Google Scholar). Characterization of such PKS systems is not only an important scientific challenge but a necessary start to the study of factors that control the specificity and extent of chain extension, which are largely unknown (4Katz L. Donadio S. Annu. Rev. Microbiol. 1993; 47: 875-912Crossref PubMed Scopus (321) Google Scholar, 9Marsden A.F.A. Caffrey P. Aparicio J.F. Loughran M.S. Staunton J. Leadlay P.F. Science. 1994; 263: 378-380Crossref PubMed Scopus (145) Google Scholar, 10Aparicio J.F. Caffrey P. Marsden A.F.A. Staunton J. Leadlay P.F. J. Biol. Chem. 1994; 269: 8524-8528Abstract Full Text PDF PubMed Google Scholar). Polyene macrolides are a group of macrocyclic polyketides that interact with membrane sterols and are, therefore, active against fungi but not bacteria (11Martı́n J.F. Annu. Rev. Microbiol. 1977; 31: 12-38Crossref Scopus (60) Google Scholar, 12Gil J.A. Martı́n J.F. Strohl W.R. Biotechnology of Antibiotics. 2nd Ed. Marcel Dekker, Inc., New York1997: 551-576Google Scholar). Pimaricin (Fig. 1) represents a prototype molecule of glycosylated polyenes (13Omura S. Tanaka H. Omura S. Macrolide Antibiotics, Chemistry, Biology, and Practice. Academic Press, Inc., NY1984: 351-404Google Scholar) important for antifungal therapy and promising for antiviral activity, stimulation of the immune response, and action in synergy with other antifungal drugs or antitumor compounds (14Bolard J. Biochim. Biophys. Acta. 1986; 719: 11-17Google Scholar). It is produced byStreptomyces natalensis and widely utilized in the food industry to prevent mold contamination of cheese and other nonsterile foods (i.e. cured meats). The macrolide rings of polyene antibiotics are larger (up to twice in size) than those of standard 14- or 16-membered nonpolyene macrolides (11Martı́n J.F. Annu. Rev. Microbiol. 1977; 31: 12-38Crossref Scopus (60) Google Scholar). Their rings include a chromophore of conjugated double bonds (the characteristic polyene structure). The genetic determination of the length of the polyene and the chromophore and the ring size of these giant macrolide rings is intriguing. Despite the general interest of polyene macrolides, very little is known about their biosynthetic routes and the gene clusters encoding them. We have, therefore, undertaken the cloning of a largeStreptomyces gene cluster involved in the biosynthesis of the polyketide backbone that forms the 26-membered tetraene macrolide ring of pimaricin. pimS1, the gene responsible for the first four rounds of chain elongation has been sequenced, and its modular nature has been established. pimS0, a gene required for chain initiation has been also studied. Gene disruption studies provide evidence for the involvement of pimS0 and pimS1in pimaricin biosynthesis. S. natalensis ATCC 27448 was used as the source of DNA in the construction of the genomic library. Escherichia coli strain XL1-Blue MR was used for obtaining SuperCos 1 cosmid (Stratagene) recombinant derivatives and also served as a host for plasmid subcloning in plasmids pBluescript (Stratagene) pUC18 and pUC19.Candida utilis (synonym Pichia jadinii) CECT 1061 was used for bioassay experiments. Phage KC515 (c + attP::tsr::vph), a ØC31-derived phage (15Rodicio M.R. Bruton C.J. Chater K.F. Gene. 1985; 34: 283-292Crossref PubMed Scopus (54) Google Scholar), was used for gene disruption experiments (16Hopwood D.A. Bibb M.J. Chater K.F. Kieser T. Bruton C.J. Kieser H.M. Lydiate D.J. Smith C.P. Ward J.M. Schrempf H. Genetic Manipulation of Streptomyces: A Laboratory Manual. John Innes Foundation, Norwich, UK1985Google Scholar). Streptomyces lividans JII 1326 (17Chater K.F. Bruton C.J. Suárez J.E. Gene. 1981; 14: 183-194Crossref PubMed Scopus (25) Google Scholar) served as a host for phage propagation and transfection. Standard conditions for culture ofStreptomyces species and isolation of phages were as described by Hopwood et al. (16Hopwood D.A. Bibb M.J. Chater K.F. Kieser T. Bruton C.J. Kieser H.M. Lydiate D.J. Smith C.P. Ward J.M. Schrempf H. Genetic Manipulation of Streptomyces: A Laboratory Manual. John Innes Foundation, Norwich, UK1985Google Scholar). S. natalensis was routinely grown in YEME medium (16Hopwood D.A. Bibb M.J. Chater K.F. Kieser T. Bruton C.J. Kieser H.M. Lydiate D.J. Smith C.P. Ward J.M. Schrempf H. Genetic Manipulation of Streptomyces: A Laboratory Manual. John Innes Foundation, Norwich, UK1985Google Scholar). Sporulation was achieved in TBO medium (2% (w/v) tomato paste, 2% (w/v) oats flakes, 2.5% (w/v) agar) at 30 °C. For pimaricin production, the strain was grown in phosphate-limited SPG medium (2.5% (w/v) soya peptone, 0.5 mmZnSO4·7H2O, 2% (w/v) glucose, pH 7.5) as described by Martı́n and Demain (18Martı́n J.F. Demain A.L. Biochem. Biophys. Res. Commun. 1975; 71: 1103-1109Crossref Scopus (34) Google Scholar). Standard genetic techniques withE. coli and in vitro DNA manipulations were as described by Sambrook et al. (19Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989Google Scholar). E. coli was transformed by the procedure of Hanahan (20Hanahan D. Glover D.M. DNA Cloning: A Practical Approach. IRL Press at Oxford University Press, Oxford1985: 109-132Google Scholar). Recombinant DNA techniques in Streptomyces species and isolation ofStreptomyces total and phage DNA were performed as described previously (16Hopwood D.A. Bibb M.J. Chater K.F. Kieser T. Bruton C.J. Kieser H.M. Lydiate D.J. Smith C.P. Ward J.M. Schrempf H. Genetic Manipulation of Streptomyces: A Laboratory Manual. John Innes Foundation, Norwich, UK1985Google Scholar). The library of size-fractionated genomic DNA was screened using probes derived from the rapamycin PKS genes from Streptomyces hygroscopicus (5Schwecke T. Aparicio J.F. Molnár Y. König A. Khaw L.E. Haydock S.F. Oliynyk M. Caffrey P. Cortés J. Lester J.B. Böhm G.A. Staunton J. Leadlay P.F. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7839-7843Crossref PubMed Scopus (399) Google Scholar, 8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar) and from the candicidin genes ofStreptomyces griseus (21Criado L.M. Martı́n J.F. Gil J.A. Gene. 1993; 126: 135-139Crossref PubMed Scopus (32) Google Scholar). A positive clone (Cos 9) containing PKS genes was used as the starting point for chromosome walking. Cosmid clones were obtained that extended over a contiguous 110-kb region of the S. natalensis chromosome, and individual inserts were subcloned into plasmid vectors. Sequencing templates were obtained by random subcloning of fragments generated by controlled partial HaeIII digestions. DNA sequencing was accomplished by the dideoxynucleotide chain termination method (22Sanger F. Nicklen S. Coulson A.R. Proc. Natl. Acad. Sci. U. S. A. 1977; 74: 5463-5467Crossref PubMed Scopus (56781) Google Scholar) on double-stranded DNA templates with an Applied Biosystems model 310 sequencer (Foster City, CA.). Subclone junctions were verified by direct sequencing of cosmid clones, using as primer, sets of synthesized internal oligonucleotides. Each nucleotide was sequenced a minimum of three times on both strands. Computer analysis of DNA and protein sequences. Primary DNA sequence data were analyzed and assembled by using software from DNAStar (Madison, Wis.). DNA and protein sequences were analyzed with the University of Wisconsin Genetics Computer Group programs (23Devereux J. Haeberli P. Smithies O. Nucleic Acids Res. 1984; 12: 387-395Crossref PubMed Scopus (12363) Google Scholar). To assay pimaricin in culture broths, 0.5 ml of culture was extracted with 2 ml of butanol, and the organic phase was diluted in water-saturated butanol to bring the absorbance at 319 nm in the range of 0.1 to 0.4 absorbance units. Control solutions of pure pimaricin (Gist-Brocades, Delft, The Netherlands) were used as control. To confirm the identity of pimaricin, a UV-visible absorption spectrum (absorption peaks at 319, 304, 291, and 281 nm) was routinely determined in a Hitachi U-2001 spectrophotometer. For routine tests, the fungicidal activity of pimaricin was tested by bioassay using C. utilis CECT 1061 as test organism. The pimaricin biosynthetic genes were identified by hybridization using as probes DNA from both the PKS genes of S. hygroscopicus that direct the biosynthesis of the macrocyclic polyketide immunosuppressant rapamycin (5Schwecke T. Aparicio J.F. Molnár Y. König A. Khaw L.E. Haydock S.F. Oliynyk M. Caffrey P. Cortés J. Lester J.B. Böhm G.A. Staunton J. Leadlay P.F. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7839-7843Crossref PubMed Scopus (399) Google Scholar, 8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar) and the PKS genes ofS. griseus that are involved in the biosynthesis of the heptaene macrolide candicidin (21Criado L.M. Martı́n J.F. Gil J.A. Gene. 1993; 126: 135-139Crossref PubMed Scopus (32) Google Scholar). A cosmid library was constructed, and a total of 110 kb of contiguous DNA was cloned. Fig.2 shows the cosmid coverage of this region. All the cosmids were mapped with restriction enzymes NotI,SphI, BamHI, and EcoRI. InternalNotI fragments were the same size as their homologous fragments of S. natalensis total DNA, suggesting that the cloned DNA was not rearranged. NotI, SphI, BamHI, and EcoRI restriction fragments of the isolated cosmids were probed with a labeled 0.7-kb NotI fragment from Cos9 (fragment a in Fig. 2 A). This probe was chosen because it codes for a KS domain of a type I PKS (the most invariant domain in modular PKSs; Ref. 8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar), as judged by sequence comparison alignments with other modular PKSs (see below). Because of the modular nature of PKS genes, the use of a highly conserved probe at moderate stringency hybridization gave us a gross indication of the extent of the PKS-coding DNA. Thus, about 13 kb of the rightmost sequences of Cos 37 (Fig. 2) did not hybridize with the probe, suggesting that the end of the PKS part of the gene cluster is located within cosmid 37. This hybridization approach also indicated that approximately 9 kb of the leftmost sequences of Cos13 (Fig. 2) were beyond the PKS portion of the cluster. In addition, no hybridization signal was observed in the central area of the cluster (Fig. 2), suggesting that a possible non-PKS region (∼10 kb) is located between sets of PKS genes in this central area. To confirm these results, all the NotI fragments within the 110 kb of cloned chromosomal DNA were subcloned into pBluescript, and their ends were sequenced. Fig. 2 B shows the function to which translated peptides from the sequence showed similarity and the direction of transcription. The first set of PKS genes pimSOand pimS1 (Fig. 2) were fully sequenced. The second set of PKS genes are transcribed in opposite orientation to the first set and appears to encode two additional modular PKSs. Several non-PKS-accompanying proteins were found to be encoded between the two large sets of PKS domains, further corroborating the above results. One of these activities showed significant sequence similarity with permeases within the ATP-dependent ABC transporter superfamily, whereas another one showed high similarity to cholesterol oxidases of other Streptomyces species. Other activities with counterparts in the protein data base included a sensory transduction protein, a ferredoxin, and a cytochrome P-450 monooxygenase. All these genes can be plausibly assigned roles in pimaricin biosynthesis. On the right side of the cluster, a cytochrome P-450 monooxygenase-like gene (pimD) was also identified and fully sequenced. 2J. F. Aparicio, A. J. Colina, and J. F. Martı́n, unpublished results. No other open reading frames with known homologies were detected on the left side of the 110-kb region studied. The involvement of the cloned DNA region in pimaricin biosynthesis was tested by gene disruption. Because S. natalensis ATCC 27448 has so far proven to be highly resistant to transformation with commonly used Streptomyces plasmids, a different approach for delivering of DNA into this strain was used. We took advantage of the ability of phage KC515, anattP-defective ØC31 derivative (15Rodicio M.R. Bruton C.J. Chater K.F. Gene. 1985; 34: 283-292Crossref PubMed Scopus (54) Google Scholar), to infect S. natalensis to introduce DNA into this strain. A 3.3-kb NotI subclone in pBluescript containing an internal coding region to PKS domains on the right side of the cluster (fragmentb in Fig. 2 A) was used as a source of DNA for subcloning into KC515 (Fig. 3). A 3.25-kbSacI fragment (containing most of the insert and theSacI site of the polylinker of pBluescript) was cloned into KC515, and the recombinant phage DNA was transfected into S. lividans protoplasts, as described elsewhere (16Hopwood D.A. Bibb M.J. Chater K.F. Kieser T. Bruton C.J. Kieser H.M. Lydiate D.J. Smith C.P. Ward J.M. Schrempf H. Genetic Manipulation of Streptomyces: A Laboratory Manual. John Innes Foundation, Norwich, UK1985Google Scholar). Further infection of S. natalensis with the recombinant phage allowed the selection for lysogen formation. Because the KC515 phage lacks attP, it can form lysogens only by homologous recombination into the chromosome (Fig. 3 A). Several lysogens of S. natalensis were obtained by selection for thiostrepton resistance, and their pimaricin production was tested by bioassay against C. utilis. No antibiotic activity was observed in the tested lysogens (Fig. 3 C) as compared with the halos of antifungal activity produced by the wild type strain. One of these disrupted mutants was selected, and its culture broth was extracted with butanol. It showed no traces of pimaricin by the spectrophotometric assay (see "Experimental Procedures"). Southern hybridization (Fig. 3 B) showed that integration of the phage DNA caused a restriction pattern change. When the 3.25-kbSacI fragment used to construct the KC515 derivative was used as a probe (shown as dashed bars in Fig. 3 A) for genomic DNA hybridization, it hybridized with the 9.3-kbSacI fragment of the wild type. However, in the disrupted mutant, a new 3.25 SacI fragment also hybridized with the probe (Fig. 3 B), indicating that a single crossover event had occurred. The observed hybridizing bands corresponded exactly to those expected according to the integration shown in Fig. 3. An approximately 35-kb area spanning the DNA fragment used for disruption (b in Fig. 2 A) was sequenced and analyzed. One of the most striking features of the sequenced region is the presence of an extraordinarily large open reading frame named pimS1 (Fig.2) encoding a protein of 6797 amino acids, including the initiating methionine. The molecular mass of the protein encoded bypimS1 (hereafter named PIMS1) was calculated to be 710,250 Da as a monomer. Comparison of protein PIMS1 with sequences of the SWISSPROT protein sequence data base revealed a significant similarity to known fatty acid and polyketide synthases. When the sequence of thepimS1 gene product was compared with itself, clear evidence was obtained for internal reiterations, indicating that PIMS1 consists of 4 modules for chain extension (Fig.4 B). Module 1 contains KS, AT, β-ketoreductase, and ACP domains. Modules 2, 3, and 4, however, contain extra dehydratase domains that account for a stretch of three contiguous enoyl groups in the final pimaricin molecule (see below). In all dehydratase domains of PIMS, there is a good alignment with the dehydratase domain from 6-deoxyerythronolide B synthase module 4 and the dehydratase domains from the RAPS proteins of the rapamycin synthase (8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar); the active-site motif HXXXGXXXXP (24Bevitt D.J. Cortés J. Haydock S.F. Leadlay P.F. Eur. J. Biochem. 1992; 204: 39-49Crossref PubMed Scopus (155) Google Scholar, 25Donadio S. Katz L. Gene. 1992; 11: 51-60Crossref Scopus (196) Google Scholar) is present. In addition, KS domains are highly conserved (70–80% identity over the whole domain) and contain the two invariant His residues located 135 and 173 amino acids C-terminal of the active-site Cys, as previously stated for other PKS ketosynthase domains (Ref. 8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar and references therein). The β-ketoreductase domains of PIMS contain a potential motif for NADP(H) binding (GXGXXGXXXA) at the N-terminal end of the domain (26Scrutton N.S. Berry A. Perham R.N. Nature. 1990; 343: 38-43Crossref PubMed Scopus (670) Google Scholar), as occurs with all the previously published β-ketoreductase domains except for those of the rapamycin synthase (8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar). ACP domains in PIMS are also highly conserved. Remarkably, ACP from PIMS module 1 has the active-site sequence QGFDS, whereas all previously identified prokaryotic ACPs have Leu adjacent to the Gly. AT domains of type I modular PKSs fall into two distinct groups, depending upon whether the substrate for the AT is 2S-methylmalonyl-CoA (9Marsden A.F.A. Caffrey P. Aparicio J.F. Loughran M.S. Staunton J. Leadlay P.F. Science. 1994; 263: 378-380Crossref PubMed Scopus (145) Google Scholar) or malonyl-CoA, as discussed in detail elsewhere (27Haydock S. Aparicio J.F. Molnár I. Schwecke T. König A. Marsden A.F.A. Galloway I.S. Staunton J. Leadlay P.F. FEBS Lett. 1995; 374: 246-248Crossref PubMed Scopus (209) Google Scholar). When PIMS1 AT domains were compared with the divergent motifs found in AT domains of type I PKSs, a convincing match was seen with the malonyl-CoA ("acetate") consensus (Pim 01, 02, 03, and 04 in Fig. 5), thus suggesting that these AT domains incorporate acetate extender units into the growing polyketide chain. Alignment of other AT domains of the cluster derived from the DNA sequence of subclone ends (when the sequence covered the stretch of 20 amino acids that corresponds to the divergent AT motifs) allowed the identification of another six acetate-incorporating ATs (Pim 00, 05, 06, 08, 09, and 10) (two others remain to be determined to complete the known chain length of pimaricin) and one AT that incorporates propionate extender units (Pim 07, Figs. 2 and 5). Based on the proposed biosynthetic units of pimaricin (11Martı́n J.F. Annu. Rev. Microbiol. 1977; 31: 12-38Crossref Scopus (60) Google Scholar) (Fig. 1), the polyketide synthase should harbor only one such propionate-specific AT, the one comprised in module 7 and responsible for the incorporation of carbons 11, 12, and the exocyclic methyl group that would undergo later oxidation to form the free carboxyl function of the aglycone, as occurs in lucensomycin biosynthesis (28Manwaring D.G. Rickards R.W. Gaudiano G. Nicolella V. J. Antibiot. (Tokyo). 1969; 22: 545-550Crossref PubMed Scopus (24) Google Scholar). The rare Leu codon TTA has been proposed to serve as a device for global regulation in Streptomyces coelicolor (29Leskiw B.K. Lawlor E.J. Fernández-Abalos J.M. Chater K.F. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 2461-2465Crossref PubMed Scopus (172) Google Scholar). There is one such codon in the pimS1 gene, which may allow this regulatory mechanism to control expression of pimS1. Lying downstream of pimS1 and convergently transcribed, there is another open reading frame (pimS0) encoding a protein of 1847 amino acids. The molecular mass of the protein encoded by pimS0 (hereafter named PIMS0) was calculated to be 193,407 Da as a monomer. PIMS0 contains in its C-terminal end a whole PKS module with no domains for reduction or dehydration (i.e. only KS, AT, and ACP; Fig.4 C). Its KS domain lacks the active-site cysteine residue (Fig. 6). Besides, the sequence of its first domain in the N-terminal region, containing about 560 amino acid residues, reveals up to 28% amino acid sequence identity to ATP-dependent carboxylic acid:CoA ligases, including acetyl-CoA synthetases from Bacillus subtilis (EMBL L17309), coumarate-CoA ligases from plants (EMBL X13325, L43362, D39405,U50845), 2,3-dihydroxybenzoate-AMP ligases from E. coli(EMBL X15058), and B. subtilis (U26444), firefly luciferases (EMBL A26772, S33788, S29354), and the long chain fatty acid-carboxylic acid:CoA ligase from Mycobacterium tuberculosis (EMBLQ10776), among others. An ACP domain following the carboxylic acid:CoA ligase domain completes the number of domains housed within this protein (Fig. 4 C). Remarkably, the C-terminal ACP has the active-site sequence MGINS instead of the signature sequence LGXDS. To rule out the possibility that pimS0 could be a remnant of a former gene with no real involvement in pimaricin biosynthesis, its disruption was accomplished. A 1.9-kb BamHI-PstI fragment (containing the coding region for most of the carboxylic acid:CoA ligase domain and the N-terminal end of the KS domain ofpimS0) was subcloned into KC515. The method to achieve disruption was as described above for pimS1. Selected disruptants showed no production of pimaricin (not shown), indicating that pimS0 is strictly required for pimaricin biosynthesis. Polyene macrolides contain very large macrolide rings (sometimes up to 38-membered rings). It is unknown if such large macrolides are synthesized by single PKSs as occurs with 14- or 16-membered macrolides. In this paper we describe the first extensive genetic characterization of a polyene macrolide biosynthetic gene cluster. Southern hybridization experiments were used as means to identify and isolate cosmid clones containing the S. natalensis gene cluster for the 26-membered ring of the macrocyclic polyketide pimaricin. The most striking feature of the DNA region cloned (110 kb) is that the polyketide synthase genes turned out to be split into two subclusters separated by genes involved in the functionalization of the polyketide backbone. This is the first example of two separate PKSs required for the biosynthesis of different fragments of a single polyketide. In the rapamycin biosynthetic gene cluster, the PKS genes are separated by a pipecolate-activating gene that is directly involved in the formation of the aglycone core of rapamycin (5Schwecke T. Aparicio J.F. Molnár Y. König A. Khaw L.E. Haydock S.F. Oliynyk M. Caffrey P. Cortés J. Lester J.B. Böhm G.A. Staunton J. Leadlay P.F. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7839-7843Crossref PubMed Scopus (399) Google Scholar). Assembly of a complex polyketide core by non-PKS enzymes occurs also in Aspergillus nidulans, where a gene for the biosynthesis of a preformed fatty acid precursor of sterigmatocystin is separated from the type II PKS genes (30Brown D.W. Adams T.H. Keller N.P. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 14873-14877Crossref PubMed Scopus (130) Google Scholar). However, this organization of a modular PKS gene cluster separated by genes involved in functionalization of the polyketide backbone is unprecedented. The arrangement in separate subclusters could suggest that this huge macrolide has been assembled by putting together genes encoding two different polyketide synthases. Future analysis of the DNA sequence composition of the pimaricin synthase subclusters and of other polyene macrolides could provide evidence for this hypothesis. In total, the giant PIMS1 protein contributes 15 catalytic functions, making this an extremely complex multienzyme, only paralleled by the RAPS proteins of the rapamycin synthase (5Schwecke T. Aparicio J.F. Molnár Y. König A. Khaw L.E. Haydock S.F. Oliynyk M. Caffrey P. Cortés J. Lester J.B. Böhm G.A. Staunton J. Leadlay P.F. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7839-7843Crossref PubMed Scopus (399) Google Scholar). The domain organization of each module is as presented in Fig. 4 B. The limits of each domain were readily assigned by comparison with modules from 6-deoxyerythronolide B synthase and the rapamycin synthase (8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar). In modular PKSs, each module includes the activities required for a round of polyketide chain elongation and reduction, if required. The order of the modules along the polypeptide reflects the order in which each module is used for polyketide chain extension (5Schwecke T. Aparicio J.F. Molnár Y. König A. Khaw L.E. Haydock S.F. Oliynyk M. Caffrey P. Cortés J. Lester J.B. Böhm G.A. Staunton J. Leadlay P.F. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7839-7843Crossref PubMed Scopus (399) Google Scholar, 6Donadio S. Staver M.J. McAlpine J.B. Swanson S.J. Katz L. Science. 1991; 252: 675-679Crossref PubMed Scopus (752) Google Scholar, 7Cortés J. Haydock S.F. Roberts G.A. Bevitt D.J. Leadlay P.F. Nature. 1990; 348: 176-178Crossref PubMed Scopus (515) Google Scholar, 8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar, 31MacNeil D.J. Occi J.L. Gewain K.M. MacNeil T. Gibbons P.H. Ruby C.L. Danis S.J. Gene. 1992; 115: 119-125Crossref PubMed Scopus (169) Google Scholar), and the order of the domains within each module is KS, AT, (dehydratase, enoyl reductase where applicable), β-ketoreductase, ACP, exactly as for fatty acid synthetase (32Wakil S.J. Biochemistry. 1989; 28: 4523-4530Crossref PubMed Scopus (702) Google Scholar), 6-deoxyerythronolide B synthase (24Bevitt D.J. Cortés J. Haydock S.F. Leadlay P.F. Eur. J. Biochem. 1992; 204: 39-49Crossref PubMed Scopus (155) Google Scholar, 25Donadio S. Katz L. Gene. 1992; 11: 51-60Crossref Scopus (196) Google Scholar), or the RAPS proteins of the rapamycin synthase (5Schwecke T. Aparicio J.F. Molnár Y. König A. Khaw L.E. Haydock S.F. Oliynyk M. Caffrey P. Cortés J. Lester J.B. Böhm G.A. Staunton J. Leadlay P.F. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7839-7843Crossref PubMed Scopus (399) Google Scholar, 8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar). The functionality of PIMS1 in pimaricin biosynthesis was established by inactivating the pimS1 gene by phage-mediated gene disruption. The mutant strain obtained showed a complete lack of pimaricin production, indicating that pimS1 directs the synthesis of pimaricin. Similarly, when we achieved the functional inactivation of pimS0, pimaricin production was completely abrogated, indicating that pimS0 was also involved in pimaricin biosynthesis. In both disrupted mutants, formation of the polyene chromophore was prevented because there was no absorption at the typical peaks of the polyene spectrum (not shown). Sequencing analysis of pimS0 indicated that in its N-terminal region, PIMS0 contains an ATP-dependent carboxylic acid:CoA ligase. The only similarity to such an adenylate-forming domain in PKS clusters is a sequence of unknown function from B. subtilis (33Albertini A.M. Caramori T. Scoffone F. Scotti C. Galizzi A. Microbiology. 1995; 141: 299-309Crossref PubMed Scopus (60) Google Scholar) and the 3,4-dihydroxycyclohexane carboxylic acid-activating domain present in RAPS 1 and responsible for the initiation of rapamycin biosynthesis (5Schwecke T. Aparicio J.F. Molnár Y. König A. Khaw L.E. Haydock S.F. Oliynyk M. Caffrey P. Cortés J. Lester J.B. Böhm G.A. Staunton J. Leadlay P.F. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7839-7843Crossref PubMed Scopus (399) Google Scholar,8Aparicio J.F. Molnár I. Schwecke T. König A. Haydock S.F. Khaw L.E. Staunton J. Leadlay P.F. Gene. 1996; 169: 9-16Crossref PubMed Scopus (202) Google Scholar). Besides, the C terminus of carboxylic acid:CoA ligase PIMS0 contains a whole PKS module with no domains for reduction or dehydration. It is noteworthy that the KS lacks the active-site cysteine residue (see Fig. 6), showing a serine instead, making this KS possibly inactive, because it cannot form the thioester bond with its substrate. It could be argued that this KS might retain some activity because of the formation of O-ester bonds between Ser and the substrate; however, site-directed mutagenesis studies in the tetracenomycin-producing strain Streptomyces glaucescens in which the active-site Cys was replaced by Ser (or Ala) completely abolished the KS activity (34Meurer G. Hutchinson C.R. J. Bacteriol. 1995; 177: 477-481Crossref PubMed Google Scholar), thus supporting the idea that this PIMS0 KS is functionally inactive. Modules 2, 3, and 4 of PIMS1 contain dehydratase domains that would catalyze a dehydration of the reduced β-keto group provided by the corresponding β-ketoreductase activities present in the same modules (6Donadio S. Staver M.J. McAlpine J.B. Swanson S.J. Katz L. Science. 1991; 252: 675-679Crossref PubMed Scopus (752) Google Scholar), yielding three consecutive double bonds in the polyketide core of pimaricin corresponding to the polyene chromophore. Module 1, however, lacks this dehydratase domain. Therefore, it can only process the β-carbon to a hydroxyl group in the growing polyketide chain (Fig.7). Thus, there is only one portion of the pimaricin polyketide backbone in which PIMS1 could be involved,i.e. the first four chain elongation steps. Module 1 would catalyze the condensation of a starter unit (acetate) to the extender unit (malonate). In modular PKSs, the starter unit is normally provided by a loading domain included at the N terminus of the first multienzyme, as occurs with 6-deoxyerythronolide B synthase 1 (10Aparicio J.F. Caffrey P. Marsden A.F.A. Staunton J. Leadlay P.F. J. Biol. Chem. 1994; 269: 8524-8528Abstract Full Text PDF PubMed Google Scholar, 35Staunton J. Caffrey P. Aparicio J.F. Robets G.A. Bethell S.S. Leadlay P.F. Nat. Struct. Biol. 1996; 3: 188-192Crossref PubMed Scopus (118) Google Scholar) or the rapamycin synthase RAPS 1 (5Schwecke T. Aparicio J.F. Molnár Y. König A. Khaw L.E. Haydock S.F. Oliynyk M. Caffrey P. Cortés J. Lester J.B. Böhm G.A. Staunton J. Leadlay P.F. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7839-7843Crossref PubMed Scopus (399) Google Scholar). PIMS1, however, lacks any such loading domain, thus making necessary the interaction of module 1 KS either with free acetyl-CoA or, more likely, with a separate loading protein that would present the starter unit to module 1. Fig. 7 shows the latter possibility. The gene product ofpimS0 could be such a loading protein. Its first domain (carboxylic acid:CoA ligase) could activate acetate, forming the acyladenylate that would be transferred by the adjacent ACP to the KS of PIMS0. This KS, being inactive, would not catalyze any condensation reaction but would allow the transfer of the growing polyketide chain to the C-terminal ACP, which would feed the acetyl moiety to module 1 KS of PIMS1 for the first elongation step (Fig. 7). The cloning and specific alteration of the biosynthesis genes for this polyene might allow the engineering of novel analogs with improved antifungal properties. We thank M. R. Rodicio for providing KC515 phage and P. F. Leadlay and J. A. Gil for the kind gift of probes from the rapamycin and the candicidin biosynthesis clusters, respectively.
Polyketide synthase
Polyene
Gene cluster
Cite
Citations (103)
Gene cluster
Heterologous expression
Heterologous
Polyketide synthase
Strain (injury)
Cite
Citations (22)
Secondary metabolites are small molecules that show large structural diversity and a broad range of bioactivities. Some metabolites are attractive as drugs or pigments while others act as harmful mycotoxins. Filamentous fungi have the capacity to produce a wide array of secondary metabolites including polyketides. The majority of genes required for production of these metabolites are mostly organized in gene clusters, which often are silent or barely expressed under laboratory conditions, making discovery and analysis difficult. Fortunately, the genome sequences of several filamentous fungi are publicly available, greatly facilitating the establishment of links between genes and metabolites. This review covers the attempts being made to trigger the activation of polyketide metabolism in the fungal model organism Aspergillus nidulans. Moreover, it will provide an overview of the pathways where ten polyketide synthase genes have been coupled to polyketide products. Therefore, the proposed biosynthesis of the following metabolites will be presented; naphthopyrone, sterigmatocystin, aspyridones, emericellamides, asperthecin, asperfuranone, monodictyphenone/emodin, orsellinic acid, and the austinols.
Sterigmatocystin
Aspergillus nidulans
Polyketide synthase
Secondary metabolism
Secondary metabolite
Cite
Citations (42)