Departamento de Ecología, Genética y Microbiología, Área de Microbiología, Facultad de Biología, Universidad de León, 24071 León, Spain1
Author for correspondence: José A. Gil. Tel: +34 987 291503. Fax: +34 987 291479. e-mail: degjgs{at}unileon.es
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Keywords: aromatic polyene antibiotic, type I PKS, genetic disruption
Abbreviations: ACP, acyl carrier protein; AT, acyltransferase; DH, dehydratase; ER, enoylreductase; KR, ß-ketoreductase; KS, ß-ketoacyl ACP synthase; mAT, methylmalonate-specific AT; PABA, p-aminobenzoic acid; PKS, polyketide synthase; TE, thioesterase
The GenBank accession numbers for the sequences reported in this paper are AJ300302 and AJ300303.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Macrolide polyketides are formed through the action of type I modular PKSs. These enzymes usually consist of several large polypeptides housing modules (sets) of enzymic activities, each module containing the enzymes required to catalyse a particular round of elongation (Cortés et al., 1990 ; Donadio et al., 1991
; Bevitt et al., 1992
; Swan et al., 1994
; Schwecke et al., 1995
; Aparicio et al., 1996
, 1999
, 2000
; Brautaset et al., 2000
). The modules are arranged in the multienzyme polypeptides in the order in which they are used, but the order of the genes encoding them may or may not be co-linear with the order of action of the polypeptides.
Polyene macrolides are a group of polyketides with lactone rings of 2044 members. They contain a chromophore formed by a system of three to eight conjugated double bonds in the macrolactone ring that contributes to the interaction with sterols present in the membrane of fungi leading to formation of transmembrane pores and cell death (Gil & Martín, 1995 ). Although polyene antibiotics are usually described as antifungal agents, some of them have antibacterial, antiviral and immunostimulating activities (Gil & Martín, 1995
). Candicidin is an aromatic polyene (heptaene) antibiotic produced by Streptomyces griseus IMRU 3570 (Fig. 1
). It was first described by Lechevalier et al. (1953)
and named antibiotic C135, although it was renamed candicidin because of its strong activity against species of Candida.
|
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
DNA isolation, manipulation and characterization.
Total DNA isolation from S. griseus and recombinant DNA techniques in Streptomyces and Escherichia coli were as described by Hopwood et al. (1985) and Sambrook et al. (1989)
. For the construction of the genomic library, S. griseus genomic DNA was partially digested with Sau3AI and fragments in the range of 3540 kb were cloned into SuperCos 1 (Stratagene) digested with BamHI and XbaI. The library was initially screened using the 4·6 kb BamHI fragment of S. griseus containing the pabAB gene (Gil & Hopwood, 1983
; Criado et al., 1993
).
DNA sequencing and analysis.
Sequencing clones were constructed with the Erase-a-Base kit (Promega), converted into ssDNA by standard procedures (Sambrook et al., 1989 ) and sequenced by the dideoxynucleotide chain-termination method of Sanger et al. (1977)
using Sequenase (Amersham). Deazanucleotides were used to eliminate compression problems. Sequence analysis was performed with DNASTAR, database similarity searches were done at the BLAST and BLITZ public server (EBI), and multiple alignments of sequences were done by using CLUSTAL W (EBI).
![]() |
RESULTS AND DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
To start the analysis of the cloned region, the inserts from cosmids Cos G33, Cos I80 and the 1·2 kb PaeI DNA fragment from Cos S54 were subcloned in E. coli plasmids and sequenced. Computer-assisted analysis of the sequenced region led to identification of the genes shown in Fig. 2(b) and listed in Table 2
. They include genes for a modular PKS as well as genes for chain-termination, modification, mycosamine biosynthesis, export and regulation.
|
|
CanP2 has not been fully studied and the N-terminal region remains unknown. So far, we have identified three modules. All of them have methylmalonate-specific AT (mAT) domains. The last module of CanP2 contains KS, mAT, KR and ACP domains, processing the ß-carbon up to a hydroxyl group in the polyketide chain. The adjacent module includes extra DH and ER domains that account for the formation of a methylene group. The last known module of CanP2 houses mAT, KR and ACP domains; a KS is most probably situated upstream from the mAT domain. If so, this module would process the ß-carbon up to a hydroxyl group in the growing polyketide chain.
CanP3 consists of five identical modules with KS, AT, DH, KR and ACP domains, responsible for the formation of five (out of seven) conjugated double bonds.
CanPF was identified by sequencing the 1·2 kb PaeI DNA fragment from Cos S54. Two incomplete ORFs were determined; one of them displays the C-terminal region of a protein similar to GDP-mannose-4,6-dehydratases and the other corresponds to the N-terminal region (KS domain) of a modular PKS polypeptide (CanPF). Therefore, CanPF presumably represents one end of the PKS gene cluster.
We propose that candicidin aglycone biosynthesis starts on CanP1, a polypeptide with two modules (a loading domain and module 1) and continues on CanP2 (modules 24) and CanP3 (modules 59). Candicidin is an aromatic molecule and the chain starter unit is PABA. Chain initiation might require activation of PABA by the loading domain, which would be followed by the first elongation step catalysed by the adjacent module (module 1). The nascent chain would be transferred by the ACP of module 1 to the KS of the first module of CanP2 (module 2). Taking into account the co-linearity of the order of modules within each polypeptide with the order of their utilization during biosynthesis in type I PKSs (Donadio et al., 1991 ; Yu et al., 1999
) and the structure of the candicidin molecule, it is very probable that complete sequencing of CanP2 would not reveal any modules beyond the three described here. CanP2 would be responsible for the incorporation of three molecules of methylmalonyl-CoA as extender units into the growing polyketide chain, which is in good agreement with candicidin structure. After the three elongation steps carried out by CanP2 the chain would be transferred to CanP3, which would catalyse formation of most of the chromophore (five out of seven conjugated double bonds). The presence of a single polypeptide responsible for the formation of five conjugated double bonds is unprecedented.
It is notable that all domains that have been studied so far contain the following active site motifs: GPXXXXXTACSS and the two invariant His residues located 135 and 175 aa on the C-terminal side of the active site Cys for KS domains (Aparicio et al., 1996 ); GHSXG complemented with three conserved active site residues Gln, Arg and His for AT domains (Aparicio et al., 1996
; Serre et al., 1995
); HXXXGXXXXP for DH domains (Bevitt et al., 1992
; Aparicio et al., 1996
); the potential NADP(H) binding GXGXXGXXXA for KR and ER domains (Scrutton et al., 1990
); and LGXDSLXXVE for ACP domains.
Genes for chain-termination and aglycone modification
The thioesterase (TE) gene (canT) included in the 4·6 kb BamHI DNA fragment described previously (Criado et al., 1993 ) has been located in this work. The deduced product (256 aa) shows a percentage of identity much higher to the TEs encoded by independent genes in the pikromycin (pikAV), tylosin (tylO), nystatin (nysE) and pimaricin (pimI) biosynthetic gene clusters of Streptomyces venezuelae, Streptomyces fradiae, Streptomyces noursei and Streptomyces natalensis, respectively (Xue et al., 1998
; Merson-Davies & Cundliffe, 1994
; Brautaset et al., 2000
; Aparicio et al., 2000
) than to the TE domains located at the C termini of many modular PKSs. All the above biosynthetic clusters include two TE activities, one of them embedded in the PKS and the other one in a discrete gene. It has been proposed that the separate additional TE proteins might be essential to unblock (proof read) the PKS complex after blockage with aberrant incomplete polyketide chains (Butler et al., 1999
; Tang et al., 1999
). If so, the candicidin biosynthetic gene cluster probably contains another TE activity located in the PKS and responsible for chain-termination and ring closure to form the candicidin aglycone.
Two genes in the analysed sequence, canC and canF, encode proteins possibly involved in modification of the candicinolide. CanC (393 aa) is a P450 monooxygenase and shows a high percentage of identity along its full length with TylH1 (40·6%), RapJ (35·9%) and EryF (33·2%), which are P450 enzymes involved in the biosynthesis of tylosin, rapamycin and erythromycin, respectively (Fouces et al., 1999 ; Molnár et al., 1996
; Haydock et al., 1991
). By analogy with a similar oxidation in the biosynthesis of nystatin (Brautaset et al., 2000
), CanC would catalyse the oxidation of a methyl group at C-18. This is based on the assumption that a propionate residue is introduced in the 13th elongation step.
The finding of a ferredoxin gene (canF) immediately downstream of the P450 monooxygenase gene strongly indicates that the deduced protein is the in vivo electron transport protein physiologically involved with the haem protein. CanF is a small acidic protein (64 aa) very similar to ferredoxins containing [3Fe4S] clusters, such as PimF (Aparicio et al., 2000 ), SuaB (OKeefe et al., 1991
), SubB (OKeefe et al., 1991
), TylH2 (Fouces et al., 1999
) and RapO (Molnár et al., 1996
).
Genes for mycosamine biosynthesis and attachment
The cluster includes at least three genes possibly involved in biosynthesis and transfer of the candicidin deoxysugar moiety mycosamine. canG and canA are located on one side of the PKS gene cluster and canM (incomplete) is on the other side.
CanG (458 aa) shows up to 27% identity with eukaryotic enzymes that belong to the UDP-glycosyltransferase family. Therefore, this enzyme might be involved in the attachment of mycosamine to the candicidin aglycone at C-21. The deduced product of canA (352 aa) shows considerable similarity to proteins that belong to the EryCI family. The members of this family are part of a large group of proteins involved in carbohydrate metabolism in either antibiotic or outer-cell-wall biosynthesis, which appear to be aminotransferase or DH enzymes dependent on pyridoxal phosphate or pyridoxamine phosphate as a cofactor, respectively (Kim et al., 1998 ). CanA includes the conserved active sites Asp (aa 154) and Lys (aa 176) that are important for pyridoxal phosphate binding. The fact that mycosamine contains an amino group suggests that CanA is an aminotransferase involved in mycosamine biosynthesis that requires pyridoxal phosphate as cofactor. CanM reveals high similarity to GDP-mannose-4,6-dehydratases and it is plausible that it is involved in one of the initial steps in mycosamine biosynthesis.
Because the candicidin biosynthetic gene cluster has not been fully sequenced, further analysis might reveal other genes involved in mycosamine biosynthesis. However, no other mycosamine biosynthesis genes different from those described in this work have been reported to be present in the completely sequenced nystatin and pimaricin biosynthetic gene clusters (Brautaset et al., 2000 ; Aparicio et al., 2000
), although both of these metabolites contain mycosamine. This suggests that the rest of the enzymes required for deoxysugar formation could be recruited from different metabolic pathways.
Regulatory region and export genes
The deduced products of orf1, orf2 and orf3 show significant sequence similarity to the transcriptional regulators found in the upstream region of the cholesterol oxidase/cytochrome P450 operon of Streptomyces sp. SA-COO (Molnár & Murooka, 1993 ). These genes lie upstream of the region containing canG-canA-canC-canF-canT-pabAB-canP1-canRA-canRB genes (pabAB sequenced by Criado et al., 1993
) and are transcribed in the opposite direction. Such putative regulatory genes have also been identified in the nystatin biosynthetic gene cluster where disruption of one of them abolished nystatin production (Brautaset et al., 2000
). In candicidin biosynthesis, the deduced products of orf1, orf2 and orf3 might have a regulatory function, although experimental confirmation is required.
The last two genes that are transcribed in the same direction as canG, canA, canC, canF, canT, pabAB and canP1 are canRA and canRB. Their deduced products display high degrees of identity to proteins belonging to the ATP-dependent ABC transporter superfamily and might be involved in candicidin export. CanRA (335 aa) exhibits up to 49% identity to ATP-binding proteins belonging to the ABC transporter superfamily, particularly several from Streptomyces species involved in the efflux of antibiotics. All these proteins, including CanRA, display single Walker A and Walker B motifs (Walker et al., 1982 ) and loop 3 (Hyde et al., 1990
). CanRB (268 aa) is very similar to transmembrane proteins described for members of the ABC transporter superfamily (up to 31% identity) and all of them, including CanRB, have a typical hydrophobicity profile. Based on sequence similarities, we believe that CanRA and CanRB constitute an ABC transporter, structurally formed by two homodimeric elements that might be involved in ATP-dependent efflux of candicidin.
Disruption experiments
It has been suggested that S. griseus IMRU 3570 is not a suitable host for molecular genetic manipulation (Martín & Gil, 1984 ). However, we decided to use new vectors that had worked in other strains with the same problems to attempt disruption of some genes in the candicidin cluster of S. griseus IMRU 3570. After several attempts to disrupt pabAB, canP1 and canG, we obtained one positive result: disruption of pabAB. PabAB is responsible for the formation of PABA, which is the immediate precursor of the aromatic moiety of candicidin (Liu et al., 1972
; Martín & Liras, 1976
; Gil et al., 1980
). The disrupted strain (S. griseus ABC28) was obtained in a conjugation experiment between E. coli ET12567(pUZ8002) (Flett et al., 1997
) transformed with pCAB and S. griseus IMRU 3570. pCAB was constructed by cloning the 4·6 kb BamHI DNA fragment containing pabAB::tsr in the BamHI site of the conjugative suicide vector pOJ260 (Bierman et al., 1992
). S. griseus ABC28 is the result of a double homologous recombination between the chromosomal BamHI fragment (containing pabAB) and the pCAB insert (containing pabAB::tsr) (Fig. 4
).
|
|
Candicidin biosynthetic pathway in S. griseus IMRU 3570
A partial analysis of the candicidin biosynthetic gene cluster led to the identification of four PKS genes and 11 more genes, all of them with a plausible role in candicidin biosynthesis. It is very probable that synthesis starts with PABA activation by the loading domain of CanP1, followed by a first-chain extension cycle catalysed by the module located in the C terminus of CanP1 (Fig. 6). Chain elongation would proceed through condensation of three methylmalonyl-CoA extender units by CanP2. The next seven steps would require condensation of seven malonyl-CoA extender units and their processing up to conjugated double bond formation. CanP3 is responsible for the formation of five (out of seven) double bonds, but we still do not know which five out of the seven. Although the rest of the PKS genes have already been cloned and localized, they have not been sequenced and further work will be necessary to complete their characterization. During chain elongation CanT is predicted to clear aberrant polyketides, preventing PKS blockage and promoting candicidin accumulation. After formation of the candicidin aglycone (candicinolide) by PKS it is probable that CanC and CanF accomplish hydroxylation of the macrolactone ring at C-18.
|
![]() |
ACKNOWLEDGEMENTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Aparicio, J. F., Colina, A. J., Ceballos, E. & Martín, J. F. (1999). The biosynthetic gene cluster for the 26-membered ring polyene macrolide pimaricin. J Biol Chem 274, 10133-10139.
Aparicio, J. F., Fouces, R., Mendes, M. V., Olivera, N. & Martín, J. F. (2000). A complex multienzyme system encoded by five polyketide synthase genes is involved in the biosynthesis of the 26-membered polyene macrolide pimaricin in Streptomyces natalensis. Chem Biol 7, 895-905.[Medline]
Bevitt, D. J., Cortés, J., Haydock, S. F. & Leadlay, P. F. (1992). 6-Deoxyerythronolide B synthase from Saccharopolyspora erythraea. Cloning of the structural gene, sequence analysis and inferred domain structure of the multifunctional enzyme. Eur J Biochem 204, 39-49.[Abstract]
Bierman, M., Logan, R., OBrien, K., Seno, E. T., Rao, R. N. & Schoner, B. E. (1992). Plasmid cloning vectors for the conjugal transfer of DNA from Escherichia coli to Streptomyces spp. Gene 116, 43-49.[Medline]
Brautaset, T., Sekurova, O. N., Sletta, H., Ellingsen, T. E., Strom, A. R., Valla, S. & Zotchev, S. B. (2000). Biosynthesis of the polyene antifungal antibiotic nystatin in Strepomyces noursei ATCC 11455: analysis of the gene cluster and deduction of the biosynthetic pathway. Chem Biol 7, 395-403.[Medline]
Butler, A. R., Bate, N. & Cundliffe, E. (1999). Impact of thioesterase activity on tylosin biosynthesis in Streptomyces fradiae. Chem Biol 6, 287-292.[Medline]
Cortés, J., Haydock, S. F., Roberts, G. A., Bevitt, D. J. & Leadlay, P. F. (1990). An unusually large multifunctional polypeptide in the erythromycin-polyketide synthase of Saccharopolyspora erythraea. Nature 348, 176-178.[Medline]
Criado, L. M., Martín, J. F. & Gil, J. A. (1993). The pab gene from Streptomyces griseus, encoding p-aminobenzoic acid synthase, is located between genes possibly involved in candicidin biosynthesis. Gene 126, 135-139.[Medline]
Donadio, S., Staver, M. J., McAlpine, J. B., Swanson, S. J. & Katz, L. (1991). Modular organization of genes required for complex polyketide biosynthesis. Science 252, 675-679.[Medline]
Flett, F., Mersinias, V. & Smith, C. P. (1997). High efficiency intergeneric conjugal transfer of plasmid DNA from Escherichia coli to methyl DNA-restricting streptomycetes. FEMS Microbiol Lett 155, 223-229.[Medline]
Fouces, R., Mellado, E., Diez, B. & Barredo, J. L. (1999). The tylosin biosynthetic cluster from Streptomyces fradiae: genetic organization of the left region. Microbiology 145, 855-868.[Abstract]
Gil, J. A. & Hopwood, D. A. (1983). Cloning and expression of a p-aminobenzoic acid synthetase gene from the candicidin producer Streptomyces griseus. Gene 25, 119-132.[Medline]
Gil, J. A. & Martín, J. F. (1995). Polyene antibiotics. In Biotechnology of Antibiotics , pp. 551-575. Edited by W. R. Strohl. New York:Marcel Dekker.
Gil, J. A., Liras, P., Naharro, G., Villanueva, J. R. & Martín, J. F. (1980). Regulation by aromatic amino acids of the biosynthesis of candicidin by Streptomyces griseus. J Gen Microbiol 118, 189-195.[Medline]
Haydock, S. F., Dowson, J. A., Dhillon, N., Roberts, G. A., Cortés, J. & Leadlay, P. F. (1991). Cloning and sequence analysis of genes involved in erythromycin biosynthesis in Saccharopolyspora erythraea: sequence similarities between EryG and a family of S-adenosylmethionine-dependent methyltransferases. Mol Gen Genet 230, 120-128.[Medline]
Hopwood, D. A. & Sherman, D. H. (1990). Molecular genetics of polyketides and its comparison to fatty acid biosynthesis. Annu Rev Genet 24, 37-66.[Medline]
Hopwood, D. A., Bibb, M. J., Chater, K. F. & 7 other authors (1985). Genetic Manipulation of Streptomyces: a Laboratory Manual. Norwich: The John Innes Foundation.
Hu, H., Bao, K., Zhou, X., Hopwood, D. A., Kieser, T. & Deng, Z. (1994). Repeated polyketide synthase modules involved in the biosynthesis of a heptaene macrolide by Streptomyces sp. FR-008. Mol Microbiol 14, 163-172.[Medline]
Hyde, S. C., Emsley, P., Hartshorn, M. J. & 7 other authors (1990). Structural model of ATP-binding proteins associated with cystic fibrosis, multidrug resistance and bacterial transport. Nature 346, 362365.[Medline]
Katz, L. & Donadio, S. (1993). Polyketide synthesis: prospects for hybrid antibiotics. Annu Rev Microbiol 47, 875-912.[Medline]
Kim, C. G., Yu, T. W., Fryhle, C. B., Handa, S. & Floss, H. G. (1998). 3-Amino-5-hydroxybenzoic acid synthase, the terminal enzyme in the formation of the precursor of mC7N units in rifamycin and related antibiotics. J Biol Chem 273, 6030-6040.
Lechevalier, H., Acker, R. F., Corke, C. T., Haenseler, C. M. & Waksman, S. A. (1953). Candicidin, a new antifungal antibiotic. Mycologia 45, 155-171.
Liu, C. M., McDaniel, L. E. & Schaffner, C. P. (1972). Studies on candicidin biosynthesis. J Antibiot 25, 116-121.[Medline]
Martín, J. F. & Gil, J. A. (1984). Cloning and expression of antibiotic production genes. Biotechnology 2, 63-72.
Martín, J. F. & Liras, P. (1976). Rapid incorporation of precursors into candicidin by resting-cells of Streptomyces griseus. J Antibiot 29, 1306-1309.[Medline]
Merson-Davies, L. A. & Cundliffe, E. (1994). Analysis of five tylosin biosynthetic genes from the tylIBA region of the Streptomyces fradiae genome. Mol Microbiol 13, 349-355.[Medline]
Molnár, I. & Murooka, Y. (1993). Nucleotide sequence analysis of a region upstream of the cholesterol oxidase-cytochrome P450 operon of Streptomyces sp. SA-COO revealing repeating units coding for putative transmembrane and DNA-binding proteins. J Ferment Bioeng 76, 257-264.
Molnár, I., Aparicio, J. F., Haydock, S. F., Khaw, L. E., Schwecke, T., König, 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.[Medline]
OKeefe, D. P., Gibson, K. J., Emptage, M. H., Lenstra, R., Romesser, J. A., Litle, P. J. & Omer, C. A. (1991). Ferredoxins from two sulfonylurea herbicide monooxygenase systems in Streptomyces griseolus. Biochemistry 30, 447-455.[Medline]
Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.
Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74, 5463-5467.[Abstract]
Schwecke, T., Aparicio, J. F., Molnár, I. & 10 other authors (1995). The biosynthetic gene cluster for the polyketide immunosuppressant rapamycin. Proc Natl Acad Sci USA 92, 78397843.[Abstract]
Scrutton, N. S., Berry, A. & Perham, R. N. (1990). Redesign of the coenzyme specificity of a dehydrogenase by protein engineering. Nature 343, 38-43.[Medline]
Serre, L., Verbree, E. C., Dauter, Z., Stuitje, A. R. & Derewenda, Z. S. (1995). The Escherichia coli malonyl-CoA:acyl carrier protein transacylases at 1·5- resolution. Crystal structure of a fatty acid synthase component. J Biol Chem 270, 12961-12964.
Swan, D. G., Rodríguez, A. M., Vilches, C., Méndez, C. & Salas, J. A. (1994). Characterisation of a Streptomyces antibioticus gene encoding a type I polyketide synthase which has an unusual coding sequence. Mol Gen Genet 242, 358-362.[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.[Medline]
Tuan, J. S., Weber, J. M., Staver, M. J., Leung, J. O., Donadio, S. & Katz, L. (1990). Cloning of genes involved in erythromycin biosynthesis from Saccharopolyspora erythraea using a novel actinomyceteEscherichia coli cosmid. Gene 90, 21-29.[Medline]
Walker, J. E., Saraste, M., Runswick, M. J. & Gay, N. J. (1982). Distantly related sequences in the - and ß-subunits ATP synthase, myosin, kinases and other ATP-requiring enzymes and a common nucleotide binding fold. EMBO J 1, 945-951.[Medline]
Ward, J. M., Janssen, G. R., Kieser, T., Bibb, M. J., Buttner, M. J. & Bibb, M. J. (1986). Construction and characterisation of a series of multi-copy promoter-probe plasmid vectors for Streptomyces using the aminoglycoside phosphotransferase gene from Tn5 as indicator. Mol Gen Genet 203, 468-478.[Medline]
Xue, Y., Zhao, L., Liu, H. W. & Sherman, D. H. (1998). A gene cluster for macrolide antibiotic biosynthesis in Streptomyces venezuelae: architecture of metabolic diversity. Proc Natl Acad Sci USA 95, 12111-12116.
Yu, T. W., Shen, Y., Doi-Katayama, Y., Tang, L., Park, C., Moore, B. S., Hutchinson, C. R. & Floss, H. G. (1999). Direct evidence that the rifamycin polyketide synthase assembles polyketide chains processively. Proc Natl Acad Sci USA 96, 9051-9056.
Zielinski, J., Borowy-Borowski, H., Golik, J., Gumieniak, J., Ziminski, T., Kolodzi-Ejczyk, P., Pawlak, J. & Borowski, E. (1979). The structure of levorin A2 and candicidin D. Tetrahedron Lett 20, 1791-1794.
Received 27 June 2001;
revised 18 September 2001;
accepted 26 September 2001.