The nysF gene encoding a putative 4′-phosphopantetheinyl transferase (PPTase) is located at the 5′ border of the nystatin biosynthesis gene cluster in Streptomyces noursei. PPTases carry out post-translational modification of the acyl carrier protein domains on the polyketide synthases (PKS) required for their full functionality, and hence NysF was assumed to be involved in similar modification of the nystatin PKS. At the same time, DNA sequence analysis of the genomic region adjacent to the nysF gene revealed a gene cluster for a putative lantibiotic biosynthesis. This finding created some uncertainty regarding which gene cluster nysF functionally belongs to. To resolve this ambiguity, nysF was inactivated by both insertion of a kanamycin (Km) resistance marker into its coding region, and by in-frame deletion. Surprisingly, the nystatin production in both the nysF::KmR and ΔnysF mutants increased by ca. 60% compared to the wild-type, suggesting a negative role of nysF in the nystatin biosynthesis. The expression of xylE reporter gene under control of different promoters from the nystatin gene cluster in the ΔnysF mutant was studied. The data obtained clearly show enhanced expression of xylE from the promoters of several structural and regulatory genes in the ΔnysF mutant, implying that NysF negatively regulates the nystatin biosynthesis.
Microbial synthesis of fatty acids, non-ribosomal peptides and polyketides requires post-translational modification of the biosynthetic enzymes by means of phosphopantetheinylation . The enzymes responsible for this modification, phosphopantetheinyl transferases (PPTases), catalyse the transfer of a phosphopantetheinyl group (Ppant) from coenzyme A to the acyl carrier proteins (ACP) in fatty acid synthases, polyketide synthases (PKS), and peptidyl carrier proteins (PCP) of non-ribosomal peptide synthetases (NRPS). Ppant functions by both providing a covalent linkage between the biosynthetic intermediate and carrier protein, and facilitating the transport of such intermediates between active sites in the enzyme complexes. All bacterial species sequenced to date contain two or more PPTases, and often the functions of these enzymes and their substrate specificity cannot be deduced from the amino acid sequences. The most conserved and widely distributed group of PPTases is represented by enzymes encoded by a separate gene and sharing high similarity with the Sfp protein of Bacillus subtilis, which is required for production of the peptide antibiotic surfactin by the latter organism . Sfp-type PPTases were also shown to be involved in biosynthesis of such secondary metabolites as siderophore, prodigiosin, and serrawettin [17,24]. Sfp homologues are frequently found in antibiotic-producing actinomycetes, and in some cases requirement of such PPTases for efficient antibiotic biosynthesis has been demonstrated [14,19,21–23].
Despite direct biochemical evidence for the involvement of several PPTases in antibiotic biosynthesis via phosphopantetheinylation of specific ACPs/PCPs [14,22], there remains a possibility that the effect observed upon inactivation of PPTase gene in certain cases might have another explanation. Considering rather broad substrate specificity of Sfp-type PPTases, it is not always clear whether a particular PPTase gene physically linked to the PKS or NRPS gene cluster is responsible for post-translational modification of ACP domains on the PKS/NRPS encoded by this cluster. The effect seen upon inactivation of such PPTase can, for example, be a consequence of involvement of this PPTase in production of a signal molecule which is involved in regulation of antibiotic biosynthesis. Biosynthesis of antibiotics by Streptomyces bacteria is a complex process, which involves many levels of regulation . It has been proposed that a large number of genes involved in this regulation reflect the necessity for the antibiotic producer to react to certain factors, thus tuning the antibiotic biosynthesis according to changing environmental conditions . In several cases it has been shown that small diffusible molecules transmit the signal to the pathway-specific regulators that directly control the expression of antibiotic biosynthetic genes [5,7,12,18]. It is worth noting that signalling molecules such as γ-butyrolactones have an acyl moiety, which is donated by an acyl-ACP in the process of their biosynthesis . Obviously, a PPTase activity is required in order for such ACP to be able to accept an acyl group in the first place, implying an involvement of PPTases in biosynthesis of the above-mentioned signalling molecules.
Streptomyces noursei ATCC 11455 produces polyene macrolide antibiotic nystatin, and the genetics and enzymology of its biosynthesis have been elucidated after cloning of the nystatin biosynthetic gene cluster . Later it has been shown that at least four pathway-specific regulatory genes are controlling nystatin biosynthesis . In this work, we disclose an unexpected role of nysF gene located at the border of the nystatin biosynthetic gene cluster and encoding a putative PPTase in regulation of nystatin biosynthesis in S. noursei ATCC 11455.
2Materials and methods
2.1Bacterial strains, plasmids, and growth conditions
Bacterial strains and plasmids used in this study are listed in Table 1. Some of the plasmids are described below in this section. S. noursei strains were maintained on ISP2 agar medium (Difco, USA), and grown in liquid TSB medium (Oxoid) for DNA isolation. Escherichia coli strains were handled using standard techniques . Conjugation from E. coli ET12567 (pUZ8002) to S. noursei and the gene replacement procedure, and cultivation of S. noursei strains for nystatin production were performed as described previously . Presence of nystatin-related polyene macrolides was assessed using DAD-HPLC, LC–MS and TOF in culture extracts as in .
Table 1. Bacterial strains and plasmids used in this study
General techniques for DNA manipulation were used as described in Sambrook et al. . Isolation of the DNA fragments from agarose gel was done with QIAEX kit (QIAGEN, Germany). Southern blot analysis was performed with DIG High Prime labelling kit (Roche Biochemicals, Germany) according to the manufacturer's manual.
Oligonucleotide primers were purchased from MWG-Biotech AG (Germany). DNA sequencing was performed at QIAGEN (Germany). The nucleotide sequence was submitted to the GenBank under Accession No. AY942707.
2.3Assay for XylE activity
Assay for XylE activity was performed as described by Sekurova et al. . Concentration of the protein was determined by the Bradford method.
2.4Construction of the plasmids for gene inactivation
A 4.78 kb Pst I–Bgl II fragment encompassing nysF and flanking genes was cloned into pGEM3Zf(-) digested with Pst I/Bam HI. The resulting plasmid was digested with Acc III, which cuts in the middle of nysF-coding region, subjected to Klenow fill-in procedure, and ligated with a blunt-ended 1.28 kb fragment from pUC4K (Pharmacia) encoding the KmR marker. A 6.06 kb Eco RI–Hin dIII fragment was isolated from the resulting construct, and ligated with 3.0 kb Eco RI–Hin dIII fragment of pSOK201 , yielding the gene replacement vector pNFK1.
A 1.52 kb DNA fragment designated NFA encompassing the part of nysF and the downstream region was amplified from the pL90X template using primers NFA1 (5′ GGCGAATTCGTGCTGGAGTTCGCGGAGCG 3′) and NFA2 (5′ GACCTGCAGACGGTCAACTCGCCGCTCC 3′). A 1.51 kb DNA fragment NFB containing the upstream region and some of the coding region of nysF was amplified from the plasmid pL90X template using primers NFB1 (5′ GACCTGCAGGTACGCCGCCTCGGTGGC 3′) and NFB2 (5′ GGCAAGCTTCAGACCCTCCAGCAGACC 3′). The NFA and NFB PCR products were digested with Eco RI/Pst I and Pst I/Hin dIII endonucleases, respectively, and ligated together with the 3.0 kb Eco RI–Hin dIII fragment from pSOK201 , yielding the nysF replacement vector pNFD1. Using this vector we obtained nysF in frame deletion mutant NFD7 by double homologous recombination in S. noursei ATCC 11455.
A 1.66 kb Sal I DNA fragment representing the internal part of orf9 was cloned into the corresponding site of pGEM3Zf(-), excised as a Eco RI–Hin dIII fragment, and ligated with the 3.0 kb Eco RI–Hin dIII fragment of pSOK201 . The resulting vector, designated pOND1, was used for disruption of orf9 via single homologous recombination in S. noursei ATCC 11455.
2.5Construction of the plasmids for expression of nysF
A 0.84 kb DNA fragment (NFEp) containing nysF gene together with an upstream part where the promoter region could be located was amplified from the pL90X template using primers NFEp1 (5′ GGACAAGCTTACGAGGCCACCAGCTCCG 3′) and NFEp2 (5′ CCAGGAATTCTCATCCGAAGGTGGGGCG 3′). The NFEp PCR-product was digested with Hin dIII and Eco RI endonucleases and ligated with pSOK804 E. coli–Streptomyces conjugative vector  digested with the same endonucleases, generating pNFE101 vector.
A 0.79 kb DNA fragment containing nysF gene coding region was amplified by PCR using primers NFE1 (5′ GGACTCTAGACTGTTCTTACCGTTCGCCGGAG 3′) and NFE2 (5′ GACAAGCTTGCAGGTTTCATCCGAAGG 3′), digested with endonucleases Xba I, Hin dIII and ligated together with Xba I/Hin dIII digested nysH gene promoter region into the Hin dIII digested plasmid pSOK804 , generating E. coli–Streptomyces conjugative vector pNFE102 for the expression of nysF gene from the nysH gene promoter.
3Results and discussion
3.1Genes presumably involved in lantibiotic biosynthesis are located downstream of nysF in S. noursei
PPTase enzymes are widely distributed among bacteria, playing an essential role in post-translational modification of large enzyme complexes such as PKS and NRPS responsible for the biosynthesis of polyketides and secondary metabolites containing peptide moieties [8,11,20]. Analysis of the nystatin biosynthetic gene cluster of S. noursei revealed the presence of a gene nysF at the 5′ border of the cluster, presumably encoding a PPTase suggested to be involved in post-translational modification of the nystatin PKS . Partial alignment of the NysF amino acid sequence with known PPTases is presented in Fig. 1. NysF apparently contains all three consensus motifs P1, P2 and P3 typical of Sfp-type PPTases, although the highly conserved Gly in P2 is replaced by Ser in NysF. According to classification for PPTases proposed by Lambalot et al. , NysF belongs to the group of Sfp-type enzymes, and thus would most likely be involved in phosphopantetheinylation of apo-ACP on PKS, or apo-PCP on NRPS. Although the most likely target for NysF appeared to be nystatin PKS proteins encoded by the genes located in the vicinity of the nysF gene, other possibilities could not be excluded. Indeed, genes encoding Sfp-type PPTases are rarely associated with polyketide antibiotic biosynthesis gene clusters, while they are frequently found in the gene clusters governing synthesis of non-ribosomally synthesized peptides. The latter consideration has prompted us to sequence a DNA region downstream of nysF to verify whether an NRPS gene cluster might be located there. Analysis of ca 10.8 kb DNA sequence suggested that a gene cluster for biosynthesis of an unknown lantibiotic, a ribosomally synthesized peptide, is located downstream of nysF (Fig. 2). Putative functions for the gene products encoded by 10 orf s identified downstream of nysF are presented in Table 2. Some of the ORFs identified clearly shared similarities with enzymes involved in lantibiotic biosynthesis, such as ORF9, which resembles a lantibiotic dehydratase SpaB. Two small orf s, orf5 and orf6, with overlapping 3′ ends found immediately downstream of nysF, could be candidates for the lantibiotic-encoding genes. orf7, orf8, and orf11 encoding a putative dienelactone hydrolase, signal peptidase, and a secreted peptidase inhibitor, respectively, could also belong to the lantibiotic gene cluster. No genes encoding peptide synthetase-like proteins could be identified among these orf s, and therefore a function of NysF as a PPTase in the synthesis of a putative lantibiotic could not be envisaged.
Table 2. Genes identified downstream of the nysF gene in the S. noursei ATCC 11455 genome
Best database match
44%, S. coelicolor Q9ACQ6
Dienelactone hydrolase homologue
45%, Caulobacter crescentus Q9AC29
Lipoprotein signal peptidase
SPase II family
64%, S. avermitilis Q82AC7
Similar to subtilin synthase SpaB
36%, S. coelicolor Q9RK93
30%, S. coelicolor Q9RCV7
Secretable peptidase inhibitor
Peptidase inhibitor family I36
30%, S. coelicolor Q9L2D6
70%, S. avermitilis Q82JE7
MerR family regulator
75%, S. coelicolor Q9L1K7
PQQP repeats, transmembrane helix
61%, Dictyostelium discoideum Q9GPR3
3.2Inactivation of nysF leads to an increased rate of nystatin production
To resolve the ambiguity in functional assignment for nysF, this gene was inactivated first by gene replacement using plasmid pNFK1 that carried a copy of nysF disrupted with KmR cassette. Surprisingly, the resulting mutant NF23 produced nystatin at an increased rate, which led to ca. 60% enhancement in the antibiotic yield compared to the WT (data not shown). This result did not correlate with the suggested nystatin PKS-specific PPTase function for the NysF protein, as one would expect reduction in the nystatin yield due to the absence of the post-translational modification of the ACP domains in the nystatin PKS.
To exclude possible polar effect of the KmR insertion, we have made an in-frame deletion within the nysF gene, and analysed nystatin production in the resulting mutant. The ΔnysF mutant, designated NFD7, overproduced nystatin in the same manner as the nysF::KmR mutant (data not shown), thus clearly establishing a negative role of nysF in the nystatin biosynthesis. One reasonable explanation for the observed phenomena would be the involvement of NysF in biosynthesis of a signalling molecule which negatively regulates nystatin biosynthesis.
pSOK804-based integrative vectors were then constructed for complementation of the NFD7 mutant. These plasmids contained a copy of nysF under its own presumed promoter (pNFE101) and under nysH p promoter (pNFE102). While introduction of pNFE101 had no effect on nystatin biosynthesis by NFD7, the NFD7 (pNFE102) strain exhibited reduced nystatin yield compared to the NFD7 strain (data not shown). These data imply that there is no promoter immediately upstream of the nysF, and this gene is most likely transcribed as a part of polycistronic mRNA, synthesis of which is initiated at the nysH p promoter.
Nystatin biosynthesis by the WT strain, NFD7, as well as WT (pNFE102) and NFD7 (pNFE102) was monitored over 96 h of fermentation in shake-flasks (Fig. 3(a)). Although in all strains nystatin biosynthesis started at ca. 18 h, the rate of antibiotic biosynthesis and its volumetric yield differed significantly. NFD7 produced nystatin at the highest rate, while it was somewhat reduced in the NDF7 complemented with pNFE102. It shall be noted that complementation of the nysF mutation by introducing an additional copy of the gene under control of nysH p promoter, which seems to drive nysF expression along with nysH and nysG ABC transporter genes, was only partially successful (Fig. 3), leading to ca. 15% reduction in the nystatin biosynthesis rate. Most significant difference was apparent between the WT and WT (pNFE102) strains, where the rate of nystatin biosynthesis by the latter was reduced by ca. 30% compared to the WT (Fig. 3(b)). This partial complementation could probably be explained by the fact that nysF gene was expressed in trans. This notion can be supported by the data obtained earlier for complementation of the nysRIV mutant, where only partial restoration of the original phenotype could be observed upon in trans expression of nysRIV. This was despite the fact that nysRIV was expressed from a strong promoter ermE∗p, and this gene's transcription was significantly enhanced (O. Sekurova, unpublished data). Partial restoration of certain phenotypes upon in trans complementation has been observed for other Streptomyces bacteria as well .
The fact that nystatin biosynthesis starts simultaneously in all strains tested suggests that putative signalling molecule is not involved in initiation of antibiotic biosynthesis. Since the difference in antibiotic synthesis rates becomes apparent after ca. 20 h, it is logical to suggest that synthesis of the signalling molecule starts before this time point.
One example indirectly supporting the hypothesis of NysF participation in the biosynthesis of a signalling molecule could be the involvement of a Photorhabdus luminescens PPTase in the production of an unknown molecule essential for growth of its symbiotic host nematode Heterohabditis bacteriophora. It has been noted that such molecule is probably unstable or active at a critical threshold level, since attempts to restore nematode growth by supplementing culture liquors of wild-type Photorhabdus luminescens have failed. If an unknown NysF-dependent signalling molecule in S. noursei has similar features, this could also explain only partial complementation of the ΔnysF mutant.
To make sure that putative lantibiotic presumably synthesized by the enzymes encoded by the genes identified downstream of nysF does not affect nystatin biosynthesis, mutation was introduced into orf9 by means of gene disruption (see Section 2). orf9 gene product is similar to subtilin synthase SpaB, and resembles lantibiotic dehydratases often encoded by lantibiotic gene clusters. It was thus reasonable to expect that orf9 is involved in biosynthesis of a putative lantibiotic. The orf9 disruption mutant, designated S916 exhibited normal levels of nystatin biosynthesis, thus excluding the possibility of involvement of a putative lantibiotic in regulation of nystatin biosynthesis.
3.3Expression from the promoters of nystatin biosynthetic and regulatory genes is differentially affected in the ΔnysF mutant
xylE-based promoter probe vectors have been constructed previously to investigate the targets for regulatory genes and the individual contribution of the latter to the control over the nystatin biosynthesis . The expression of xylE reporter gene under control of different promoters from the nystatin gene cluster has been investigated in the NFD7 mutant. Several promoter-probe vectors were introduced into both wild-type S. noursei and ΔnysF mutant. The XylE activity assays revealed that expression from nysA p (initiation of nystatin biosynthesis), nysDI p (glycosylation of the nystatin precursor), and nysRI p (regulation of nystatin biosynthesis) were up to 10-fold enhanced in the mutant (Fig. 4). At the same time, nysF deletion had no such effect on the expression of xylE from the nysI p, nysH p, nysRIV p and nysDIII p promoters. Moreover, in the case of both nysH p, nysRIV p, and nysDIII p promoters a decrease (up to 85% in case of nysRIV p) in XylE activity was observed, suggesting that expression from this promoters might be negatively affected.
In the previous study on nystatin regulatory genes we have shown that NysRIV is a regulator which most probably directly affects the expression from nysA p and nysH p promoters . Interestingly, nysRIV p promoter seems to be down-regulated in the ΔnysF mutant, suggesting that some other regulatory protein(s) responding to the absence of a putative signalling molecule synthesized with the help of NysF is regulating the expression of nystatin genes. This notion is supported by the apparently differential effect of nysF deletion on expression from nysA p and nysH p promoters, both of which were shown to be dependent on NysRIV on the nysF-positive genetic background .
3.4Inactivation of the cryptic regulatory genes orf2 and orf3 in the nystatin gene cluster on ΔnysF background
Two more genes might be involved in the regulation of nystatin biosynthesis process-orf3, located downstream of nysRIV, which encodes a putative transcriptional regulator of the DeoR family, and orf2, encoding a polypeptide similar to the transcriptional regulators of the AsnC type . orf3 and orf2 were previously inactivated in S. noursei, and these mutations had no significant effect on nystatin biosynthesis . At the same time, it seemed possible that the products of these genes can somehow interact with NysF-mediated regulatory network, and therefore can be indirectly involved in the regulation process. To investigate this possibility, the nysF deletion was introduced into the strains DNR609 and NRD5 carrying orf2 and orf3 mutations, generating mutants NFD72 (ΔnysF orf2::KmR) and NFD73 (ΔnysFΔorf3). Both mutants exhibited elevated levels of nystatin production similar to NFD7 (data not shown). Therefore, it was concluded that neither orf2 nor orf3 gene products are involved in the NysF-mediated regulation of nystatin biosynthesis in S. noursei.
We thank R. Aune for help with analysis of nystatin production. This work was supported by the Research Council of Norway.