A pathway-specific transcriptional regulatory gene for nikkomycin biosynthesis in Streptomyces ansochromogenes that also influences colony development

Authors

  • Gang Liu,

    1. State Key Laboratory of Microbial Resources, Institute of Microbiology, Chinese Academy of Science, Beijing 100080, China.
    Search for more papers by this author
  • Yuqing Tian,

    1. State Key Laboratory of Microbial Resources, Institute of Microbiology, Chinese Academy of Science, Beijing 100080, China.
    Search for more papers by this author
  • Haihua Yang,

    1. State Key Laboratory of Microbial Resources, Institute of Microbiology, Chinese Academy of Science, Beijing 100080, China.
    Search for more papers by this author
  • Huarong Tan

    Corresponding author
    1. State Key Laboratory of Microbial Resources, Institute of Microbiology, Chinese Academy of Science, Beijing 100080, China.
      E-mail tanhr@sun.im.ac.cn; Tel. (+86) 106 255 8248; Fax (+86) 106 265 4083.
    Search for more papers by this author

E-mail tanhr@sun.im.ac.cn; Tel. (+86) 106 255 8248; Fax (+86) 106 265 4083.

Summary

DNA sequence analysis of a 7.5 kb XhoI DNA fragment from the region flanking the nikkomycin biosynthesis gene cluster in Streptomyces ansochromogenes revealed one 3.3 kb open reading frame (ORF), designated sanG. The deduced product of sanG (1061 amino acids), which is similar to PimR of Streptomyces natalensis, contains an OmpR-like DNA binding domain in its N-terminal portion and A- and B-type nucleotide binding motifs in the middle of the protein. Disruption of sanG abolished nikkomycin biosynthesis, reduced sporulation and led to brown pigment accumulation. All aspects of this complex phenotype were complemented by a single copy sanG which was integrated into the chromosome. The introduction of multiple copies of sanG resulted in increased nikkomycin production. S1 mapping results indicated that sanG is transcribed from at least three promoters (P1, P2 and P3), P1 being strongly upregulated when production of nikkomycins starts. Two putative transcription units for nikkomycin biosynthesis, starting from sanN and sanO, are dependent on the expression of sanG, whereas a putative transcription unit starting from sanF was not regulated by sanG. These results suggested that sanG encodes a transcriptional activator important for nikkomycin biosynthesis that, unusually, also has pleiotropic effects on secondary metabolism and development.

Introduction

Streptomycetes are filamentous soil bacteria with a complex life cycle. Spores germinate to form a substrate mycelium (vegetative growth) that goes on to develop aerial hyphae, the tips of which form chains of spores. The production of a variety of secondary metabolites is considered to be closely coordinated with this morphological differentiation (Chater, 1993). In liquid medium, antibiotic production in Streptomyces is generally dependent on the growth phase, and involves the expression of clustered biosynthetic genes. Pathway-specific regulatory genes are typically located in these gene clusters (Martin and Liras, 1989). Expression of the pathway-specific regulatory genes is influenced by unlinked pleiotropic regulatory genes, some of which also control morphological differentiation (Ueda et al., 2000).

Nikkomycins are a group of peptidyl nucleoside antibiotics that are structurally similar to the chitin synthase substrate UDP-N-acetylglucosamine (Fiedler et al., 1982). These antibiotics show potent activity against phytopathogenic fungi and against human pathogens (Hector et al., 1990). Nikkomycins are produced by both Streptomyces ansochromogenes (Chen et al., 2000) and Streptomyces tendae (Brillinger, 1979). Their biosynthesis begins with the formation of the nucleoside moieties and the peptidyl moiety (aminohexuronic acid moiety) (Isono and Suzuki, 1979; Bormann et al., 1989). The biosynthetic precursors of the nucleoside moieties of nikkomycin X and Z are ribose and histidine, or ribose and uracil with phosphoenolpyruvate (Isono et al., 1978; Schuz et al., 1992). The biosynthesis of the peptidyl moiety is thought to be similar to that of the related polyoxins, and begins with l-lysine (Bruntner and Bormann, 1998). Molecular analysis of the gene cluster revealed that more than 20 genes are involved in nikkomycin biosynthesis in S. ansochromogenes (Chen et al., 2000; Li et al., 2000; Zeng et al., 2002). However, no report has been published so far on the regulation of nikkomycin biosynthesis. A better understanding of the molecular regulation mechanisms is of utmost interest in both academic research and industrial applications, as it will contribute insights into the fundamental issue of temporal regulation of differentiation and secondary metabolism in Streptomyces and the construction of overproducing strains.

In this study, physical and functional analysis of a 7.5 kb DNA fragment located downstream of sanV (Li and Tan, 2003) revealed one antibiotic pathway-specific regulatory gene in Streptomyces.

Results

The sanG gene product resembles positive regulators involved in antibiotic biosynthesis

A 2.0 kb XhoI-BamHI fragment containing sanV gene was used as a probe to identify a recombinant cosmid, COS19, which contains 18 kb of the nikkomycin biosynthesis gene cluster plus about 10 kb downstream of sanV. A 7.5 kb XhoI DNA fragment containing sanV and the downstream DNA was inserted into the XhoI site of pBluescript KS+ to generate pGL101 (Fig. 1).

Figure 1.

Organization of the gene cluster for nikkomycin biosynthesis. The solid arrow shows sanG and its orientation. The dashed arrows indicate the different mRNAs transcribed by specific transcription units. The DNA fragment containing most of sanG and its promoter region was replaced by a kanamycin resistance gene (aph II) in GLD020. The 0.9 kb BamHI fragment in sanG was replaced by aph II in GLD030. A 7.3 kb XhoI-EcoRI DNA fragment was inserted into the same sites of pKC1139 to generate pKCG, and a 5.1 kb BamHI-EcoRI DNA fragment was inserted into the same sites of pSET152 to generate pSEG.

The cloned fragment was sequenced. It included one large open reading frame (ORF) designated sanG (Fig. 1). sanG contained 3186 nucleotides and started with an ATG codon at position 3171 of the sequenced fragment, preceded immediately upstream by a possible ribosome-binding site. sanG was separated from the nearest upstream gene (a diverging gene, orf3) by about 2 kb of apparently non-coding DNA. The overall G + C content of the ORF was about 73.4%, which is typical in the genes of Streptomyces (Bibb et al., 1984). The deduced SanG protein consists of 1061 amino acids and has a relative mass of 116 kDa.

In searches of databases, the deduced protein (SanG) of sanG showed end-to-end similarity to several proteins: 89% identity amino acids with the deduced product of orfR (nucleotide sequence accession number AJ250878), annotated as a putative nikkomycin regulatory gene from S. tendae; 32% identity with PimR of Streptomyces natalensis (Anton et al., 2004) and 33% identity with PteR of Streptomyces avermitilis (Ikeda et al., 2003). PimR is the positive regulator for pimaricin biosynthesis and PteR is a putative regulatory protein involved in the pentane filipin biosynthesis. Like PimR, SanG is composed of two portions (Fig. 2A). Notably, the N-terminal portion of SanG showed significant sequence similarity to so-called Streptomyces antibiotic regulatory proteins (SARPs) (Wietzorrek and Bibb, 1997), containing one trans-Reg-C domain (transcriptional regulatory protein, C terminal) and one BTAD domain (the bacterial transcriptional activator domain). The trans-Reg-C domain resembles the helix–turn–helix DNA binding domain at the C-terminus of the Escherichia coli activator OmpR (Martinez-Hackert and Stock, 1997). Thus, this region of SanG showed 32% identity to ActII-ORF4 of Streptomyces coelicolor (Fernandez-Moreno et al., 1991), 31% identity to DnrI of Streptomyces peucetius (Madduri and Hutchinson, 1995) and 31% identity to RedD of S. coelicolor (Narva and Feitelson, 1990) (Fig. 2B). The C-terminal portion of SanG showed the characteristics of ATPases associated with diverse cellular activities, and contains an ATP/GTP-binding domain with Walker A and B motifs (Fig. 2C). It showed high similarity to several regulators of the LuxR-family, including a putative transcriptional regulator PteR (32% identity) in S. avermitilis (Ikeda et al., 2003).

Figure 2.

Domain structure and amino acid alignment of parts of the SanG protein.
A. The predicted domain structure of SanG. The N-terminal (trans-Reg-C) region resembles the DNA binding domain of OmpR; BATD, bacterial transcriptional activator domain; Ad_Cycl_assoc, Adenylate cyclase-associated domain.
B. Alignment of the N-terminal portion of the SanG, ORFR, PimR, PteR, ActII-ORF4, DnrI and RedD proteins. ORFR, the deduced product of orfR (AJ250878) from S. tendae; PimR, Pimaricin biosynthesis regulator from S. natalensis; PteR, a putative regulatory protein from S. avermitilis; ActII-ORF4, a transcriptional activator for actinorhodin biosynthesis of S. coelicolor; DnrI, a transcriptional activator for daunorubicin biosynthesis of S. peucetius; RedD, a transcriptional activator for undecylprodigiosin biosynthesis of S. coelicolor. The arrowheads over the alignment indicate the residues encoded by the rare TTA codon in Streptomyces.
C. Comparison of the Walker A and B motifs of SanG with those of ORFR, PimR and PteR. Amino acid residues that are identical are shown in black, and similar residues are shaded.

Interestingly, two leucines in the N-terminal portion of SanG were encoded by the rare TTA codon. TTA codons are also found in some other SARP genes including ActII-ORF4 (Fernandez-Moreno et al., 1991). This suggests that the translation of sanG is controlled by bldA (Leskiw et al., 1991; White and Bibb, 1997), which determines the tRNA for this codon.

Deletion mutants of sanG have lost the ability to produce nikkomycin

In order to identify the function of sanG, disruption mutants were constructed via homologous recombination. sanG was completely replaced in the S. ansochromogenes chromosome by a kanamycin resistance gene (aph II) as described in Experimental procedures. Seven individual sanG disruption mutants were selected randomly and confirmed by restriction digestion and Southern hybridization. To assess nikkomycin production, duplicate cultures from the same time-course experiments were subjected to bioassay against Alternaria longipes. Culture filtrates from the wild-type strain after 24 h incubation showed clear inhibition zones, whereas no inhibition was shown by culture filtrates of the sanG deletion mutant (Fig. 3). High-performance liquid chromatography (HPLC) analysis revealed no peaks of nikkomycin X and Z in culture filtrates of sanG deletion mutants (Fig. 4), in contrast to the culture filtrates of the wild-type strain, even though the biomasses were similar. These results suggested that sanG plays an important role in nikkomycin biosynthesis.

Figure 3.

Effect of sanG disruption on production of nikkomycin. Nikkomycin bioassay of fermentation filtrates with different incubation times and different strains. (A) S. ansochromogenes 7100; (B) sanG disruption mutant GLD020; (C) sanG complementation strain; (D) sanG truncation mutant GLD030. Alternaria longipes was used as the indicator strain.

Figure 4.

Fermentation products from specifically disrupted mutants and the wild-type strain. HPLC analysis of culture filtrates from (A) wild-type strain, (B) sanG deletion mutant GLD020, and (C) sanG truncation mutant GLD030. X, nikkomycin X; Z, nikkomycin Z.

Removal of the 3′-terminus of sanG abolishes nikkomycin production

Many SARPs lack extended C-terminal regions such as that found in SanG. To evaluate the importance of this region, a strain was constructed that contained a 3′-truncated copy of sanG. Thus, by homologous recombination, a 0.9 kb BamHI internal fragment in the 3′-terminus of sanG gene was replaced by the aph II gene as described in Experimental procedures. Two individual sanG disruption mutants were selected randomly and confirmed by Southern blotting, and one (GLD030) was selected for further study. The growth of GLD030 was similar to that of the wild-type strain. When supernatants from SP medium cultures at different times during growth were analysed for nikkomycin production by bioassay (Fig. 3) or quantified by comparing peak areas obtained following HPLC analysis (Fig. 4), the wild-type strain showed production of nikkomycin X and Z, but the truncation mutant GLD030 did not even after 168 h (data not shown), indicating that the 3′-terminus is essential for the function of sanG.

sanG complementation restores nikkomycin production to a sanG disruption mutant

All seven sanG replacement mutants and both truncation mutants were nikkomycin-deficient, and there are no downstream genes potentially co-transcribed with sanG that could be subject to polar effects, making it highly probable that SanG deficiency was responsible for the mutant phenotypes. To confirm that the disruption of sanG was responsible for the abolition of nikkomycin production, a 5.1 kb DNA fragment containing sanG and its promoters was reintroduced into the sanG disruption mutant GLD020 on pSET152, a vector that integrates at the φC31 prophage attachment site in the chromosome. The recombinant strain restored nikkomycin production in the SP medium, albeit at a very low level (Fig. 3). These results showed that sanG is essential for nikkomycin biosynthesis, and suggest that the location of sanG in the chromosome may affect the function of sanG significantly.

Pleiotropic effects of sanG on colonial morphology

Mutants of the GLD020 deleted for sanG formed colonies that grew at the same rate as the wild-type strain and formed aerial mycelium at the same time. However, the mutant colonies showed little of the grey pigmentation which was associated with sporulation in wild-type strain, even after prolonged incubation for 7–10 days on MM solid medium with either mannitol or glucose as the carbon source. The reduction of their spore numbers were observed by using phase-contrast microscopy. Meanwhile, they produced a large number of diffusible brown pigment in contrast to the wild-type strain, which produced a small amount of brown pigment. As a control, a sanN disruption mutant constructed exactly in the same way displayed the phenotype of the wild-type strain (H. Tan, unpublished). In complementation experiments, the cloned fragment containing sanG and its promoter region rescued the morphological deficiency in these mutants and also resulted in loss of the brown pigment (Fig. 5). These results suggested that sanG has pleiotropic functions for both nikkomycin biosynthesis and morphological differentiation in S. ansochromogenes.

Figure 5.

Effect of sanG disruption on morphological differentiation. (A) S. ansochromogenes 7100; (B) sanG disruption mutant GLD020; (C) sanG complementation strain. Note that the phenotype of GLD020 mutant is much lighter than the wild-type and the complementation strain, and shows more diffusible brown pigment on the reverse side of the plate.

It is interesting that 3′-truncation of sanG did not change the morphology or pigmentation of S. ansochromogenes (data not shown), even though this mutation eliminated nikkomycin production. This indicated that sporulation and pigmentation were only affected by the N-terminal region of SanG.

Increasing the copy number of sanG results in increased nikkomycin production

In many species of Streptomyces, antibiotic biosynthesis is precisely controlled by regulatory proteins, especially by transcriptional activators. Overexpression of these transcriptional activators is often associated with a concomitant increase in titres of the corresponding antibiotics. When pKCG (containing sanG on the multicopy vector pKC1139) was introduced into the wild-type strain, nikkomycin production was increased, even though the biomass of the transformants was similar to that of the wild-type strain (Fig. 6). This result reinforced the evidence that sanG is an important activator gene for nikkomycin production.

Figure 6.

Growth and nikkomycins production of S. ansochromogenes containing sanG on a multicopy plasmid. Growth curves: (◆) wild-type strain; (▪) wild-type strain with pKC1139; (▴) transformants with pKCG. Nikkomycin production: (◊) wild-type strain; (□) wild-type strain with pKC1139; (▵) transformants with pKCG.

Three transcription start points are located unusually far upstream of sanG

To determine the transcription start point of sanG, S1 nuclease protection assay was performed as described in Experimental procedures. Initial attempts with probes covering the region close to the coding region revealed only full-length protection making it necessary to investigate regions further upstream. In the end, three signals were detected (Fig. 7A). The transcription start points of the corresponding promoters sanG-P1, sanG-P2 and sanG-P3, were localized at the nucleotides C (1016 base position), A (1064 base position) and G (1182 base position) in the middle of the long non-coding region separating sanG from orf3 and about 1 kb upstream of the sanG translation start codon (Fig. 7B).

Figure 7.

Transcriptional analysis of sanG.
A. High resolution S1 nuclease mapping of sanG. Three sanG transcripts P1, P2 and P3 were detected at 15, 18, 24 and 48 h in the wild-type strain. The arrowheads indicate the transcription start points. The S. ansochromogenes principal sigma factor gene (hrdB-l) (AY628703) was used as an internal control.
B. The nucleotide sequence covering the promoter region of sanG is shown, together with its N-terminal amino acid sequence.

The temporal regulation of these three signals was not the same, sanG-P3 transcript being rather constant throughout the time-course, while sanG-P1 was strongly induced when the strain started to produce nikkomycin. sanG-P2 had a similar profile to that of sanG-P1, but was much weaker (Fig. 7A). After the start of nikkomycin production, the transcripts of sanG from P1 and P2 declined significantly. As a control, the transcript of the hrdB-like gene (nucleotide sequence accession number AY628703) expected to encode the principal sigma factor of S. ansochromogenes was essentially constant during the time-course (Fig. 7A).

SanG as a transcriptional activator for nikkomycin biosynthetic genes

As the deduced product of sanG showed similarities to pathway-specific regulators of the SARP family, SanG was expected to act as a transcriptional activator of the nikkomycin biosynthetic genes. We therefore evaluated the effects of sanG deletion on promoters of the three putative transcription units encoding the enzymes of nikkomycin biosynthesis including the sanO, sanN (Wang et al., 2003) and sanF promoters (G. Liu, unpublished). S1 mapping experiments were carried out according to the method of Kieser et al. (2000).

Transcripts from the divergently oriented promoters in the sanN–sanO intergenic region (Lauer et al., 2001; Wang et al., 2003) were abolished in the sanG mutant, but surprisingly, the sanF transcript was found in both the wild-type strain and sanG disruption mutants (Fig. 8). These results confirmed the regulatory importance of SanG, but also showed that there are differences in the transcriptional requirements of different transcription units in the san gene cluster.

Figure 8.

Transcriptions of sanN, sanO and sanF genes in S. ansochromogenes wild-type strain (wt), sanG disruption mutants GLD020 and GLD030. S. ansochromogenes hrdB-l was used as an internal control. RNAs were extracted from strains for 48 h incubation in SP medium.

Discussion

So far as we know, the nikkomycin biosynthesis genes are located in three putative transcription units in which sanN, sanO and sanF are the respective first genes. Pathway-specific regulatory genes are usually clustered with the antibiotic biosynthetic genes. In this study, a large pathway-specific regulatory gene, sanG, was found next to the nikkomycin biosynthetic genes. The deduced product of sanG showed end-to-end similarity to the pimaricin biosynthetic regulator PimR in S. natalensis and the putative biosynthetic regulator PteR involved in filipin biosynthesis in S. avermitilis. Pimaricin and filipin belong to the polyene antibiotics, but nikkomycins are peptidyl nucleoside antibiotics. Although their structures and biosynthesis pathways are different, all of them can be used against fungi. As the regulators showed high similarities, the regulation mechanisms may be similar. It seems plausible that these regulators might have been acquired in the process of evolution against fungi. Very few is known about the regulation of pimaricin and filipin biosynthesis. In this paper, we showed that SanG regulates nikkomycin production by controlling the transcription of the sanO and sanN operons, but interestingly, does not seem to be needed for expression of the sanF operon.

Translation of sanG may be controlled by the bldA gene

In S. coelicolor, mutation of bldA, which encodes the tRNA for the rare leucine codon UUA, causes pleiotropic deficiencies in both morphological differentiation and antibiotic production on some media. bldA mutants are generally unable to translate UUA codon-containing mRNA. A TTA codon of the adpA gene is the principal target through which bldA influences morphological differentiation of S. coelicolor (Takano et al., 2003); and a TTA codon has been proved to be involved in translational control of actII-ORF4, the pathway-specific regulatory gene for actinorhodin biosynthesis in S. coelicolor (Fernandez-Moreno et al., 1991). Genes controlling undecylprodigiosin biosynthesis and methylenomycin biosynthesis also contain TTA codons (White and Bibb, 1997; S. O’Rourke, unpublished). Phenotypically similar bldA mutants have also been found in S. griseus and may be widespread in actinomycetes (Leskiw et al., 1991). We also found that some bald mutants of S. ansochromogenes have lost nikkomycin production (G. Liu, unpublished). The presence of two rare TTA codons in sanG suggested that they may be targets through which bldA influences nikkomycin biosynthesis. However, the TTA-containing ccaR gene for clavulanic acid and cephamycin C biosynthesis can be translated efficiently in a bldA mutant of Streptomyces clavuligerus (Trepanier et al., 2002). Thus, the translational regulation by bldA is complex in Streptomyces.

Transcription of sanG is growth phase-regulated

Antibiotic production is generally dependent on growth phase. Nikkomycin was first detected after S. ansochromogenes was cultured in SP medium for 24 h. In S. tendae, transcriptions of nikA-G and nikP1-V genes reached the maximal level at stationary phase after 25 h incubation (Bruntner et al., 1999; Lauer et al., 2001). Our data showed that sanG was regulated at the level of transcription during S. ansochromogenes growth in SP medium and sanG-P1 exhibited the strongest signal for 24 h incubation and then reduced quickly with further incubation, indicating that transcription of sanG gene is growth phase-regulated, but what kinds of factors triggered metabolic changes are still unknown.

sanN and sanO genes are controlled by SanG

The N-terminal portion amino acid sequence of the SanG has a high similarity with proteins of SARP family, such as DnrI (Madduri and Hutchinson, 1995), ActII-ORF4 (Fernandez-Moreno et al., 1991) and RedD (Narva and Feitelson, 1990). DnrI binds to promoters of the daunorubicin biosynthesis genes (Tang et al., 1996) and ActII-ORF4 target sites were also found in the act cluster (Arias et al., 1999). It is reasonable to predict that they may maintain the precise control of antibiotic biosynthesis by a similar mechanism of DNA binding and transcriptional activation. As transcriptions of sanN and sanO genes depending on the expression of sanG, their promoter regions may contain binding sites for SanG. It has been recognized that proteins of SARP family contain an OmpR-like DNA binding domain which binds to direct repeat sequences within the regulatory region of target genes (Sheldon et al., 2002). As expected, an intergenic region between the sanN and the sanO genes carries the divergently oriented promoters and contains two intact sets of tandem five-base pair repeat sequences of the kind proposed specific region (Wang et al., 2003) for DNA recognition and binding by SARPs (Fig. 9). Meanwhile, the promoter region of sanF gene which transcribes independently does not contain the similar tandem repeat sequence. So the tandem repeat sequences in sanN–sanO intergenic region are candidates for such binding sites by SanG.

Figure 9.

The intergenic region between the sanN-I and sanO-V operons within the nikkomycin biosynthesis gene cluster. The arrowheads indicate the transcription start points of sanN and sanO. The boxes showed tandem five-base pair repeat sequences preceding the promoters of sanN and sanO.

How might SanG activity be regulated?

Analysis of the nucleotide sequence of sanG promoter region revealed an A-T rich region (5′-GACAATATCCAC ATTTGTTCTGTTTTGTT-3′), which are partially palindromic with A and T rows at ends and share several highly conserved residues of the experimentally verified autoregulatory elements (AREs), including scbR (Takano et al., 2001), ccaR (Pérez-Llarena et al., 1997), spbR (Folcher et al., 2001) and sabR (Li et al., 2003). In Streptomyces,γ-butyrolactone autoregulators and the corresponding receptors play a crucial role in controlling the production of antibiotics and the morphological differentiation (Chater and Horinouchi, 2003). In Streptomyces pristinaespiralis, the pristinamycin biosynthetic pathway-specific regulator PapR1 was regulated by a butyrolactone receptor SpbR. In our previous studies, SabR, a γ-Butyrolactone receptor homologous gene, was identified in S. ansochromogenes (Li et al., 2003), and nikkomycin biosynthesis was delayed conditionally in the sabR disruption mutant. These suggested that sabR positively controls nikkomycin production. Thus, sanG promoter region may be potentially the ultimate target for DNA binding by the SabR protein.

A developmental role for SanG

Remarkably, the complete elimination of SanG resulted in a markedly reduced efficiency of sporulation, and in the appearance of an uncharacterized brown pigment (Fig. 5). These aspects of the mutant phenotype were complemented by the introduction of a wild-type copy of sanG, and are therefore verified to be causally associated with the mutant locus. We are not aware of any other example in the literature of such a pleiotropic effect of a pathway-specific regulator of antibiotic biosynthesis, though there are numerous examples of mutations in genes unlinked to pathway gene sets that have effects on both secondary metabolism and development (notably, most of the so-called bld mutants: Chater, 2001). This effect is unlikely to be caused indirectly through the switching off of nikkomycin production, because the loss of nikkomycin production through a mutation in a biosynthetic gene sanN (which encodes acetyl dehydrogenases involved in nikkomycin biosynthesis) did not have the same pleiotropic effects. Moreover, the deletion of the C-terminal domain of SanG eliminated nikkomycin biosynthesis, but did not have any associated pigmentation or morphoplogical phenotype. Thus, it seems that the region including the SARP domain influences multiple distinct characteristics, whereas the C-terminal domain, which contains a nucleotide binding motif, is needed only for the regulation of nikkomycin biosynthesis. The interesting fact that the pathway-specific regulators of the biosynthetic pathways of two other antifungal antibiotics (but in those cases, polyene compounds) have end-to-end similarity to SanG raises the question of whether all three producing organisms use such regulators to permit them to respond to particular environmental circumstances, associated with fungal competitors, by both producing the antifungal antibiotics and by sporulating efficiently.

Experimental procedures

Strains, plasmids and growth conditions

Streptomyces ansochromogenes 7100 (wild-type strain), a nikkomycin producer, was grown at 28°C and also as a host strain for gene propagation and gene disruption. Sporulation was achieved on minimal medium (MM) using mannitol as sole carbon source. For nikkomycin production, SP medium (3% mannitol, 1% soluble starch, 0.75% yeast extract, and 0.5% soy peptone, pH 6.0) was used as described previously (Zeng et al., 2002). Streptomyces liquid medium YEME and solid medium R2YE were prepared as described (Hopwood et al., 1985) and used for the regeneration of protoplasts and for the selection of the transformants. For routine subcloning, E. coli DH5α and JM109 were grown at 37°C in Luria–Bertani (LB) medium containing ampicillin or apramycin when necessary for propagating plasmids. E. coli ET12567 (dam dcm hsds) was used to propagate non-methylated DNA when it was to be introduced into S. ansochromogenes (MacNeil et al., 1992). E. coli ET12567 (pUZ8002) was used for conjugal transfer of DNA from E. coli to Streptomyces (Paget et al., 1999).

Plasmids pBluescript KS+ (Stratagene) and pUC19 were used for routine cloning and subcloning experiments. Cosmid, COS19 containing part of the nikkomycin biosynthetic gene cluster was used to isolate sanG (Li et al., 2000). pSET152 (Bierman et al., 1992), which can integrate into the Streptomyces chromosome by site-specific recombination at the bacteriophage φ C31 attachment site (attB) (Kuhstoss and Rao, 1991), was used to introduce a single copy of sanG into S. ansochromogenes. E. coliStreptomyces shuttle vector pKC1139, which contains a Streptomyces temperature-sensitive origin of replication, was used for gene expression and gene disruption, as it can be used efficiently in gene replacement by homologous recombination at the non-permissive temperature (Kieser et al., 2000).

DNA manipulation and sequencing

Plasmid and chromosomal DNA were isolated according to the standard techniques from Streptomyces (Kieser et al., 2000) or E. coli (Sambrook et al., 1989). Preparation of protoplasts and transformation of S. ansochromogenes were performed as described previously (Li et al., 2003). Intergeneric conjugation from E. coli ET12567 (pUZ8002) to S. ansochromogenes was carried out as described previously (Kieser et al., 2000). Southern hybridization was taken place with probes labelled with a digoxigenin DNA labelling kit (Roche Biochemicals).

DNA sequencing was performed by Takara Biotechnology Cooperation (Dalian, China). Database searching and sequence analysis were made using Artemis program (Sanger, UK), FramePlot 2.3 (Ishikawa and Hotta, 1999) and the program psi-blast (Altschul et al., 1997).

Construction of sanG disruption mutants

To construct sanG disruption mutants, a 7.5 kb XhoI DNA fragment containing sanG was cloned into the SalI site of a pUC18 derivate from which the BamHI site had been deleted. The resulting plasmid was digested with BamHI, and the largest fragment containing the vector and the flanking DNA fragment of sanG was recovered after agarose gel electrophoresis and self-ligated to yield pGLD018, it was subsequently digested with BamHI and ligated with the BamHI fragment containing the kanamycin resistance cassette (aph II) from pUC119::KanR to generate pGLD019, in which sanG was replaced by aph II. A 4.3 kb insert of pGLD019 was isolated after HindIII and EcoRI digestion, and ligated into the same sites of pKC1139 to generate pGLD020. The same strategy was used to construct pGLD030, in which an internal 0.9 kb BamHI fragment of sanG was replaced by aph II. Subsequently, pGLD020 and pGLD030 were introduced into S. ansochromogenes, and the transformants were confirmed by plasmid isolation and restriction digestion. The spores of transformant containing pGLD020 or pGLD030 were harvested and spread on agar MM containing kanamycin (Kan). After growing for 4 days at 39°C, colonies were replicated on MM containing apramycin (Apr). The sanG disruption mutants were selected by both apramycin sensitivity (Aprs) and kanamycin resistance (Kanr). Their chromosomal DNAs and wild-type DNA were isolated and digested with XhoI. Southern blotting experiments were performed using a 7.5 kb XhoI fragment containing sanG as a probe. A 7.5 kb hybridizing band was found for the wild-type strain, whereas a 4.3 kb positive signal was found for sanG disruption mutant GLD020. GLD030 was confirmed by the same way except that the chromosomal DNA was digested with XhoI-BamHI. The 3.5 kb, 1.7 kb, 1.4 kb and 0.9 kb hybridizing bands were found for the wild-type strain, whereas only 6.1 kb and 1.4 kb positive signals were found for sanG disruption mutant GLD030 (Fig. 1). This result indicated that a double cross-over event had occurred.

Complementation of specifically disrupted strains

For complementation analysis, the integrative vector pSET152 was used. A 3.3 kb BamHI fragment containing a partial sanG gene and its flanking DNA was isolated from pGL101 and inserted into pSET152 to give pSET152::3.3 kb, which was then digested with Bgl II and EcoRI, and subsequently ligated with a 4.1 kb Bgl II-EcoRI fragment containing a partial sanG gene from pGL101. The resulting recombinant plasmid, pSEG, contained sanG and its flanking DNA sequence together with the apramycin resistance gene, aac3 (IV). Then, pSEG was integrated into the chromosomal attB site of S. ansochromogenes after conjugal transfer from E. coli.

Overexpression of sanG

A 7.5 kb XhoI DNA fragment carrying sanG from pGL101 was subcloned into the SalI site of pUC19, and the resulting plasmid was digested with HindIII and EcoRI to give a 7.3 kb HindIII-EcoRI DNA fragment containing sanG. This was ligated into the same sites of pKC1139 to give pKCG, which was then introduced into S. ansochromogenes 7100.

Transcriptional analysis

To investigate transcription during nikkomycin biosynthesis, total RNA was isolated from strains grown in SP medium for different times. Mycelium was collected, frozen quickly in liquid nitrogen and ground into a fine white powder. Ground mycelial samples were frozen at −20°C until the time-course was completed. RNA was then extracted using the Trizol reagent (Invitrogen, USA) according to the manufacturer's protocol. S1 protection assays were performed using the hrdB-like gene (hrdB-l) (AY628703) probe as a control. The hrdB-l probe was generated by polymerase chain reaction (PCR) using the unlabelled oligonucleotide 5′-GGGTACGCC CCGTCAGTG-3′ and the radiolabelled oligonucleotide 5′-AGCCTTTCCCCGCTCAAT-3′, which was uniquely labelled at its 5′ end with [32p]-ATP using T4 polynucleotide kinase. The probe to detect sanG transcripts was generated by PCR using the radiolabelled oligonucleotide 5′-CTTTCGTCACGG TCTCGGA-3′. The sequence ladder was made using an fmol DNA cycle sequencing kit (Promega, USA) with the same labelled primer. For sanO, the probe was generated by PCR using the radiolabelled oligonucleotide 5′-CGACCTGGGCG GCGAACA-3′ and the unlabelled oligonucleotide 5′-CGACG AGGGACTGGATGC-3′. For sanN, the probe was generated by PCR using the radiolabelled oligonucleotide 5′-CGACG AGGGACTGGATGC-3′ and the unlabelled oligonucleotide 5′-CGACCTGGGCGGCGAACA-3′. For sanF, the probe was amplified using the radiolabelled oligonucleotide 5′-TACTG CTTCTCGTGCTTCGGGT-3′ and the unlabelled oligonucleotide 5′-CGCGCAGGTCGGCCAGGT-3′.

Nikkomycin bioassay and HPLC analysis

Nikkomycins produced by S. ansochromogenes 7100 were measured by a disk agar diffusion method using Alternaria longipes as indicator strain. Nikkomycins in culture filtrates were identified by HPLC analysis. For HPLC analysis, Agilent 1100 HPLC and RP C-18 were used. The detection wavelength was 290 nm, the reference wavelength was 350 nm. Chemical reagent, mobile phase and gradient elution process were as described by Fiedler (1984).

Nucleotide sequence accession number

The nucleotide sequence of sanG determined in this study has been submitted to the GenBank database under accession number AY631852.

Acknowledgements

We thank Prof Keith Chater (John Innes Centre, Norwich, UK) for providing E. coli ET12567 (pUZ8002), plasmids pKC1139 and pSET152, and for helpful discussion during this work and critical reading in preparation of this paper. We would like to thank Dr Brenda Leskiw (University of Alberta, Canada) for the gift of apramycin. This work was supported by grants from the National Basic research Program of China (Grant no. 2003CB114205) and the National Natural Science Foundation of China (Grant nos. 30430010, 30370016 and 30270028).

Ancillary