Introduction

Streptomyces coelicolor is the soil-dwelling filamentous bacterium with a complex developmental life cycle. The life cycle begins with a spore, and it germinates and gives rise to the vegetative mycelium at presence of proper conditions and nutrients (Flärdh and Buttner 2009). In response to nutrient depletion, the vegetative mycelium grows from the colony surface into the air to form an aerial mycelium along with the onset of secondary metabolism (Chater 2011; Flärdh and Buttner 2009). Then, the aerial mycelium differentiates into chains of spores, which is an important event of morphological differentiation and tightly regulated by a complex regulation system such as whi genes (Chater 2011; McCormick and Flärdh 2012). The genome of S. coelicolor contains 29 secondary metabolite biosynthetic gene clusters. Four of them, responsible for the biosynthesis of the blue-pigmented polyketide antibiotic actinorhodin (ACT) (Malpartida and Hopwood 1984), the red oligopyrrole prodiginine antibiotics (RED) (Cerdeño et al. 2001), the lipopeptide calcium-dependent antibiotics (CDA) (Hojati et al. 2002), and the yellow-pigmented type I polyketide (yCPK) (Gottelt et al. 2010), have been characterized. The regulation of secondary metabolism is a complex network (Liu et al. 2013), and the primary metabolites are the source of precursors for secondary metabolic biosynthesis. The onset of morphological differentiation and secondary metabolic biosynthesis is also affected by the limitation of phosphate (Sola-Landa et al. 2003; Martín 2004), nitrogen sources (Hesketh et al. 2002; Tiffert et al. 2008), glucose, or other easily utilized carbon sources (Borodina et al. 2008). The acetyl-CoA which is the initiator of tricarboxylic acid cycle is also used as the precursor of ACT biosynthesis (Ryu et al. 2006).

Myo-inositol is widely distributed in freshwater and soil; it can be used as carbon and energy source by many bacterial species (Bzymek et al. 2007; Turner et al. 2002). Myo-inositol catabolic pathway has been identified in many different bacteria, and the catabolic genes involved in the pathway are extensively conserved (Boutte et al. 2008; Yebra et al. 2007; Yoshida et al. 1997, 2008). These genes are generally clustered in bacterial genome, and usually a regulatory gene is situated in the cluster. In Bacillus subtilis, the myo-inositol catabolism is repressed by a DeoR family protein which binds to each of the iol and iolRS promoter regions (Yoshida et al. 1999). In Sinorhizobium meliloti, the regulator of myo-inositol catabolism belongs to the RpiR family, and it binds to the upstream regions of idhA, iolY, iolR, and iolC (Kohler et al. 2011). Both proteins recognize a conserved palindromic sequence, and the repression can be released by a myo-inositol catabolic intermediate (Kohler et al. 2010, 2011; Yoshida et al. 1999, 2008).

Myo-inositol is the precursor of phosphatidylinositol which is the crucial component of cellular membrane in Streptomyces (Chouayekh et al. 2007; Hoischen et al. 1997; Michell 2008). Inhibition of myo-inositol biosynthesis results in myo-inositol-dependent growth and development in S. coelicolor (Zhang et al. 2012). The mutant can grow at low concentration of myo-inositol, but cannot complete the morphological differentiation and is arrested at intermediate stages (Zhang et al. 2012). Thus, it is possible that the accumulation of myo-inositol and its derivatives is required for the sporulation septation of aerial hyphae in S. coelicolor.

Besides its biosynthesis, myo-inositol catabolism also affects the accumulation of myo-inositol in cells. However, there is no report about the myo-inositol catabolism and its regulation in Streptomyces as we know. In this study, the myo-inositol catabolic gene cluster was predicted in S. coelicolor based on the extensive studies of myo-inositol catabolism in other bacteria (Boutte et al. 2008; Yebra et al. 2007; Yoshida et al. 1997, 2008) and further verified by gene disruption. A cluster-situated regulatory gene SCO6974 was identified and found to be required for the normal sporulation of S. coelicolor. Further studies demonstrated that the regulator SCO6974 affected Streptomyces morphological differentiation by controlling the myo-inositol catabolism.

Materials and methods

Bacterial strains, plasmids, and growth conditions

Bacterial strains and plasmids used in this study are shown in Table S1. Mannitol soya flour (MS) medium or minimal medium (MM) supplemented with mannitol or myo-inositol as carbon source was used for the growth and sporulation of S. coelicolor M145 (SCP1− SCP2−) and its derivatives, RNA preparation, as well as for the intergeneric conjugation of S. coelicolor (Kieser et al. 2000). R2YE medium was used for the production of actinorhodin (ACT) and undecylprodigiosin (RED). Difco Nutrient Agar (DNA) medium supplemented with 0.5 % NaCl (w/v) was used for the production of calcium-dependent antibiotics (CDA) (Kieser et al. 2000). Yeast extract-malt extract (YEME) liquid medium was used for the growth of S. coelicolor (Kieser et al. 2000). Staphylococcus aureus was used as the indicator of CDA production (Anderson et al. 2001). For detecting myo-inositol catabolism in S. coelicolor, the modified minimal medium (3.8 mM (NH4)2SO4, 2.9 mM K2HPO4, 0.8 mM MgSO4, 0.036 mM FeSO4, and 27 mM myo-inositol) was used. Escherichia coli strains were cultured at 37 °C in Luria–Bertani (LB) medium supplemented with antibiotics (100 μg ml−1 for ampicillin, 100 μg ml−1 for kanamycin, 100 μg ml−1 for apramycin, 100 μg ml−1 for spectinomycin, 25 μg ml−1 for chloramphenicol, and 15 μg ml−1 for tetracycline) when necessary for propagating plasmids.

DNA manipulation, sequencing, and conjugation

Chromosomal DNA and plasmids were isolated from Streptomyces or E. coli according to the standard techniques (Kieser et al. 2000; Sambrook et al. 1989). DNA sequencing was performed by Biosune Company (Beijing, China). Database searching and sequence analysis were performed using the online program PSI-BLAST (Altschul et al. 1997). Plasmids or cosmids were firstly introduced by transformation into the methylation-deficient E. coli ET12567 (pUZ8002) and then transferred to S. coelicolor by intergeneric conjugation as described previously (Kieser et al. 2000).

Primers and PCR

All primers used in this study are listed in Table S2. The PCRs were carried out using EasyTaq DNA polymerase (TransGen Biotech), EasyPfu DNA polymerase (TransGen Biotech), KOD FX (TOYOBO), or KOD-Plus- (TOYOBO). An initial denaturation at 94 °C for 5 min was followed by 30 cycles of amplification (94 °C for 30 s, 60 °C for 30 s, 72 °C for 1 min), and additional 10 min at 72 °C (or 68 °C for KOD FX and KOD-plus-). Considering different DNA templates and primers, the annealing temperature and the elongation time were changed in some cases.

Construction of the recombinant strains

Disruptions of the myo-inositol catabolic genes were based on the PCR targeting method with some modifications (Datsenko and Wanner 2000; Gust et al. 2003, 2004). A cosmid library of S. coelicolor M145 genomic DNA was constructed using SuperCos 1 vector kit (Stratagene) according to its protocol. Three cosmids containing the putative myo-inositol catabolic gene cluster SCO2726, SCO6255, and SCO6974-6985 were isolated by screening the cosmid library with primers 2726YF1/2726YR1, 6255YF1/6255YR1, and 6984YF1/6984YR1, respectively. The plasmids pIJ773 and pIJ778 were used as the templates for PCR amplification of the cassettes aac(3)IV-oriT and aadA-oriT which were used to disrupt the target myo-inositol catabolic genes in these three cosmids. The amplified cassette was introduced into E. coli BW25113/pIJ790 carrying the target cosmids with corresponding myo-inositol catabolic genes by electroporation. The recombinants were obtained through the spectinomycin/apramycin resistance screening. The resulting mutagenized cosmids were transferred into E. coli ET12567/pUZ8002 and then introduced into S. coelicolor M145 by conjugal transformation. Subsequently, the corresponding myo-inositol catabolic gene disruption mutants were selected from exconjugants by spectinomycin (100 μg ml−1) or apramycin (50 μg ml−1) resistance and confirmed by PCR using primers listed in Table S2.

To construct the SCO7254 disruption mutant, a DNA fragment corresponding to the upstream region of SCO7254 was amplified by using primers 7254UF/UR. Then, the fragment was inserted into the HindIII site of pBluescript KS+::Kan r, giving pD1. The other DNA fragment corresponding to the downstream region of SCO7254 was amplified by using primers 7254DF/DR and ligated into the XbaI site of pD1 to give pD2. The DNA fragment was isolated with digesting the pD2 at HindIII-XbaI sites, and then it was inserted into the corresponding sites of pKC1132, giving pD7254. Subsequently, pD7254 was introduced into E. coli ET12567/pUZ8002 and then it was transferred to S. coelicolor M145 by intergeneric conjugation. After growing on MS agar supplemented with kanamycin (10 μg ml−1) for 6 days at 28 °C, the colonies were transferred to MM medium with mannitol as the carbon source containing apramycin (10 μg ml−1) or kanamycin (10 μg ml−1). The kanamycin-resistant and apramycin-sensitive strains were selected and were confirmed by PCR using primers 7254YF1/YR1 and 7254YF2/YR2.

To construct the complemented strains of SCO2726, SCO6975, and SCO6984, these genes with their own promoters were amplified from genomic DNA using the corresponding primers 2726HF/HR, 6975HF/HR, and 6984HF/HR. Then, the amplified DNA fragments were digested with XbaI/EcoRI and inserted into the same sites of pSET152 (Kieser et al. 2000). The resulting complemented plasmids were introduced into the corresponding disruption mutants by conjugation to generate the complemented strains. To construct the complemented and overexpressed strain of SCO6974, the entire SCO6974 was amplified using the corresponding primers 6974HF/HR and inserted into the XbaI site of pSET152::P rrnF (Pan et al. 2013). The resulting plasmid was introduced into the SCO6974 disruption mutant and S. coelicolor M145 by conjugation.

RNA isolation and real-time RT-PCR

RNA was isolated from S. coelicolor M145 and the SCO6974 disruption mutant (SCO6974DM) grown in MS or MM medium supplemented with or without myo-inositol for 24, 48, 72, 96, and 120 h as described previously (Liu et al. 2005). RNA was treated with DNase I (Promega) to remove the contaminated DNA, and then it was reverse-transcribed to complementary DNA by using PrimeScript™ RT reagent kit (TaKaRa).

Real-time PCR was carried out in realplex2 Master cycler (Eppendorf, Germany) using Ultra SYBR Mixture (CWBIO) with primers listed in Table S2. The conditions are used as follows: 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s, and 60 °C for 1 min. The hrdB was used as an internal control. The relative transcriptional levels of tested genes were normalized to hrdB and determined by using the 2-∆∆CT method (Livak and Schmittgen 2001). The values were presented as fold change in comparison with the relative expression levels for each gene at the first test time point in the wild-type strain. Data are presented as the averages of three independent experiments conducted in triplicate.

Heterologous expression and purification of SCO6974

The SCO6974 coding region was amplified from genomic DNA using primers 6974OF and 6974OR. The amplified DNA fragment was digested with NdeI/XhoI and ligated into the corresponding sites of pET28a (Novagen) to generate the expressed plasmid pET28a:: SCO6974. After confirmed by DNA sequencing, the plasmid was introduced into E. coli C43 (Miroux and Walker 1996) for gene expression. E. coli C43 carrying pET28a:: SCO6974 was cultured at 37 °C in 200 ml LB with 100 μg ml−1 kanamycin to an OD600 of 0.6. Isopropyl β-d-thiogalactopyranoside (IPTG) was added to a final concentration of 1 mM, and the cultures were incubated for additional 3 h at 37 °C. After centrifugation (6000×g, 5 min, 4 °C), the cells were collected and washed with binding buffer (20 mM Tris–HCl, 500 mM NaCl, 5 mM imidazole, pH 7.9), and re-suspended in 20 ml of the same buffer. After sonication on ice, the cell suspension was centrifugated (12,000×g, 20 min, 4 °C) and the supernatant was recovered. After the supernatant was treated with Ni-NTA agarose chromatography (Novagen), the purified His6-SCO6974-His6 was separated from the whole-cell lysate. The concentration of the purified proteins was determined using BCA protein assay reagent (Novagen). Protein purity was determined by Coomassie blue staining after SDS-polyacrylamide gel electrophoresis (PAGE) on a 10 % polyacrylamide gel. The purified protein was stored with 5 % glycerol at −70 °C until used in the subsequent experiments.

Electrophoretic mobility shift assays

The electrophoretic mobility shift assays (EMSAs) were performed as described previously (Pan et al. 2009). DNA probes containing the promoter regions of SCO2727, SCO6978-SCO6979, and SCO6985 were generated by PCR using the primers listed in the Table S2. The probe of hrdB was obtained by PCR using primers hrdBPF/hrdBPR and was used as negative control. EMSAs were carried out using DIG Gel Shift Kit, 2nd Generation (Roche) according to the manufacturer’s instruction with some modifications. A mixture of the DIG-ddUTP labeled oligonucleotide probes and varying quantities of His6-SCO6974-His6 (0, 40, 80, 200, 400, and 2000 nM protein) was incubated at 25 °C for 20 min in the reaction buffer which contained 1 μg of poly-(dI-dC) (Sigma), 20 mM Tris–HCl (pH 7.5), 1 mM dithiothreitol (DTT), 10 mM MgCl2, 54 mM KCl, 0.5 mg ml−1 calf BSA, and 5 % glycerol in a total volume of 20 μl. Then, the mixture was separated on a 5 % native polyacrylamide gel with a running buffer containing 45 mM Tris–HCl (pH 8.0), 45 mM boric acid, and 1 mM EDTA at 150 V for 1 h. After electrophoresis, the DNA-protein complexes were blotted onto a nylon+ membrane by electroblotting, fixed by UV light, then blocked and washed. Finally, the shifted oligonucleotide was detected by anti-digoxigenin-AP conjugate and the chemiluminescent substrate CSPD® provided in the kit.

DNase I footprinting assays with fluorescence-labeled primers

DNase I footprinting assays were performed as described previously (Zianni et al. 2006). The forward primers were labeled with 6-carboxy fluorescein and the reverse primers were labeled with Hexachloro Fluorescein. The promoter regions of SCO2727, SCO6978-SCO6979, and SCO6985 were amplified by primers used in EMSAs with labeling fluorescein. These probes were purified by using a Cycle-pure kit (Omega) and quantified with a NanoDrop 2000 (Thermo, USA). For this assay, 100 ng probes were incubated individually with varying quantities of His6-SCO6974-His6 in the same buffer as used in EMSAs in a total volume of 50 μl. After incubation for 20 min at 25 °C, 1 μl of the solution containing 0.08 U of DNase I (Promega) and 5.5 μl of 10 × DNase I reaction buffer [400 mM Tris–HCl (pH 8.0), 100 mM MgSO4, 10 mM CaCl2] were added, followed further incubation at 37 °C for 1 min. The reaction was stopped by adding 50 μl DNase I stop solution (20 mM EGTA, pH 8.0). Samples were extracted with phenol-chloroform and precipitated with ethanol, and the pellets were dissolved in 10 μl Milli-Q water. Then, the samples were sequenced by a company (Genolab, China), and the results were analyzed on GeneMarker V1.80 (http://www.softgenetics.com).

Alterations of the SCO6974-binding sequence

To evaluate the importance of the identified SCO6974 binding sites, the conserved sequence (A/T)TGT(A/C)N(G/T)(G/T) ACA(A/T) at the promoter region of SCO2727, SCO6978-SCO6979, and SCO6985 were changed into (A/T)TAG(A/C)N(G/T) (G/T) GC T(A/T) using Easy Mutagenesis System (TransGen Biotech) with primers listed in Table S2. The binding capacity of His6-SCO6974-His6 to the mutagenized probes was measured by EMSAs as described above.

Mutational analysis of the SCO6978 binding sites of SCO6974

The wild-type and the mutagenized promoter regions of SCO6978 were obtained using primers listed in Table S2. Then, the fragments were digested with EcoRI/BamHI, and inserted into the same sites of pSET152::xylE (Pan et al. 2009). The resulting plasmids were introduced into S. coelicolor M145 and the SCO6974 disruption mutant by conjugation. These strains were cultured in MM medium supplemented with myo-inositol as carbon source for 24, 48, 72, 96, and 120 h. The activity of catechol dioxygenase in these strains was detected described previously (Pan et al. 2013).

Determination of antibiotic production

In order to assess the production of ACT and RED, 1 × 107 spores of the various strains were plated on cellophane discs laid on the surface of R2YE solid medium, and incubated at 28 °C for 24, 48, 72, 96, 120, 144, and 168 h. Then, it was evaluated as described previously (Kieser et al. 2000). CDA production in DNA medium was assayed as described previously (Anderson et al. 2001).

Microscopy

The experiments of scanning electron microscopy were performed exactly as described previously (Pan et al. 2009).

Results

Identification of myo-inositol catabolic genes in S. coelicolor

Myo-inositol biosynthesis is important for the growth and morphological differentiation of S. coelicolor (Zhang et al. 2012), but little is known about its catabolism in Streptomyces. In search of the S. coelicolor genomic database, a whole set of putative myo-inositol catabolic genes were predicted based on the extensive studies of myo-inositol catabolic pathways in other bacteria (Boutte et al. 2008; Yebra et al. 2007; Yoshida et al. 1997, 2008). With the exception of SCO2726 which encodes the putative methylmalonate semialdehyde dehydrogenase (homolog of IolA), the myo-inositol catabolic genes are situated in a single cluster (Fig. 1a). Three transcription units (SCO6978-6974, SCO6979-6983, and SCO6985-6984) were predicted in this cluster using an operon prediction program (www.microbesonline.org). Based on the prediction, SCO6984 encodes a myo-inositol dehydrogenase (homolog of IdhA), SCO6982 encodes a 2-keto-myo-inositol (2KMI) dehydratase (homolog of IolE), SCO6975 encodes a 3-d-(3,5/4)-trihydroxycyclohexane-1,2-dione (THcHDO) hydrolase (homolog of IolD), SCO6976 encodes a 5-deoxy-d-glucuronic acid (5DG) isomerase (homolog of IolB), SCO6978 encodes a 2-deoxy-5-keto-d-gluconic acid (DKG) kinase (homolog of IolC), SCO6979-6981 encodes three sugar transport proteins, SCO2727 and SCO6977 encode the hypothetical proteins, SCO6983 encodes a (4-hydroxy-3-nitro) benzylated polystyrene (PhnB)-like protein, and SCO6985 encodes a phytanoyl-CoA dioxygenase. Besides the predicted myo-inositol catabolic genes, SCO6255 and SCO7254 also encode the paralogues of myo-inositol dehydrogenase. Thus, myo-inositol dehydrogenase encoded by SCO6255, SCO6984, and SCO7254 may catalyze the first step of myo-inositol catabolism to form 2KMI. SCO6982 is responsible for the second step to give THcHDO which will be cleaved in the C2–C3 bond by SCO6975 to yield 5DG. 5DG will be further catalyzed by SCO6976, SCO6978, and an unknown protein to form dihydroxyacetone phosphate (DHAP) and malonate semialdehyde (MSA) which will be degraded by SCO2726 at the last step of myo-inositol catabolic pathway (Fig. 1b).

Fig. 1
figure 1

Organization of the myo-inositol catabolic genes and the putative catabolic pathway of myo-inositol in S. coelicolor. a The myo-inositol catabolic genes are clustered and their coding regions are indicated by arrows. b The putative catabolic pathway of myo-inositol in S. coelicolor. Functional assignments of the myo-inositol catabolic genes are based on the biochemical work in B. subtilis (Yoshida et al. 2008)

Full size image

The catabolic genes are required for the bacterial growth on myo-inositol as a single carbon source

To confirm that these genes were responsible for myo-inositol catabolism in S. coelicolor, the following genes SCO2726, SCO6255, SCO6975, SCO6976, SCO6978, SCO6982, SCO6984, and SCO7254 were disrupted and the mutants were confirmed by PCR (Fig. S1). When myo-inositol was used as the sole carbon source in the modified minimal medium as described in “Materials and methods” section, the disruption mutants of SCO6975, SCO6976, SCO6978, SCO6982, and SCO6984 did not grow (Table 1), whereas the wild-type strain (WT) grew well. In the control experiment, all the strains grew well when glucose was used as the sole carbon source. Unexpectedly, the SCO2726 disruption mutant still grew although slower than WT. However, the disruption mutants of SCO6255 and SCO7254 still grew well on myo-inositol as a single carbon source, suggesting that SCO6255 and SCO7254 are not essential for the myo-inositol catabolism in S. coelicolor. After complemented by the corresponding genes, the disruption mutant of SCO2726, SCO6975, and SCO6984 restored to the normal growth on myo-inositol as a single carbon source (Table 1). Taken together, the gene cluster containing SCO6975, SCO6976, SCO6978, SCO6982, and SCO6984 is the main myo-inositol catabolic gene cluster in S. coelicolor.

Table 1 The catabolic genes are required for the bacterial growth on myo-inositol (MI) as the single carbon source
Full size table

Transcription of myo-inositol catabolic genes in S. coelicolor

Reverse transcriptase (RT) PCR confirmed that the myo-inositol catabolic gene cluster contains three transcription units which have been predicated by the operon prediction program (www.microbesonline.org) (Fig. S2). After addition of myo-inositol in the minimal medium with mannitol as the carbon source, the transcriptional level of SCO2726, SCO6975, SCO6976, SCO6978, SCO6979, SCO6980, SCO6981, SCO6982, and SCO6984 were remarkably increased, especially the transcription of SCO6984 increased 140-fold at the early stage (24 h) of growth (Fig. 2). However, the transcriptions of SCO6255 and SCO7254 (paralogues of SCO6984) did not show significant changes upon myo-inositol addition (Fig. 2). Considering that disruption of SCO6255 or SCO7254 did not affect the bacterial growth when myo-inositol was used as the single carbon source, our results clearly indicated that SCO6984 is the only enzyme responsible for the first step of the myo-inositol catabolism.

Fig. 2
figure 2

Transcriptions of the catabolic genes are induced by myo-inositol. The transcriptional levels of myo-inositol catabolic genes were detected in WT by real-time RT-PCR after grown for 24, 48, 72, 96, and 120 h in MM medium with mannitol as the carbon source. All real-time RT-PCR data were normalized according to the abundance of hrdB in each sample. The open columns represent the gene transcriptions without induction and the solid columns represent the gene transcriptions with induction of 5 mM myo-inositol. Data are presented as the averages of three independent experiments conducted in triplicate

Full size image

SCO6974 is a conserved GntR protein and represses myo-inositol catabolism

A putative regulatory gene SCO6974 was identified in the myo-inositol catabolic gene cluster (Fig. 1). The transcription of SCO6974 in the minimal medium with mannitol as the carbon source almost remained the same level (Fig. 3), but it was clearly induced by myo-inositol (Fig. 2). The deduced protein SCO6974 shows high similarity to the metabolite-responsive GntR family regulators which are widely distributed in prokaryotes (Hoskisson and Rigali 2009). Like most of GntR proteins, SCO6974 contains a DNA-binding domain at its N-terminus and a ligand-binding domain at its C-terminus.

Fig. 3
figure 3

Transcriptional analysis of the myo-inositol catabolic genes in WT and SCO6974DM. The transcriptional levels of myo-inositol catabolic genes were detected in WT or SCO6974DM by real-time RT-PCR after grown for 24, 48, 72, 96, and 120 h in MM medium with myo-inositol as the sole carbon source. All real-time RT-PCR data were normalized according to the abundance of hrdB in each sample. The open columns represent the gene transcriptional levels in WT and the solid columns represent the gene transcriptional levels in SCO6974DM. Data are presented as the averages of three independent experiments conducted in triplicate

Full size image

To investigate its function in myo-inositol catabolism, SCO6974 was disrupted via PCR targeting system and the SCO6974 disruption mutant was confirmed by PCR (Fig. S1). Disruption of SCO6974 resulted in the transcriptional increase of the following myo-inositol catabolic genes: SCO2726, SCO6975, SCO6976, SCO6978, SCO6979, SCO6980, SCO6981, and SCO6982 (Fig. 3), suggesting that SCO6974 is a key repressor of myo-inositol catabolism. However, the transcriptional level of SCO6984 responsible for the first step of myo-inositol catabolism did not increase in SCO6974DM (Fig. 3).

Repression of myo-inositol catabolism by SCO6974 via direct protein-DNA interaction

To further understand the regulatory role of SCO6974 in transcription of the myo-inositol catabolic genes, EMSAs were performed. SCO6974 was overexpressed in E. coli, and the His6-tagged SCO6974 was purified and used for EMSAs. Probes covering the promoter regions of SCO2727, SCO6978-SCO6979, and SCO6985 were used, respectively (Fig. 1a). EMSAs demonstrated that SCO6974 (from 40 to 2000 nM) bound specifically to the SCO2727 promoter region where it formed three stable complexes and SCO6974 bound to the intergenic promoter region of SCO6978-SCO6979 to form four stable complexes and SCO6974 bound specifically to the SCO6985 promoter region where it formed only one stable complex (Fig. 4). Specificity of binding was examined with the addition of 10- or 100-fold excess unlabelled probe, which abolished the binding ability of SCO6974 (200 nM) to the corresponding labeled fragments while addition of 100-fold excess unlabelled probe of hrdB could not abolish the binding ability (Fig. 4). These results revealed that SCO6974 represses myo-inositol catabolism through interaction with the catabolic gene promoter regions directly.

Fig. 4
figure 4

EMSAs for detecting the binding of SCO6974 to the promoter regions of these catabolic genes. a Binding of SCO6974 to the promoter region of SCO2727. b Binding of SCO6974 to the intergenic promoter region between SCO6978 and SCO6979. c Binding of SCO6974 to the promoter region of SCO6985. The probes containing the upstream regions of myo-inositol catabolic genes were incubated with increasing amounts of His6-SCO6974-His6 (lanes 1–6 contain 0, 40, 80, 200, 400, and 2000 nM proteins, respectively). EMSAs of 200 nM His6-SCO6974-His6 with 10- and 100-fold excess of unlabelled specific probe are shown in lanes 7–8 and 100-fold excess of non-specific competitor hrdB are shown in lane 9. The arrows indicate the free probes and the braces show His6-SCO6974-His6-DNA complexes

Full size image

To identify the specific binding sites of SCO6974, DNase I footprinting experiments were performed (Fig. 5). Protected regions were observed upstream of SCO2727, SCO6978-SCO6979, and SCO6985, respectively (Fig. 5). Sequence analysis showed that the SCO6974 binding sites is constituted by a palindromic sequence (A/T)TGT(A/C) N(G/T)(G/T)ACA(A/T). To assess the importance of the identified SCO6974 binding sites, the conserved binding sequence of SCO6974 was mutagenized to (A/T)TAG(A/C)N(G/T)(G/T)GCT(A/T). The base substitution analysis demonstrated that the binding capacity of SCO6974 to the mutagenized probes was completely abolished in comparison with their corresponding wild-type targets (Fig. S3). These results implied that this conserved sequence is important for SCO6974 binding activity. To test whether the SCO6974-binding to the conserved sequence was essential for the transcriptional repression in vivo, the xylE reporter gene, encoding catechol 2,3-dioxygenase, was placed under the control of SCO6978 promoter region with either the wild-type or the mutagenized SCO6974 binding sites. As shown in Fig. S3, the activity of XylE was detected in both strains. The activity of XylE was significantly increased when the SCO6974 binding sequence was mutagenized. The increment of xylE expression under the control of SCO6978 promoter with the mutagenized binding site in M145 was in agreement with that of xylE expression under the control of SCO6978 wild-type promoter in SCO6974DM. Thus, SCO6974 represses the transcription of myo-inositol catabolic genes through binding to the specific sequence in the promoter regions of these genes.

Fig. 5
figure 5

Identification of the SCO6974 binding sites by DNase I footprinting assays in the promoter regions of SCO2727 (a), SCO6978-6979 (b), and SCO6985 (c). The binding sequence of SCO6974 is underlined. The control reaction is shown in the upper electropherogram in each panel

Full size image

Disruption of SCO6974 reduces the sporulation of S. coelicolor

Myo-inositol is the precursor of phosphatidylinositol, so its biosynthesis is required for the normal sporulation septation of aerial hyphae in Streptomyces (Zhang et al. 2012). Disruption of SCO6974 significantly enhanced the transcription of myo-inositol catabolic genes likely leading to a reduction of the intracellular pool of myo-inositol that is necessary for sporulation septation. Indeed in MS medium, SCO6974DM grew normally, but produced few spores in comparison with WT and the complemented strain (SCO6974CM) (Fig. 6). However, SCO6974DM was almost restored to normal sporulation septation when myo-inositol was added to the MS medium (Fig. 6), demonstrating that the sporulation defect in SCO6974DM was due to a shortage of myo-inositol. SCO6974 is thus likely to regulate Streptomyces spore formation via its role in the control of the myo-inositol content of the cells.

Fig. 6
figure 6

Disruption of SCO6974 impaired sporulation and the deficiency of sporulation in SCO6974DM was restored by addition of myo-inositol. a Morphological phenotype of the WT, SCO6974DM, SCO6974CM, and SCO6974OE in MS medium with or without addition of 40 mM myo-inositol for 72 h growth. b Scanning electron micrographs of the colony surfaces of WT, SCO6974DM, and SCO6974CM in MS medium, and SCO6974DM in MS medium with addition of 40 mM myo-inositol for 72 h growth

Full size image

Disruption of SCO6974 enhances antibiotic production in S. coelicolor

As the morphological differentiation is closely associated with secondary metabolism and some regulators of morphological differentiation also affect antibiotic production in Streptomyces (Flärdh and Buttner 2009), the productions of ACT, RED, and CDA were determined in SCO6974DM. In contrast with the deficiency of sporulation septation in MS, the ACT production of SCO6974DM was significantly increased in R2YE medium compared with that of WT or SCO6974CM (Fig. 7a). SCO6974DM also produced more CDA in DNA medium than WT or SCO6974CM (Fig. 7b). However, there was no significant difference in the production of RED between SCO6974DM and WT (data not shown). It is possible that the enhanced myo-inositol catabolism yields DHAP and acetyl-CoA that are indirect or direct precursors of the overproduced antibiotics.

Fig. 7
figure 7

Disruption of SCO6974 stimulated ACT and CDA production in S. coelicolor. a ACT production of WT, SCO6974DM, and SCO6974CM in R2YE medium. Cell cultures at each time point were treated with KOH (final concentration, 1 N) and the OD640 corresponding to 1 g of mycelium was determined. b CDA production of WT, SCO6974DM, and SCO6974CM after grown for 48 and 72 h in DNA medium. S. aureus was used as the indicator strain in the presence or absence of Ca(NO3)2

Full size image

Discussion

In S. coelicolor, the biosynthesis of myo-inositol was recently found to be essential for the growth and morphological differentiation as the precursor of phosphatidylinositol which is the crucial component of cell membrane (Zhang et al. 2012). However, a signaling/regulatory role for this molecule cannot be excluded. During sporulation septation of aerial hyphae, large amount of phosphatidylinositol and its derivatives were shown to be required (Zhang et al. 2012). Besides their structural role as constituent of membrane lipids, myo-inositol derivatives might have important signaling/regulatory functions necessitating a tight regulation of the myo-inositol levels. If myo-inositol is in limiting amount, the sporulation septation will be impaired but if myo-inositol excesses bacterial needs, its degradation will be promoted. inoR is the repressor of myo-inositol biosynthesis in S. coelicolor, and the defect of inoR increases the amount of myo-inositol in vivo (Zhang et al. 2012). As expected, the transcriptions of myo-inositol catabolic genes were remarkably increased in the inoR mutant (Fig. S4). SCO6974 may play a key role in the maintenance of myo-inositol balance during the morphological differentiation of S. coelicolor since it negatively controls the expression of most genes of the catabolic operons identified.

The transcriptions of genes in the cluster of myo-inositol catabolism were increased after the addition of myo-inositol (Fig. 2), and SCO6974 repressed the transcriptions of most of the genes in the cluster (Fig. 3). Therefore, a regulatory mechanism might function in the presence of myo-inositol. Myo-inositol catabolism maintains a basal level in S. coelicolor due to its capable myo-inositol biosynthesis. While adding myo-inositol in medium, excess myo-inositol dissociates the repressor SCO6974 from the promoter regions of myo-inositol catabolic genes indirectly, so the expressions of myo-inositol catabolic genes and SCO6974 itself are increased which lead to the enhancing myo-inositol catabolism and consumption of myo-inositol. When myo-inositol is depleted, excess SCO6974 binds to the promoters of myo-inositol catabolic genes again and myo-inositol catabolism returns to basal level.

Our results indicate that SCO6984, rather than SCO6255 and SCO7254, is the real myo-inositol dehydrogenase catalyzing the first step of myo-inositol catabolism. However, its expression, that is strongly inducible by myo-inositol, is not under the control of the GntR regulator SCO6974 that was shown to control negatively the expression of most of the other myo-inositol catabolic genes. This suggests that SCO6984 expression is also controlled by other regulators or the association of SCO6974 and other proteins. Considering SCO6984 is responsible for the first step of myo-inositol catabolism, it is easy to understand that expression of SCO6984 is highly controlled to keep the myo-inositol balance in Streptomyces. Our results demonstrated that the transcriptions of SCO6255 and SCO7254 were not induced by addition of myo-inositol in medium, and disruption of SCO6255 and SCO7254 did not inhibit myo-inositol catabolism in S. coelicolor. It will be interesting to know the function of SCO6255 and SCO7254 in future work. In B. subtilis, three inositol dehydrogenases (IolG, IolX, and IolW) have been found to catalyze different inositol stereoisomers (myo-, d-chiro-, and scyllo-inositols). IolG shows the catalytic activity to myo-inositol and d-chiro-inositol, IolX catalyzes scyllo-inositol, and IolW converts 2KMI to scyllo-inositol with NADPH oxidation (Morinaga et al. 2010). Combining the evidence that SCO6255, SCO6984, and SCO7254 are scattered in the different region of S. coelicolor genome, it is very possible that SCO6255, SCO6984, and SCO7254 play different roles in bacterial growth and differentiation, and they may have specific activity to the different inositol isomers.

The deduced protein SCO6974 shows high similarity to GntR family proteins. The metabolite-responsive GntR family proteins are widely distributed in bacteria and regulate various biological processes (Hoskisson and Rigali 2009). Among 57 putative GntR family regulators which respond to nutritional and/or environmental signals in S. coelicolor, DasR is well-known for its critical function in antibiotic production and morphological differentiation (Rigali et al. 2006). Like disruption of dasR resulted in the deficiency of aerial mycelium and spore formation, disruption of SCO6974 significantly reduced the sporulation septation of S. coelicolor. We examined the transcription of some developmental genes (whiA, ftsZ, whiB, crgA, ssgA, sigF, whiJ, whiD, and whiI); however, except ssgA, other genes remained the same level between WT and SCO6974DM (Fig. S5). Addition of inositol did not affect the transcription of these genes (Fig. S5). It indicated that the few spores are produced by SCO6974DM mainly because of the shortage of myo-inositol.

Based on our results, disruption of SCO6974 enhanced the productions of ACT and CDA, suggesting that myo-inositol catabolism facilitates antibiotic production. Although myo-inositol is the structural basis for a number of secondary messengers in eukaryotic cells, it has not been proved the same in bacteria. As presence of many stereo- and regio-isomers of the inositol moiety and few of them has been studied, it could not exclude the possibility of inositol and its derivatives as signaling molecules during antibiotic production in S. coelicolor. As myo-inositol is a carbohydrate, its catabolites could also influence antibiotic production through changing bacterial physiological development. As the production of myo-inositol catabolite acetyl-CoA is also the precursor of ACT, enhancement of myo-inositol catabolism will produce more acetyl-CoA and in turn increase ACT production. Disruption of SCO6974 also increased CDA production; it is speculated that myo-inositol metabolites or its derivatives stimulate CDA production either through the regulation or the precursors’ accumulation.

In conclusion, the cluster of myo-inositol catabolism is identified in S. coelicolor, and SCO6974 directly regulates the catabolism of myo-inositol. SCO6974 also plays diverse roles in spore formation and the biosynthesis of antibiotics of S. coelicolor.