Background

Prokaryotic gene expression is a process that is adjusted to the growth phase and to the changes in environmental conditions. As bacterial gene expression is predominantly regulated at the transcriptional level, bacterial genomes encode numerous proteins that control transcription initiation. Among them, the key players are DNA-binding proteins such as sigma factors, which determine promoter recognition by RNA polymerase (RNAP), as well as other transcription factors (TFs), acting as repressors or activators, which may affect the binding of RNAP to a promoter [6, 39]. However, non-DNA binding proteins such as anti-sigma factors, proteases and other proteins can also control the accessibility of direct regulators to the DNA, thus acting indirectly and playing a critical role in transcriptional regulation. Therefore, the identification of all the components of a regulatory system is a challenging task.

Most often, studies on the regulation of gene expression have been limited to searching for promoters bound and controlled by certain regulatory proteins [8, 31]. To date, several powerful methods for the determination of the DNA-binding sites of known TFs and the genes regulated by them have been developed, such as SELEX, ChIP-chip, ChIP-seq and RNA-seq [5, 30, 33, 45, 78]. All these tools aim to identify all the putative targets of a certain regulatory protein [2]. However, on the other hand, there are only limited strategies to identify TFs of a given promoter of interest [71]. Currently available techniques to identify TFs that bind specific regions include a modified bacterial one-hybrid reporter system [25] and in vitro DNA capture strategies [7, 49, 50, 68]. To search for gene expression regulators in vivo, the combination of random transposon mutagenesis with reporter genes (predominantly lacZ or antibiotic resistance cassettes) was successfully developed. This strategy has been applied in a number of bacterial species (Pseudomonas chlororaphis, Proteus mirabilis, Staphylococcus aureus and Vibrio cholerae) [7, 40,41,42, 67]. An approach based on the combination of random mutant library construction and the lux reporter gene was used for the identification of regulatory proteins of the lecA in Pseudomonas aeruginosa [16] and the acs gene in Escherichia coli [2]. Such approaches may be highly beneficial for the identification of global regulatory factors and the dissection of complex regulatory networks, such as those controlling secondary metabolite synthesis in Streptomyces.

Streptomyces are soil-dwelling bacteria that undergo morphological differentiation, which encompasses vegetative growth and sporulation [20]. They are used as producers of numerous biologically active secondary metabolites, such as antibiotics (approximately 60% of the world’s natural antibiotics are Streptomyces-obtained), immunosuppressants and cytostatics [11]. The pathways for the synthesis of secondary metabolites are encoded by gene clusters that are activated only at specific growth phases or physiological conditions [22, 66]. Thus, the production of secondary metabolites is tightly controlled by complex regulatory systems, many of which remain uncharacterized. In silico predictions revealed that the genome of any Streptomyces species may encode up to 1100 transcriptional regulators [58], a large fraction of which fall into one of the two main clades: cluster situated regulators (CSRs) and pleiotropic/global regulators [32, 38, 44, 79]. CSRs are regulatory proteins (such as ActII-orf4, RedD and CdaR or TetR, LacI, MerR, and LuxR family regulators [12, 36, 44, 48, 74, 76, 79]) that are usually situated in secondary metabolite biosynthetic gene clusters and directly control the expression of their nearby genes, while pleiotropic regulators (e.g., AdpA [77, 82]), AfsR [21, 28], BldD [15], and DasR [56] are scattered throughout the chromosome and positioned distantly from the genes they regulate. While the identification of CSRs is relatively straightforward, the identification of global regulators that control a particular gene of interest may be challenging [27]. A deep understanding of all aspects of Streptomyces gene expression, particularly transcription, is crucial to better exploit these bacteria as producers of widely used compounds.

Notably, in Streptomyces, similar to other studied bacteria (Streptococcus pneumoniae, Haemophilus influenzae, E. coli, Salmonella enterica), DNA supercoiling also plays a role in global gene regulation by directly affecting the transcriptional activity of a number of promoters [18, 19, 23, 53, 59, 65, 73]. In Streptomyces, chromosome supercoiling is a global regulatory factor that controls the transcription of 3–7% of genes [65]. Proper DNA supercoiling in the cell is controlled by a set of enzymes called topoisomerases. The opposing activities of topoisomerase I (TopA), which removes negative supercoils, and gyrase, which can introduce negative supercoils, maintain topological homeostasis in bacterial cells [10]. Inhibition of topoisomerase activity or alteration of their level leads to changes in chromosome topology and affects DNA transactions, including replication and transcription. One of the most important mechanisms that maintains the balance of topoisomerases activity is transcriptional control of their cellular level [46, 69]. In contrast to many model bacterial species, TopA is the only type I topoisomerase in S. coelicolor; thus, it is essential and must be precisely regulated to maintain the proper level of chromosomal supercoiling [64]. TopA depletion in Streptomyces causes severe growth retardation, including increased DNA supercoiling and altered gene expression, including the expression of secondary metabolite genes [17, 62]. As in other bacteria, in S. coelicolor, the TopA level is predominantly regulated by the transcriptional control of the topA gene [1, 19, 64, 69]. In S. coelicolor, transcription of topA is driven from at least two promoters, with equal contributions of both promoters during both vegetative growth and spore production. The p1 promoter was shown to be supercoiling sensitive, which corroborates the shortened distance between motifs – 10 and – 35 [64]. On the other hand, the comparison of the p2 promoter to other known promoter sequences did not identify any known recognition site for sigma factors or other transcriptional regulators. Moreover, apart from transcriptional regulation, no other mechanism regulating TopA activity has been described in Streptomyces.

Here, to identify regulators of the topA promoter in S. coelicolor, we used an approach based on Tn5 minitransposon (mini-Tn5)-driven random transposon mutagenesis combined with the lux reporter system. As a proof of concept, we established that disruption of the sco4804 gene lowers the TopA level, while its overexpression results in enhanced topA and gyrase gene transcription. Thus, our approach allowed us to identify a new component of the chromosome supercoiling-related regulatory network in S. coelicolor.

Results

Application of transposon mutagenesis combined with the lux reporter gene system identifies potential regulators of topA promoter activity

Our previous studies have shown that S. coelicolor topA promoter activity is highly dependent on chromosome supercoiling, but neither negative nor positive protein regulators of the topA promoter have been identified [64]. Previously, to detect topA activity changes, we used a lux reporter plasmid (pFLUXptopA) in which the topA promoter controls the transcription of the luxCDABE reporter genes [64]. The luxCDABE operon encompasses the luxAB genes that encode luciferase (a heterodimer of LuxA and LuxB) and the luxCDE genes that encode enzymes necessary for luciferase substrate (tetradecanal) biosynthesis [14]. Here, to search for unknown regulators of topA promoter activity, we combined lux reporter genes and random transposon mutagenesis.

The S. coelicolor WT-lux strain (pFLUXptopA in the wild-type background) was subjected to random transposon mutagenesis (Fig. 1, stage 1), performed using a transposon plasmid (pHL734) containing the mini-Tn-5 transposon [80]. pHL734 harbours a codon-optimized, highly efficient Tn5 transposase (under the control of the ermE promoter), which inserts mini-Tn-5 transposons randomly along the chromosome. The conjugation was repeated 4 times to deliver a WT-lux-tn library consisting of at least 8300 single colonies (Fig. 1, stage 2). Since transposition with mini-Tn5 occurs once per genome (as pHL734 cannot replicate in Streptomyces), each of the 8300 obtained colonies was assumed to carry a single transposon insertion in the genome. The luminescence of all obtained single colonies was measured during their growth on plates (Fig. 1, stage 3). Next, clones with altered luminescence intensity compared to the paternal WT-lux strain were selected (23 colonies with decreased luminescence and 18 colonies with increased luminescence) and re-streaked on fresh MS plates, and their altered luminescence in comparison to the paternal strain was verified (Fig. 1, stage 4). Subsequently, the clones were cultured in liquid medium, and the luminescence of selected clones was measured again (Fig. 1, stage 5). After the second round of selection, we obtained 12 colonies with significantly lowered or abolished luminescence signals and 2 colonies exhibiting elevated luminescence intensity. The presence of intact lux genes in clones with diminished fluorescence was confirmed by PCR (Additional file 1: Fig. S1). Since modification of the topA promoter activity detected by changes in lux gene activity was also expected to affect the TopA protein level (although earlier we observed that high activity of the topA promoter may not lead to high protein level; [64], we next aimed to estimate the TopA level in 14 transposant clones. To this end, the selected clones were cultivated in liquid medium, and the TopA level in the cell lysate was detected using Western blotting with anti-TopA antibodies and compared to the wild-type strain (Fig. 1, stage 6). Lower TopA protein levels (in comparison to those in the wild type strain) were observed in 2 clones, but we did not observe a significant increase in TopA protein level in any of the mutants with elevated luminescence levels. The mutants with altered TopA protein levels were used for further analysis.

Fig. 1
figure 1

Scheme of random Tn5 transposon mutagenesis in the S. coelicolor WT-lux reporter strain. 1 Random transposon mutagenesis of the WT-lux strain (MG03) with a mini-Tn5 transposon. 2 The WT-lux-tn mutant library consisted of approximately 8300 single colonies obtained on MS agar plates. 3 Measurement of the luminescence of WT-lux-tn library single colonies. 4 Selection of colonies with altered light emission compared to the WT-lux paternal strain and TopA-depleted lux strain (MG04, high activity of topA promoter). 5 The luminescence of selected colonies from the WT-lux-tn library measured in liquid culture compared with the WT-lux strain and TopA-depleted lux strain (high activity of the topA promoter). RU—relative luminescence units. 6 Western blot analysis of TopA protein level in cell lysates of selected colonies from the WT-lux-tn library with anti-TopA polyclonal antibodies. M—molecular mass marker

Full size image

In summary, among the transposon mutants, we detected clones with both increased and decreased reporter gene activity. This indicates that the application of transposon libraries in combination with the lux reporter genes may be used to identify both positive and negative transcriptional regulators.

Transposon mutation leads to modified transcription of the topA gene but does not affect chromosome supercoiling

One of the WT-lux strain transposants, named lux-tn66, emitted very weak luminescence when cultured on solid and in liquid media (Fig. 2a, b). Western blot analyses showed that the level of TopA protein in lux-tn66 was approximately 50% of the protein level in the wild-type strain (Fig. 2c, left), which was also verified by the measurement of topA transcript level (Fig. 2c, right). RT-qPCR showed that the topA transcript level was approximately 70% of the topA transcript level in the wild-type strain. The observed discrepancy between the very low luminescence and moderate lowering of the topA transcript level and protein level could suggest either the complex posttranscriptional regulation of the TopA protein level or diminished luciferase activity in the transposon mutant. The RT-qPCR analysis of luxC transcript level in the mutant strain compared to the wild type strain confirmed the latter (Fig. 2c, right panel), showing the decrease of luxC transcript similar to decrease of topA transcript level and much less profound than reduction of luminescence signal. Thus, we infer, that the extensive drop of luminescence in transposon mutant may be caused by transposition-triggered, limited supply of the cofactors for light production. However, the modified level of the topA transcript and its protein product confirmed the analysis of the reporter gene activity.

Fig. 2
figure 2

Phenotype of the lux-tn66 transposon strain. a Growth and luminescence of the lux-tn66 transposon strain on solid MS medium in comparison to the WT-lux strain (MG03), TopA-depleted lux strain (MG04) and the negative control—wild type strain with empty pFLUXH vector (MG01). Left panel: plate view (after 5 days of growth), right panel: luminescence intensity (after 48 h). b Luminescence of mutant reporter strains after 24 h of growth in liquid 79 medium compared to the WT-lux (MG03) and the negative control—wild-type strain with empty pFLUXH vector (MG01). c Western blot analysis of TopA protein level (left panel) and the relative transcription of the native topA gene, as well as luxC reporter gene, in the mutant lux-tn66 strain determined using RT-qPCR analysis performed on 24-h 79 medium cultures, compared to the WT-lux strain (right panel). d The growth curves of the lux-tn66 strain (79 medium, Bioscreen C, measurements every 20 min) compared to the WT-lux (MG03) and TopA-depleted lux strain (MG04), as well as to MG04 with restored TopA protein level (after induction with 0.5 µg/ml thiostrepton). e Supercoiling density of the reporter plasmids pWHM3Hyg or pWHM3Spec isolated from the transposon mutant lux-tn66 derivative (lux-tn66_RP) strain, the wild-type strain derivative (MS10) and the TopA-depleted (MS11) strain (representative image of two independent experiments). The figure shows topoisomers detected in agarose gel as well as band intensity measurements performed using ImageJ software. f The level of gyrB transcript in lux-tn66 strain determined using RT-qPCR analysis performed on 24-h 79 medium cultures, compared to the WT-lux strain (MG03)

Full size image

Based on our earlier studies, which showed that lowering the TopA protein level slows the growth of S. coelicolor in liquid and solid medium [62], we measured the rate of growth of the lux-tn66 strain. The growth rate analysis in liquid culture showed a slight retardation of transposon strain growth compared to the WT-lux paternal strain, but the growth rate of the transposon strain was still significantly faster than that of the TopA-depleted strain (in which the TopA protein level was approximately 20-fold lower than the wild-type TopA level [62] (Fig. 2d).

Since the severe TopA depletion increases chromosome supercoiling [62], we checked whether decreased TopA protein levels in the transposon lux-tn66 strain caused any changes in global DNA supercoiling. To this end, we determined the level of global DNA supercoiling using modified strains containing the reporter plasmids pWHM3Hyg or pWHM3Spec. We compared the supercoiling of plasmids isolated from modified lux-tn66 (lux-tn66-RP), the wild-type strain derivative (MS10) and the TopA-depleted strain derivative (MS11), which was used as a positive control. DNA supercoiling in transposon mutant lux-tn66-RP was found to be unaffected by decreased TopA protein level (Fig. 2e). Since chromosome supercoiling is maintained by concerted action of gyrase and TopA, we expected that unaltered DNA supercoiling despite the lowered TopA protein level in the lux-tn66 strain may result from changes in gyrase protein level. To test this hypothesis, we determined the activity of the gyrB gene encoding one of the gyrase subunits. RT-qPCR analysis of the gyrB transcript level in the lux-tn66 transposon mutant showed significantly lower transcription level than wild-type gyrB transcription (Fig. 2f). This observation suggests that the lowered gyrase protein level compensated for the undesirable supercoiling changes triggered by the lowered TopA protein level in the lux-tn66 transposon mutant.

Based on our experiments, we infer that the modified level of topA gene transcription may either be linked to inactivation of the direct regulator of topA gene expression or to the indirect regulation of topA involving changes in the gyrase protein level in the mutant strain.

SCO4804 is a potential candidate for a topA promoter activator/regulator

Analysis of the transposon insertion site in the lux-tn66 strain (performed as described by [80] showed that transposition occurred within the sco4804 gene, 85 bp downstream of its predicted start codon (Fig. 3a). The sco4804 gene encodes a hypothetical protein, SCO4804, composed of 815 amino acids (predicted molar mass 86.04 kDa), that is rich in glycine and proline residues, and that is conserved in Streptomyces species. Structural prediction was performed using Robetta software [55] and indicated the presence of putative alpha-helical structures in the central region of the protein and unstructured regions at both the C- and N-termini. Another analysis performed using PredictProtein [81], also showed three possible DNA-binding regions within the SCO4804 protein structure, which indicates that this protein may act as a transcriptional regulator (Additional file 2: Fig. S2). Comparative analysis in the HOGENOM database [52] showed only a few homologues in other bacterial families (such as Alphaproteobacteria, Bacteroidetes and Cyanobacteria), however, no annotated role was provided for the protein in any species.

Fig. 3
figure 3

Genomic localization and supercoiling-dependent transcription of sco4804. a Transposition site in lux-tn66 strain. ME (dark green)—the mosaic end sequence; ori (yellow)—origin of replication from pUC vector for DNA replication in E. coli; apraR (dark violet)—apramycin resistance gene; UpS—primer used for recognition of mini-Tn5 insertion site. The red arrow shows the identified site of the mini-Tn5 insertion. The black arrows at the bottom of the scheme show the distance between neighbouring genes. b The transcription profile of sco4803-sco4808 genes in wild type strain (WT) and TopA- depleted strain (PS04, TopA↓) based on RNA-seq experiment data [65], visualized by IGV Viewer. c RNA-Seq-based analysis of the expression level of sco4803-sco4806in the TopA-depleted (PS04) and control wild-type (M145) strains performed for 18-h YEME/TSB cultures, normalized by the upper quartile [65]. The error bars correspond to standard deviations calculated for two independent biological replicates. d The relative transcription of sco4804 and sco4805 in the TopA-depleted lux (MG04) and WT-lux reporter (MG03) strains calculated using RT-qPCR analysis performed for 24-h cultures in 79 medium

Full size image

Positioning of the sco4804 gene (103 bp and 171 bp of non-coding regions upstream and downstream of the sco4804 gene, respectively) suggests that it may not form an operon with adjacent genes (sco4803 and sco4805); however, its genomic location is conserved within the Streptomyces genus. SCO4803 and SCO4805 are annotated as hypothetical proteins while SCO4806 as secreted protein. Prediction of SCO4803 and SCO4805 structure [81] showed possible DNA binding domains, but also domains responsible for interactions with other proteins (Additional file 2: Fig. S2). RNA-seq experiments performed previously using S. coelicolor wild-type and TopA-depleted strains [65] showed significant, eightfold induction of the sco4804 gene, as well as adjacent genes, under TopA-depleted conditions (Fig. 3b, c). This result was confirmed for sco4804 and sco4805 by RT-qPCR experiments using the WT-lux strain (MG03) and TopA-depleted lux strain (MG04) (Fig. 3d). This strongly suggests that transcription of these genes is supercoiling-dependent, corroborating their potential function in controlling topoisomerase levels.

Markedly, RT-qPCR analysis revealed significant decrease of sco4805 transcription in the lux-tn66 transposon strain suggesting a polar effect of transposition within sco4804 gene. This indicates that either product of one of those two genes is a topA promoter regulator or both their products act in cooperation to fulfill this role.

SCO4804 overproduction increases topA promoter activity

Since sco4804 transcription was shown to be induced by TopA protein depletion and disruption of the sco4804 gene lowered TopA protein level, we predicted that SCO4804 acts as a positive regulator of topA transcription. To confirm this hypothesis, we constructed a strain overexpressing the sco4804 gene in the WT-lux strain background.

In the obtained strain (MG66), sco4804 (as a second gene copy in the integrative vector pIJ6902) was controlled by a thiostrepton-inducible tipA promoter. Overexpression of sco4804 in MG66 cells was confirmed by RT-qPCR, revealing significantly elevated sco4804 transcript levels in comparison to the WT-lux strain background (Fig. 4a). However, we also observed significant induction of sco4805 gene transcription in response to sco4804 overexpression (Additional file 3: Fig. S3). Th observation that SCO4804 controls sco4805 transcription reinforces the notion that both proteins act in cooperation.

Fig. 4
figure 4

Phenotype of the SCO4804 overproducing strain. a RT-qPCR analysis of the sco4804 transcript level in the MG66 strain (non-induced and induced with 10 µg/ml thiostrepton) compared to the WT-lux (MG03) and TopA-depleted lux (MG04) strains performed for 24 h cultures grown in 79 medium. b The growth curves of the non-induced MG66 strain and MG66 induced with 10 µg/ml thiostrepton (79 medium, Bioscreen C, measurements every 20 min) compared to the WTØ strain (M145_pIJ6902). c Measurement of luminescence of the MG66 strain in the absence or in the presence of the sco4804 inducer (0 and 10 μg/ml of thiostrepton) indicating the changes of topA promoter activity, performed in liquid 79 medium after 24 h of growth and compared to the control strain (WTØ-lux, MG03_pIJ6902) that contains empty pIJ6902 plasmid (and TopA- depleted lux (MGO4) strains. d Luminescence of MG66 indicating the activity of the topA promoter controlling the lux reporter genes after 48 h of growth on solid MS agar plates, without and with the inducer (10 µg/ml thiostrepton), as compared to the control strain (WTØ-lux, MG03_pIJ6902) that contains empty pIJ6902 plasmid (and TopA-depleted lux (MGO4) strain, negative control (the wild-type strain with empty pFLUXH vector (MG01)) and positive control (the wild type strain with pFLUXH_permE (MG02)). e RT-qPCR analysis of the relative transcription of the topA gene in the SCO4804 overproducing strain (MG66) cultured in 79 medium for 24 h and induced with 10 µg/ml thiostrepton for 30 min, 60 min or cultured for 24 h in the presence of the inducer. The data were compared to the non-induced control and WT-lux strain (MG03) grown for 24 h in 79 medium. f RT-qPCR analysis of the relative transcription of the gyrB gene in the SCO4804-overproducing strain (MG66) induced after 24 h of growth with 10 µg/ml thiostrepton for 30 min and 60 min and/or cultured for 24 h in the presence of the inducer. The data were compared to the non-induced control and WT-lux strain (MG03) grown for 24 h in 79 medium

Full size image

Having confirmed the induction of sco4804 in the MG66 strain, we set out to analyse the influence of SCO4804 on growth and topA promoter activity, as well as DNA supercoiling. Induction of sco4804 led to slight retardation of growth compared to the control WTØ strain (with pIJ6902 empty vector) (Fig. 4b). The topA promoter activity, measured using the lux reporter genes, was significantly increased by sco4804 induction either in 24-h liquid or 48-h plate cultures of the MG66 strain cultured in the presence of inducer (10 µg/ml thiostrepton) (Fig. 4c, d). We excluded the effect of thiostrepton on the activity of topA promoter and luxCDABE operon in induced sco4804 overexpressing strain by checking the luminescence as well as topA transcription in response to induction in the WTØ-lux strain (Fig. 4c, d and Additional file 4: Fig. S4). Thus, the obtained results indicated that overexpression of sco4804 caused significant activation of the topA promoter.

Next, to confirm that sco4804 is a positive regulator of the topA promoter, we performed RT-qPCR analysis of topA transcript level in the MG66 strain with induced sco4804 overexpression. The results showed that increased topA transcript level was observed immediately after induction of sco4804 (30 min after the addition of thiostrepton to the medium), but after 60 min of incubation in the presence of inducer, topA transcript level decreased to the wild-type level (Fig. 4e). The discrepancy between long-term elevated lux activity after sco4804 induction and topA transcript elevation only during a very short period of time after induction suggests other post-transcriptional regulation modifications of topA transcript level, reinforcing our previous experiments [65]. Interestingly, the level of gyrB transcript exhibited similar changes, with an increase 30 min after sco4804 induction and a decrease 60 min after induction (Fig. 4f). This indicates that the balance between gyrase and TopA activities was established and that the native supercoiling level could be restored. Indeed, the analysis of the reporter plasmid supercoiling showed no changes in the MG66 strain upon sco4804 induction (Additional file 5: Fig. S5).

To check whether SCO4804 is a direct regulator of topA promoter activity, we tested its binding to the topA promoter in vitro. To this end, we purified the 6His-SCO4804 protein using the E. coli BL21 (DE3) groEL-groES strain [24, 51] (Supplementary info, Additional file 6: Fig. S6A) and performed an electrophoretic mobility shift assay (EMSA) using a 458 bp DNA fragment encompassing the topA promoter and 632 bp promoter of the sco4697 gene, as well as 654 bp of a part of the sco3928 gene as the negative controls. While 6His-SCO4804 bound all tested DNA fragments at a concentration of 1 µM, it was non-specific towards the topA promoter (Additional file 6: Fig. S6B). Moreover, the addition of poly(dIdC) competitor DNA eliminated all non-specific interactions. A further pull-down assay and topoisomerase activity tests [63] in the presence of 6His-SCO4804 excluded the possibility of a direct interaction between SCO4804 and TopA (Additional file 7: Figs. S7 and Additional file 8: S8). These experiments suggest that SCO4804 influences topA promoter activity in an indirect manner.

Discussion

Our approach combining lux reporter genes and a random transposon library allowed us to perform high-throughput screening for potential regulatory proteins that control TopA protein level in S. coelicolor. Genome-wide transposition, as a powerful genetic tool, is widely used for systematic genetic studies of different bacterial species, including the construction of random insertion Streptomyces mutants with IS6100, Tn4560, IS493, Tn5, and Himar1 transposons [4, 43, 70, 72, 75]. The Tn5 minitransposon together with the codon-optimized Tn5 transposase displays high efficiency, less codon bias and lower host specificity than other transposases [80]. In Streptomyces, random mutagenesis has previously been used to find repressors or activators of genes of interest, the products of which are easy to monitor within the cell, such as antibiotic or pigment production. This technique has been successfully applied for the identification of actinorhodin and landomycin E negative regulators [13, 29]. Moreover, transposon mutagenesis combined with a reporter system based on an antibiotic resistance cassette was previously applied to search for repressors for daptomycin production in S. roseosporus [40]. However, this approach, based on antibiotic resistance genes, limited the screening to negative transcriptional regulators. The advantage of our, in comparison to the abovementioned approaches, is its suitability for high-throughput searches of both negative and positive regulators in Streptomyces transposon libraries. The strategy described here may be used for challenging identification of global regulators that control secondary metabolite synthesis pathways. Thanks to the application of lux reporter genes fused to promoter(s) of either regulatory or biosynthetic gene(s) from secondary metabolite cluster, identification of its global regulator(s) may not rely on detection of the product of interest but on easily performed luminescence measurements. Moreover, the changes in gene of interest expression can be readily monitored in both liquid and solid cultures over time and in different environmental conditions, which also enables the identification of regulators active only in particular environmental conditions. However, it must be considered that this method is limited to non-essential regulators. The other disadvantage of using lux reporter genes is the formation of artefacts due to metabolic influences on luciferase activity (affected by changes in oxygen, ATP, Mg ions levels), but this can be overcome by using a second reporter system, for example, based on gfp, gusA, lacZ expression [2, 9, 35, 47], which, although not that convenient as lux reporter in terms of measurements, are less dependent on metabolic state. Nevertheless, because of the compatibility of all genetic elements, we believe that the strategy tested here for the identification of regulators may be widely used in Streptomyces.

By applying our screening approach, we expected to find any proteins that influence TopA protein level with either transcriptional or translational/post-translational modes of action. It was shown earlier that the topA promoter was activated by increased chromosome supercoiling and was inhibited due to chromosome relaxation after novobiocin treatment [64]. In addition to supercoiling sensitivity, no other factor controlling promoter activity has been identified to date,thus, the identification of either a topA activator or repressor was of interest. We found that SCO4804 (possibly in cooperation with SCO4805) acts as a topA transcriptional activator, since its elimination decreased topA promoter activity and protein level, while induction of sco4804 (and related to this sco4805 overexpression) resulted in higher activity of the topA promoter. Notably, our previous studies showed that elevated topA transcript level does not correspond with elevated TopA protein level, as well as or with significant changes in DNA supercoiling levels [64]. However, the fact that the topA transcript level increased and subsequently diminished shortly after SCO4804 induction suggests that other mechanisms of maintaining TopA protein level are also activated. We previously suggested that the topA transcript and TopA protein levels are controlled by multiple regulatory strategies that act concertedly to preserve constant supercoiling level in Streptomyces [64, 66]. Additionally, we observed that sco4804 induction influences not only TopA but also gyrase genes expression, indicating that the newly identified protein may be a component of the chromosome supercoiling maintenance system. The fact that transcription of sco4804 is activated in response to increased negative supercoiling corroborates its potential function as the regulator of topoisomerase activity. Moreover, the supercoiling activated transcription of the neighbouring genes (sco4803 and sco4805) suggest that their products may act in a cooperation with SCO4804.

Conclusions

To summarize, our screening approach is optimized for Streptomyces and allows the identification of both positive and negative regulators that control the expression of genes of interest by either direct or indirect mechanisms. As proven by our concept, the protein SCO4804 was found to be a component of a complex regulatory network involved in S. coelicolor chromosome supercoiling maintenance. Since the production of secondary metabolites is regulated by chromosomal topology, understanding complex transcriptional regulation in Streptomyces is crucial for the industrial application of these bacteria.

Methods

Bacterial strains, plasmids, and growth conditions

Basic DNA manipulation procedures were performed according to standard protocols [60]. Unless otherwise stated, all enzymes and isolation kits were obtained from Thermo Fisher Scientific (Waltham, MA) and NEB (Ipswitch, MA). Bacterial media and antibiotics were purchased from Difco Laboratories (Detroit, MI) and Carl Roth (Karlsruhe, Germany), respectively. The S. coelicolor growth conditions and antibiotic concentrations, as well as the conjugation procedure, followed the general protocols described by [34]. For induction, thiostrepton at concentrations of 0.5–10 µg/ml was added. Growth curves of the S. coelicolor strains were determined using the Bioscreen C device (Oy Growth Curves Ab Ltd., Helsinki, Finland). Cultures were grown in triplicate in 79 medium [54] (300 µl/well), inoculated with 0.01 U/ml spores (1 U is defined as the volume of spore stock solution diluted up to 1 ml with OD600nm = 1). The S. coelicolor and E. coli strains used in this study are shown in Table 1. The plasmids used in this study are shown in Additional file 9: Table S1.

Table 1 Strains used in this study
Full size table

Transposon mutagenesis

To perform random transposon mutagenesis in S. coelicolor MG03, we used the pHL734 vector. Mutagenesis was performed according to a procedure described earlier [80]. Briefly, E. coli ET12567 pUZ8002 harbouring the pHL734 plasmid was grown to OD600nm = 0.5 and conjugated with 5 U of S. coelicolor MG03 (WT-lux) spores. The conjugated cell mixture was diluted (10–4–10–6) and plated on MS agar supplemented with 10 mM MgCl2 and 60 mM CaCl2 to obtain single colonies. After 17 h of growth at 30 °C, the plates were overlaid with 20 µl per plate of each antibiotic: nalidixic acid (25 mg/ml), apramycin (50 mg/ml) and hygromycin B (50 mg/ml). The obtained colonies were tested for luminescence intensity (see below), re-streaked on fresh MS agar plates supplemented with apramycin and hygromycin and used to establish liquid cultures. The positions of mini-Tn5 insertions in the S. coelicolor MG03 chromosome were identified using the rescue plasmid method. First, chromosomal DNA was isolated from a 24-h culture in 79 medium of S. coelicolor transposon strains. Subsequently, 2 µg of chromosomal DNA was digested with the ApaI restriction enzyme (at 37 °C overnight, 50 µl total reaction volume), and then DNA was purified using a CleanUp kit (A&A Biotechnology, Gdynia, Poland) and eluted with 15 µl of ultrapure water. Then, 100–200 ng of the ApaI-digested DNA was re-ligated (at 4 °C overnight, 20 µl total volume of the reaction) using 1 µl of T4 DNA ligase (NEB) to allow formation of a mini–E. coli replicative plasmid. Electrocompetent E. coli DH5α cells were transformed with the ligation products (using half of the reaction volume) and selected on apramycin LB agar plates. Plasmid DNA was isolated from single E. coli colonies using a Plasmid Screening Kit (Syngen Biotech, Wrocław, Poland) according to the manufacturer’s instructions. The isolated plasmids were digested with the ApaI restriction enzyme and analysed using gel electrophoresis, and plasmids exhibiting different digestion patterns were picked for subsequent DNA sequencing. Sequencing (Sigma-Aldrich, Saint Louis, MO) using UpS oligonucleotides identified the sites of mini-Tn5 insertion.

Strain construction

For inducible overexpression of the sco4804 gene, the pIJ6902_sco4804 plasmid was constructed. A DNA fragment encompassing the sco4804 gene with flanking EcoRI and NdeI restriction sites was synthesized and cloned into the pUC57 mini plasmid, yielding pUCmini_4804 (GenScript Biotech Corporation, New Jersey, US). The pIJ6902_sco4804 plasmid was obtained by restriction cloning of the sco4804 insert into the pIJ6902 plasmid using NdeI and EcoRI sites. The construct was then conjugated from E. coli ET12567 pUZ8002 into S. coelicolor MG03 (WT-lux strain), apramycin-resistant exconjugants were selected, and the plasmid presence in the obtained strain MG66 was confirmed by PCR using M13pUCr and sco4804_rv oligonucleotides.

To analyse DNA supercoiling in the S. coelicolor transposon mutant lux-tn66 strain, we modified the pWHM3Hyg reporter plasmid [62] by substituting the hygromycin resistance cassette with the spectinomycin resistance gene using the Redirect system [26], oligonucleotides spect_fwd_2 and spect_rv and plasmid pIJ778 as a template, yielding pWHM3Spec. Next, we introduced the pWHM3Spec plasmid into the lux-tn66 transposon mutant strain via conjugation with E. coli ET12567 pUZ8002 [34]. The MG66_RP strain, which was also used for analysis of DNA supercoiling, was obtained by conjugation of the pIJ6902_sco4804 plasmid into the MS10 strain (WT harbouring the pWHM3Hyg reporter plasmid).

As a control for the analysis of the MG66 strain induced with thiostrepton, we also constructed the M145_pIJ6902 and MG03_pIJ6902 strains, in which the empty plasmid pIJ6902 was introduced via conjugation into the M145 and MG03 S. coelicolor strains, respectively.

Reporter gene activity assays

To measure luciferase activity in liquid culture, strains containing the luxCDAEB operon (under the control of the topA promoter or under the control of the erm promoter) in the pFLUXH ΦBT1 integrating vector were grown in liquid 79 medium for 24 h at 30 °C (in three biological replicates for each strain). Subsequently, the mycelium was collected by centrifugation, wet weight was determined, and mycelium was resuspended in 300 µl of 79 medium. Measurement of the luciferase activity was performed in triplicate directly from the mycelium suspension for each biological sample in a 100 µl volume in 96-well microplates (Perkin Elmer, Waltham, MA) using the Infinite PRO Multimode Plate Reader (Tecan, Männedorf, Switzerland). The luminescence intensity was normalized against wet weight (units/100 mg of mycelium). Luminescence visualization on solid medium was performed on MS agar plates after 48 h of growth at 30 °C, and luminescence detection was performed using a ChemiDocXRS + device (Bio-Rad, Hercules, CA).

RNA isolation and RT-qPCR

For gene expression analysis, RNA was isolated from S. coelicolor cultures grown for 18 h (unless otherwise stated) in liquid 79 medium. Before harvesting, a 1/10 culture volume of 95% EtOH saturated with phenol was added at a 5% final concentration to stabilize cellular RNA [57], and then mycelium was harvested by centrifugation and frozen in liquid nitrogen. Next, total RNA was isolated using Tri-Reagent (Sigma-Aldrich) according to the manufacturer’s protocol. The RNA solution was transferred to a Total RNA Mini column (A&A Biotechnology) and processed according to the manufacturer’s instructions. The RNA samples were digested with Turbo DNase I (Invitrogen, Carlsbad, CA) to remove traces of chromosomal DNA and then purified and concentrated using Clean-Up RNA Concentration (A&A Biotechnology). Five hundred micrograms of RNA was used for cDNA synthesis with the Maxima First Strand cDNA Synthesis Kit (Thermo Fisher Scientific). cDNA samples were diluted 5 times and used as templates for quantitative PCR (qPCR, each reaction performed in triplicate) using PowerUp SYBR Green Master Mix (Thermo Fisher Scientific). The level of the topA, gyrB, sco4804, sco4805 and luxC transcript was quantified using hrdB as a reference gene (ΔΔCT method) (StepOnePlus Real-Time PCR system; Applied Biosystems, Foster City, CA) (Additional file 9: Table S2). Isolated RNA was tested for DNA contamination by qPCR with oligonucleotides complementary to the S. coelicolor hrdB gene. The difference > 5 Ct after 30 PCR cycles between the RNA sample and the corresponding cDNA sample as a template showed that the RNA samples were DNA-free.

DNA supercoiling assay

Global DNA supercoiling in S. coelicolor strains was quantified using the pWHM3Hyg reporter plasmid [62] or its modified version, pWHM3Spec (Table S1). The plasmids were isolated using alkaline lysis and column purification based on a modified version of the manufacturer’s (Plasmid Screening Kit, Syngen) procedure. After 48 h of growth in liquid 79 medium supplemented with hygromycin or spectinomycin, S. coelicolor mycelium was collected by centrifugation, resuspended in PZ buffer containing 25 mg/ml lysozyme and incubated at 30 °C for 5 min. The subsequent steps followed the manufacturer’s protocol. The isolated reporter plasmids were resolved in 0.8% agarose in Tris–acetate-EDTA (TAE) buffer (40 mM Tris, 20 mM acetic acid, and 1 mM EDTA, pH 8.3) in the presence of 4.6 µM chloroquine at a voltage of 20 V. To visualize topoisomers, the gel was stained with ethidium bromide for 30 min at room temperature. The experiment was repeated twice. The topoisomer distribution was analysed using ImageJ software.

TopA level quantification using Western blotting

For TopA level quantification, S. coelicolor 5 ml liquid cultures in 79 medium were cultivated for 24 h. Next, the cell pellet was collected by centrifugation, resuspended in phosphate-buffered saline (PBS), sonicated and centrifuged. The cell lysates (5 µg of total protein) were separated by 10% SDS-PAGE according to standard procedures [37]. After electrophoresis, the resolved proteins were stained overnight with PageBlue Protein Staining Solution (Thermo Fisher Scientific) or transferred to a nitrocellulose membrane and blocked with 2% milk in TBST (20 mM Tris, 150 mM NaCl, 0.1% Tween 20, pH 7.4–7.6). The blots were subsequently incubated with rabbit polyclonal TopA antiserum (1:10,000 in TBST,1-h incubation; [62] and visualized using alkaline phosphatase-conjugated goat anti-rabbit antibodies (1:5000) (Sigma-Aldrich). The band intensities were analysed using ImageJ software, comparing the TopA band intensity of particular mutants to the wild-type reference.

Structure prediction and homologue analysis

Protein structure predictions were performed using the Robetta Web Server and “TrRefineRosetta” modelling method [55, 61], as well as using PredictProtein [81]. Homologue searching was performed using HOGENOM [52].