Abstract
Most secondary metabolism in Actinobacteria is controlled by multi-layered, gene-regulatory networks. These regulatory mechanisms are not easily identified due to their complexity. As a result, when a strong transcriptional regulator (TR) governs activation of biosynthetic pathways of target antibiotics such as actinorhodin (ACT), additional enhancement of the biosynthesis is difficult in combination with other TRs. To find out any “synergistic transcriptional regulators (sTRs)” that show an additive effect on the major, often strong, transcriptional regulator (mTR), here, we performed a clustering analysis using the transcriptome datasets of an mTR deletion mutant and wild-type strain. In the case of ACT biosynthesis in Streptomyces coelicolor, PhoU (SCO4228) and RsfA (SCO4677) were selected through the clustering analysis, using AfsS (SCO4425) as a model mTR, and experimentally validated their roles as sTRs. Furthermore, through analysis of synergistic effects, we were able to suggest a novel regulation mechanism and formulate a strategy to maximize the synergistic effect. In the case of the double TR mutant strain (ΔrsfA pIBR25::afsS), it was confirmed that the increase of cell mass was the major cause of the synergistic effect. Therefore, the strategy to increase the cell mass of double mutant was further attempted by optimizing the expression of efflux pump, which resulted in 2-fold increase in the cell mass and 24-fold increase in the production of ACT. This result is the highest ACT yield from S. coelicolor ever reported.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
Microorganisms are an important source for producing various secondary metabolites useful for the human such as antibiotics, herbicides, and antifungal and anticancer agents (Baltz 2008). Due to the high industrial value of these secondary metabolites, many studies were undertaken to understand and exploit the secondary metabolism. Secondary metabolites are not essential for cell growth and development but help the producing cells to adapt and survive under the surrounding environment. Biosynthetic pathways in the secondary metabolism are tightly regulated by environmental changes such as nutrient depletion, growth phase, and signal transduction to optimize the metabolism in response to various environmental changes (Patra et al. 2013). Therefore, in order to improve production yield of such secondary metabolites, it is important to understand the detailed and phase-dependent regulations of the cell according to environmental changes.
Transcriptional regulator (TR) is a DNA-binding protein that regulates gene expression in response to environmental changes by binding to a specific DNA sequence as an activator or repressor (Patra et al. 2013). Since secondary metabolism is often regulated by a number of TRs (Ohnishi et al. 2005), the metabolism can be more effectively reconstructed to increase the production of target molecules through simultaneous regulation of expression of several TRs. However, due to the complexity of the TR network, it is very difficult to obtain synergistic effect of TRs when a strong TR primarily dominates biosynthetic activation of target secondary metabolism.
In general, there are two groups of TRs: a pathway-specific TR that controls a specific metabolic pathway gene cluster and a global TR that exerts pleiotropic effects on multiple gene clusters by either having the same consensus binding sequence or following the same signal transduction pathway. Secondary metabolism is under the control of both types of TRs in multi-layered, complex, regulatory networks (Lv et al. 2014). Owing to the nature of these multi-layered regulation networks, the interactions of the TRs in hidden layers are not easily predictable, unless the interaction is quite significant. A number of approaches to understanding these networks are common, such as chromatin-immunoprecipitation coupled with massive parallel sequencing (ChIP-Seq) for genome-wide profiling of DNA-protein interactions (Valouev et al. 2008) and DNA-affinity capture assay (DACA) to identify the TRs binding to the promoter region of specific genes (Park et al. 2009). Despite these methods, elucidation of all the networks of TRs in a given organism is difficult, labor-intensive, and time-consuming tasks.
In this work, we performed a clustering analysis using DNA chip data to find a combination of TRs that can exert synergistic effects in complex TR networks without the need to understand TR network or prior knowledge. Clustering analysis enables us to separate and classify the genes with the same expression patterns (Eisen et al. 1998). This method assumes that the TR genes classified into the same cluster are likely to be under the same or similar control networks (Eisen et al. 1998). Thus, clustering analysis can be used to select candidate TRs that have synergistic effects among them. Here, Streptomyces coelicolor and actinorhodin (ACT) were selected as a model strain and target molecule, respectively, since the most abundant transcriptomic DNA chip data from public database are available for its physiological and metabolism studies. There are two assumptions to find a synergistic combination of TRs; first, important genes related to secondary metabolism such as major TR (mTR) show highly dynamic transcriptome profile changes during all growth phases and second, the changes in the expression profile of the specific mTR do not affect the expression of synergistic TR (sTR), as sTRs are likely to belong to an upper layer than the specific mTR, or can play an independent role in the same layer. Based on these assumptions, clustering analysis was performed using time-series transcriptome data of afsS deletion mutant (ΔafsS) and parental wild-type strain (WT) to find sTR for afsS, known as the mTR of ACT production (Lee et al. 2002). Through the following experiments, it was confirmed that the TRs selected as sTR for AfsS show a synergistic effect on ACT production when combined with AfsS. In addition, causal analysis of synergistic effects enabled us to further improve the synergistic effect of mTR and sTR and to propose new regulatory mechanisms.
Materials and methods
Bacterial strains and culture conditions
All the strains used in this experiment are listed in Table 1. Escherichia coli cells were cultured at 37 °C in Luria Bertani (LB) medium, supplemented with appropriate antibiotics when necessary. E. coli DH5α was used for DNA amplification and manipulation, and E. coli JM110 was used for non-methylated DNA propagation for transformation into S. coelicolor A3(2) M145 was used as parental strain. R5− complex media was used for cell growth and protoplast transformation (Kieser et al. 2000), and SMM, TSB, and R5MS media were used for antibiotic, i.e., ACT, production (Okamoto et al. 2009). For evaluation of secondary metabolites, spores were inoculated into a 250-ml baffled flask containing 50 ml of media with appropriate concentrations of antibiotic markers, and then shaken at 30 °C for 8 days. Samples (1 ml) were taken at various time points for ACT quantification and measurement of dry cell weight. pIBR25 was used as a cloning vector.
Microarray data analysis
Time-series microarray data were obtained from Gene Expression Omnibus (GEO) datasets (Barrett et al. 2007; Edgar et al. 2002). The GEO accession numbers of wild-type M145 and its afsS deletion mutant were GSE8084, GSE8086, and GSE8107 and GSE8110 and GSE8160, respectively (Lian et al. 2008). The raw intensities at every time point in each dataset were transformed to log2 scales and normalized using the quantile normalization method (Bolstad et al. 2003). For the normalized dataset, log2 values at specified time points were computed by subtracting the intensity of the initial time point from those of the sample time points to generate a fold-change matrix. The same process was done separately for all the five datasets, resulting in five fold-change matrices. To make the distribution of the five matrices identical, the five matrices were concatenated to create a new matrix, which was then normalized again using the quantile normalization method (Bolstad et al. 2003). The values of the five matrices were substituted by the values obtained from the above quantile normalization. Finally, the normalized matrices were smoothed using moving average method (window size = 3).
To identify the differentially expressed genes (DEGs) of the wild-type strain during growth phase changes, a modified time-series transcriptome analysis method (Hwang et al. 2005) was used as follows: in a matrix for the wild-type strain, (i) forward- and backward-smoothed fold-change profiles for each gene were numerically integrated over the time course to obtain the area under the smoothed fold-change profiles; (ii) reference fold-change profiles were generated through the permutation-based random sampling from the smoothed fold-change matrix; (iii) a null distribution was empirically generated using the trapezoidal method for the areas of reference fold-change profiles; (iv) P values for each gene were calculated using the null distribution by one-tailed test; (v) three P values for each gene in the three datasets of the wild-type strain were combined using Stouffer’s method; (vi) the genes showing significantly altered values of expression in all the three wild-type datasets were identified with a combined P value < 0.01. The DEGs identified from the above procedure are likely to be related to the genes for secondary metabolism in the wild-type cell.
For the clustering analysis of the gene expression profiles, the normalized time-series fold-change values of identified DEGs in the M145 and those of the same ID genes in afsS deletion mutant were collected, and a new DEG matrix with time-series fold-change values was generated. To find sTR genes for AfsS, i.e., mTR, the genes with similar profiles between the M145 and the afsS deletion mutant were identified from the newly generated DEG matrix through hierarchical clustering using the Pearson correlation.
RNA-seq data processing
We used time-series RNA-seq data of S. coelicolor A3(2) M145 obtained from GSE69350 (Jeong et al. 2016) for the screening of inducible promoters. This dataset was generated at four sampling time points depending on the growth phase, and the dataset was generated in duplicate at each time point (EE early exponential, ME mid exponential, LE late exponential, S stationary phase). The raw data of RNA-seq were normalized using spike-in controls in R, to compare the absolute expression values among the time-series data (Bar-Joseph et al. 2012). In this method, we used eight housekeeping genes as spike control (Li et al. 2015). After data processing, the global transcript profiles were clustered based on Pearson correlation coefficient to evaluate the similarity of the gene expression pattern along the growth phases (D’Haeseleer 2005). The hierarchical clustering was performed with “hclust” function in R using the Pearson method.
Measurement of undecylprodigiosin (RED) and actinorhodin (ACT)
The amounts of RED and ACT were measured at 530 and 640 nm, respectively, using a Multiskan Spectrum plate reader (Thermo Fisher Scientific, USA) following standard procedures (Kieser et al. 2000). For dry cell weight determinations, 2 mL of sampled cell broth (in triplicate) was collected in pre-weighted Eppendorf tube and dried at 70 °C for 4 days.
qPCR
The qPCR was performed with TOPreal SYBR Green PCR Kit (Enzynomics) on the Roche LightCycler® 480 real-time platform to quantify the absolute expression of specific mRNAs. Gene expression was normalized relative to hrdB (SCO5820) (Nieselt et al. 2010). All reactions were performed in technical triplicates. The hrdB was amplified from strain M145 using primers hrdB_f and hrdB_r, and the actAB was amplified from strain M145 using primers actAB_f and actAB_r. Table S1 shows the primers used for these experiments.
Results
Identification of sTR candidates for AfsS
The procedure of clustering analysis is elaborated in the “Microarray data analysis” subsection in the “Materials and methods” section. To identify sTRs for AfsS, we first selected differentially expressed genes (DEGs) from the three datasets of wild-type M145 to catalog all the genes related to the secondary metabolism (Fig. 1a). As a result, 1010 genes among 7343 genes were selected as DEGs. Subsequently, clustering analysis between the M145 and afsS deletion mutant (M145(ΔafsS)) was performed to select genes showing the same profile in both strains among the DEGs (Fig. 1b). The basic premise of this analysis was that the genes classified into the same cluster in the M145 and the mutants would be related to the secondary metabolism but would not be affected by AfsS. We chose a cluster showing an increasing pattern during the stationary phase to screen the TRs related to ACT biosynthesis. The increasing cluster consisted of 109 genes, where 11 TRs were found and assigned as sTR candidates for AfsS (Table 2).
The objective of this experiment is to find sTRs for AfsS, which have a synergistic effect on the biosynthesis of ACT. Therefore, among the 11 sTR candidates, five TRs (SCO0608 (SlbR), SCO3579 (WblA), SCO4677 (RsfA), SCO7252 (NsdB), and SCO4228 (PhoU)) appearing to be related to ACT biosynthesis were selected. According to the literature survey, four of five candidates of sTR, excluding PhoU, were reported as repressor of ACT production (Hindra et al. 2014; Lee et al. 2013; Yang et al. 2012; Zhang et al. 2007). Among them, when the repression is removed by gene deletion, the sTR with the largest increase in ACT production was RsfA. Therefore, RsfA was finally selected as sTR because it was expected to maximize the synergistic effect on ACT production along with the overexpression of AfsS. In the case of PhoU, the impact on the production of ACT has not been reported. However, it is strongly anticipated that it is related to the PhoR-PhoP system that affects the production of ACT. The main advantage of clustering analysis is that it can find sTRs without knowing the TR network. If the double mutant (The strain that combines the effects of mTR and sTR is hereinafter referred to as “double mutant”), PhoU deletion and AfsS overexpression, shows a synergistic effect, the robustness of the clustering analysis could be confirmed. Therefore, PhoU and RsfA were finally selected as potential sTRs and subjected to experiments to demonstrate and confirm their synergistic effects on ACT production by the combination with AfsS.
Confirmation of the synergistic effect of PhoU on AfsS
PhoU has been considered as a TR associated with a two-component PhoR-PhoP system, but the exact function of PhoU in S. coelicolor remains unclear (Santos-Beneit et al. 2009). To investigate the role of PhoU in ACT biosynthesis, a phoU-amplified mutant (BGPhoU) and a phoU deletion mutant (M145 ΔphoU) were constructed. Since the PhoR-PhoP system controls both primary and secondary metabolism in the phosphate limitation environment (Sola-Landa et al. 2003), ACT productions of BGPhoU, M145 ΔphoU, and M145 were firstly measured on SMM solid agar plates with varying phosphate concentrations to approximately figure out what is the optimum concentration of phosphate for ACT production (Fig. S1). As a result, SMM media without phosphate was selected for ACT production, and all strains were cultured in SMM liquid media without phosphate for 5 days. When the ACT yields of the two mutants were compared to that of BG25 (M145 harboring pIBR25), no significant differences were observed between BGPhoU (12.9 mg/L) and BG25 (11.8 mg/L), but the ΔphoU mutant showed an increased ACT yield (69.1 mg/L) (Fig. 2), suggesting that PhoU is a negative TR on ACT production under such conditions. After confirming the role of PhoU, the double-mutant BG4228AfsS (ΔphoU pIBR25::afsS) was constructed to confirm the synergistic effect of phoU deletion and afsS overexpression on the production of ACT. The ACT yield of the double mutant was then compared with that of the single mutants, BG4228 (ΔphoU pIBR25) and BGAfsS (M145 pIBR25::afsS). The double-mutant BG4228AfsS yielded the highest titer (151.4 mg/L), which was 9 times higher than that of BGAfsS (M145 pIBR25::afsS; 16.3 mg/L) and 63 times higher than that of BG4228 (2.37 mg/L) (Fig. 3a). Since AfsS is a pleiotropic TR for other antibiotics biosynthesis as well in S. coelicolor, RED production was also measured in the same strains shown above, and the trends were similar (Fig. S2 A). These results indicate that the effect of afsS overexpression and phoU deletion on ACT and RED production is mutual and additive.
Confirmation of the synergistic effect of RsfA on AfsS
RsfA (SCO4677), previously known as anti-sigma factor F, exerts a negative effect on ACT production and cell morphological differentiation (Kim et al. 2008). When BG4677(ΔrsfA pIBR25) and BGAfsS were grown in R5− complex media, they showed 1.5- and 3.2-fold increases in ACT yield compared to that of BG25, respectively. Whereas the double mutant, BG4677AfsS (ΔrsfA pIBR25::afsS), showed an 8-fold increase in ACT yield under the same condition, confirming that afsS overexpression and rsfA deletion have a synergistic effect on ACT production (Fig. 3c). In terms of RED production, BG4677AfsS displayed an 8-fold increase in RED titer compared to that of BG25, whereas both BG4677 and BGAfsS did not show any significant positive effects on RED production (Fig. S2 B). These results also indicate that the effect of afsS overexpression and rsfA deletion on ACT and RED production is mutual. The above results could validate our clustering analysis to find sTRs for the target mTR, AfsS.
The 8-fold increase in ACT yield of BG4677AfsS compared to that of BG25 was resulted from the combination of 1.6-fold increase in cell mass and 4.8-fold increase in its specific yield (Fig. 3c, d). It was somewhat expected that inhibition of cell growth might occur in the case of BG4677AfsS owing to the ACT secretion and/or toxicity of ACT itself. Surprisingly, however, an opposite result was acquired. BG4677AfsS showed ca. 60% increase in cell growth compared to that of BG25 despite higher ACT production (Fig. 3c). One possible explanation would be that actAB induction by ACT or its three-ring intermediate (S)-DNPA synthesis, which may confer additional resistance to the cell by ACT overproduction, resulting in cell growth enhancement (Xu et al. 2012). When comparing the expression of actAB in BG25 and BG4677AfsS, it was confirmed that the expression level of actAB in BG4677AfsS increased about 3.5 times than that of BG25. (Fig. 3e). This result agrees well with our assumption that the additional overexpression of actAB can further improve the ACT yield of BG4677AfsS.
Inducible overexpression of actAB increases not only ACT production but also final cell mass
In order to prove whether the increase in cell mass of BG4677AfsS by overexpression of actAB can increase ACT yield, actAB was overexpressed in the M145. To construct the strain overexpressing actAB, transformation of pCActAB25 having a constitutive strong ermE promoter into BG25 was performed (Table 1). However, the BGCActAB (BG25 pIBR25::actAB) strain showed very sluggish cell growth in R5− media. On R5− agar plate, BG25 colonies appeared within 2 days, whereas the colonies of BGCActAB appeared after 7 days. In R5− liquid medium, the BG25 strain followed a normal growth curve, but the growth of BGCActAB did not start for 4 days (Fig. S3). Surprisingly, the cell growth of BGCActAB starts to recover after 5 days and the mutant outgrew BG25 by 150% after 7 days (Fig. S3) and its ACT yield was 4-fold higher than that of BG25 at 7 days (data was not shown). These results clearly demonstrated that overexpression of actAB, the putative efflux pumps of ACT, has a positive effect on final cell mass and ACT production but does prolong the lag phase. We hypothesized that excessive actAB expression in earlier cell growth stage might be the cause of initial sluggish cell growth, and the possible solution to overcome this would be expression of actAB using inducible promoter at late exponential phase.
Promoter used for overexpression of efflux pumps requires high expression level of target protein starting from late exponential phase. To discover such additional inducible promoters, another clustering analysis was performed using our own experimental RNA-seq data taken at four different time points. The raw data of RNA-seq were processed with spike-in normalization and clustered based on Pearson correlation coefficient. All the genes were grouped into 12 clusters (Fig. 4). In particular, we could confirm that the ACT cluster including the putative efflux pumps belong to cluster 7. The genes in cluster 7 were listed according to the fold changes in the ratio of the corresponding gene expression level of the stationary phase to that of the early exponential phase (i.e., S/EE), to discover any promoters with the desired expression profile. Among the selected candidates, the promoter of SCO4677 was chosen as the inducible promoter to express the efflux pump since its function is well characterized and its fold-change ratio (S/EE) was very high (Table 3). pIActAB25 was constructed for inducible expression of efflux pump (ActA-ActB) and transformed into S. coelicolor M145 strain by protoplast fusion to yield BGIActAB. In the case of BGIActAB grown in R5− liquid media containing 10 μg/ml thiostrepton for 5 days, its dry cell weight and specific yield of ACT increased 2- and 2.1-fold, respectively, compared to those of BG25 (Fig. 5a, b). BGIActAB exhibited an c.a. 4-fold increase in ACT yield, compared to that of BG25 (Fig. 5c). This result demonstrates that the optimal overexpression of actAB under the control of the screened inducible promoter is able to increase not only ACT production but also the final cell mass.
Inducible overexpression of actAB for high ACT producers with the combination of sTR and afsS
Since the cell mass of BGIActAB achieved in R5− liquid media was higher than that of BG4677AfsS, we reasoned that the additional overexpression of actAB in the background of BG4677AfsS would give an improved production of ACT. First, pAfsSIA25, in which actAB and afsS genes were inserted in opposite directions, was constructed to avoid interference of the two gene expressions. Then, the plasmid was introduced into the BG4677 strain by protoplast transformation to yield BG4677AfsSIA. After 4-day growth of BG4677AfsSIA, its cell mass and ACT yield were 9.57 mg/ml of DCW and 480 mg/L, respectively. The final cell mass was more than two times higher than that of its parental strain BG25 (Fig. 5a–c), and ACT titer was the highest ever reported.
Discussion
Identification of sTRs while the overexpression of strong mTR is present is a formidable task and very important for industrial mass production of antibiotics, since secondary metabolism is often simultaneously regulated by various TRs. However, due to the complexity of the TR network, it is difficult to find such sTRs for any specific mTR (Lv et al. 2014). In this study, we applied a clustering analysis to discover such sTRs by comparing time-series transcriptome data of wild-type and specified mTR deletion mutant (here, afsS) without the need to understand the regulatory network of corresponding antibiotic biosynthesis (here, ACT).
The major advantage of clustering analysis is its accurate screening capability of sTRs for a specified target mTR. In the case of ACT, PhoU (SCO4228) is a good example. Despite the uncertainty of the function and regulatory network of PhoU, PhoU was identified as a sTR for AfsS (SCO 4425). Although each of the effect of phoU deletion and afsS overexpression on ACT production was negligible, a significant increase in the production of ACT was observed in the double mutant, BG4228AfsS. This result did not appear to be caused by a simple additive effect but by a synergistic effect via unknown interactions. We reasoned that PhoU and AfsS might influence each other within its TR network, resulting in a higher production of ACT. In S. coelicolor, the transcription of afsS is repressed by the phosphorylated form of PhoP, whose phosphorylation is controlled by PhoR (Santos-Beneit et al. 2009). Thus, the biosynthesis of ACT is very likely to be repressed in the presence of phosphorylated PhoP, despite the overexpression of afsS. Especially, it has been reported that the kinase activity of PhoR is controlled by PhoU in Escherichia coli (Gardner et al. 2014) and Aquifex aeolicus (Oganesyan et al. 2005). Therefore, it is possible to assume that the PhoU in S. coelicolor can control the PhoR-PhoP system and may affect the transcription of afsS. In fact, a synergistic effect on ACT yield was observed in BG4228AfsS strain, which can be strongly anticipated as a result of relieving AfsS repression by blocking PhoP phosphorylation through phoU deletion. This is consistent with the assumption that PhoU of S. coelicolor controls the PhoR-PhoP system. Furthermore, it suggests that PhoU, a selected sTR for AfsS, is an upper layer TR of AfsS that regulates the negative interaction of AfsS and PhoP by controlling phosphorylation of PhoP.
In the case of the double mutant, BG4677AfsS, the specific yield of ACT was increased by 4.8-fold compared to that of BG25. The specific yield of BG4677 and BGAfsS showed 1.5- and 3.2-fold increase compared to that of BG25, respectively. This result can be interpreted as the increase in yield of ACT in BG4677AfsS is due to the simple addition of rsfA deletion and afsS overexpression effect on ACT production. Therefore, it is reasonable to conclude that RsfA and AfsS independently affect ACT production in the TR network. This result demonstrates that the selected sTR for the mTR through the clustering analysis actually functions independently.
Furthermore, a synergistic effect was observed in terms of ACT yield of BG4677AfsS, mainly due to an increase in cell mass. To our surprise, the increase in cell mass was hypothesized to be somewhat related to the increased expression of the actAB induced by the increased amount of ACT in the cell (Xu et al. 2012). We attempted to demonstrate the effect of efflux pump overexpression on increase in cell mass via overexpression of actAB. The overexpression level and expression timing of actAB appeared to greatly affect cell growth. When they were overexpressed in early growth phase, very long lag time was observed in the cell growth, whereas, when they were overexpressed in a late exponential phase, increase in cell growth was observed. The BGIActAB which overexpresses the actAB using an inducible promoter (promoter of SCO4677) in the M145 greatly increased not only the cell mass but also the specific yield of ACT. Since the cell mass of the BGIActAB was even higher than that of BG4677AfsS, fine control of the efflux pump overexpression in BG4677AfsS mutant strain was expected to result further increase in the productivity of ACT, which was confirmed through our study.
In addition, since the efflux pump, in general, has the function of pumping out toxic substances (Poole 2007), it was expected that the strain which overexpresses the efflux pump (ActA-ActB) show higher secretion of ACT. Unlike our prediction, higher ACT secretion was not observed in our experiment, and the specific secretion ability of the BGIActAB was reduced by 2.8-fold compared to that of parental strain BG25 (Fig. S4 A). Instead, the strain which overexpresses the actC (one of the three putative efflux pumps of ACT biosynthesis cluster) using an inducible promoter used in the aforementioned experiment (BGIActC) showed that the specific secretion ability of the cell increased 2.2-fold compared to that of the parental strain BG25 (Fig. S4 A). These results suggest that ActC functions as an actual efflux pump of ACT, and ActA-ActB has other functions that enhance ACT resistance. This is consistent with the two-step model for ACT export and resistance, in which ActA-ActB is required for efficient production of ACT and increases resistance to ACT (Xu et al. 2012).
In summary, we could find sTRs of mTR (e.g., AfsS) for overproduction of desired antibiotics through clustering analysis using transcriptome data of the wild-type strain and the mTR deletion mutant without prior understanding of complex regulatory network of TR. Even though the synergistic effects of all five sTRs on mTR have not been tested, it was well demonstrated that the clustering analysis is suitable for selecting sTR candidates for mTR. Analysis of the synergistic effects of the sTR candidates allowed us to suggest an unknown regulatory mechanism among the TRs and predict which to optimize in order to further increase the production of ACT. The ACT yield of the final strain, BG4677AfsSIA, is the highest ever reported, and this study is the first successful report of antibiotic overproduction by the combination of two TRs (here, sTR and mTR). These strategies can be applied to increase the production of any secondary metabolites in Actinobacteria.
References
Baltz RH (2008) Renaissance in antibacterial discovery from actinomycetes. Curr Opin Pharmacol 8(5):557–563. https://doi.org/10.1016/j.coph.2008.04.008
Bar-Joseph Z, Gitter A, Simon I (2012) Studying and modelling dynamic biological processes using time-series gene expression data. Nat Rev Genet 13(8):552–564. https://doi.org/10.1038/nrg3244
Barrett T, Troup DB, Wilhite SE, Ledoux P, Rudnev D, Evangelista C, Kim IF, Soboleva A, Tomashevsky M, Edgar R (2007) NCBI GEO: mining tens of millions of expression profiles—database and tools update. Nucleic Acids Res 35(Database issue):D760–D765. https://doi.org/10.1093/nar/gkl887
Bolstad BM, Irizarry RA, Astrand M, Speed TP (2003) A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 19(2):185–193
D’Haeseleer P (2005) How does gene expression clustering work? Nat Biotechnol 23(12):1499–1501. https://doi.org/10.1038/nbt1205-1499
Edgar R, Domrachev M, Lash AE (2002) Gene expression omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res 30(1):207–210
Eisen MB, Spellman PT, Brown PO, Botstein D (1998) Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci 95(25):14863–14868
Gardner SG, Johns KD, Tanner R, McCleary WR (2014) The PhoU protein from Escherichia coli interacts with PhoR, PstB, and metals to form a phosphate-signaling complex at the membrane. J Bacteriol 196(9):1741–1752. https://doi.org/10.1128/jb.00029-14
Gust B, Challis GL, Fowler K, Kieser T, Chater KF (2003) PCR-targeted Streptomyces gene replacement identifies a protein domain needed for biosynthesis of the sesquiterpene soil odor geosmin. Proc Natl Acad Sci U S A 100(4):1541–1546. https://doi.org/10.1073/pnas.0337542100
Hindra MMJ, Jones SE, Elliot MA (2014) Complex intra-operonic dynamics mediated by a small RNA in Streptomyces coelicolor. PLoS One 9(1):e85856. https://doi.org/10.1371/journal.pone.0085856
Hwang D, Rust AG, Ramsey S, Smith JJ, Leslie DM, Weston AD, de Atauri P, Aitchison JD, Hood L, Siegel AF, Bolouri H (2005) A data integration methodology for systems biology. Proc Natl Acad Sci U S A 102(48):17296–17301. https://doi.org/10.1073/pnas.0508647102
Jeong Y, Kim JN, Kim MW, Bucca G (2016) The dynamic transcriptional and translational landscape of the model antibiotic producer Streptomyces coelicolor A3(2). Nat Commun 7:11605. https://doi.org/10.1038/ncomms11605
Kieser T, Bibb MJ, Buttner MJ, Chater KF, Hopwood DA (2000) Practical streptomyces genetics. The John Innes Foundation, Norwich
Kim ES, Song JY, Kim DW, Chater KF, Lee KJ (2008) A possible extended family of regulators of sigma factor activity in Streptomyces coelicolor. J Bacteriol 190(22):7559–7566. https://doi.org/10.1128/jb.00470-08
Lee PC, Umeyama T, Horinouchi S (2002) afsS is a target of AfsR, a transcriptional factor with ATPase activity that globally controls secondary metabolism in Streptomyces coelicolor A3(2). Mol Microbiol 43(6):1413–1430
Lee HN, Kim JS, Kim P, Lee HS, Kim ES (2013) Repression of antibiotic downregulator WblA by AdpA in Streptomyces coelicolor. Appl Environ Microbiol 79(13):4159–4163. https://doi.org/10.1128/aem.00546-13
Li S, Wang J, Li X, Yin S, Wang W, Yang K (2015) Genome-wide identification and evaluation of constitutive promoters in streptomycetes. Microb Cell Factories 14(1):172. https://doi.org/10.1186/s12934-015-0351-0
Lian W, Jayapal KP, Charaniya S, Mehra S, Glod F, Kyung YS, Sherman DH, Hu WS (2008) Genome-wide transcriptome analysis reveals that a pleiotropic antibiotic regulator, AfsS, modulates nutritional stress response in Streptomyces coelicolor A3(2). BMC Genomics 9:56. https://doi.org/10.1186/1471-2164-9-56
Lv Q, Cheng R, Shi T (2014) Regulatory network rewiring for secondary metabolism in Arabidopsis thaliana under various conditions. BMC Plant Biol 14:180. https://doi.org/10.1186/1471-2229-14-180
Nieselt K, Battke F, Herbig A, Bruheim P, Wentzel A, Jakobsen ØM, Sletta H, Alam MT, Merlo ME, Moore J, Omara WAM, Morrissey ER, Juarez-Hermosillo MA, Rodríguez-García A, Nentwich M, Thomas L, Iqbal M, Legaie R, Gaze WH, Challis GL, Jansen RC, Dijkhuizen L, Rand DA, Wild DL, Bonin M, Reuther J, Wohlleben W, Smith MCM, Burroughs NJ, Martín JF, Hodgson DA, Takano E, Breitling R, Ellingsen TE, Wellington EMH (2010) The dynamic architecture of the metabolic switch in Streptomyces coelicolor. BMC Genomics 11:10–10. https://doi.org/10.1186/1471-2164-11-10
Oganesyan V, Oganesyan N, Adams PD, Jancarik J, Yokota HA, Kim R, Kim S-H (2005) Crystal structure of the “PhoU-like” phosphate uptake regulator from Aquifex aeolicus. J Bacteriol 187(12):4238–4244
Ohnishi Y, Yamazaki H, Kato J-Y, Tomono A, Horinouchi S (2005) AdpA, a central transcriptional regulator in the A-factor regulatory cascade that leads to morphological development and secondary metabolism in Streptomyces griseus. Biosci Biotechnol Biochem 69(3):431–439. https://doi.org/10.1271/bbb.69.431
Okamoto S, Taguchi T, Ochi K, Ichinose K (2009) Biosynthesis of actinorhodin and related antibiotics: discovery of alternative routes for quinone formation encoded in the act gene cluster. Chem Biol 16(2):226–236. https://doi.org/10.1016/j.chembiol.2009.01.015
Park SS, Yang YH, Song E, Kim EJ, Kim WS, Sohng JK, Lee HC, Liou KK, Kim BG (2009) Mass spectrometric screening of transcriptional regulators involved in antibiotic biosynthesis in Streptomyces coelicolor A3(2). J Ind Microbiol Biotechnol 36(8):1073–1083. https://doi.org/10.1007/s10295-009-0591-2
Patra B, Schluttenhofer C, Wu Y, Pattanaik S, Yuan L (2013) Transcriptional regulation of secondary metabolite biosynthesis in plants. Biochim Biophys Acta 1829(11):1236–1247. https://doi.org/10.1016/j.bbagrm.2013.09.006
Poole K (2007) Efflux pumps as antimicrobial resistance mechanisms. Ann Med 39(3):162–176. https://doi.org/10.1080/07853890701195262
Santos-Beneit F, Rodriguez-Garcia A, Sola-Landa A, Martin JF (2009) Cross-talk between two global regulators in Streptomyces: PhoP and AfsR interact in the control of afsS, pstS and phoRP transcription. Mol Microbiol 72(1):53–68. https://doi.org/10.1111/j.1365-2958.2009.06624.x
Sola-Landa A, Moura RS, Martín JF (2003) The two-component PhoR-PhoP system controls both primary metabolism and secondary metabolite biosynthesis in Streptomyces lividans. Proc Natl Acad Sci U S A 100(10):6133–6138. https://doi.org/10.1073/pnas.0931429100
Thuy ML, Kharel MK, Lamichhane R, Lee HC, Suh JW, Liou K, Sohng JK (2005) Expression of 2-deoxy-scyllo-inosose synthase (kanA) from kanamycin gene cluster in Streptomyces lividans. Biotechnol Lett 27(7):465–470. https://doi.org/10.1007/s10529-005-2222-y
Valouev A, Johnson DS, Sundquist A, Medina C, Anton E, Batzoglou S, Myers RM, Sidow A (2008) Genome-wide analysis of transcription factor binding sites based on ChIP-Seq data. Nat Methods 5(9):829–834. https://doi.org/10.1038/nmeth.1246
Xu Y, Willems A, Au-Yeung C, Tahlan K, Nodwell JR (2012) A two-step mechanism for the activation of actinorhodin export and resistance in Streptomyces coelicolor. MBio 3(5):e00191–e00112. https://doi.org/10.1128/mBio.00191-12
Yang Y-H, Song E, Kim J-N, Lee B-R, Kim E-J, Park S-H, Kim W-S, Park H-Y, Jeon J-M, Rajesh T, Kim Y-G, Kim B-G (2012) Characterization of a new ScbR-like γ-butyrolactone binding regulator (SlbR) in Streptomyces coelicolor. Appl Microbiol Biotechnol 96(1):113–121. https://doi.org/10.1007/s00253-011-3803-4
Zhang L, Li WC, Zhao CH, Chater KF, Tao MF (2007) NsdB, a TPR-like-domain-containing protein negatively affecting production of antibiotics in Streptomyces coelicolor A3 (2). Wei Sheng Wu Xue Bao 47(5):849–854
Acknowledgments
This research was supported by the National Research Foundation of Korea(NRF) funded by the Ministry of Science, ICT and Future Planning (2016953757), and by Korea Institute of Planning and Evaluation for Technology in Food, Agriculture, Forestry and Fisheries (IPET) through Agri-Bio industry Technology Development Program, funded by Ministry of Agriculture, Food and Rural Affairs (MAFRA)(116139-03-1-SB010), and by the Institute for Basic Science (IBS-R13-G1).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Ethical approval
This article does not contain any studies with human participants or animals performed by any of the authors.
Electronic supplementary material
ESM 1
(PDF 614 kb)
Rights and permissions
About this article
Cite this article
Kim, M.W., Lee, BR., You, S. et al. Transcriptome analysis of wild-type and afsS deletion mutant strains identifies synergistic transcriptional regulator of afsS for a high antibiotic-producing strain of Streptomyces coelicolor A3(2). Appl Microbiol Biotechnol 102, 3243–3253 (2018). https://doi.org/10.1007/s00253-018-8838-3
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00253-018-8838-3