Background

Saccharopolyspora erythraea has been used for the production of erythromycin A (Er A), a broad-spectrum macrolide antibiotic that is effective against pathogenic gram-positive bacteria [1, 2]. In recent years, Er A has been in vast demand owing to its good antibacterial effect and the therapeutic potential of its semi-synthetic derivatives [3]. Erythromycin production is a world-wide, large-scale industrial process, and methylation of erythromycin C (Er C) to Er A is one of the main manufacturing bottlenecks, therefore, it is necessary that methods used to promote the production and purity of Er A, such as regulatory strategy, are improved to better meet market demand. Er A biosynthesis involves the formation of a macrolide intermediate, 6-deoxyerythronolide B (6-DEB), and its subsequent modifications [3,4,5]. These modifications include hydroxylation of the C-12 site at the aglycone and O-methylation of the C-3 site at the macrose, which are catalyzed by the P450 hydroxylase (EryK) and S-adenosyl-methionine (SAM)-dependent O-methyltransferase (EryG), respectively [6,7,8]. The synthesis of Er A is accompanied by the synthesis of Er C and erythromycin B (Er B), which can affect the yield and purity of Er A during fermentation [3, 9]. Methylation is the key and final step in the synthesis of Er A from Er C, which is catalyzed by the EryG. As a catalytic substrate for EryG, increasing the supply of SAM is conducive to increasing the synthesis of antibiotics requiring methylation modification [10,11,12,13,14,15,16]. In addition, SAM plays an important role in intracellular processes and a variety of biosynthetic machineries; it is an active methyl donor and regulatory factor [17, 18]. Therefore, enhancing S-adenosyl-methionine synthetase (MetK) expression is a common strategy for improving antibiotic synthesis [10,11,12,13,14,15,16]. In addition, it is well known that the expression of gene clusters related to erythromycin synthesis is regulated by the phosphate-sensing factor PhoP and the nitrogen-sensing factor GlnR [19, 20]. Interestingly, the EryG gene exists in the middle of the erythromycin synthesis gene cluster in S. erythraea and is regulated by PhoP [19]. However, to date, the mechanism by which PhoP and GlnR regulate the MetK-encoding gene metK and its affect upon the supply of intracellular SAM has not yet been resolved. Investigating the regulatory relationship between PhoP and GlnR and the expression of metK is conducive to enhancing the understanding of the nutritional response of actinomycetes, the supply of SAM, and the synthesis of methylation-modified antibiotics [1, 12, 19, 21].

PhoP and GlnR have been reported to respond to phosphate and nitrogen stress and to regulate antibiotic production in species such as in Streptomyces coelicolor [22], S. erythraea [19], and Streptomyces natalensis [23]. PhoP- and GlnR-mediated regulation of phosphate and nitrogen metabolism affects the expression of genes related to antibiotic synthesis, precursor supply, post-modification of antibiotics, and so on. PhoP senses the change in phosphorus signal and directly or indirectly regulates the biosynthesis of undecylprodigiosin in S. coelicolor [22], actinorhodin in Streptomyces lividans [24], and polyene macrolide pimaricin in S. natalensis [23]. In S. erythraea, the biosynthesis of Er A is also sensitive to phosphate concentration in the growth media [19]. Phosphate limitation strongly induces expression of the ery gene cluster and further promotes the biosynthesis of Er A [19]. Similarly, GlnR also regulates antibiotic biosynthesis in Streptomyces [20, 25]. In S. coelicolor A3(2), the deletion of glnR led to a remarkable increase in actinorhodin production [25]. The same phenomenon was also found in the rifamycin producer Amycolatopsis mediterranei [26] and erythromycin producer S. erythraea [20]. In summary, since PhoP and GlnR have a relatively general regulatory effect on antibiotic synthesis gene clusters, the question becomes, do they have a similar transcriptional regulation effect on metK and the supply of SAM?

In this work, we performed a series of systematic studies to analyze the PhoP- and GlnR-mediated regulation of metK expression. EMSAs confirmed that PhoP and GlnR interact with the operator interval of the metK gene promoter. A transcription difference further verified that metK expression level was influenced by PhoP and GlnR, and confirmed the direct regulatory mechanism of PhoP and GlnR on metK expression. Finally, by assessing Er A production we provide evidence that PhoP and GlnR-mediated regulation of metK expression might be an effective target for influencing the synthesis of engineered methyl-modified antibiotics in actinobacteria.

Results

PhoP and GlnR interact with the regulatory region of SACE_3900 gene encoding S-adenosylmethionine synthetase

Previous studies have shown that PhoP and GlnR have a regulatory effect on erythromycin synthesis and the expression of central metabolism-related genes [19, 20, 27]. To investigate the influence of PhoP and GlnR on the expression of the SAM-donor for the erythromycin side chain methyl group, it was first necessary to identify the metK in S. erythraea [17]. Base on the annotation from the KEGG database and genome sequence analysis, the genome of S. erythraea possesses two putative metK genes (SACE_2103 and SACE_3900) [17]. The genes shared high sequence similarity at the amino acid level (approximately 88.9%) (Additional file 1: Fig. S1). Therefore, the promoter regions of SACE_2103 and SACE_3900 were selected for in vitro experiments to investigate the binding effects of PhoP and GlnR (Fig. 1A). In this study, a classic EMSAs experiment was used to determine whether PhoP and GlnR interact with the promoter region of each putative target gene. The 5-Biotin-labeled SACE_2103 and SACE_3900 DNA probes were mixed with purified His6-tagged PhoP or GlnR. For SACE_3900, shifted bands were found in the labeled SACE_3900 probe lane for both PhoP and GlnR (Fig. 1B, C). Unlabeled specific probe (S) and non-specific competitor DNA (N) were included to confirm the specificity of the binding. These data verify that the PhoP and GlnR proteins specifically bind to the promoter region of SACE_3900, which is consistent with our bioinformatic prediction. In contrast, neither His6-tagged PhoP nor GlnR caused a band shift when incubated with SACE_2103, which suggests that these proteins do not interact with the promoter of this gene. Utilizing MAST/MEME tools and previously reported conserved binding sequence [2], no potential PhoP or GlnR binding motifs in the upstream promoter region of SACE_2103 were detected.

Fig. 1
figure 1

Interaction between PhoP or GlnR and the promoter region of putative metK genes. A Schematic representation of the genetic positioning of putative metK genes in S. erythraea, and the possible regulatory effect of PhoP and GlnR. B Interaction assessment for His6-PhoP and the upstream promoter region of SACE_2103 (left) and SACE_3900 (right). C Interaction assessment for His6-GlnR and the upstream promoter region of SACE_2103 (left) and SACE_3900 (right). Concentration of His6-tagged PhoP and GlnR protein was (Lane 1, 0; Lane 2, 0.2; and Lane 3, 0.4 μM). Arrow heads show the free probes without bound protein. Unlabeled specific probe, S; non-specific competitor DNA, N; DNA/pantothenate metabolism flavoprotein, dfp; hydrolase (secreted trypsin-like serine protease), hyd; S-adenosyl-L-homocysteine hydrolase, sahH; 5-methyltetrahydrofolate–homocysteine methyltransferase, metH; adenosine kinase, adoK; S-adenosylmethionine synthetase, metK; glycine sarcosine N-methyltransferase, met1; acetyl-CoA synthetase (AMP-dependent synthetase/ligase), acs.

Full size image

EMSAS results showed that regulators PhoP and GlnR specifically bind to the upstream promoter region of SACE_3900 (Fig. 1B, C). To precisely determine the PhoP and GlnR binding positions in the SACE_3900 promoter region, the upstream promoter region of SACE_3900 was divided into three sections (referred to as SACE_3900-1/-2/-3) (Additional file 1: Fig. S2) and each DNA fragment was mixed with various concentrations of His6-tagged PhoP or GlnR. SACE_3900-3 was found to harbor predicted PhoP and GlnR binding sites (Fig. 2A, B). The sectioned EMSAs further suggest that PhoP and GlnR possibly regulate SACE_3900 expression via specific interaction with the SACE_3900-3 region. Furthermore, putative GlnR-binding (Fig. 2C, a-site: GGATC; b-site: GAAAC) and PhoP-binding motifs (Fig. 2C, GTTCACGAGTG), were validated using MEME prediction software [28]. In summary, these results demonstrated that PhoP and GlnR have binding boxes in the promoter region of SACE_3900 and have potential regulatory functions for its expression.

Fig. 2
figure 2

Analysis of the PhoP- and GlnR-binding region for the SACE_3900 promoter. A 5-biotin labeled SACE_3900-1/-2/-3 DNA fragments were reaction with His6-tagged GlnR and were analyzed using EMSA. B 5-biotin labeled SACE_3900-1/-2/-3 DNA fragments were reaction with His6-tagged PhoP and analyzed using EMSA. C Speculated binding motif of GlnR and PhoP protein in the SACE_3900 upstream promoter. His6-tagged PhoP (0, 0.3 and 0.6 μM) and GlnR (0, 0.4 and 0.8 μM) were added in reaction system. Arrowheads show the free probes without bound protein

Full size image

PhoP and GlnR positively regulate the transcription of SACE_3900

It has been reported that exogenous addition of SAM [13] or overexpression of the metK markedly enhances the production of antibiotics [12]. The specific binding sites of PhoP and GlnR on the SACE_3900 promoter were confirmed in vitro (Figs. 1 and 2). To determine if the interaction between PhoP or GlnR and the SACE_3900 promoter influences gene expression, we further investigated the transcription of SACE_3900 in liquid culture. To assess the effect of PhoP, the wild-type S. erythraea strain (WT), and the phoP overexpression strain (OphoP) were cultivated in phosphate-limiting modified Evans medium (see “Materials and methods” section). To assess effect of GlnR, WT, the glnR deletion strain (ΔglnR), the glnR complemented strain (ΔglnR::glnR), and the glnR overexpression strain (OglnR) were cultivated in nitrogen-limiting Evans (see “Materials and methods” section) [19]. The quantitative reverse transcription polymerase chain reaction (RT-qPCR) results showed that the expression of SACE_3900 in OphoP was upregulated 2.3- and 6.7-fold at 36 h and 72 h, respectively, when compared to WT (Fig. 3A). The observations demonstrated that PhoP specifically bound to the promoter region of SACE_3900 (Fig. 1) and likely activated its transcription (Fig. 3A). Similarly, the expression of SACE_3900 was upregulated in OglnR at 36 h and 72 h by 8- and 5.3-fold respectively, when compared to WT. Conversely, ΔglnR showed reduced SACE_3900 expression compared with the WT strain (Fig. 3B). Moreover, the ΔglnR::glnR strain restored the phenotype to WT levels. These results indicated that PhoP and GlnR positively induced SACE_3900 expression under phosphate and nitrogen limited conditions, respectively.

Fig. 3
figure 3

PhoP and GlnR positively regulate the expression of SACE_3900. A RT-qPCR analysis of phoP and SACE_3900 in WT and OphoP following 36 h or 72 h of growth. B Gene transcription of glnR and SACE_3900 in WT, ΔglnR, ΔglnR::glnR, OglnR strain at 36 and 72 h. The WT and OphoP were grown in phosphate-limiting Evans (40 μM KH2PO4 and C and N as basic), while WT, ΔglnR, ΔglnR::glnR and OglnR were grown in nitrogen-limiting Evans [1 mM (NH4)2SO4 and C and P as basic]. Error bars indicate standard deviations from three independent biological replicates

Full size image

PhoP and GlnR increase intracellular SAM content

Overexpression of phoP and glnR induced the expression of SACE_3900 in liquid culture (Fig. 3). Next, we determined the intracellular SAM levels of the strains using HPLC. As shown in Fig. 4, the production of SAM was increased in the OphoP and OglnR strains. The SAM levels in OphoP and OglnR strains were 1.4- and 1.5-fold (P < 0.05) higher than those in the WT strains, respectively (Fig. 4A, B). In contrast, the ΔglnR strain resulted in the downregulation of the SAM concentration compared with the WT strain (Fig. 4B), while the ΔglnR::glnR strain restored the WT level. The results suggested that the global regulators PhoP and GlnR influenced the SAM content, which is consistent with the influence of each protein upon SACE_3900 transcription levels.

Fig. 4
figure 4

PhoP and GlnR promote SAM production in S. erythraea. A Content of SAM in WT and OphoP strains. B Content of SAM in WT, ΔglnR, ΔglnR::glnR and OglnR strains. The WT and OphoP were grown in phosphate-limiting Evans, while WT, ΔglnR, ΔglnR::glnR and OglnR were grown in nitrogen-limiting Evans. Error bars indicate standard deviations from three independent biological replicates. *P < 0.05 by One-way ANOVA and Tukey’s multiple comparisons test using GraphPad Prism 8

Full size image

Effect of GlnR and PhoP on the transformation of Er C to Er A

Previous studies have demonstrated that the PhoP response to phosphate starvation induces the expression of the ery cluster and further promotes Er A production [19]. Moreover, overexpression of metK, eryK, and eryG significantly increased Er A content and decreased the Er B component in S. erythraea [3]. Since it has been shown that GlnR and PhoP could regulate and promote the expression of the putative metK gene SACE_3900 (Figs. 1, 2 and 3), we further investigated the relative concentrations of Er C and Er A, and examined the regulation of methylation by GlnR and PhoP from the perspective of the transformation of Er C to Er A, which depends on SAM (coded for by metK). The Er A produced by OphoP strains increased by 17% compared with that of the WT strain (Fig. 5A). Moreover, the conversion ratio of Er C to Er A increased by about 10%, which further indicated that PhoP promoted Er A production via the methylation of Er C, and this was possibly due to upregulation of metK, which is consistent with previously reported studies [19]. However, no significant improvement in Er A production and Er C conversion was observed in OglnR strains (Fig. 5B), and surprisingly Er A production in the ΔglnR strain was significantly improved by 31% (P < 0.05), and the conversion from Er C to Er A increased by 13% compared with the WT strain. Although the intracellular SAM level was increased in OglnR, Er A production and biotransformation from Er C to Er A did not increase. The influence of GlnR on erythromycin biosynthesis promotes the expression of metK (Figs. 1, 2 and 3); however, it may also inhibit the erythromycin biosynthesis [19]. Many studies have reported that erythromycin production and biosynthetic gene expression are strongly inhibited by ammonium in over-producing and low-producing S. erythraea strains [29]. Since GlnR is a global factor related to nitrogen induction, its detailed mechanism for SAM and erythromycin biosynthesis is worthy of further investigation in subsequent studies. In summary, PhoP and GlnR activate the expression of SACE_3900 and promote SAM production likely via sensing phosphorus and nitrogen signal changes. Only OphoP strains activated the production of Er A and conversion of Er C to Er A.

Fig. 5
figure 5

Quantification of Er C and Er A in different S. erythraea strains. A The concentration of Er C and Er A produced by the WT strain and OphoP was measured using HPLC. B Er C and Er A concentration of WT, ΔglnR, ΔglnR::glnR, and OglnR strains. Strains A and B grown in phosphate or nitrogen limited medium. Supernatants were harvested after being grown for 6 days. Three independent replicates were operated to calculate mean and the standard deviations

Full size image

Discussion

It has been previously reported that the phosphate regulator PhoP plays a vital role in primary and secondary metabolite biosynthesis [2]. The PhoP sensing signal changes through its pairing with the response regulator PhoR, and it regulates gene expression, often via direct binding to the promoter region of target genes [19]. For example, PhoP-ery and PhoP-BldD-ery nutrient-sensing signal transduction routes have been reported for the regulation of erythromycin biosynthesis [19]. Moreover, the methylation of Er C is the key final step in Er A biosynthesis. Importantly, the methylation process is SAM dose-dependent. Therefore, exogenous SAM or overexpression of metK genes should increase Er A production. We speculated that manipulating PhoP to overexpress metK might be a new strategy to promote the synthesis of SAM. Studies have also shown a pleiotropic regulatory relationship between PhoP and GlnR (Fig. 6) [19]. The crosstalk between the systems makes it possible for GlnR to regulate the synthesis of erythromycin and SAM. Interestingly, GlnR is a nitrogen regulator has also been shown to have a significant impact on the ery gene cluster [29]. Research has shown that erythromycin high-producing strains display lower expression levels of the glnR gene than the WT strain [19, 29, 30]. Moreover, erythromycin production was shown to be repressed by ammonium concentration in growth medium [20, 31]. Our study confirmed that the nitrogen regulator GlnR positively regulates metK expression (Fig. 6), however it did not ultimately increase the yield of Er A.

Fig. 6
figure 6

Regulatory model of metK mediated by PhoP and GlnR in S. erythraea. Solid lines indicate transcriptional regulation. Arrows represent positive influence. Lines represent negative control. S-adenosyl-L-homocysteine, SADH;  homocysteine, HMC; methionine, Met; S-adenosyl-methionine, SAM; cysteine, Cys.

Full size image

The methyl donor SAM participates in the methylation modification of the side chain of antibiotics under the action of methyltransferases, such as EryG [19, 29, 30]. It is known that the methyltransferase-encoding gene eryG is directly or indirectly regulated by PhoP and GlnR [19], yet the regulatory mechanism of PhoP and GlnR in SAM synthesis is unknown. In addition to metK, shown to be involved in this study, the SAM synthesis cycle also requires the supply of important amino acids, such as cysteine and methionine, as precursors (Fig. 6). Cells can acquire such amino acids by absorption or de novo synthesis and it is well known that GlnR and PhoP in Streptomyces [32,33,34,35] regulate the absorption and synthesis of amino acids. Therefore, for follow-up research, the effect of GlnR and PhoP on the transport and synthesis of cysteine and methionine should be investigated. GlnR and PhoP-mediated nitrogen and phosphorus metabolism processes respond to changes in the nitrogen source [20, 36] and phosphorus [37, 38] in the culture medium; conversely, in the anti-biosynthesis approach, the concentration of nitrogen and phosphorus sources can be adjusted to intervene with the functions of GlnR [29, 39] and PhoP [19, 23]. As a result, GlnR and PhoP-mediated regulation for chemical synthesis can be achieved, especially in the context of antibiotic methylation and the regulation of precursor SAM supply. From the perspective of metabolic engineering, through overexpression or deletion of global regulatory factors PhoP and GlnR, the conversion of antibiotic precursors to final products is achieved. The above research will promote the development of SAM-dependent antibiotic synthesis, and the discovery of efficient strategies for the synthesis and transformation of methylated antibiotics.

Materials and methods

Strains, plasmids, and culture conditions

Strains and plasmids mentioned in this experiment are shown in Table 1. S. erythraea strains were grown in TSB or Evans media. Evans media were used as described previously [19, 30], including phosphate-limiting Evans (40 μM NaH2PO4 and basic C and N), nitrogen-limiting Evans [1 mM (NH4)2SO4 and basic C and P) and basic Evans [7.5 mM (NH4)2SO4, 8 mM NaH2PO4] medium with 1% glucose. Escherichia coli was grown in Luria–Bertani (LB) liquid media or on LB plates at 37 °C. All media were autoclaved at 121 °C for 20 min. Glucose was autoclaved at 115 °C for 30 min and then added to Evans medium.

Table 1 Strains and plasmids used in this study
Full size table

Computational analysis

The predicted PhoP and GlnR binding box in the upstream promoter regions of metK genes were identified using the MEME/MAST tool and PREDetector software program. The details of the prediction methods have been described previously [40].

Electrophoretic mobility shift assay (EMSAs)

Predicted promoters of putative metK genes (SACE_2103 and SACE_3900) were amplified by PCR using gene-specific and 5′-biotin-labeled universal primers (Table 2). PCR probes were verified using agarose gel electrophoresis and purified with a PCR purification kit (Beijing TIANGEN Biotech Co., Ltd., China). The probe concentration was measured using a microplate reader (Biotek, USA). EMSAs were performed as described by Xu et al. [19]. The samples were separated on a 6% non-denaturing PAGE gel in an ice-bath of 0.5 × Tris–borate-EDTA at 120 V, and band shifts were tested by BeyoECL P.

Table 2 Oligonucleotides used in this study
Full size table

RNA preparation and quantitative reverse transcription polymerase chain reaction (RT-qPCR)

WT and mutant strains were all activated in TSB culture for 48 h at 30 °C and then transferred into phosphate- or nitrogen-limiting Evans media. Cells were collected by centrifugation (8000 rpm for 10 min at 4 °C) at 36 and 72 h and were washed with saline thrice. RNA preparation methods and integrity analysis were performed as described by Xu et al. [2]. RNA concentration was measured using a microplate reader (BioTek). cDNA was synthesized using a PrimeScript reverse transcription (RT) reagent kit with gDNA Eraser (TIANGEN, Beijing). The primers used for qRT-PCR are shown in Table 2. The qRT-PCR experiments were executed using a StepOne plus real-time system (ABI, CA), and the reaction conditions were described by Xu et al. [19]. The 16S rRNA gene was used for reference, and variations in transcription were calculated using the reported method [19].

SAM (S-adenosyl-methionine) determination

WT and Ophop were grown in phosphate-limited Evans medium, whereas WT, ΔglnR, ΔglnR::glnR, and OglnR were grown in nitrogen-limited Evans medium. Cells were harvested by centrifugation (8000 rpm, 10 min, 4 °C), followed by incubation with 20 mL 10% trichloroacetic acid for 2 h at 25 °C with continuous stirring. The supernatant was collected (8000 rpm, 4 °C) and freeze-dried. The dried samples were then dissolved in 1 mL of sterile water. Intracellular SAM content was determined by high-performance liquid chromatography (HPLC) as described by Payne [41]. Dissolved samples were first obtained through a 0.22-μm pore-size filter and tested with a Hypersil BDS-C18, 5-μm, 4.6 × 250 mm column (Agilent HP1260). The testing conditions were shown below: mobile phase [(10 mM HCOONH4): C2H3N 1: 9]  and detection wavelength of 254 nm (UV–VIS).

Fermentation and erythromycin determination

Fermentation (50 mL) was harvested by centrifugation (8000 rpm, 10 min, 4 °C), and the supernatant was freeze-dried. Dried samples were dissolved in 1.5 mL acetonitrile through a 0.22-μm pore-size filter and tested with HPLC. The testing conditions were as follows: mobile phase [(50 mM K2HPO4; pH 6.8): C2H3N 60:40], detection wavelength (215 nm), chromatographic column (5 μm, 4.6 × 250 mm), and rate (1 mL min−1) [19].