Novel mechanism of hydrogen peroxide for promoting efficient natamycin synthesis in Streptomyces

ABSTRACT The mechanism of regulation of natamycin biosynthesis by Streptomyces in response to oxidative stress is unclear. Here, we first show cholesterol oxidase SgnE, which catalyzes the formation of H2O2 from sterols, triggered a series of redox-dependent interactions to stimulate natamycin production in S. gilvosporeus. In response to reactive oxygen species, residues Cys212 and Cys221 of the H2O2-sensing consensus sequence of OxyR were oxidized, resulting in conformational changes in the protein: OxyR extended its DNA-binding domain to interact with four motifs of promoter p sgnM . This acted as a redox-dependent switch to turn on/off gene transcription of sgnM, which encodes a cluster-situated regulator, by controlling the affinity between OxyR and p sgnM , thus regulating the expression of 12 genes in the natamycin biosynthesis gene cluster. OxyR cooperates with SgnR, another cluster-situated regulator and an upstream regulatory factor of SgnM, synergistically modulated natamycin biosynthesis by masking/unmasking the −35 region of p sgnM depending on the redox state of OxyR in response to the intracellular H2O2 concentration. IMPORTANCE Cholesterol oxidase SgnE is an indispensable factor, with an unclear mechanism, for natamycin biosynthesis in Streptomyces. Oxidative stress has been attributed to the natamycin biosynthesis. Here, we show that SgnE catalyzes the formation of H2O2 from sterols and triggers a series of redox-dependent interactions to stimulate natamycin production in S. gilvosporeus. OxyR, which cooperates with SgnR, acted as a redox-dependent switch to turn on/off gene transcription of sgnM, which encodes a cluster-situated regulator, by masking/unmasking its −35 region, to control the natamycin biosynthesis gene cluster. This work provides a novel perspective on the crosstalk between intracellular ROS homeostasis and natamycin biosynthesis. Application of these findings will improve antibiotic yields via control of the intracellular redox pressure in Streptomyces.

specialized metabolites, including half of all known antibiotics (1).Streptomyces have had a long evolutionary association with fungi, with fungal cell walls being a key source of nutrition (2,3), and Streptomyces spp.play key roles in the natural recycling of materials from cell walls of fungi and plants (4,5).Sterols are present in membranes of organisms from unicellular fungi to animals and have versatile functions as structural elements and signaling molecules (6).
Previous studies have shown that the peroxide-sensing transcriptional regulators OxyR, SoxR, PerR, and OhrR are involved in oxidative stress defense responses in bacteria (26).In each case, oxidative modification of the regulator alters its DNA-bind ing properties.The genome of S. gilvosporeus F607 contains a gene-encoding OxyR (GenBank no.NZ_CP020569.1)(17).OxyR has an N-terminal DNA-binding domain (DBD) and a C-terminal regulatory domain (RD) and is a member of the LysR family (38).Several studies demonstrated that OxyR was a switch that modulated the transcription of antioxidant genes dependent on the intracellular oxidizing conditions (26,39,40).Reduced OxyR has two redox-sensing cysteine residues that can be oxidized to form a disulfide bond, or overoxidized to the S-sulfonate conformation (Cys-SO 3 H), at specific H 2 O 2 concentrations (26,41,42).In this way, the OxyR tetramer functions as a "threestate redox switch, " whereby its extended conformation and DNA-binding sites change to alter its regulatory function to turn on or off the transcription of its target antioxidant genes (43,44).
Recently, we found that overexpression of SgnE (cholesterol oxidase) enhanced the production of natamycin in S. gilvosporeus (24).Here, S. gilvosporeus SgnE catalyzed the production of H 2 O 2 from sterols from plants, fungi, and vertebrates (ergosterol, stigmasterol, and cholesterol, respectively).The H 2 O 2 acted as signaling molecule to trigger redox-state-dependent gene transcription, protein expression, DNA binding, and enzyme interactions that regulate natamycin biosynthesis.

Strains and cultures
For analyses of sterols as elicitors, S. gilvosporeus F607 was cultured in a seed medium containing 10 g/L glucose, 5 g/L peptone, 3 g/L yeast extract, and 3 g/L malt extract for 24 h.Then, 5% of the seed culture was added to TC medium (2.0 g/L soy peptone, 0.45 g/L yeast extract, 2.0 g/L NaCl, 1.0 g/L MgSO 4 , 60.0 g/L glucose, 0.2 g/L stigmasterol, and 0.045 g/L ergosterol, pH 7.5); TH medium (2.0 g/L soy peptone, 0.45 g/L yeast extract, 2.0 g/L NaCl, 1.0 g/L MgSO 4 , 60.0 g/L glucose, 2.0 g/L stigmasterol, and 0.45 g/L ergosterol, pH 7.5); or TF medium (2.0 g/L soy peptone, 0.45 g/L yeast extract, 2.0 g/L NaCl, 1.0 g/L MgSO 4 , 60.0 g/L glucose, 4.0 g/L stigmasterol, and 0.9 g/L ergosterol, pH 7.5).S. gilvosporeus F607 was used as the sgnE and oxyR gene donor.It was routinely cultured in seed medium containing 10 g/L glucose, 5 g/L peptone, 3 g/L yeast extract, and 3 g/L malt extract.For natamycin production, the strain was cultured in a fermenta tion medium (20.0 g/L soy peptone, 4.5 g/L yeast extract, 2.0 g/L NaCl, 1.0 g/L MgSO 4 , and 60.0 g/L glucose).The same media were supplemented with thiostrepton when used to culture S. gilvosporeus strain DE13.Escherichia coli ET12567 (pUZ8002) was used as a donor in intergeneric conjugations.E. coli BL21 (DE3) was used as a host for protein expression.The E. coli strains were cultured in lysogeny broth, containing 100 µg/mL kanamycin when necessary.The strains, plasmids, and primers used in this work are listed in Table S1.

Transcriptomic analyses
For transcriptome sequencing and analysis, mycelia of S. gilvosporeus F607 cultured in different media were harvested after 24, 60, and 120 h.The transcriptomes were sequenced using a HiSeq 3000 sequencer (Illumina) at RibBio Corporation (Shenzhen, China).The transcript level of each gene was normalized by the number of reads per kilobase of transcriptome per million mapped reads (RPKM).Differentially expressed genes were selected using the Audics program with parameters |log 2 fold-change| > 1 and q-value < 0.001.

Construction of sgnE and oxyR deletion mutants and complementation strains
sgnE and oxyR deletion mutants strains were constructed by homologous recombination as previously (45).The whole sgnE gene was amplified with primers sgnE R-F and sgnE R-R using S. gilvosporeus F607 genomic DNA as the template (Table S2).The region downstream of the sgnE gene was amplified with primers sgnE L-F and sgnE L-R.Then, the two fragments were digested with XbaI and joined together using T4 DNA ligase.The fusion gene was cloned into plasmid pJTU1278 to construct pJTU1278-ΔsgnE using Gibson Assembly Master Mix (New England BioLabs, Ipswich, MA, USA).pJTU1278-ΔsgnE was transformed into E. coli ET12567 containing pUZ8002 and conjugated into S. gilvosporeus F607, as described by Wang et al. (24) Positive conjugants were selec ted using 10 µg/mL thiostrepton and 20 µg/mL streptomycin as resistance screening markers.Successful conjugants were confirmed by colony PCR with primers sgnE V-F1 and sgnE V-R1 and by DNA sequencing; the resulting strain was named S. gilvosporeus DE13.The ermE promoter region (p ermE ) was amplified from pUC57, which is preserved in our laboratory (24).The p ermE and sgnE fragments were both digested with XbaI and joined together using Gibson Assembly Master Mix, and then, the joined fragment was cloned into pMS82; the resultant plasmid was transferred into S. gilvosporeus 607 to obtain sgnE-complemented S. gilvosporeus strain CE13, which was selected using 50 µg/mL hygromycin as a resistance screening marker.The oxyR deletion mutant S. gilvosporeus DO7 and its complemented mutant S. gilvosporeus CO7 were constructed using analogous procedures.Primers for oxyR mutation and complementation were listed in Table S2.

Expression and purification of SgnE and OxyR
The sgnE gene was amplified from S. gilvosporeus F607 genomic DNA using the primer pair sgnE His-F/R (Table S2) and then cloned into pMD18T to generate pMD18T-sgnE.After confirmation by DNA sequencing, sgnE from pMD18T-sgnE was cloned into the expression vector pET-15b to construct pET-sgnE, which was transformed into E. coli BL21 (DE3) to obtain E. coli BL21/pET-sgnE.When E. coli BL21/pET-sgnE was cultured to an OD 600 nm value of 0.6-0.8,0.1 mM isopropyl-D-1-thiogalactopyranoside was added to induce SgnE expression.His 6 -SgnE was purified by Ni-NTA chromatography (Sangon Biotech, Shanghai, China) (46).Construction of recombinant E. coli BL21/pET-OxyR and expression and purification of OxyR were performed using analogous procedures.Two fragments (amplified using the primers C212D-F and C212D-R) and pMD18T were linked using Gibson Assembly Master Mix and assembled with pET-15b to obtain the Cys212Asp variant of OxyR, following the same procedures.

Enzyme assay and H 2 O 2 quantification
The enzyme activities of SgnE from E. coli and S. gilvosporeus swjs-801 were determined by a colorimetric method with cholesterol, stigmasterol, and ergosterol as substrates as described by Wang et al. (24).Solution A (3 mL; 1 mmol/L 4-aminoantipyrine, 6 mmol/L phenol, 7,000 U/L peroxidase, 25 mmol/L potassium phosphate buffer, pH 7.5) was heated at 37°C for 3 min and then mixed with 0.02-0.10mmol/L H 2 O 2 for 5 min.The absorbance of the mixture was determined at 500 nm to establish a standard curve.Aliquots of shaken flask or bioreactor cultures were centrifuged at 10,000 × g for 20 min, and then, 1 mL of the supernatant was retained for enzyme assays.Solution A (3 mL) and 150 µL solution B (8.26 g/L cholesterol, stigmasterol, or ergosterol and 4.26% Triton X-100, dissolved in isopropanol) were heated at 37°C for 3 min and mixed with 50 µL fermentation supernatant, and the mixture was incubated for 5 min.The reaction mixture was boiled and then cooled on ice before determining the absorbance at 500 nm.Enzyme activity was calculated as enzyme activity (U/mL) = C × (V1/V2) × N ÷ T, where C is the H 2 O 2 concentration determined from the standard curve, V1 is the total volume of the enzymatic reaction mixture, V2 is the volume of fermentation supernatant, N is the dilution ratio of fermentation supernatant, and T is the reaction time.

Effect of H 2 O 2 on natamycin production
The intracellular H 2 O 2 concentrations in S. gilvosporeus strains DE13, CE13, and F607 were determined every 12 h for 120 h during fermentation using a Hydrogen Peroxide Assay Kit (Beyotime Biotech, Shanghai, China).Bioassays and high-performance liquid chromatography (HPLC) analysis of natamycin production were performed as follows.S. gilvosporeus DE13, CE13, and F607 were each cultured in seed medium at 29°C with shaking at 220 rpm for 16 h.Then, 1.5 mL of the culture was inoculated into a 250-mL flask containing 30 mL fermentation medium and cultured at 29°C with shaking at 220 rpm.The wet cell weight and natamycin production were determined every 12 h, as described previously (24,32).H 2 O 2 (100 µM) was added into the fermentation broth of S. gilvosporeus DE13 at 60 h, and natamycin production was determined 6 and 12 h later.Quantitative determination of natamycin was performed by HPLC analysis as described by Wang et al. (24)

Real-time quantitative PCR analysis
Total RNA isolation and RT-qPCR were performed as described previously (47).Mycelia of S. gilvosporeus F607 and S. gilvosporeus DO7 were harvested from fermentation liquid at 24, 72, 120, 168, and 216 h and then rapidly frozen in liquid nitrogen.Total RNA was extracted from mycelia using an RNA extraction kit (SBSBIO, Beijing China) according to the manufacturer's protocol and then treated with Turbo DNA-free reagents (ABI Ambion, Austin, TX, USA) to remove residual chromosomal DNA.cDNAs were synthe sized using random hexamer primers (pdN6, Amersham Pharmacia Biotech, Bucking hamshire, England), M-MLV reverse transcriptase (Invitrogen, Carlsbad, CA, USA), and dNTPs (Roche, Basel, Switzerland).RT-qPCR assays were performed on a Roche Light Cycler 480 instrument using SYBR Green Mix (Toyobo, Osaka, Japan).Relative quantities of cDNA were normalized to the amounts of the 16S rRNA gene.For RT-qPCR assays, experiments were conducted in triplicate.The primers used are listed in Table S3.

Analysis of OxyR redox status in vivo and in vitro
Aliquots of S. gilvosporeus CO7 cultures (OD 600 nm = 0.4) were taken with prewarmed pipettes and mixed with 1 / 10 th vol of 100% trichloroacetic acid (TCA).Precipitated proteins were collected by centrifugation at 10,000 × g for 10 min.After the complete removal of the supernatant, the pellet was dissolved in AMS buffer to an OD 600 nm value of 10.Alkylation was performed at 37°C for a minimum of 2 h.Subsequently, aliquots (10 µL) of the alkylated samples were subjected to non-reducing SDS-PAGE.After electrophoresis, the separated proteins were blotted onto nitrocellulose membranes and probed with polyclonal antibodies against OxyR.The bound antibodies were detected using a western blotting kit.
For in vitro assays, samples of His 6 -OxyR were incubated in transcription buffer (pH 7.0) containing dithiothreitol at room temperature for a minimum of 24 h in an anaerobic chamber (Coy Laboratory Products, Ann Arbor, MI, USA).An aliquot (1.0 mL) was removed and acidified with 1 / 10 th vol of 100% TCA.The remainder of the sample was treated with H 2 O 2 , and then, aliquots were precipitated with TCA at the indicated time points.The TCA-precipitated samples were removed from the anaerobic chamber and collected by centrifugation.The precipitated proteins were washed once with cold 10% TCA and collected by centrifugation.The pellet was redissolved and alkylated as described above.Samples were subjected to non-reducing SDS-PAGE.

Molecular simulation analyses
Protein sequences were aligned using CLUSTALW2 at http://www.expasy.ch.Sequences were obtained using the protein search algorithm at the National Centre for Biotechnol ogy Information (NCBI).Conserved domains of SgnR were analyzed by CD-Searches of the NCBI.A homology model of OxyR was; constructed using Robetta (https:// robetta.bakerlab.org/).Multiple sequence comparison was carried out using Clustal Omega (48) and ESPript software (49).A homology model of OxyR (and the Cys212Asp variant) was constructed using Robetta and Discovery Studio 2.0 (50).The model stereochemical structure of OxyR and the DNA sequences of all intergenic regions in the natamycin biosynthesis gene cluster were used as the accepter and ligands, respectively, in molecular docking analyses.Molecular docking analyses of OxyR and DNA were performed using the ZDOCKER protocol of Discovery Studio 2.0 (50).Allosteric sites were predicted using AlloSitePro (2016) (51).

EMSAs
A fragment of 150-300 bp was amplified from the intergenic region of sgnM and sgnR using S. gilvosporeus F607 genomic DNA as the template.The amplified DNA fragments were labeled at the 3′-end with biotin-11-UTP using a Biotin 3′ End DNA Labeling Kit (Thermo Scientific, Waltham, MA, USA).Non-specific cold probes (poly dI/dC) were added to control reaction mixtures as competitors.Electromobility shift assays (EMSAs) were carried out as described previously (52).To confirm the conserved DNA sites binding with OxyR, mutations of p sgnM were made (Table S2).

Microscale thermophoresis
p sgnM was used as ligands, and different redox status OxyR [prepared by excess dithiothreitol and H 2 O2 (42, 52)] were used as targets.The RED-NHS protein labeling kit was used to label the empty proteins.The DNA fragment was continuously diluted for 16 gradients by reaction buffer [50 mM HEPES (pH 7.4), 0.05% Tween 20].Mixture was incubated at room temperature for 20 min.Thermal swim signal of the mixture was measured by Monolith NT.115 Capillary.Affinity constant Kd was analyzed by using Nano Temper analysis software.

Co-immunoprecipitation assays
For co-immunoprecipitation (Co-IP) assays, proteins were incubated with anti-OxyR antibody, anti-SgnR antibody, or control IgG overnight with the protein A/G agarose beads.The complexes were washed three times and resuspended in 2 × SDS loading buffer.The immunoprecipitated proteins were eluted from the beads by incubation at 95°C for 5 min.The eluted proteins were detected by immunoblotting after separation by SDS-PAGE.

sgnE and oxyR show maximum transcription level change during sterolstimulated natamycin biosynthesis in S. gilvosporeus
Cholesterol, the main sterol in vertebrates, has been reported to be an effective elicitor of natamycin production (24).Here, we investigated whether the main sterols in plants and fungi, ergosterol and stigmasterol respectively, also trigger natamycin production in S. gilvosporeus.When S. gilvosporeus F607 was cultured in TC medium (defined in the Materials and Methods), which contained 2.0 g/L stigmasterol and 0.45 g/L ergosterol, the maximum production of natamycin was 27.92 mg/g mycelium.The production of natamycin reached its highest level, 91.3 mg/g mycelium, in TH medium, which contained 20 g/L ergosterol and 4.5 g/L stigmasterol.However, doubling the concentra tion of sterols to produce TF medium did not further promote natamycin production compared with the TH group (Fig. 1A).
To investigate how sterols stimulate natamycin production in S. gilvosporeus F607, we conducted transcriptomic and real-time quantitative PCR (RT-qPCR) analyses of samples collected at three points during natamycin synthesis after the initiation of culture (prophase at 24 h, rapid phase at 60 h, and late phase at 120 h).Transcriptome analyses (SRA data Accession NO: PRJNA949275) revealed that, in TC medium, which contained low concentrations of ergosterol and stigmasterol, 60 genes were upregulated and 283 genes were downregulated at 24 h compared with their respective transcript levels in TH medium (Fig. 1B).When natamycin began to be synthesized rapidly in the TC group at 60 h, the numbers of up-and downregulated genes increased to 1,941 and 2,816 compared with the TH group (Fig. 1C).When the natamycin synthesis rate slowed down at 120 h, there were 831 and 597 up-and downregulated genes, respectively, in the TC group compared with the TH group (Fig. 1D).
The differentially expressed genes were subjected to Gene Ontology (GO) term enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses.The GO terms significantly enriched among differentially expressed genes were mainly in the categories "proton transmembrane transport" at 24 h (Fig S1A through C), "organoni trogen compound biosynthetic process" at 60 h, and "translation" at 120 h.The KEGG pathway analysis showed that many of the differentially expressed genes at 24 h were involved in "oxidative phosphorylation, " while those at 60 and 120 h were involved in "ribosome" (Fig S1D through F).The KEGG analysis showed that the term "type I polyke tide structures, " which includes polyketide synthase (PKS) genes in the natamycin biosynthesis gene cluster, was enriched in differentially expressed genes at 24 and 60 h but not at 120 h.Moreover, genes in the "oxidation-reduction process" term and all the genes in the natamycin biosynthesis gene cluster showed significant differences in transcript levels between the TC and TH groups (Fig S1).
Thus, compared with the TC group, the TH group showed upregulation of genes associated with natamycin biosynthesis and the antioxidant system (Table S4).The transcript levels of PKS genes (sgnS0-sgnS4) were higher in the TH group than in the TC group (by 6.09-to 16.72-fold at 24 h, by 25.40-to 157.76-fold at 60 h, and by 1.54-to 4.67-fold at 120 h).Genes related to mycosamine formation and attachment to macrocyclic aglycones (sgnJ, sgnC, and sgnK) were also upregulated, especially in the rapid phase of natamycin biosynthesis (Table S4).Antioxidant system gene oxyR, in the TH group, upregulated to 2.54-fold at 24 h and 1.34-fold at 60 h but downregulated to 0.8-fold at 120 h (Table S4).This variation tendency is consistent with changes in natamycin synthesis rate in F607 [studied previously (45)].To confirm the transcriptomic results, RT-qPCR was performed for representative genes.The RT-qPCR results were generally consistent with the transcriptomic results, except for RS10405 at 24 h (Fig. S2).

SgnE catalyzes the production of H 2 O 2 from sterols which provide a ROS signal
Much debate has focused on the function of the cholesterol oxidase encoded by sgnE in the natamycin biosynthesis gene cluster in Streptomyces.To investigate its function in detail, we compared SgnE from S. gilvosporeus F607 with other cholesterol oxidases from Streptomyces (Fig. S3).A TAT (Tat/SPI)-type signal peptide for secretion (orange box), three repeating glycine residues (GXGXG) for interactions with FAD cofactor (53) (green box), and three residues (Ala403, Pro491, and Asn529) for catalysis (blue shading) were conserved in these cholesterol oxidases.To determine whether Streptomyces cholesterol oxidases degrade sterols as components of the "sterol cycle" in the natural environment, we selected three sterols, ergosterol, stigmasterol, and cholesterol (the main sterols in plants, fungi, and vertebrates, respectively) (6), for molecular simulation analyses.The tertiary structure of S. gilvosporeus F607 SgnE was modeled using structure PDB 4XWR (54) as the template (Fig. S4A).Ergosterol, stigmasterol, and cholesterol are structural analogs.The active site of the enzyme was buried in an amphipathic environment (Fig. S4B) that was large enough to accommodate the four-ring sterol structures.The enzymatic reaction cavity was connected to the loop cavity of the substrate-binding domain to form a central channel in SgnE, from the C-terminus to the N-terminus (Fig. 2A).In SgnE, the amino acid residues Tyr63, Val161, Val260, Ile261, Gly331, Tyr489, His490, Pro491, Pro529, and Ile533 formed the substrate-binding cavity.In this model, the 3β-hydroxyl group of the sterols was oxidized and interacted with the ketone of glycine (Gly331), and His490 acted as a general base catalyst during oxidation and an electrophilic catalyst during isomerization (Fig. S4C-E).
To verify the relationship between SgnE activity and H 2 O 2 generation, SgnE from S. gilvosporeus F607 (as positive control) was cloned and expressed in S. gilvosporeus swjs-801 (24) and Escherichia coli BL21 (DE3) (Fig. 2B).The recombinant S. gilvosporeus-SgnE and E. coli-SgnE could successfully oxidize cholesterol, ergosterol, and stigmas terol to generate H 2 O 2 (Fig. 2C).Thus, catalytic ability of SgnE in degrading sterols to form H 2 O 2 , which is a signaling molecule in organisms ( 26), provides a possibility of regulating natamycin biosynthesis via its enzymatic product indirectly, rather than protein directly.
During the natamycin synthesis stage of culture in fermentation medium (36-120 h), S. gilvosporeus F607 produced intracellular H 2 O 2 at 5.62-19.80µmol/g wet cell.In the sgnE-deleted strain S. gilvosporeus DE13, the H 2 O 2 concentration was decreased to 1.41-2.83µmol/g.S. gilvosporeus CE13, the sgnE complementation strain, produced H 2 O 2 at a similar concentration to that produced by strain F607 (5.17-19.65 µmol/g) (Fig. 3D).When the intracellular H 2 O 2 concentration in S. gilvosporeus F607 and CE13 presented a sharp increase at about 120 h of culture, natamycin production was decreased compar ing with the earlier time points when the H 2 O 2 concentration was lower.When the intracellular H 2 O 2 concentration was 5.62 µmol/g at 36 h in S. gilvosporeus F607, or 5.17 µmol/g in S. gilvosporeus CE13 at 48 h, which were lower than 6.0 µmol/g, the biosynthesis of natamycin was inactivated (Fig. 3C and D).
To verify the relationship between natamycin synthesis and the intracellular H 2 O 2 concentration, 100 µM H 2 O 2 was added into the fermentation broth of sgnE-deleted strain S. gilvosporeus DE13 at 60 h.The intracellular H 2 O 2 concentration increased to 11.25 µmol/g after 30 min, and natamycin could be detected with a yield of 0.08 mg/L after 6 h.Natamycin then continuously accumulated and its yield reached 0.17 g/L after 12 h.Similar phenomenon was not detected in the DE13 strain (Fig. 3E).These results indicate that the inhibitory effect of sgnE deletion on natamycin production could be partly alleviated by the addition of exogenous H 2 O 2 , suggesting that SgnE-mediated intracellular H 2 O 2 production regulates natamycin biosynthesis in S. gilvosporeus F607.

H 2 O 2 -sensing transcriptional regulator OxyR controls sgnM gene expression
The intracellular H 2 O 2 concentration was closely related to natamycin biosynthesis, and therefore, the role of the H 2 O 2 -sensing transcriptional regulator OxyR in natamycin production was investigated in S. gilvosporeus.OxyR from S. gilvosporeus F607 was precipitated with TCA, alkylated with fresh AMS buffer [0.1% (wt/vol) SDS, 1 M Tris, 15 mM 4-acetamido-49-maleimidylstilbene-2,29-disulfonic acid (AMS), pH 8.0] to alkylate free thiols, and then subjected to non-reducing SDS-PAGE and immunoblotting using an anti-OxyR antibody.The OxyR from S. gilvosporeus F607 was visible as three bands after reacting with AMS (Fig. 4A), indicative of the three redox states of the protein.
An oxyR-knockout strain, S. gilvosporeus DO7, was constructed (Fig. 4B).S. gilvosporeus DO7 showed decreased natamycin production, 0.53 g/L at 120 h, 11.8% of that produced by the parental strain S. gilvosporeus F607 (Fig. 4C).Then, S. gilvosporeus strain CO7 was constructed, in which strain DO7 was complemented with oxyR using the integrative vector pMS82.In strain CO7, natamycin biosynthesis was restored to 3.24 g/L at 120 h, slightly less than that produced by S. gilvosporeus F607 (Fig. 4C).The transcript levels of each transcriptional unit were detected by RT-qPCR in S. gilvosporeus strain DO7 to investigate differences in the expression of the natamycin biosynthesis gene cluster upon oxyR deletion.
The transcript levels of the natamycin efflux pump gene sgnH and PI factor secretion gene sgnT showed no significant difference between S. gilvosporeus strains DO7 and F607.However, the transcript level of the pathway-specific regulatory factor gene sgnM was decreased in S. gilvosporeus DO7 to 16.3% of the level in the parental strain S. gilvosporeus F607, while the transcript level of the other regulatory factor gene sgnR was only slightly decreased to 86.5% of that in S. gilvosporeus F607.Other genes also showed decreased transcription upon disruption of oxyR.For example, the transcript levels of the type I modular PKS-encoding genes (sgnS0, sgnS1, and sgnS2) were dramatically decreased to only 6.4%-11.2% of their respective levels in S. gilvosporeus F607.The transcript levels of other genes, including sgnK, sgnI, sgnJ, sgnA, sgnE, and sgnD in gene clusters controlled by SgnM, were also obviously decreased in strain DO7 to 9.5%-27.6% of their respective levels in strain F607 (Fig. 4D).These results indicate that OxyR activity is mediated by H 2 O 2 , and it regulates the expression of sgnM.
The predicted DNA-binding sequence of OxyR in the p sgnMR region, which is similar to the binding site for ideR in S. avermitilis, covered −30 to −85 bp upstream of the transcription start site of sgnM (Fig. 5A).To confirm the binding of OxyR to the pro moter region of the natamycin pathway-specific transcriptional regulator gene sgnM, the binding of OxyR with p sgnM was analyzed using EMSAs.We tested biotin-labeled probes of the sgnMR promoter region, p sgnM-1 -p sgnM-3 , and, as negative controls, the promoter regions of sgnS0, sgnS1, sgnS2, and sgnD (p sgnS0 , p sgnS1 , p sgnS2 , and p sgnD , respectively).Probe p sgnM-2 , containing the predicted binding sequence of OxyR to p sgnM , was shifted in EMSA by the addition of purified oxidized OxyR (Fig. 5B).To evaluate the specific interaction between OxyR and the promoter region of sgnM, competition tests were performed using oligonucleotide probe p sgnM-2 .A small amount of probe sgnM2 was shifted by 0.5 µg OxyR; most of the probe was shifted by 1.0 µg OxyR, and 1.5 µg OxyR was enough to shift all the sgnM2 probe (Fig. 5C).
Sequence analyses revealed that a sequence in p sgnM-2 was similar to a previously reported OxyR-binding site in the promoter of ideR in S. avermitilis.They shared the conserved sequence "GGT/ACC-15 nt-CGGC-15 nt-GCCG-9/8 nt-GGA/TCC." The four parts of this palindromic sequence were named motifs 1-4.To validate the OxyR consensus-binding sequence, oligonucleotide probes of the predicted recognition sequences containing motif mutations were tested in EMSAs.Mutations of motif 1 (M1), motif 2 (M2), or motif 3 (M3), respectively, dramatically decreased the binding affinity of OxyR for p sgnM compared with the native sequence (Fig. 5D, lanes M1-M3).Importantly, mutation of motif 4 (M4) abrogated OxyR binding with p sgnM (Fig. 5D, lane M4).When the tests were conducted with mutations in both motifs 1 and 2 (M5) (Fig. 5D, lane M5), or in all of motifs 1, 2, and 3 (M7) (Fig. 5D, lane M7), binding was weaker than that detected with the wild-type sequence.The simultaneous mutation of all four motifs (M6) abrogated OxyR binding with p sgnM (Fig. 5D, lane M6).These findings suggest the importance of these motifs in the interaction between OxyR and p sgnM .

Oxidized OxyR shows higher motif-binding affinity with p sgnM than reduced or overoxidized OxyR
Allosteric site predictions showed that the asymmetrical tetrameric OxyR contained four allosteric sites (Table S5).All the allosteric sites were in close proximity to the DBDs of OxyR.Eight cysteines (two in each monomer) are wrapped in the center of the tetramer and form an "H 2 O 2 reaction pocket." The frame of this H 2 O 2 reaction pocket was formed by eight α-helixes (four α6 and four α8) (Fig. S8).On oxidation, the diameter of the whole protein is shortened, and α4 rotated by 31.6°relative to its position in the reduced protein (Fig. 6A).In reduced OxyR, Cys212 and Cys221 are located at the ends of an α-helix in the RD, 14.4 Å apart (C α -C α ).After OxyR was oxidized by H 2 O 2 , disulfide bond formation unfolded the helix, resulting in the formation of a short β-strand, and the C α -C α distance between Cys212 and Cys221 shortened to 4.7 Å (Fig. 6B).Conformational changes in OxyR from reduction to oxidation are presented in Video S1.Overoxidized state of OxyR resulted in a conformation of helix α3 that was almost identical to that in reduced OxyR; the Cys212Asp mutation recovered the distance between Cys212 and Cys221 to 14.4 Å (Fig. 6B).The conformational change between the oxidized and overoxidized states of OxyR is presented in Video S2.
To investigate the effect of conformational changes of reduced, oxidized, and peroxidated OxyR on sgnM regulation, EMSAs were performed to investigate the binding of OxyR in different oxidation states with p sgnM DNA.Reduced and oxidized OxyR were prepared by dithiothreitol and H 2 O 2 treatment, respectively.The overoxidized state of OxyR was mimicked by using a Cys212Asp mutant of OxyR.Reduced, oxidized, and overoxidized OxyR could all bind to the DNA sequence of p sgnM .However, the oxidized OxyR showed a higher affinity for p sgnM than the reduced and overoxidized states (Fig. 6C).This binding was confirmed by microscale thermophoresis.Kd of reduced, oxidized, and overoxidized OxyR binding with p sgnM were 135.2 ± 3.5 nM, 45.9 ± 0.5 nM, and 323.5 ± 5.2 nM, respectively (Fig. 6D).Affinity of different oxidation states binding with p sgnM is oxidized OxyR > reduced OxyR > overoxidized OxyR.Therefore, OxyR presented low, high, and low DNA-binding affinities with the sgnM promoter region in the reduced, oxidized, and peroxidated states, respectively.This indicates that the conformation of OxyR regulates natamycin biosynthesis and is dynamically variable in S. gilvosporeus.

SgnR-OxyR complex cooperatively regulates sgnM transcription to control natamycin synthesis in response to H 2 O 2 concentration
SgnR binds to the sgnM promoter to activate its transcription, and then, SgnM activates transcription of eight genes in the natamycin biosynthesis cluster (14).The DNA-binding sequence of OxyR in the sgnM promoter region was "GGTCC-15 nt-CGGC-15 nt-GCCG- 9 nt-GGACC." The four parts of this palindromic sequence were named motifs 1-4, respectively.Motif 1 of this sequence covered the −35 bp region upstream of sgnM (Fig. 7A).Reduced and overoxidized OxyR adopt an extended conformation and turn off P sgnM by binding to elongated sequences masking the −35 region, while the contracted architecture of oxidized OxyR unmasks the −35 region of target gene sgnM.This is known as the "OxyR regulatory mode" (42).Notably, on sgnM promoter region, the OxyR-bind ing sequence was located 7 bp downstream of the SgnR-binding site, via which sgnM transcription was strictly controlled (55) (Fig. 7A).
Considering that the DNA-binding site of SgnR to sgnM was not contained in the −35 region of sgnM, which is not consistent with the regulatory mode of other (Strep tomyces antibiotic regulatory protein) SARP proteins (56), we conducted molecular simulation analyses to explore the interaction between SgnR and OxyR.First, the conserved domains and architecture of SgnR were predicted.SgnR is an SARP regulator consisting of 1,184 amino acid residues in S. gilvosporeus F607.It was found to contain five domains, namely, an N-terminal DNA-binding domain, a bacterial transcriptional activation domain consisting of three-tetratricopeptide repeats (TPRs), a random coil, an The result showed that PDCD4 interacts with DTL.
ATP/GTP-binding domain, and a TPR-like domain (Fig. 7B).Molecular docking analyses showed that SgnR could combine with the OxyR tetramer.Interaction residues, located in DBD domain of DBD and the RD of OxyR and random coil of SgnR, respectively, combined four regions when the two regulatory proteins interact (Fig. S9).The random coil of SgnR was inserted into the cavity located between the DBD and the RD of OxyR, this combination surface presents a long and narrow shape (Fig. 7C).This combination made the DBDs of OxyR and SgnR arrange into a linear structure, resulting in the formation of an extended groove (Fig. 7C).The extended groove provided a gap to allow DNA containing the binding sites for SgnR and OxyR to pass through (Fig. 7D).Co-IP demonstrated that OxyR could interact with SgnR (Fig. 7E).Functional complementation of the SgnR-OxyR complex synergistically regulate the transcription of sgnM by masking or unmasking the −35 region of the sgnM promoter to control natamycin synthesis in response to ROS concentration although sequential order of OxyR-SgnR complexes needs a further investigation.

DISCUSSION
In natural ecosystems, animals, plants, and microbes have evolved symbiotic interactions (57).The genus Streptomyces is an abundant source of secondary metabolites, includ ing antibiotics, anticancer compounds, immunosuppressants, pigments, herbicides, and enzymes.Ecologically, as saprophytic soil bacteria, Streptomyces play key roles in the natural recycling of the cell walls of plants, fungi, and insects (4,5).Sterols are a type of lipid that exist in membranes of diverse organisms from unicellular fungi and protists to multicellular animals and plants (6).In many multicellular organisms, sterols have another very important role in the production of signaling molecules such as steroids in vertebrates, brassinosteroids in plants, and ecdysteroids and molting hormones in invertebrates (6).Ecological and environmental conditions are believed to influence the metabolic capacity of bacteria to produce secondary metabolites (58).However, the role of Streptomyces in the "sterol cycle" and the influence of sterols on antibiotic metabolism in Streptomyces have rarely been investigated.Here, we show the function of sterols as elicitors that activate natamycin biosynthesis in S. gilvosporeus.Our results also show that the signaling molecule H 2 O 2 , generated from the degradation of sterols by cholesterol oxidase from S. gilvosporeus, regulates natamycin biosynthesis via the peroxide-sensing transcriptional regulator OxyR and pathway-specific regulator SgnR.
Natamycin is a 26-membered tetraene macrolide antifungal antibiotic that can be produced by S. gilvosporeus, S. chattanoogensis (10), S. natalensis (14), and S. lydicus (18).These strains contain natamycin biosynthesis gene clusters with different arrange ments and members, but all of them contain a gene-encoding cholesterol oxidase, an enzyme that oxidizes cholesterol to produce H 2 O 2 (19).Natamycin can be oxidized by H 2 O 2 (35,36).This contradiction indicates that a precise regulatory network balances natamycin biosynthesis and ROS homeostasis in natamycin-producing strains.Crosstalk between ROS homeostasis and natamycin metabolism was first proposed for S. natalensis ATCC 27448 (34).Consistent with previous evidence (19,23), our results show that the cholesterol oxidase-encoding gene sgnE in the natamycin biosynthesis gene cluster is indispensable for natamycin biosynthesis in S. gilvosporeus.SgnE can oxidize at least three sterols, ergosterol, stigmasterol, and cholesterol (the main sterols in plants, fungi, and vertebrates, respectively) (6), to generate H 2 O 2 .In previous studies, the enzymatic activity of cholesterol oxidase was shown to be a key factor for natamycin production by deletion (19) or overexpression (24) analyses.Another study showed that no natamycin was produced when H 2 O 2 or cholestenone was addedto the culture broth of a choles terol oxidase gene mutant (23).However, according to our results, the concentration of H 2 O 2 added in those experiments (10 pM to 10 nM) was insufficient to activate natamycin production.Additionally, in some previous reports on natamycin production, cholesterol oxidase was unable to interact with sterols because the latter were absent from the PMM medium that was used (which lacks yeast extract or malt extract).However, PMM contains diverse metal ions such as copper, manganese, and molybde num, which can induce intracellular oxidative stress (59)(60)(61).
Cholesterol oxidase-encoding genes are also found in gene clusters for the biosyn thesis of small polyene macrolide antibiotics including tetramycin (20), filipin (21), and rimocidin/CE-108 (22).The biosynthesis of polyene macrolide antibiotics via gene clusters that contain a cholesterol oxidase-encoding gene may be a pathway to regulate H 2 O 2 homeostasis in the various strains that produce these compounds.Although H 2 O 2 is better known for its cytotoxic effects, it is also an important regulator of eukaryotic (25) and prokaryotic (62) signal transduction.Our data indicate that S. gilvosporeus functions as a sterol-degrading bacterium in the "sterol cycle" to generate H 2 O 2 , which then elicits natamycin biosynthesis.This may be a common characteristic among these strains although more research is required to confirm this.
When ROS levels exceed safe limits, bacteria mount an inducible response, resulting in increased expression of ROS-detoxification enzymes and protective systems that repair oxidative damage (62).Global regulators, such as sigma factors, two-component regulatory systems, SoxR, OxyR, OhrR, and CatR-CatA, appear to be the main factors that respond to ROS in Streptomyces (26,62).OxyR, a LysR family H 2 O 2 -sensing transcriptional regulator, is generally found in Gram-negative bacteria and also in a few Gram-positive bacteria (62).It is activated by disulfide bond formation between two cysteine residues.OxyR fulfills the definition of a peroxide receptor; in E. coli, it is activated by H 2 O 2 at concentrations that just exceed cellular physiological levels (20 nM) but is below the threshold toxic level (estimated at >1 µM) (26,41).In V. vulnificus, OxyR functions as a three-state redox switch that tightly regulates the expression of prx2, preventing the futile production of Prx2 in cells exposed to levels of H 2 O 2 that are sufficient to inactivate it.Bang et al. reported that there are reduced, oxidized, and overoxidized states of OxyR.Cys206 in OxyR Vvu1 and OxyR Vvu2 (equivalent to Cys212 in S. gilvosporeus OxyR) was identified as the sensing residue that is overoxidized to S-sulfonated cysteine (Cys-SO 3 H) by high H 2 O 2 concentrations in vitro and in vivo (42).The two cysteines of OxyR have the H 2 O 2 -sensing function in Streptomyces, and our experiments verified the different oxidation states of OxyR in S. gilvosporeus.We show that SgnE affects the production of natamycin via the OxyR oxidative stress-response system.Thus, we conclude that natamycin biosynthesis, a way to accumulate a reduced product, is one of the pathways to regulate H 2 O 2 homeostasis.
In E. coli, there are >20 target genes of OxyR, including genes involved in H 2 O 2 detoxification, heme biosynthesis, reductant supply, thiol-disulfide isomerization, Fe-S center repair, iron binding, and repression of iron and manganese import, as well as the small regulatory RNA OxyS (62).In Streptomyces, the target genes of OxyR include those encoding components of the alkyl hydroperoxide reductase system in S. coelicolor A3(2) (63), antioxidant enzymes, oxidative stress regulators (64), and the iron-response regulator IdeR in S. avermitilis (65).In the present work, we show that the DBD of OxyR can grasp the DNA of p sgnM .Together, our results show that the natamycin pathwayspecific transcriptional regulator gene sgnM is a target of OxyR in S. gilvosporeus and that the redox state of OxyR affects its binding to the promoter region of sgnM.The binding sequence of reduced and overoxidized OxyR, GGTCC-15 nt-CGGC-15 nt-GCCG-9 nt-GGACC, is similar to that of the ideR promoter region in S. avermitilis (65) and the "ATAGntnnnanCTAT-7 nt-ATAGntnnnanCTAT" sequences in the promoters of oxyRS, katG, and ahpC in E. coli and Salmonella typhimurium.According to previous results, the distance between DBD1 and DBD3 of OxyR in the reduced state is 120 Å, and this distance is shortened to 75 Å in the oxidized state; that is, the DBD1-DBD2 dimer and the DBD3-DBD4 dimer get closer by 45 Å following the transition from the reduced to the oxidized state (66).Thus, when OxyR is oxidized, it releases the "GGACC" motif and exposes the whole −35 region of the sgnM promoter in S. gilvosporeus.While, precise mechanism about OxyR regulation on sgnM and other ROS homeostasis targets should be verified by DNase I footprinting assay or protein-DNA complex crystallographic structural analysis in further studies.SgnM, SgnR, and their homologs are pathway-specific transcriptional regulators.They share a bidirectional promoter region and are encoded within their respective biosynthe sis gene clusters.Oxidation of (a) cysteine residue(s) in a protein prevents its alkylation by AMS (42,43).Alkylation of one cysteine residue by AMS increases the molecular mass of the protein by 0.5 kDa (and alkylation of two increases it by 1 kDa), which can be detected by SDS-PAGE.In reduced S. gilvosporeus OxyR, two cysteine residues can be alkylated by AMS.If only one cysteine residue is alkylated by this reagent, it indicates that one of the two cysteine residues in OxyR is already oxidized, indicative of the overoxi dized state of OxyR.The regulatory effect of OxyR on natamycin biosynthesis does not depend on its expression level, but on its oxidation state, because that affects its binding affinity to the target DNA.The oxidation state of OxyR is determined by the intracellular H 2 O 2 concentration.Thus, the DNA (promoter) binding affinity and binding coverage of different redox states of OxyR control the expression of sgnM and contribute to the dynamic regulation of natamycin biosynthesis in response to ROS in S. gilvosporeus.
The SgnR homolog PimR binds to a main operator that contains three heptameric direct repeats of the consensus CGGCAAG sequence with 4 bp spacers, which seems to be the DBD-binding pattern that is conserved among SARP regulators (56).PimR also binds to a secondary operator with two heptameric repeats of the consensus sequence separated by a 3 bp spacer.The optimum number of repeats of the heptamer recognized by PimR is three, and at least one triplet is required at the operator to activate pimM transcription (56).Unlike SARP regulators that bind DNA overlapping the −35 hexamer of target promoters, SgnR binds to a sequence located upstream of that hexamer (55).Surprisingly, we found that the OxyR DNA-binding sequence includes the two heptameric repeats that were predicted to be the SgnR-binding site.Protein-protein interaction analyses indicated that SgnR cooperates with OxyR to regulate the expres sion of sgnM.The formation of an SgnR-OxyR protein-protein complex can mask or unmask the −35 region of the sgnM promoter depending on the oxidation state of OxyR, ultimately regulating the expression of sgnM.
In summary, Streptomyces participates in the "sterol cycle" via cholesterol oxidase SgnE, whose coding gene sgnE is located in the natamycin biosynthesis gene cluster.SgnE is secreted and oxidizes sterols from cell walls to generate H 2 O 2 , which acts as a signaling molecule to regulate H 2 O 2 homeostasis.Oxidative stress is controlled via the natamycin biosynthesis pathway in S. gilvosporeus.The intracellular H 2 O 2 concen tration determines the oxidation state of the H 2 O 2 -sensing transcriptional regulator OxyR.By interacting with another transcriptional regulator SgnR, OxyR directly binds to the promoter region of the natamycin biosynthesis pathway-specific transcriptional regulator gene sgnM to regulate the expression of genes including sgnR, sgnK, sgnS2, sgnI, sgnJ, sgnA, sgnE, sgnS1, sgnD, sgnT, and sgnH in the natamycin biosynthesis gene cluster.SgnR-OxyR complexes by cooperation control natamycin biosynthesis by masking or unmasking the −35 region of the sgnM promoter depending on the oxidation state of OxyR (reduced, oxidized, or overoxidized), which responds to the intracellular H 2 O 2 concentration.On the basis of these findings, we propose a dynamic regulation mechanism of natamycin biosynthesis by Streptomyces in response to oxidative stress (Fig. 8).This work provides a novel perspective on the crosstalk between intracellular ROS homeostasis and natamycin biosynthesis.Application of these findings will improve antibiotic yields via control of the intracellular redox pressure in Streptomyces.

FIG 1
FIG 1 Influence of sterols on natamycin production.(A) Natamycin production by S. gilvosporeus with different sterols as substrates in the media (TH, control; TC, chopped sterols; TF, fertile sterols).(B-D) Volcano maps of differentially expressed genes in TC group compared with TH at 24, 60, and 120 h of culture, respectively.Red dots represent up-regulated genes, blue dots represent down-regulated genes, and gray dots represent non-differentially expressed genes.

FIG 3
FIG 3 Exogenously addition of H 2 O 2 restores natamycin production in the sgnE-deletion strain.(A) Schematic diagrams of construction of mutant strains with deleted sgnE gene (S.gilvosporeus DE13) and complemented sgnE gene (S.gilvosporeus CE13) in the S. gilvosporeus F607 background.F, forward primer; R, reverse primer.(B) PCR verification of DE13 and CE13 mutant strains.Lane 1, DE13; lane 2, CE13; lane 3, control F607.(C) Natamycin production by S. gilvosporeus F607 and its mutant strains DE13 and CE13 during fermentation in liquid medium.(D) Intracellular H 2 O 2 concentrations in S. gilvosporeus F607 and its mutant strains DE13 and CE13 during fermentation in liquid medium.In (C) and (D), results are the mean ± standard deviation (SD) from triplicate experiments.(E) Restoration of natamycin production by S. gilvosporeus DE13 with sgnE deletion by addition of H 2 O 2 .

FIG 4
FIG 4 H 2 O 2 -sensing transcriptional regulator OxyR controls expression of genes in natamycin biosynthesis gene cluster.(A) Amounts of reduced, oxidized, and overoxidized OxyR at different stages of fermentation.(B) Schematic diagrams of oxyR mutant construction.F, forward primer; R, reverse primer.(C) Natamycin production at indicated times in cultures grown in the liquid fermentation medium.Results are the mean ± SD from triplicate experiments.(D) Transcript levels of genes from each transcriptional unit in S. gilvosporeus (parental strain), DO7 (oxyR deletion mutant), and CO7 (oxyR complementation mutant).

FIG 7
FIG 7 OxyR regulates sgnM in cooperation with SgnR.(A) Sequence arrangement upstream of sgnM.Putative −10 and −35 hexanucleotides are underlined.Transcription start point is indicated by a bent arrow and bold-type letter.Box labeled "RBS" indicates ribosome-binding sites.Start codon is shown in a red box.Shaded box indicates SgnR-binding site on sgnM.OxyR-binding site inverted repeat motifs are shown in bold and labeled Motif 1, 2, 3, and 4. (B) Tetrameric structure of SgnR from S. gilvosporeus.DBD, BTA, random coil, ATP/GTP-binding domain TPR-like domain are shown in red, green, gray, blue, and carmine, respectively.(C) Protein-protein complex comprising OxyR (brown) and SgnR (blue-green) and interaction surface.(D) Protein-protein-DNA complex comprising OxyR (brown), SgnR (blue-green), and p sgnM .(E) proteins were co-immunoprecipitated by OxyR and SgnR antibodies with IgG as negative control.

FIG 8
FIG 8 Dynamical regulation mechanism of efficient natamycin biosynthesis in response to oxidative stress by Streptomyces.