Nitrogen oxide cycle regulates nitric oxide levels and bacterial cell signaling

Nitric oxide (NO) signaling controls various metabolic pathways in bacteria and higher eukaryotes. Cellular enzymes synthesize and detoxify NO; however, a mechanism that controls its cellular homeostasis has not been identified. Here, we found a nitrogen oxide cycle involving nitrate reductase (Nar) and the NO dioxygenase flavohemoglobin (Fhb), that facilitate inter-conversion of nitrate, nitrite, and NO in the actinobacterium Streptomyces coelicolor. This cycle regulates cellular NO levels, bacterial antibiotic production, and morphological differentiation. NO down-regulates Nar and up-regulates Fhb gene expression via the NO-dependent transcriptional factors DevSR and NsrR, respectively, which are involved in the auto-regulation mechanism of intracellular NO levels. Nitrite generated by the NO cycles induces gene expression in neighboring cells, indicating an additional role of the cycle as a producer of a transmittable inter-cellular communication molecule.

Scientific RepoRts | 6:22038 | DOI: 10.1038/srep22038 are produced in S. antibioticus to form NO 2 − under conditions that are different from those known in many other bacteria. Given that production of NO 2 − synchronized with rapid cell growth and was inhibited by glucose or glycerol, we suggested that the "NO 2 − -forming pathway" is an energy-producing metabolic reaction, although this conclusion requires further investigation.
This study investigated NO 2 − production by a model actinobacterium, S. coelicolor A3 (2). Although S. coelicolor A3(2) produces and excretes NO 2 − like S. antibioticus, the NO-producing mechanism is likely to be different from the latter since the S. coelicolor genome 11 does not encode a gene for NOS. In addition, no gene encoding the dissimilatory, NO-generating nitrite reductase was found in the genome. Here we demonstrated a NO-formation via NO 2 − produced from organic nitrogen, and homeostatic regulation of cellular NO in S. coelicolor. The endogenously formed NO is controlled by the nitrogen oxide cycle and acts as a signaling molecule for antibiotic production and morphological differentiation.

Results
S. coelicolor nitrogen-oxide cycle tunes endogenous NO concentration. We found that S. coelicolor A3 (2) M145 (M145) excreted NO 2 − into the medium when cultured with organic nitrogen as the nitrogen source under NO 2 − -producing conditions 6 (Fig. 1a). The NO 2 − _ production was also observed when M145 was grown on minimal medium containing L-asparagine as the sole nitrogen source (chemically defined medium) (supplementary Fig. 1), showing that M145 can convert organic nitrogen to NO 2 − . NO 2 − concentration increased for 72 h, indicating that S. coelicolor produces NO 2 − from organic nitrogen during the early vegetative growth phase. The S. coelicolor A3(2) has three Nar enzyme homologs (Nar1, Nar2, and Nar3 encoded by narGHJI, narG2H2J2I2, and narG3H3J3I3, respectively). We constructed mutants lacking combinations of the three narG paralogs, each of which encodes the catalytic subunits of Nar. (∆narG, ∆narG2, ∆narG3, ∆narG2/G3, and ∆narG/G2/G3). The wild-type M145 strain cultured under the NO 2 − -producing conditions showed Nar activity in cell membrane fractions (Fig. 1b). The ∆narG2/G3 and ∆narG/G2/G3 mutants showed drastic decrease in this activity (Fig. 1b) accompanied by decreased NO 2 − production (Fig. 1c). The ∆narG/G2/G3 mutant excreted NO 3 − instead of NO 2 − (Fig. 1a), indicating that Nar reduces NO 3 − to NO 2 − . Introducing the narG2H2J2I2 operon into the ∆narG/G2/ G3 mutant restored the NO 2 − -production (Fig. 1c). ∆narG mutant produced almost the same level of Nar activity as the wild type M145 strain (Fig. 1c). The result indicated that narG2 and narG3 but not narG are responsible for NO 2 − formation from organic nitrogen in the culture medium. Recently, Fischer et al. proposed that all three Nars are synthesized in S. coelicolor during the aerobic growth independent of the presence of NO 3 − , and they further indicated that narG is mainly working in spore, narG2 and narG3 are mainly working in mycelium, respectively. Thus, present results of narG2 and narG3 expression correspond with the results reported by Fischer et al. 12,13 .
Since NO 2 − as well as NO 3 − are produced as endogenously oxidative metabolites of NO, we investigated NO accumulation by using a fluorometric method and observed that NO production activity was present in M145 (Fig. 2a). The gene knockout of hmpA (SCO7472, encodes FhbA) (∆hmpA) accumulated more NO than did the M145 strain (Fig. 2a), and complementation with the hmpA gene (∆hmpA:: hmpA) attenuated NO accumulation to the same level as that in M145 (Fig. 2a). Western blot analysis detected the hmpA gene product (flavohemoglobin, Fhb) in the NO 2 − -producing cells (Fig. 2b), indicating that the NO dioxygenase activity of Fhb 8 led to lowered intracellular NO levels.
When grown under the NO 2 − -producing conditions, the ∆narG/G2/G3 cells accumulated little NO, whereas exogenous addition of NO 2 − to the cells restored the NO formation (Fig. 2a). This showed that NO was formed from NO 2 − , and that the ∆narG/G2/G3 cells could not form NO because they could not produce NO 2 − (Fig. 1a,c). The ∆narG/G2/G3 cells produced little Fhb, and addition of an NO donor (50 μ M NO 2 − ) to the medium restored its production (Fig. 2b). Transcription of hmp is known to be negatively regulated by NsrR 14,15 , which loses the ability to repress hmp upon exposure to NO. Disruption of nsrR in the ∆narG/G2/G3 cells also recovered the defect of FhbA production (Fig. 2c). These results are consistent with the notion that NO is formed from NO 2 − that is produced by Nar and derepresses Fhb production through NsrR transcription control. The present results demonstrate that Nar encoded by narG2 and narG3 reduces NO 3 − to NO 2 − , which is converted to NO, while Fhb oxidizes NO to NO 3 − . These enzymes constitute the metabolic cycle of nitrogen oxides, which can participate in the cellular processes without exogenous supply of either of the nitrogen oxides.
The expression of FhbA was observed in ∆narG2/G3 mutant only at early stage of growth (24 h) (Fig. 2b). This suggests that narG supported the expression of FhbA, confirming that narG acts during at spore germination stage and also confirms that narG2 and narG3 are working during vegetative growth phase to supply NO via NO 2 − . Endogenous NO promotes antibiotic production and regulates differentiation. Addition of NO donors to S. coelicolor cultures up-regulated transcription of redD, which encodes a positive regulator of undecylprodigiosin (Red) synthesis genes 16 and Red production (Fig. 3a). Direct supply of NO by NOC5 induced redD transcription more effectively than NO 2 − (Fig. 3a, left), which indicates that NO directly regulates redD transcription. The ∆hmpA strain produced more Red than did the M145 strain (Fig. 3b) whereas ∆narG2 and ∆narG/G2/G3 mutants produced little Red (Fig. 3c) in accordance with the roles of Fhb and Nar in sequestering and producing NO, respectively. Addtion of NO donor (100 μ M NO 2 − ) restored the Red production by the ∆narG/G2/G3 mutant (Fig. 3c), showing that NO 2 − production from NO 3 − by Nar is necessary for NO formation. Higher concentration of NO 2 − increased Red production by the M145 and ∆narG/G2/G3 strains, but redD gene knockout confirms no NO-dependent Red production (Fig. 3a, right). These results indicated that the bacterial Nar and FhbA control cellular NO levels and regulate Red production. Moreover, the ∆narG/G2/G3 mutant developed aerial mycelia after cultivation for 120 h, which is much earlier than differentiation of aerial mycelia in Scientific RepoRts | 6:22038 | DOI: 10.1038/srep22038 M145. Addition of 100 μ M NO 2 − to the medium recovered this phenotype (Fig. 3c), indicating that endogenously generated NO delays aerial mycelia development.

DevSR controls nitrogen oxide cycle and cellular NO levels.
DevS is a heme-containing, NO-sensing histidine kinase 17,18 that transduces signals to a transcriptional regulator DevR, constituting a two-component system (TCS) with DevS. Deletion of the S. coelicolor orthologs devS (SCO203) or devR (SCO204) (Supplementary Fig. 2) decreased transcripts of narG2 in S. coelicolor along with reducing Nar activity (Fig. 4a,b). Both the deletion mutants produced low levels of Red as observed in the ∆narG2 mutant and were recovered by a high concentration of NO 2 − (Fig. 4c). In addition, DevR (rDevR) protein could bind to the upstream region of narG2 in vitro (Fig. 4d). These results revealed that the S. coelicolor DevSR regulates the expression of narG2 and are integral components in the NO-forming pathway.
In vitro phosphorylation of rDevS, induced by its autokinase activity (Fig. 5a), was decreased in the presence of more than 1 μ M NOC5. Exposure to high concentrations of exogenous NO 2 − or NOC5 also inhibited cellular transcripts of narG2 ( Fig. 5b) and Nar activity ( Supplementary Fig. 3). These results indicate that NO negatively regulates the DevSR TCS and transcription of the narG2 operon. The ∆hmpA mutant which accumulated more NO in the cell (Fig. 2a) excreted less NO 2 − into the medium after cultivation for 48 h (Fig. 5c), and the defect was complemented by introduction of hmpA gene (Fig. 5c). This observation supports the inhibitory effect of excess intracellular NO on narG2 expression since NO 2 − is a product of Nar reaction. It can be concluded from these results that the DevSR TCS system regulates the concentration of endogenous NO by controlling the expression of the nar2 gene cluster and that NO itself acts as the negative regulator depending on its intracellular concentration.

NO 2
− is an intercellular signaling molecule. When the ∆narG/G2/G3 strain was cultured on a plate in which a single colony was surrounded by eight colonies of the parent strain M145, ∆narG/G2/G3 that had lost its Fhb-producing ability ( Fig. 2) began to produce FhbA again (Fig. 6). This indicated that excreted NO 2 − or NO derived from NO 2 − acts as a signaling molecule for communication between cells, and confirmed that NO is the end product of the nitrogen oxide cycle and is a hormone-like molecule. We found that Red-producing ability of the ∆narG/G2/G3 mutant strain was not restored ( Supplementary Fig. 4), probably because NO 2 − transmitted to the mutant could not supply sufficient NO to trigger Red synthesis.

Discussion
This study proposed a nitrogen oxide cycle that regulates cellular NO levels in S. coelicolor and its underlying metabolic and morphogenic mechanisms (Fig. 7). The homeostatic regulation of NO in cells is crucial to understanding the complex life of organisms. To date, endogenous production of NO without any exogenous nitrogen species is known to be achieved by transiently controlled production of NO synthase 19 . The S. coelicolor mechanism is unique in that Nar and Fhb play key roles in NO homeostasis. Conventional roles ascribed to Nar and Fhb are dissimilation of NO 3 − and detoxification of NO, respectively, both of which are mechanisms for responding to exogenously added nitrogen oxides in most bacteria. Recent studies showed a similar Nar-dependent NO 2 − production in human pathogen bacterium, Mycobacterium tuberculosis (Mtb). The NO 2 − production is believed to be a system for survival of Mtb in host 20,21 . We here demonstrated the role of Nar and Fhb in balancing the levels of cellular NO to control cell signaling in the proposed nitrogen oxide cycle mechanism (Fig. 7). The mechanism  of S. coelicolor does not require exogenous nitrogen oxides, which highly suggests that it is not a mechanism for environmental responses, but is a constitutive housekeeping one.
In this study, we could not identify the mechanism involved in the production of NO 3 − in S. coelicolor. Our results indicated that NO 2 − and NO were generated after the production of NO 3 − (Fig. 2), indicating that NO 3 − is the first nitrogen oxide produced among the three nitrogen oxide species (NO 3 − , NO 2 − , and NO) as the precursor of NO 2 − and NO. To date, few NO 3 − producing enzymes are known except for the enzymes that convert NO and NO 2 − to NO 3 − , thereby, indicating that NO 3 − generated in S. coelicolor was produced by some unidentified metabolic pathway. Therefore, we propose an extremely interesting topic regarding the production of NO 3 − in S. coelicolor.
While NOS and NO-generating nitrite reductase were previously known as the only enzymes to produce NO, recent studies have disclosed several other NO generation mechanisms. Nitric oxide generation from NO 2 − is known to depend on ether enzymatic or nonenzymatic reaction in bacterial cells. It was suggested that Escherichia coli and Salmonella typhimurium produce NO from NO 2 − by periplasmic cytochrome nitrite reductase 22 and Nar 23 , respectively. Moreover, other heme-containing [24][25][26] proteins (such as hemoglobins and NOS) and molybdenum proteins 27,28 (such as aldehyde oxidase and xanthine oxidase) were shown to convert NO 2 − to NO. Further, nonenzymatic formation of NO from NO 2 − is also known 29,30 . Here we demonstrated the NOS-independent NO production in S. coelicolor (Fig. 2), its mechanism is to be elucidated.
Only links have been suggested between NsrR or DevSR and tolerance against stress 8,31-33 . In S. coelicolor, both proteins up-and down-regulate Fhb and Nar gene expressions respectively, to control cellular NO level in response to endogenously produced NO in an auto-regulation mechanism (Fig. 5). Despite NO production from early stages of culture, this mechanism can explain that NO performed as a signaling molecule at the later stage. Disruption of the genes devS or devR had a significant influence on secondary metabolism (Fig. 4c), indicating high stringency of this regulation system and the importance of endogenous NO. Generally, bacterial cells synchronously start new metabolic processes including secondary metabolism, which is called as quorum sensing 34 . Here, the excreted NO 2 − (or NO derived from NO 2 − ) influenced FhbA production by neighboring cells (Fig. 6) showing that S. coelicolor cells communicate with one another via extracellular NO 2 − or NO derived from NO 2 − . Intracellular and extracellular NO 2 − are in equilibrium thus when extracellular NO 2 − concentrations increase, endogenous NO can overcome FhbA, and the cell can start the new metabolic process which can shared between all cells. Thus, NO 2 − acts as an autoinducer in quorum sensing. So it should be NO 2 − that is also the purported signaling molecule.
Furthermore, we found a NO 2 − removal system (Fig. 5c). This NO homeostatic regulation system can explain the gradual decrease of accumulated NO 2 − (Fig. 1a). Thus, the NO 2 − removal system may be important not only to regulate NO homeostasis but also to complete the NO homeostatic regulation system in S. coelicolor. However, this notion needs to be proven by additional genetic studies.
The identification of NO as a signaling molecule in Streptomyces bacteria and the novel regulation system now allows us to take a step towards a better understanding of the regulation of synthesis of biologically active agents in the producer. In fact, our results show that the production of the antibiotic Red drastically increases depending on the concentration of exogenous NO (Fig. 3). The regulation of NO homeostasis in accordance with various systems continues to be an important subject for further investigation in all organisms to provide a new perspective on NO biology and to contribute towards human welfare.

Methods
Bacterial strains, plasmids, and culture conditions. Strains used in this study are listed in Supplementary Table S1. Streptomyces coelicolor A3(2) M145 strain (wild-type) was obtained from the John Innes Centre, UK. Mannitol soya flour agar (2% mannitol, 2% soya flour, 2% agar) was used for sporulation, and the strains were routinely grown on YEME-gln (glutamine) solid medium [0.3% yeast extract, 0.5% Bacto-peptone, 0.3% malt extract, 1% glucose, 50 mM L-glutamine (pH 7.2)] at 30 °C. The medium contained no detectable (by ion chromatography or colorimetric analysis) amount of nitrate or nitrite. Minimal solid medium (0.05% L-asparagine, 0.05% K 2 HPO 4 , 0.02% MgSO 4 · 7H 2 O, 0.001% FeSO 4 · 7H 2 O, 1% Glucose, 2% agar) was used for detection of NO 2 −_ production ability in M145. E. coli DH5α (Takara, Kyoto, Japan) was used as the host for routine cloning. Media, culture conditions, and DNA manipulations for Streptomyces and E. coli were performed as described by Kieser et al. 35 and Green and Sambrook 36 , respectively. Media and culture conditions for the strains used for gene disruption followed a protocol of the REDIRECT PCR-targeting method 37 . E. coli HST04 dam-/dcm-(Takara) was used as a non-methylating cosmid and plasmid donor strain. E. coli Origami 2 (DE3) was used for recombinant DevS and DevR production. The plasmids, cosmids (kindly provided by the John Innes Centre, UK), and primers used for this study are listed in Supplementary Tables 2 and 3.

Construction of mutants and complementation. The open reading frames in the chromosomes
were replaced with drug resistance cassettes by using REDIRECT PCR targeting 37 . Each drug resistance cassette flanked by Flippase recognition target (FRT) sites was amplified by PCR using PrimeSTAR GXL DNA polymerase (Takara), and each primer set is listed in Table S2. To obtain a target gene-disrupted version of the mutant cosmids by the λ Red system, amplified cassettes were introduced into E. coli BW25113/pIJ790 37 harboring an appropriate cosmid (Supplementary Table 2). The resulting construct was confirmed by PCR and introduced into E. coli HST04 dam-/dcm-(Takara) to obtain a non-methylating cosmid, and each mutated cosmid was introduced into S. coelicolor A3(2) M145 or its derivatives by protoplast transformation. Drug-resistant recombinants (Supplementary Table 1) were screened, and successful recombination was checked by PCR using appropriate primer sets and a complementation study.
To obtain a marker-less mutant, the drug-resistance cassette was eliminated from the corresponding disrupted cosmid by introduction into E. coli strain BT340 in which recombination between both FRT mutagenesis cassette-flanking regions was induced by Flippase. In these new cosmids, only 81 base pairs (SCAR) remained in frame with the adjacent ORFs. Each resulting cosmid was introduced into the corresponding mutant and then the drug-sensitive mutant was screened and the replacement of drug-resistant cassette with SCAR was checked by PCR, using the appropriate primer set. Integration plasmids, pTYM19-narG2H2J2I2, pKU460-nsrR-hmpA, and pKU460-devS-R, used for genetic complementation of knockout mutants, were prepared in the following manner. Spores of M145 or narGs mutant were inoculated onto the YEME-gln plate (a total of 9 spots) and incubated for 48 h. Only cells growing in the center spot were collected and subjected to western blotting to detect the production of Fhb. The M145/M145, M145 strain was inoculated at all 9 spots; ∆narGG2G3/M145; M145 was inoculated in all spots except the center spot in which the ∆narG/G2/G3 mutant was inoculated; the ∆narGG2G3/∆narGG2G3, ∆narG/G2/G3 mutant was inoculated at all spots. Each gene coding region, narG2H2J2I2 (SCO0216-0219), nsrR-hmpA (SCO7427-7428), and devS-R (SCO0203-0204), containing each upstream region (150-300 bp), was amplified by PCR, using the primer sets listed in Supplementary Table 3. The resulting narG2H2J2I2 fragment was cloned into the HinDIII site of pTYM19 38 . The nsrR-hmpA and devS-R fragments were cloned into the EcoRI/HinDIII site of pKU460 39 , respectively, and the resulting plasmids were introduced into each disruptant.

Determination of NO 3
− and NO 2 − . NO 3 − was determined by ion chromatography, using a 761 Compact IC (Metrohm) 6 . NO 2 − was determined by Griess reagent assay 40 . Both nitrogen oxides were extracted from the medium for determination. Cells were grown for the indicated periods on a cellophane membrane covering the surface of a YEME-gln agar plate. After cultivation, cells on the cellophane were removed, five blocks of 1 × 1 cm were cut out from separate positions on the plate, and the blocks were combined and homogenized in 5 ml distilled water. After centrifugation, the supernatant was further filtered through a 0.45-μ m cellulose acetate filter and subjected to determination of nitrogen oxides in the culture medium. The determination depended on a standard curve made using medium containing 0, 1, 10, 25, 50, and 100 μ M NO 2 − or NO 3 − .
Enzyme assay. Nar activity was assayed using dithionite/methylviologen as an electron donor, as previously In situ detection of NO. Strains were grown at 30 °C on YEME-gln. Endogenously formed NO was detected using DAF-2DA (Dojindo) as described previously 6 . Photographs were taken with excitation at 495 nm and emission at 515 nm using FLUOVIEW FV300 System (Olympus).
Isolation of total RNA and qPCR. Total RNA was isolated using RNeasy Kit (Qiagen) from strains grown on cellophane-covered solid medium under several culture conditions, according to the manufacturer's instructions. Conditions for each culture are given in the figure legends. cDNA was generated using a PrimeScript ® RT reagent Kit with gDNA Eraser (Takara) and served as a template for qPCR. The primers used for qRT-PCR are indicated in Supplementary Table 3. qPCR was performed in a Thermal Cycler Dice Real Time System (Takara). PCR mixture (total 25 μ l) contained 0.1 μ g of generated cDNA, 10 pmol of an appropriate primer set (Table S3), and SYBR ® Premix Ex Taq TM II (Takara). The hrdB gene of S. coelicolor was used as an internal control 41 .
Determination of RED. Undecylprodigiosin (RED) was determined as described 42 with some modifications. Spores were inoculated onto a cellophane-covered YEME-gln plate at intervals of 1 cm (total of 45 spots) with a toothpick. After culture at 30 °C for 48 h, cells on cellophane were transferred to a fresh YEME-gln medium plate containing 0, 100, 500, or 1000 μ M NO 2 − . Before the transfer, 300 μ l distilled water was added onto the plate to allow tight contact of cellophane with the agar plate. Cells were incubated at 30 °C for 48 h and then collected and submitted for the determination of RED production. To remove the blue-pigmented antibiotic actinorhodin (ACT) from the cells, 1 M KOH was added and after incubation at 25 °C for 1 h, cells were centrifuged at 8,000 × g for 10 min, and the supernatant containing ACT was removed. For RED, the cells were washed twice with 0.9% NaCl after ACT extraction. The resulting pellet was extracted with methanol (pH 2.0, adjusted with HCl) overnight at 25 °C, followed by centrifugation at 8,000 × g for 5 min, and absorbance at 530 nm was measured. A molecular extinction coefficient of ε 530 = 100,500 M −1 cm −1 was used for the determination of RED.
Overexpression and purification of recombinant proteins. hmpA (SCO7428) encoding FhbA, devR (SCO7428), and devS (SCO7428) genes were amplified with primeSTAR GXL DNA polymerase, using primers listed in Supplementary Table 3. Each amplified gene was treated with NdeI and EcoRI and cloned into the corresponding sites of pET28b. Each expression construct was then transformed into E. coli Origami 2(DE3). For the overexpression of proteins, 1% of each overnight culture was inoculated into a fresh LB medium (1000 ml) containing 50 μ M kanamycin, and after 1 h growth at 37 °C, IPTG [0.2 mM (for FhbA and DevR) or 1 mM (for DevS)] was added to the medium. Each culture was then incubated at 30 °C (FhbA) or 16 °C (DevR and DevS) for 24 h. Grown cells were harvested and resuspended in Tris buffer [20 mM Tris-HCl (pH 8.0); containing 300 mM NaCl, 0.1 mM DTT, 20 mM imidazole, and 10% glycerol] and broken by sonication on ice. After centrifugation at 4 °C, the supernatant was loaded onto a Ni-NTA Agarose column (Qiagen) and washed with the same buffer. Each target protein was eluted with Tris buffer containing 500 mM imidazole. Purified FhbA was used to produce rabbit polyclonal antibody.
Other analytical methods. Cellular protein was determined using a protein assay reagent (Bio-Rad) after homogenization of cells with an ultra-sonicator as described previously 6 .