We previously demonstrated different spacial expression profiles of the neuronal nitric oxide (NO) synthase (nNOS) splice variants nNOS-µ and nNOS-α in the brain; however, their exact functions are not fully understood. Here, we used electron paramagnetic resonance to compare the electron-uncoupling reactions of recombinant nNOS-µ and nNOS-α that generate reactive oxygen species (ROS), in this case superoxide. nNOS-µ generated 44% of the amount of superoxide that nNOS-α generated. We also evaluated the ROS production in HEK293 cells stably expressing nNOS-α and nNOS-µ by investigating these electron-uncoupling reactions as induced by calcium ionophore A23187. A23187 treatment induced greater ROS production in HEK293 cells expressing nNOS-α than those expressing nNOS-µ. Also, immunocytochemical analysis revealed that A23187-treated cells expressing nNOS-α produced more 8-nitroguanosine 3′,5′-cyclic monophosphate, a second messenger in NO/ROS redox signaling, than did the cells expressing nNOS-µ. Molecular evolutionary analysis revealed that the ratio of nonsynonymous sites to synonymous sites for the nNOS-µ-specific region was higher than that for the complete gene, indicating that this region has fewer functional constraints than does the complete gene. These observations shed light on the physiological relevance of the nNOS-µ variant and may improve understanding of nNOS-dependent NO/ROS redox signaling and its pathophysiological consequences in neuronal systems.
Nitric oxide (NO) mediates signal transduction in various processes, in which nitric oxide synthase (NOS) mediates NO production [1,2]. Typically, NOSs utilize NADPH as an electron donor, and accepted electrons are transferred through a P450 reductase-like domain to a heme-containing oxidase domain, where l-arginine is oxidized to produce NO and l-citrulline . The major NO signaling pathway is considered to be mediated by activation of soluble guanylate cyclase (sGC), which leads to production of guanosine 3′,5′-cyclic monophosphate (cGMP) . However, we recently found that a nitrated derivative of guanosine cyclic monophosphate [8-nitroguanosine 3′,5′-cyclic monophosphate (8-nitro-cGMP)] is formed endogenously under conditions of excess NO production coupled with reactive oxygen species (ROS) production [5–7]. 8-Nitro-cGMP is an extremely potent signaling molecule in biological systems because of its dual roles in signal transduction, i.e. in the canonical NO/cGMP pathway and in the noncanonical electrophilic signaling, including unique post-translational modification and protein S-guanylation [5,6,8–14]. It is now clear that ROS functions as not only a toxicant but also a signaling molecule; therefore, 8-nitro-cGMP is considered a downstream messenger of NO/ROS redox signaling [5,6,8–14].
Although NOSs were identified as the enzymes responsible for NO synthesis [1,2], all NOS isoforms including nNOS have been known to produce superoxide in addition to NO because of the naturally occurring electron uncoupling of the enzymes [13,15–18]. Under certain conditions, NADPH oxidation takes place independently of NO generation, and uncoupled electrons are transferred to molecular oxygen, which is then reduced to superoxide . For instance, nNOS-α utilizes 50–80% of the electrons from NADPH for NO generation and the remaining electrons for generation of ROS, even in the presence of l-arginine . Superoxide, via the uncoupling reaction, is reportedly enhanced through the P450 reductase-like domain, by endogenous redox-active or electrophilic compounds such as 8-nitroguanosine [20,21] and 8-nitro-cGMP . Although the ROS generation from NOSs appears to contribute to the functional diversity of NOSs, the exact mechanism of nNOS uncoupling as related to ROS production and the pathophysiological significance of ROS produced by nNOS remain unexplained.
One of the important processes regulating nNOS expression and function may be formation of its translational splice variants — nNOS-α, nNOS-β, nNOS-γ, nNOS-µ, and nNOS-2 [22–25]; the principal isoform expressed in the brain is nNOS-α. Although nNOS-µ was first detected in skeletal muscles and the heart , we later identified the expression of nNOS-µ in rat brain . In the cerebellum, the expression pattern of nNOS-μ differed from that of nNOS-α in that it occurred predominantly in granule cells, which suggests that nNOS-μ may play unique roles in different neurons . However, the functional differences between nNOS-α and nNOS-µ have not yet been fully characterized.
In the present study, to elucidate the physiological and pathological functions of nNOS-µ, we investigated the generation of nNOS-μ-dependent superoxide. We used electron paramagnetic resonance (EPR) to analyze superoxide production induced by purified recombinant nNOS-α and nNOS-µ. We also studied ROS production induced by calcium ionophore A23187 in HEK293 cells that stably expressed nNOS-α and nNOS-μ, by which we demonstrated formation of the nitrated second messenger 8-nitro-cGMP, a downstream messenger of NO/ROS signaling, mostly in cells expressing nNOS-α.
Materials and methods
8-Nitro-cGMP was synthesized as described previously . Horseradish peroxidase (HRP)-conjugated anti-nNOS and anti-nNOS-µ monoclonal antibodies , and anti-8-nitro-cGMP and anti-S-guanylated protein antibodies  were prepared as described previously. Anti-sGC antibody was obtained from Cayman Chemical (Ann Arbor, MI, U.S.A.). P450 reductase was prepared and purified according to the literature . Bovine brain calmodulin (CaM), bovine serum albumin (BSA), Cu,Zn-superoxide dismutase (SOD), tetrahydrobiopterin (BH4), RPMI 1640 medium, anti-actin polyclonal antibody, and HRP-conjugated antirabbit secondary antibody were purchased from Sigma–Aldrich (St Louis, MO, U.S.A.). 2,3-Diaminonaphthalene (DAN), NG-nitro-l-arginine methyl ester (l-NAME), and 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) were from Dojindo Laboratories (Kumamoto, Japan). 5-(2,2-Dimethyl-1,3-propoxycyclophosphoryl)-5-methyl-1-pyrroline-N-oxide (CYPMPO) was obtained from Radical Research, Inc. (Tokyo, Japan). 2′,5′-ADP-Sepharose and Cy3-labeled goat antimouse IgG antibody were purchased from GE Healthcare (Little Chalfont, U.K.). Catalase, hemoglobin, and protease inhibitor cocktails were from Nacalai Tesque (Kyoto, Japan). 7-Nitroindazole (7NI) was obtained from Tocris (Ellisville, MO, U.S.A.). Dulbecco's modified Eagle's medium (DMEM) was purchased from Wako Pure Chemical Industries (Osaka, Japan). Fetal bovine serum (FBS) was obtained from MultiSer (Cytosystems, Castle Hill, NSW, Australia). Dihydroethidium (DHE) was from ABD Qioquest (Sunnyvale, CA, U.S.A.). All other chemicals and reagents were from common suppliers and were of the highest grade commercially available.
nNOS preparation and purification
cDNAs encoding nNOS-α and nNOS-µ  were cloned into the pCW vector. Recombinant rat nNOS-α and nNOS-µ were expressed in Escherichia coli BL21 cells and purified by using 2′,5′-ADP-Sepharose, as described previously . Enzyme preparation was >95% pure, as judged with sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Protein concentrations were determined via the Bradford method with BSA as the standard .
NO synthesis was detected by using a spectrophotometric assay in which the conversion of oxyhemoglobin to methemoglobin was monitored at 401 nm (extinction coefficient, 39 mM–1 cm–1) [13,30]. Unless otherwise stated, assays were carried out at 25°C in 50 mM sodium phosphate buffer (pH 7.4) containing 10 µM oxyhemoglobin, 0.1 mM NADPH, 10 µg/ml CaM, 1 mM CaCl2, 100 units/ml catalase, 10 units/ml SOD, 5 µM BH4, varying concentration of l-arginine, and 31 nM nNOS. The rate of NADPH oxidation was determined spectrophotometrically, in the absence of oxyhemoglobin, as a decrease in absorbance at 340 nm with an extinction coefficient of 6.22 mM–1 cm–1 [13,30].
nNOS activity in HEK293 cells was measured by the DAN assay, as described previously [13,14,31]. To study NO generation under the condition of Ca2+-dependent nNOS activation, cells were stimulated with or without A23187 (5 µM) for 6 h in the presence or absence of NOS inhibitor l-NAME (0.2 mM) at 37°C. The nitrite levels in the culture supernatants were measured using DAN. The cells were washed three times with PBS and lysed in PBS containing 0.1% Triton X-100. Lysate was assayed for protein concentration by the Bradford method.
EPR detection of superoxide
Superoxide generation induced by nNOS-α, nNOS-µ, and P450 reductase was quantified via EPR spectroscopy [5,13,21]. The reactions catalyzed by nNOS-α and nNOS-µ were initiated by adding NADPH (final concentration, 0.2 mM) to reaction mixtures containing 50 nM enzymes, 0.5 mM Ca2+, 10 µg/ml CaM, 0.1 mM diethylenetriaminepentaacetic acid (DTPA), and 45 mM DMPO or 2 mM CYPMPO in 50 mM sodium phosphate buffer (pH 7.4) at room temperature. For EPR measurement of 7NI-induced superoxide production from P450 reductase, the reaction mixture contained 50 nM P450 reductase, 0.1 mM DTPA, 0.2 mM NADPH, and 45 mM DMPO with 1 mM 7NI in 50 mM sodium phosphate buffer (pH 7.4). This reaction mixture was left to sit undisturbed for 2 min, after which EPR spectroscopy was performed. Spin adducts of superoxide generated by the NOSs were identified and quantified by using an X-band EPR spectrometer (JES-RE1X; JEOL, Tokyo, Japan). The DMPO adduct of superoxide (DMPO-OOH) was quantified by normalizing the intensity of the EPR signal obtained for DMPO-OOH to that of the standard reference signal of manganese oxide. We used CYPMPO in addition to DMPO because the CYPMPO-OOH adduct shows much better stability than does DMPO-OOH under present spin trapping conditions . For CYPMPO, EPR measurement of CYPMPO-OOH (the CYPMPO adduct of superoxide) was repeated five times, and these spectra were averaged to obtain the signal. Spectra were recorded at room temperature under the following conditions: modulation frequency, 100 kHz; modulation amplitude, 0.079 mT; scanning field, 335 ± 5 mT; receiver gain, 250; response time, 0.1 s; sweep time, 2 min; microwave power, 10 mW; and microwave frequency, 9.421 GHz.
Production of stable transformants
Rat nNOS-α and nNOS-µ were stably transfected into HEK293 cells by using a ViraPower Lentiviral Expression System (Thermo Fisher Scientific, Waltham, MA, U.S.A.) . In brief, the nNOS cDNAs  were subcloned into the pENTR vectors and were then transferred into the pLenti6/V5-DEST vector by using the LR Clonase enzyme mix kit (Thermo Fisher Scientific). The resultant plasmids and a plasmid containing the LacZ gene (control vector) were co-transfected into HEK293FT cells with ViraPower packaging mix. The transfected cells were cultured for 3 days, after which culture supernatants were collected and added to the HEK293 cells. These cells were then cultured at 37°C for 12 days in the presence of blasticidin (1 µg/ml), and surviving cells that expressed nNOSs at appropriate levels were used for subsequent analyses. Selected HEK293 cells were then cultured in DMEM containing 10% FBS at 37°C in a humidified atmosphere of 5% CO2 and 95% air. nNOS expression was confirmed by western blotting with monoclonal antibodies specific for total nNOS (recognizing all the isoforms of nNOS including nNOS-α and nNOS-µ) and for nNOS-µ .
Measurement of intracellular ROS generation
Intracellular ROS generation was analyzed by using fluorescence microscopy  and fluorescence-HPLC (high-performance liquid chromatography)  using the oxidant-sensitive fluorescent probe DHE. For fluorescence microscopy, cells were seeded on 8-well coverglass chamber (AGC Techno Glass, Shizuoka, Japan) and treated with or without A23187 (5 µM) for 30 min in the presence or absence of l-NAME (0.2 mM) at 37°C. Cells were then washed twice with PBS and incubated in 10 µM DHE at 37°C for 30 min, and observed with a confocal microscope (FV1200, Olympus, Tokyo, Japan). For fluorescence-HPLC, the unique product (2-hydroxyethidium, 2-OH-E+) formed from the reaction between superoxide and DHE in cells was extracted and resolved by HPLC equipped with a fluorescence detector. After treatment with A23187, HEK293 cells were washed twice with PBS and incubated in 10 µM DHE at 37°C for 30 min. Thereafter, cells were washed three times with PBS and lysed in PBS containing 0.1% Triton X-100. Lysate was assayed for protein concentration by the Bradford method. n-Butanol was then added to cell lysates and mixed and further incubated on ice for 10 min. After centrifugation, the top layer of butanol containing 2-OH-E+ and ethidium (E+) was dried under vacuum, dissolved in 0.1% formic acid, and analyzed by HPLC using a C18 reverse-phase column (Mightysil RP-18; Kanto Chemical, Tokyo, Japan) with a fluorescence detector (excitation at 510 nm and emission at 595 nm). The mobile phase was 40% methanol in 0.1% formic acid and was delivered at a flow rate of 0.5 ml/min.
Cells were seeded on circular coverglasses and treated with A23187 at 37°C for 6 h. Immunostaining was achieved by using a monoclonal mouse antibody specific for 8-nitro-cGMP (clone 1G6), as reported in our recent publications [5,13,14], and staining was then observed with a confocal microscope (FV1200, Olympus). Adobe Photoshop v. 7.0 (Adobe Systems, Waltham, MA) was used for additional image processing and quantification.
Comparison of human and rat nNOS sequences
Sequences of human nNOS (GenBank accession numbers NM_000620 and AJ 004918) and rat nNOS (GenBank accession number NM_052799) were compared via Genetyx software (Genetyx, Tokyo, Japan). The rate of synonymous and nonsynonymous nucleotide substitutions was estimated by using the method of Miyata and Yasunaga .
All experiments were performed at least three times. Values for individual experiments are presented as means ± SE. Statistical significance was determined with one-way analysis of variance (ANOVA) or Student's paired t-test, as indicated, by using the GraphPad Prism program (GraphPad, Inc., San Diego, CA).
Activity of purified recombinant rat nNOS-µ and nNOS-α
To analyze nNOS-µ- and nNOS-α-dependent superoxide production, we first examined the activity of purified recombinant rat nNOS-µ and and nNOS-α expressed in E. coli cells. The recombinant nNOSs were purified to near homogeneity from E. coli cells by affinity chromatography (Figure 1A). With regard to NO production, the Km values of l-arginine for nNOS-α and nNOS-µ were 3.8 and 3.0 µM, and the Vmax values for nNOS-α and nNOS-µ were 199.1 and 203.5 nmol/min/mg, respectively (Figure 1B). Thus, nNOS-µ activity was almost identical with that of nNOS-α for NO production. These results are consistent with previous findings . The initial rates of NADPH oxidation in the reaction with nNOS-α or nNOS-µ were 2066.1 and 896.1 nmol/min/mg protein in the absence of l-arginine, respectively; these values reduced with the micromolar level of l-arginine (0–10 µM; Figure 1C). These results indicate that micromolar levels of l-arginine are critical for regulating NO production and NADPH oxidation. The oxidation rate of nNOS-µ was significantly slower than that of nNOS-α, regardless of l-arginine concentration. These data suggest that nNOS-α uncouples more than nNOS-µ in the presence of l-arginine.
NO formation and NADPH oxidation by nNOS-α and nNOS-µ.
Generation of superoxide by recombinant rat nNOS-µ and nNOS-α
To determine whether nNOS-µ and nNOS-α generate superoxide, we performed EPR spin trapping studies with DMPO as the spin trap (Figure 2). Several earlier studies demonstrated that nNOS-α catalyzed superoxide formation [13,15,35]. In the present analysis, DMPO-OOH spectra, which indicate superoxide trapping [13,20,21], were detected in reactions with both nNOS-µ and nNOS-α in the absence of l-arginine (Figure 2A). The addition of SOD to the reaction mixture nullified these signals (Figure 2A), which indicated that superoxide was the primary radical responsible for the EPR signals observed. No signals were detected in the full reaction mixture in the absence of enzyme (data not shown). The amount of superoxide generated in the reaction with nNOS-µ was ∼44% of that generated in the reaction with nNOS-α (Figure 2B). Superoxide adduct generation induced by nNOS-µ, similar to that induced by nNOS-α, was attenuated by the nNOS substrate l-arginine and the nNOS cofactor BH4 (Figure 2A).
Superoxide production by nNOS-α and nNOS-µ.
Superoxide production by nNOS-µ and nNOS-α was inhibited when KCN was added to the reaction mixture (Figure 3A). Because the cyanide anion binds tightly to heme and suppresses uncoupling at the oxygenase domain , this finding indicated that the oxygenase domains of nNOS-α and nNOS-µ were responsible for superoxide formation. Moreover, the presence of an l-arginine analog, the NOS inhibitor l-NAME, stopped superoxide production by nNOS-α and nNOS-µ (Figure 3A). l-NAME also inhibits oxygenase domain-dependent electron uncoupling , so this result supported the idea that the oxygenase domains of nNOS-α and nNOS-µ were responsible for superoxide generation.
Effects of NOS inhibitors on superoxide production by nNOS-α and nNOS-µ.
However, treatment with another NOS inhibitor, 7NI, which is thought to compete for the l-arginine-binding site in NOS [36,37], enhanced superoxide production by nNOS-α and nNOS-µ (greater for nNOS-α than for nNOS-µ), in the absence of l-arginine (Figure 3A). Even in the presence of l-arginine (Figure 3B), however, 7NI enhanced superoxide production by both nNOSs (greater for nNOS-α than for nNOS-µ; data not shown). Because a 10-fold molar excess of l-arginine did not inhibit 7NI-induced superoxide production, we speculated that the site of action of 7NI differed from that of l-arginine. In fact, 7NI markedly enhanced superoxide formation from P450 reductase (Figure 3C); this finding strongly suggested that a redox-active property of this inhibitor serves to uncouple electrons for P450 reductase.
This unexpected redox-active behavior of 7NI was similar to the action of 8-nitro-cGMP, which we previously identified as a novel second messenger of redox signal and that possessed redox-active properties . Similar to 7NI, we had previously reported the redox-active property of 8-nitro-cGMP as an electrophile and shown that 8-nitro-cGMP stimulates superoxide production via P450 reductase and all isoforms of human NOS . However, because the effect of 8-nitro-cGMP on different nNOS variants in terms of superoxide generation remained unclear, we performed EPR spin trapping with physiological concentration (10 µM) of 8-nitro-cGMP  as a redox uncoupler for nNOS-µ and nNOS-α. EPR spin trapping with CYPMPO, which gives comparable results to those of DMPO, as the spin trap revealed superoxide formation by both nNOSs (Figure 4); even in the presence of l-arginine, 8-nitro-cGMP enhanced superoxide production, more so for nNOS-α than for nNOS-µ.
Effects of 8-nitro-cGMP on superoxide production by nNOS-α and nNOS-µ.
ROS production in nNOS-expressing cells
For cellular analysis, we produced stable transformants (HEK293 cells stably expressing nNOS-α and nNOS-µ). Western blotting revealed identical levels of nNOS-α and nNOS-µ proteins in the each stable HEK293 transformant. Control cells showed no endogenous background expression (Figure 5A). This result was supported by the NOS enzyme activity assay, which showed that the amount of nitrite, a stable reaction product of NO with molecular oxygen, in the culture supernatants was approximately equal in cells expressing either isoform (Figure 5B). Nitrite production from nNOS-expressing cells was completely inhibited by pretreatment of the cells with NOS inhibitor l-NAME (Figure 5B). We used these stable transformants in later analysis.
Characterization of nNOS-expressing HEK293 cells.
To analyze superoxide production in nNOS-expressing cells treated with A23187, we used fluorescence microscopy and fluorescence-HPLC with ROS-sensitive fluorescent dye DHE (Figure 6). Although A23187 treatment did not alter the superoxide level in control cells, it markedly raised superoxide levels in cells expressing either nNOS-α or nNOS-µ and was completely inhibited by pretreatment with l-NAME (Figure 6A,B). Superoxide production was significantly greater in cells expressing nNOS-α than those expressing nNOS-µ (Figure 6A,B).
Ca2+-induced ROS generation in nNOS-expressing HEK293 cells.
It has been reported that 2-OH-E+ is the specific product of superoxide reaction with DHE in cells, and E+ is formed as a nonspecific product . Owing to their similar fluorescence characteristics, it is difficult to distinguish fluorescent signals from 2-OH-E+ and E+ by microscopic analysis . Therefore, we confirmed the superoxide generation in cells by analyzing 2-OH-E+ production using fluorescence-HPLC (Figure 6C,D). Figure 6C shows the typical HPLC traces of extracts from HEK293 cells treated with or without A23187 in the presence or absence of l-NAME. Although A23187 treatment did not alter the 2-OH-E+ level in control cells, it raised 2-OH-E+ levels in cells expressing either nNOS-α or nNOS-µ and was completely inhibited by pretreatment with NOS inhibitor (Figure 6D). Superoxide production was significantly greater in cells expressing nNOS-α than in those expressing nNOS-µ (Figure 6C,D).
8-Nitro-cGMP formation in nNOS-expressing cells
We previously reported that ROS play an important role in 8-nitro-cGMP production [7,13]. Therefore, we speculated that ROS production from nNOS splice variants might affect the formation of 8-nitro-cGMP and analyzed 8-nitro-cGMP formation as a downstream consequence of NO/ROS signaling in A23187-treated HEK293 cells using immunocytochemistry (Figure 7A,B). A23187-treated cells expressing nNOS-α manifested strong immunoreactivity of the monoclonal antibody specific for 8-nitro-cGMP, whereas nNOS-µ expressing cells and control cells showed only marginal immunoreactivity (Figure 7A,B). The signal in A23187-treated cells expressing nNOS-α was significantly inhibited by pretreatment with the NOS inhibitor (Figure 7A,B).
Ca2+-induced 8-nitro-cGMP production in nNOS-expressing HEK293 cells.
Molecular evolutionary analysis: synonymous and nonsynonymous substitutions in human and rat nNOS genes
Molecular evolutionary analysis is useful not only for phylogenetic studies but also for studies of the functional constraints of proteins, which can provide important insights into the biological reason for conservation of proteins such as the unique nNOS splice variant nNOS-µ. Within nucleotide sequences, synonymous sites are more variable than are nonsynonymous sites, because less functional restriction occurs in the former . Therefore, determining the ratio of synonymous to nonsynonymous substitutions may be a useful method for identifying functional constraints of proteins.
Hence, we analyzed sites of synonymous and nonsynonymous substitutions in human and rat nNOS genes, with a focus on the nNOS-µ-specific insertion (Figure 8). The conservation of nucleotide and amino acid sequence identities in the complete NOS1 gene between the two species was 86.6 and 93.7%, respectively. For the nNOS-µ-specific insertion, the nucleotide sequence identity between the two species was 89.2%, which was comparable to that for the complete nNOS gene. The amino acid sequence identity, however, was 76.5%, which was markedly lower than that for the complete nNOS protein (Figure 8A).BCJ-2016-0999F9
Nucleotide and amino acid substitutions in human and rat nNOS-µ.
A predictive model of the NO/ROS signaling system involving nNOS.
To obtain additional information about the molecular evolution of the nNOS-µ-specific insertion, we determined the number of nucleotide substitutions per the synonymous site (Ks) and nonsynonymous site (Ka) in the nNOS-µ-specific insertion of human and rat nNOSs. Of 12 nucleotide substitutions identified in the insertion, 8 were synonymous. The Ks value for the nNOS-µ-specific region was 0.12, which was lower than that for the complete NOS1 gene (Ks = 0.44). The Ka value for this region, however, was 0.10, which was higher than that for the complete NOS1 gene (Ka = 0.03). The Ka/Ks value for the nNOS-µ-specific region was considerably higher than that for the rest of the gene (Figure 8C). The averages of the Ka/Ks value for the nNOS-µ-specific region and for the rest of the gene were 0.85 and 0.07, respectively. This molecular evolution analysis therefore indicates that the nNOS-µ-specific region has fewer functional constraints than does the complete gene.
We previously reported expression of nNOS-μ in rat brain and its different localization in the cerebellum compared with that of nNOS-α, which are the predominant nNOS isoforms expressed in the brain. We therefore speculated that nNOS-μ performs a unique function that is distinct from nNOS-α in the brain . In the present work, we focused on comparing ROS production by nNOS-μ and nNOS-α. In the absence of l-arginine, nNOS-α utilizes NADPH to provide electrons to molecular oxygen, to generate superoxide as the product of an uncoupling reaction . In nNOS-µ, a unique 34-amino acid insertion occurs within a flavin mononucleotide (FMN)-binding subdomain (Figure 8B), which autoregulates electron transfer in the molecular structure of the protein [39,40]; that this insertion affects electron transfer seems plausible. We thus expected that the amount of superoxide produced by nNOS-µ would be less than that produced by nNOS-α; this expectation was confirmed by our EPR spin trapping studies, which revealed that purified nNOS-µ generated 44% less superoxide than did nNOS-α.
nNOS was previously shown to produce superoxide under certain conditions in cells [13,16,41,42]. Recently, we reported that nNOS-dependent ROS production induced by nicotine in rat cerebellar granule neurons (CGNs) and nNOS-transfected PC12 cells . The nicotine-induced ROS production led to expression of heme oxygenase 1 in CGNs and exerted cytoprotective effects, which were mediated by ROS and 8-nitro-cGMP . In the present study, we found that A23187 treatment of HEK293 cells stably expressing nNOS-α or nNOS-µ induced ROS production, with greater effects in cells expressing nNOS-α than in those expressing nNOS-µ.
8-Nitro-cGMP is a unique nitrated second messenger of NO/ROS redox signaling that our group identified in 2007 . This compound not only mimics cGMP but also possesses other chemical properties such as redox activity associated with electrophilicity [5–10,12,43]. 8-Nitro-cGMP stimulates production of superoxide by P450 reductase and by the reductase domain of NOS isoforms . This redox activity was observed by nitrated guanine-containing compounds, not by cGMP . The ability of 8-nitro-cGMP to enhance superoxide production by NOS is believed to be a biologically important property. In fact, our study here showed that 8-nitro-cGMP enhanced superoxide production, especially that by nNOS-α, even in the presence of l-arginine. Although 8-nitro-cGMP is induced in NO-mediated signal transduction, recent studies indicated that ROS also plays an important role in 8-nitro-cGMP production [5–7,13], which may support our interpretation again that 8-nitro-cGMP is a downstream messenger of NO/ROS signaling. The formation of 8-nitro-cGMP depends completely on production of both NO and ROS. We indeed found that 8-nitro-cGMP was produced more in nNOS-α-expressing cells than in nNOS-µ-expressing cells; this finding can be attributed to differences in ROS production, suggesting that the NO/ROS redox signal, including 8-nitro-cGMP production, may be regulated by the nNOS splicing variants. We have demonstrated that 8-nitro-cGMP regulates various biological events, such as antioxidative adaptive responses [5,6,8,13], cellular senescence , autophagy , and SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) complex formation . nNOS-α and nNOS-µ may differentially regulate these biological signaling functions.
Our in vitro and cell culture studies indicated that the amount of superoxide produced by nNOS-µ was less than that produced by nNOS-α. Because superoxide has been considered to be toxic to living organisms, the following questions arise. Why is nNOS-α the predominant isoform expressed in the brain? Does superoxide produced by nNOS-α in the brain have beneficial effects? In an attempt to answer these questions, we examined synonymous and nonsynonymous substitutions in the gene containing an nNOS-µ-specific insertion, which is involved in inhibitory regulation of superoxide formation. Our molecular evolution analysis showed that the nNOS-µ-specific insertion had fewer functional constraints than did the gene as a whole. The lower number of functional constraints for the nNOS-µ-specific insertion region suggests that maintaining a low level of superoxide generation is actually not unfavorable, at least in the brain. Paradoxically, considerable superoxide production by nNOS-α may exert beneficial effects on the brain during various physiological and pathological processes. In this regard, the physiological importance of ROS is now increasingly recognized in diverse areas of cell signaling research [9,44–47].
In summary, we herein clarified that nNOS-µ activity, although almost identical with that of nNOS-α with regard to NO production, differed quantitatively from nNOS-α activity related to production of superoxide. The difference in ROS production may be crucial for pathophysiological functions mediated by 8-nitro-cGMP, a downstream messenger of NO/ROS signaling, as demonstrated in the present study (Figure 9). Additional experiments comparing the NO/ROS signaling system involving nNOS-α and nNOS-µ are currently underway. Our identification of the substantial pathophysiological functions of nNOS-µ may help advance understanding of NO- and ROS-mediated cell signaling in neuronal systems.
8-nitroguanosine 3′,5′-cyclic monophosphate
analysis of variance
bovine serum albumin
guanosine 3′,5′-cyclic monophosphate
cerebellar granule neurons
CYPMPO adduct of superoxide
Dulbecco's modified Eagle's medium
DMPO adduct of superoxide
electron paramagnetic resonance
fetal bovine serum
high-performance liquid chromatography
NG-nitro-l-arginine methyl ester
neuronal nitric oxide synthase
reactive oxygen species
soluble guanylate cyclase
soluble N-ethylmaleimide-sensitive factor attachment protein receptor
This work was supported in part by a Grant-in-Aid for Young Scientists B [15K20855 to S.K. and 15K20876 to T.I.], a Grant-in-Aid for Scientific Research A [25253020 to T.A.], a Grant-in-Aid for Scientific Research B [15H03115 to T.S. and 16H04674 to H.I.], a Grant-in-Aid for Challenging Exploratory Research [16K15208 to T.A. and 16K13089 to H.I.], a Grant-in-Aid for Scientific Research on Innovative Areas (Research in a Proposed Area) [26111008 to T.A., Y.W., and T.S. and 26111011 to H.I.], and a Grant-in-Aid for the Strategic Research Foundation at Private Universities [S131101 to Y.W.] from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.
We thank Judith B. Gandy for her excellent editing of the manuscript. We also wish to thank Mrs Naoko Wada-Ando and Dr Kohei Kunieda for technical assistance.
The Authors declare that there are no competing interests associated with the manuscript.