In addition to nitric oxide (NO) synthases, molybdenum-dependent enzymes have been reported to reduce nitrite to produce NO. Here, we report the stoichiometric reduction in nitrite to NO by human sulfite oxidase (SO), a mitochondrial intermembrane space enzyme primarily involved in cysteine catabolism. Kinetic and spectroscopic studies provide evidence for direct nitrite coordination at the molybdenum center followed by an inner shell electron transfer mechanism. In the presence of the physiological electron acceptor cytochrome c, we were able to close the catalytic cycle of sulfite-dependent nitrite reduction thus leading to steady-state NO synthesis, a finding that strongly supports a physiological relevance of SO-dependent NO formation. By engineering SO variants with reduced intramolecular electron transfer rate, we were able to increase NO generation efficacy by one order of magnitude, providing a mechanistic tool to tune NO synthesis by SO.
In mammals, two biochemical pathways are known for the de novo generation of nitric oxide (NO): an oxidative pathway catalyzed by the NO synthases (NOS) , and a reductive pathway described as the inorganic nitrate–nitrite–NO pathway . The oxidative pathway utilizes oxygen (O2) and a guanidine nitrogen derived from arginine to produce citrulline and NO . On the other hand, the reductive nitrate–nitrite-NO pathway is independent of O2 and is thought to compensate for NO homeostasis at conditions of compromised NOS activity . The reductive pathway requires both nitrate and nitrite in vivo, with nitrite being efficiently converted to NO. Both nitrate and nitrite are abundant in the serum and tissue of mammals, although tissue concentrations of nitrite surpass serum concentrations by one order of magnitude [4–6]. Nitrate–nitrite homeostasis is dependent on dietary uptake and the re-oxidation of NO to nitrite and nitrate [7,8].
In the last decades, enzymatic nitrite reduction has been reported for a variety of metal-depending enzymes [9–14]. Among these, all five eukaryotic molybdenum-containing enzymes (i.e. xanthine oxidoreductase, aldehyde oxidase, sulfite oxidase (SO), nitrate reductase and the mitochondrial amidoxime reductase complex, mARC) have been shown to be effective nitrite reductases that generate NO [10,11,13,15,16]. In all these enzymes, nitrite reduction takes place at the molybdenum center, where the metal is complexed in a square-pyramidal manner by a unique tetrahydropyranopterin-based cofactor known as molybdenum cofactor (Moco) . In humans, a defect in Moco biosynthesis leads to loss of activity of all Moco-dependent enzymes resulting in severe neurodegeneration and childhood death . This is mainly attributed to the loss of SO activity, a Moco- and heme-dependent enzyme of the intermembrane space of mitochondria , which converts sulfite by oxidation to sulfate as a terminal step in cysteine catabolism. Loss of SO activity leads to accumulation of sulfite and S-sulfonated cysteine, both leading to different forms of cytotoxicity .
SO is a homodimer with each subunit divided into three distinct domains . The N-terminal heme domain carries a b5-type cytochrome linked via a flexible 11-residue tether to the molybdenum-containing domain, where Moco is deeply buried in the protein backbone. The C-terminal dimerization domain promotes homo-dimerization . The two-electron oxidation of sulfite to sulfate occurs at the molybdenum center, which is reduced from MoVI to MoIV. Re-oxidation occurs via two intramolecular electron transfers from Mo to heme and from there to the final electron acceptor cytochrome c. Both steps depend on protein structural rearrangements enabled by the flexible tether connecting the heme- and molybdenum-containing domains that bring heme transiently into close proximity to the molybdenum center . Following the transfer of a first sulfite-derived electron to heme, the molybdenum center remains reduced by one electron in the MoV oxidation state . Electron transfer to heme is accompanied by deprotonation of the molybdenum center, resulting in the formation of a hydroxyl moiety bound to MoV. The MoV is paramagnetic (with a single d electron) and can be visualized via EPR spectroscopy . Transfer of the electron at heme to cytochrome c allows the transfer of the remaining electron in the molybdenum center, leading to completely re-oxidized MoVI and completion of the catalytic cycle.
We recently reported the reduction in nitrite by SO . We found that the heme domain negatively influenced NO production as illustrated by an SO variant lacking the heme domain, which exhibits enhanced nitrite reduction activity. This suggests a competition between nitrite and heme for sulfite-derived electrons. Interestingly, EPR spectroscopy has shown that sulfite-reduced MoIV but not MoV is able to reduce nitrite to NO . In cells, the presence of SO has been shown necessary to promote nitrite-dependent cGMP formation, a proxy for cellular NO levels .
In this study, we provide novel insights into the biochemistry of the reduction in nitrite by SO. Using a NO analyzer, we demonstrate stoichiometric conversion of nitrite to NO by sulfite-reduced MoIV in both full-length SO as well as the molybdenum-containing domain alone. In the presence of the physiological electron acceptor cytochrome c, we are able to complete the catalytic cycle showing steady-state NO release. Furthermore, the inert nature of MoV can be accounted for—on the basis of nitrite reduction—occurring through an inner-sphere electron mechanism at the molybdenum center, which is possible only in the MoIV oxidation state. Moreover, interfering with heme domain mobility shifts the equilibrium from heme to nitrite reduction within SO, providing a means by which the nitrite-dependent formation of NO can be tuned.
Cloning and site-directed mutagenesis of human SO
PCR fragments of human SO cDNA (NM_000456) were cloned into pQE80L (Qiagen, Hilden, Germany) without mitochondrial targeting sequence (Δ1-79; full-length SO) using SacI and SalI as restriction sites and introducing an N-terminal His6-tag. Heme-deficient SO (SO–Mo) (Δ1-166) was introduced into pQE80L using SacI and HindIII additionally introducing a PreScission protease cleavage site composed of LEVLFQ/GP residues to specifically cleave-off the N-terminal His6-tag. For site-directed mutagenesis, SO cDNA was used as template and fusion PCR was utilized in order to introduce deletions.
Expression and purification
E. coli TP1004  were transformed with pQE80L-SO constructs. Cells were cultured in 2 l LB medium supplemented with 1 mM sodium molybdate, 100 µg/ml ampicillin, 25 µg/ml kanamycin and grown at 37°C until OD600 of 0.6. Expression was induced with the addition of 250 µM IPTG. Cells were grown at 18°C for 72 h and harvested by centrifugation at 5000×g. Pellets were re-suspended in buffer containing 50 mM Tris/Ac pH 8.0, 300 mM NaCl and 15 mM imidazole. Cell disruption was achieved by the use of an Emulsiflex-C5 (Avestin) applying a maximum pressure of 2 × 103 psi. Cell debris was pelleted via centrifugation at 20 000×g for 45 min and soluble SO was affinity purified via Ni-NTA gravity flow (Macherey-Nagel). Heme-deficient SO was proteolytically cleaved-off the Ni-NTA matrix by applying 1 mg PreScission protease/ml Ni-NTA matrix in 50 mM Tris/Ac pH 8.0 and 100 mM NaCl in batch for 1 h. GST-tagged PreScission protease was removed from the elution by glutathione-affinity purification. Eluted SO was re-buffered in 50 mM Tris/Ac pH 8.0, 30 mM NaCl using PD10 columns (GE Healthcare), frozen in liquid nitrogen and stored at −80°C until further use.
Determination of cofactor saturation
The concentration of heme-saturated SO was determined by measuring heme absorption at 413 nm (ε413 = 113 000 M−1 cm−1). For Moco quantification, 100 pmol SO (based on A413) was used to oxidize Moco to Form-A, which can be quantified by HPLC as described previously .
Sulfite-dependent steady-state SO activity
Sulfite-dependent SO activity was determined by either using 50 µM porcine cytochrome c (ε550 = 19 630 l/mol cm) (BioChemica) or 400 µM ferricyanide (FeIII(CN)6) (ε420 = 1020 l/mol cm) as an electron acceptors. Reduction was followed in a time-dependent manner by use of an ELISA reader (EL808 Biotec). The reaction took place in 200 µl 20 mM Bis–Tris/Ac pH 6.5 or 100 mM Tris/Ac pH 8.0 containing SO. The reaction was started with 100 µl sulfite solution of varying concentrations in the appropriate buffer.
Rapid kinetics were performed with an Applied Photophysics Inc.SX-18MV stopped-flow instrument, equipped with a photo diode array detector. Experiments were performed anaerobically at 20°C in 100 mM Tris/Ac pH 8.0. Enzymes and sulfite solutions were prepared anaerobically and mounted on the stopped-flow. Reduction in heme was followed at 423 nm.
NO formation was monitored via a NO analyzer (Model 280i, GE Healthcare) detecting NO by a chemiluminescent reaction with O3. Via the purge vessel, the NO analyzer was calibrated using defined amounts of sodium nitrite injected in a fully reducing triiodide (I3−) solution with argon as carrier gas as described previously . Peak analysis was carried out via Origin software. Nitrite reduction took place in 6 ml 20 mM Bis–Tris/Ac pH 6.5 and 200 or 400 pmol of SO based on Moco concentration. Depending on the experimental setup, sulfite, as well as nitrite concentrations, varied. In the steady-state NO release assay, 5 µM cytochrome c was added. Photo-enhanced NO release was achieved by a portable 360 nm UV lamp, which was mounted next to the purge vessel.
Pre-reduction in Moco domain
Heme-deficient SO was pre-reduced anaerobically with 10-fold excess of sulfite (1 mM) in 20 mM Bis–Tris/Ac pH 6.5. The reaction was re-buffered using DextraSEC-PRO2 columns into 20 mM Bis–Tris/Ac pH 6.5. The protein concentration was determined photometrically and 400 pmol of the pre-reduced enzyme was injected into the NO analyzer containing 6 ml 20 mM Bis–Tris pH 6.5, 20 µM sulfite and 2.5 mM nitrite.
Electron paramagnetic resonance spectroscopy
X-band EPR spectra were recorded with the EWWIN 6.0 acquisition and lineshape analysis software (Scientific Software Services) on a Bruker ER 300 spectrometer equipped with an ER035 m gaussmeter and an HP 5352B microwave frequency counter. Temperature was controlled at 150 K using a Bruker ER 4111VT liquid N2 cryostat. The spectra were acquired by sweeping the magnetic field between 330 and 360 mT. EPR samples were prepared in 20 mM Bis–Tris/Ac pH 6.5 under argon. For EPR spectra of SO, 200 µl 50 µM of the enzyme were incubated in buffer with stirring under an atmosphere of Ar to achieve anaerobicity. Subsequently, 3 µl sulfite and nitrite were added to final concentrations of 300 µM and 20 mM, respectively. Samples were frozen in an ethanol/dry ice bath and EPR spectra were recorded.
Stoichiometric NO release by human SO
Previously, we allocated nitrite reduction to the molybdenum center of SO, where sulfite-reduced MoIV was shown to reduce nitrite by one electron, while the resulting MoV was inert to nitrite, suggesting a stoichiometric conversion of nitrite to NO by SO . In order to investigate whether the enzyme produces NO stoichiometrically, full-length human SO was recombinantly expressed and purified (Figure 1A). Based on Moco saturation of the enzyme, 400 pmol full-length SO was injected into the NO analyzer in the presence of 20 µM sulfite and increasing concentrations of nitrite (0.25–5 mM). Traces of enzymatically produced NO (Figure 1B) revealed a peak-like shape, which indicated a single enzymatic turnover rather than a steady-state reaction. Integration of these peaks allowed the determination of the amount of NO release using a NO calibration curve. The stoichiometry of Moco to NO, produced by the full-length enzyme, was found to be ∼1 : 2 as indicated by plotting the amount of NO formed against the amount of enzyme used under saturating nitrite concentrations (Figure 1C). Based on these results, we concluded that full-length SO released two molecules of NO per monomer.
Stoichiometric NO release of full-length and heme-deleted SO.
To investigate whether the 1 : 2 stoichiometry derived from the presence of heme as a second electron acceptor in the full-length enzyme, an N-terminally truncated variant lacking the heme domain (SO–Mo) was recombinantly expressed and purified (Figure 1D). Sulfite- and nitrite-dependent NO formation of SO–Mo revealed a much higher production of NO at low nitrite concentrations (Figure 1E). However, saturation was reached much earlier than with full-length SO leading to a stoichiometry of 1 : 1, which indicated the release of NO and formation of MoV in a single cycle of sulfite oxidation and nitrite reduction (Figure 1E). These results suggest, that the second NO molecule, which is released by full-length enzyme required the presence of heme, which is able to accept an electron by IET from MoV following the reduction in one molecule of nitrite by sulfite-reduced SO. The resulting re-oxidized MoVI is able to react with another sulfite molecule in a second cycle to produce again fully reduced MoIV, which accepts a second nitrite molecule to produce NO. The resulting FeII/MoV intermediate is prohibited from further reaction, thus explaining the 1 : 2 Moco to NO stoichiometry observed in full-length SO.
To verify the FeII/MoV state following the 1 : 2 conversion of nitrite to NO by full-length SO, we used UV–vis spectroscopy to detect ferrous iron as well as EPR spectroscopy to detect MoV. The presence of FeII was validated by a sharp peak at 555 nm, which is characteristic for reduced b5-type heme in SO (Figure 1F), confirming the proposed resting state of the enzyme following two cycles of sulfite oxidation and nitrite reduction. EPR spectroscopy revealed the presence of MoV in the same reaction tube (Figure 1G). In aggregate, these results demonstrate that human SO is able to quantitatively convert nitrite to NO.
Towards the reaction mechanism of nitrite reduction by SO
In the re-oxidation of sulfite-reduced SO, the sulfite-derived electrons are transferred sequentially to the heme, a process that involves the formation of an MoV intermediate between the two individual electron transfer events (Figure 2A). As we have shown here and in a previous study , the MoV species is inert towards nitrite. In order to elucidate the mechanism of nitrite reduction by MoIV, SO–Mo was pre-reduced anaerobically with excess sulfite (1 mM) to produce fully reduced MoIV. Following the reaction, the enzyme was re-buffered to remove the reaction product sulfate and any remaining sulfite. The concentration of the pre-reduced SO–Mo was determined photometrically and the enzyme then injected into the NO analyzer together with 2.5 mM nitrite. In parallel, fully oxidized SO–Mo was injected with 1 mM sulfite and 2.5 mM nitrite together into the NO analyzer. When comparing the amount of released NO, we found that pre-reduction resulted in a significant decrease in detected NO compared with the detected NO, which was produced by an oxidized enzyme following the treatment with sulfite and nitrite simultaneously (Figure 2B,C). Hence, an enzyme in the MoIV state which is re-oxidized in this way is less prone to reduce nitrite than enzyme that undergoes sulfite oxidation in the presence of nitrite. The likeliest explanation for such a finding suggests an inner-sphere electron transfer mechanism of nitrite reduction as the pre-reduced enzyme presumably co-ordinates H2O in the equatorial position of molybdenum, which interferes with efficient NO production.
NO release depends on nitrite coordination and protonation.
In line with this hypothesis, the previously reported pH dependence of NO formation , confirmed here, shows an order of magnitude lower level of NO production at pH 8 than is observed at pH 6.5 (Figure 2D). This suggests a protonation-sensitive nitrite reduction mechanism. In order to discriminate if nitrite protonation is a prerequisite for NO release or whether the enzyme itself responds to pH differences during nitrite reduction, we used UV-light excitation (360 nm) assuming an accelerated NO release in case of the formation of a proximal Mo-NO2H bond (Figure 2E). As result, we found a significantly increased NO production following UV-light exposure, consistent with the idea that the SO-catalyzed nitrite reduction requires protonation of a molybdenum-co-ordinated oxygen atom of nitrite as catalytic intermediate towards MoV and NO formation (Figure 2F).
Steady-state NO release in the presence of cytochrome c
So far only stoichiometric NO release had been shown, which is not expected to be physiologically relevant. In order to demonstrate enzyme-dependent steady-state NO production, a second electron acceptor is necessary to accept the sulfite-derived second electron that otherwise would rest on MoV. We therefore added the physiological electron acceptor cytochrome c to the reaction mix to determine whether any steady-state NO production was generated. In comparison to a control with inactive SO (Figure 3A) and to a control without cytochrome c (Figure 3B), the addition of 10 nmol cytochrome c (5 µM) evoked a steady-state NO production by full-length SO that was dependent on the sulfite concentration (Figure 3C). The amount of produced NO was plotted versus the sulfite concentration, which indicated that the highest NO production occurred between 10 and 37.5 µM sulfite, with a dose-dependent inhibition of NO formation at higher sulfite concentrations. At the highest rate, ∼1600 pmol NO was formed with 200 pmol enzyme suggesting eight catalytic cycles per enzyme molecule (Figure 3D).
Steady-state NO release of full-length SO in presence of cytochrome c.
Sulfite and nitrite both act at the molybdenum center
The requirement of low sulfite concentrations to allow steady-state nitrite reduction in SO suggested that sulfite competes with nitrite, resulting in increased transfer of both sulfite-derived electrons to cytochrome c. In order to probe this possibility, the competition of sulfite and nitrite at the molybdenum center of full-length SO was analyzed. First, sulfite-dependent cytochrome c reduction was monitored in the absence and presence of 20 mM nitrite (Figure 4A). Nitrite was found to inhibit sulfite-dependent cytochrome c reduction at sulfite concentrations ranging from 10 to 100 µM, representing a sulfite concentration range in which steady-state NO production by SO was observed. We also monitored nitrite-dependent NO production in the NO analyzer at two different sulfite concentrations (20 and 100 µM) and found decreased NO production at the higher sulfite concentration (Figure 4B), consistent with competition of sulfite and nitrite at the molybdenum center. Double-reciprocal plots of these data yielded lines intersecting on the y-axis, indicative of competitive inhibition of sulfite on NO formation (Figure 4C). Conversely, monitoring NO production in a sulfite-dependent manner at a fixed nitrite concentration showed a reduced NO production at higher sulfite concentrations (Figure 4D). Our results collectively indicate that sulfite and nitrite compete for the molybdenum center and explain the observed decrease in steady-state NO production at higher sulfite concentrations.
Competition of sulfite and nitrite at the molybdenum center.
Reduction in heme- to Moco domain distance leads to increased nitrite affinity
Since nitrite reduction requires a fully reduced MoIV center, NO production and electron transfer from MoIV to the heme compete with each other in full-length SO. It is for this reason that heme domain-deleted SO–Mo exhibited much higher levels of NO production at 100 µM nitrite than did full-length SO enzyme (Figure 5A).
Increased nitrite reduction efficacy by decreasing heme flexibility.
In a previous study, electron transfer within SO was slowed significantly upon truncation of the inter-domain tether by three to five residues . We therefore generated variants in which the tether between heme and Moco domain was truncated even further by 5–11 residues in order to decrease heme mobility and electron transfer (Figure 5B). In total, four truncated SO variants (Δ5, Δ6, Δ8, Δ11) were expressed and purified to homogeneity (Figure 5B) and the impact of tether truncation was analyzed. Using ferricyanide (FeIII(CN)6) as electron acceptor, which acts at the molybdenum center rather than the heme as does cytochrome c, all four truncated variants exhibited activities similar to wild-type enzyme (Figure 5C). This indicated that the tether truncation had no effect on sulfite affinity and oxidation at the molybdenum center. The impact of tether truncation on electron transfer from MoIV to FeIII was analyzed via stopped-flow rapid kinetics. Upon reduction in the molybdenum center by sulfite, subsequent electron transfer to the heme was monitored by the large spectral change in the heme Soret region. All four variants exhibited markedly reduced rates of heme reduction, which slowed as the length of the tether was progressively shortened, from 49.2 ± 6.1 s−1 in wild-type SO to 4.1 ± 1.4 s−1 in the Δ11 variant having the shortest tether (Figure 5D). The velocity of cytochrome c reduction, which depends on heme flexibility and IET rates, was comparably reduced, from 2.5 ± 0.1 µM cytochrome c/min for wild-type enzyme to 0.2 ± 0.01 µM cytochrome c/min for Δ11 SO (Figure 5E) strongly correlating with the observed changes in internal heme reduction rate.
After demonstrating that progressive tether shortening resulted in ∼10-fold reduction in sulfite-dependent cytochrome c reduction velocities with an underlying loss in IET for the MoIV-heme couple, we next recorded sulfite-dependent nitrite reduction for all SO tether variants (Figure 5F). During stoichiometric NO formation, the height as well as the slope of the NO peak allowed calculation of KM, kred and kred/KM values for the individual SO tether variants. We found that the KM for nitrite decreased gradually with decreasing tether length by more than one order of magnitude from 37.0 ± 6.7 mM nitrite in full-length SO to 1.5 ± 0.6 mM nitrite in Δ11 SO thus closely resembling heme-deleted SO–Mo with a KM of 1.0 ± 0.1 mM nitrite. In line, the enzymatic efficacy kred/KM was increased proportionally to tether length ranging from 3.2 M−1 s−1 in full-length SO to 28.2 M−1 s−1 in SO–Mo (Figure 5F). In aggregate, our results support the notion that restriction of IET reduces the rate of sulfite-dependent cytochrome c reduction, which in turn favors the alternative path of SO catalysis by reducing nitrite to NO.
Identification of SO as fourth molybdenum-dependent nitrite reductase in animals underscores the importance of molybdenum-containing enzymes in NO homeostasis . However, the molecular mechanism of nitrite reduction by Moco enzymes remains poorly understood. Here, we investigated nitrite reduction by human SO, demonstrating that nitrite competes with sulfite at the molybdenum center. We provide evidence for a reaction mechanism involving enzyme reduction by sulfite, followed by electron transfer to nitrite and cytochrome c followed by re-reduction by sulfite, thus allowing a steady-state catalytic NO formation. Our data are consistent with nitrite coordination to the molybdenum center, presumably at the equatorial position, supporting a computational analysis of nitrite reduction by human mARC, another Moco enzyme .
In SO, the one-electron reduction in nitrite to NO by the fully reduced molybdenum center leads to the formation of MoV, which cannot reduce a second equivalent of nitrite. With full-length enzyme, however, electron transfer to the heme fully re-oxidizes the molybdenum center, which can then be reduced by a second equivalent of sulfite, generating a second equivalent of NO. The SO–Mo domain alone, however, cannot be re-oxidized and thus produces only a single equivalent of NO. Strikingly, progressive shortening of the polypeptide tether between the molybdenum- and heme-containing domains reduced electron transfer from molybdenum to heme, which in turn allows these variants to generate NO with a much higher efficacy. These findings are in line with previous reports demonstrating domain movement in SO as an important mechanistic feature in the catalytic cycle.
In addition to stoichiometric NO release by SO, we demonstrate steady-state production of NO in the presence of cytochrome c with full-length SO. Our finding that this steady-state NO formation is inhibited at increasing sulfite concentrations suggests a competition between sulfite and nitrite for the SO active site. Indeed, we observe nitrite-induced inhibition of sulfite-dependent cytochrome c reduction. Strikingly, this inhibition is observed in the same concentration range as that required for steady-state NO synthesis. This indicates that nitrite is an effective alternate electron acceptor in the sulfite-dependent cytochrome c reduction, even at elevated concentrations of the physiological electron acceptor .
The mechanism for NO production by SO presented here is fully consistent with a computational analysis of nitrite reduction by mARC in which a barrier-less decay of the Mo–nitrite complex to form both MoV and NO occurs via a proton-induced disproportionation. The protonation of nitrite, which forms nitrous acid mitigated N–O bonding proximal to the molybdenum, promoting NO release . Nitrite reduction by FeII-polyoxotungstate complexes occurs in a similar fashion, with proton-dependent NO formation in a first-order manner at or below the pKA of nitrite . However, the reaction converts to second-order at pH values above the pKA, which indicates obligate protonation of nitrite as a prerequisite for complex decay and NO release .
Similar to the model compound chemistry described above, we find a decline of SO-dependent NO production by one order of magnitude on increasing the pH from 6.5 to 8.0, being consistent with the proton-dependent formation of nitrous acid at a molybdenum-co-ordinated nitrite in the course of NO formation. Interestingly, both nitrite and nitrous acid have absorption maxima between 340 and 370 nm . However, while photolysis of nitrous acid with UV-A light is reported to release a hydroxyl radical and NO in aqueous solution, photolysis of the non-protonated nitrite anion is hampered due to mesomeric stabilization [31,32]. This indicates that proton-dependent nitrous acid formation is a prerequisite for nitrite decay and NO release. Thus, we propose that irradiation-enhanced NO release from SO is dependent on nitrous acid formation at the molybdenum-co-ordinated nitrite. These findings are collectively well in line with pH-dependent nitrite reduction in other metal-centered enzymes [10,13,14,33].
We note that the pH-dependent nitrite reduction in SO is fully compatible with the intracellular localization of the enzyme in the intermembrane space of mitochondria, which is reported to have a pH of 6.8 . Thus, the nitrite reduction capacity of SO at pH 6.5 is reasonably close to the enzyme's physiological environment, which favors proton-dependent nitrite reduction. It might be added that sub-compartmentalization of the mitochondrial intermembrane space itself might lead to local variations in pH , which depend on the specific molecular environment. It has been shown that the pH in the vicinity of the proton-pumping complex IV is 0.3 pH units below that at F0F1-ATP-synthase . In this respect, the obligate proximity of SO to cytochrome c and thus also to complex IV ensures a proton-rich environment. A pH somewhat lower than 6.8 is thus reasonable for the immediate environment of SO. This implies protonation of Mo-co-ordinated nitrite as physiologically more likely than for other mammalian Moco enzymes as well as heme proteins such as hemoglobin, myoglobin and neuroglobin, for which pH sensitivity of nitrite reduction was reported in vitro [10,13,14,33].
The physiological significance of SO-derived NO is not yet clear, although the proximity of the enzyme to the respiratory chain raises the possibility of SO being a source of NO to regulate mitochondrial function. It is well known that at least complexes I, III and IV of the respiratory chain are inhibited by NO [36–38], which is considered as a basic mechanism to suppress mitochondria-derived ROS formation. During prolonged ischemia, for example, O2 restriction prohibits electron flux within the respiratory chain  and following tissue re-oxygenation, uncontrolled release of accumulated electrons leads to ROS formation what is commonly referred to as ischemia/reperfusion (IR) injury . ROS formation is, at least in part, controlled by NO-induced inhibition of the respiratory chain, which slows regeneration of respiratory chain activity upon tissue reperfusion [41,42]. In eNOS−/− mice, sensitivity of heart tissue to IR injury increases significantly compared with control mice. However, the infarct area of these mice is reduced by 50%, when nitrite is added to the drinking water . In humans, nitrite infusion prior to coronary restriction-induced ischemia significantly protects from ROS-induced endothelial dysfunction upon reperfusion . This nitrite-induced pre-conditioning is attributed to nitrite-derived NO formation, which targets mitochondria and protects cells from IR injury. To date, the enzymatic source(s) for nitrite-derived NO remain a matter of debate. Its proximity to the respiratory chain and the low pH of its surrounding make SO a promising candidate to fulfill a protective role during IR injury and underlines the importance of this enzyme within the nitrate–nitrite–NO pathway.
D.B. designed and performed experiments, analyzed results and wrote the paper. A.T.K., D.N. and R.H. performed experiments and analyzed results. G.S. designed the study, analyzed results and wrote the paper.
This work was funded by the Center for molecular medicine Cologne, Germany [Grant number: CMMC2017-C13]. This work was also supported by the US Department of Energy grant DE-SC0010666 (to R.H.).
We acknowledge Monika Laurien, Simona Jansen and Joana Stegemann for technical assistance.
The Authors declare that there are no competing interests associated with the manuscript.