Reactivity and endogenous modification by nitrite and hydrogen peroxide: does human neuroglobin act only as a scavenger?

Abstract

NGB (human neuroglobin), a recently discovered haem protein of the globin family containing a six-co-ordinated haem, is expressed in nervous tissue, but the physiological function of NGB is currently unknown. As well as playing a role in neuronal O₂ homoeostasis, NGB is thought to act as a scavenger of reactive species. In the present study, we report on the reactivity of metNGB (ferric-NGB), which accumulates in vivo as a result of the reaction of oxyNGB (oxygenated NGB) with NO, towards NO₂⁻ and H₂O₂. NO₂⁻ co-ordination of the haem group accounts for the activity of metNGB in the nitration of phenolic substrates. The two different metNGB forms, with and without the internal disulfide bond between Cys⁴⁶ (seventh residue on the inter-helix region between helices C and D) and Cys⁵⁵ (fifth residue on helix D), exhibit different reactivity, the former being more efficient in activating NO₂⁻. The kinetics of the reactions, the NO₂⁻-binding studies and the analysis of the nitrated products from different substrates all support the hypothesis that metNGB is able to generate an active species with the chemical properties of peroxynitrite, at pathophysiological concentrations of NO₂⁻ and H₂O₂. Without external substrates, the targets of the reactive species generated by the metNGB/NO₂⁻/H₂O₂ system are endogenous tyrosine (resulting in the production of 3-nitrotyrosine) and cysteine (oxidized to sulfinic acid and sulfonic acid) residues. These endogenous modifications were characterized by HPLC-MS/MS (tandem MS) analysis of metNGB after reaction with NO₂⁻ and H₂O₂ under various conditions. The internal S–S bond affects the functional properties of the protein. Therefore metNGB acts not only as scavenger of toxic species, but also as a target of the self-generated reactive species. Self-modification of the protein may be related to or inhibit its postulated neuroprotective activity.

INTRODUCTION

NGB (human neuroglobin) was first identified in 2000 as a third globin-type protein, the other types of globins are Mb (myoglobin) and Hb [1]. NGB is expressed at low levels (in the micromolar range) in various regions of the brain, but at much higher levels (∼100 μM) in retinal cells [1,2]. NGB is composed of 151 amino acids and has a high sequence similarity to mouse neuroglobin (94% identity), but little homology with vertebrate Mbs and Hbs (<21% and <25% identity respectively) [1]. NGB differs from the other vertebrate globins in the binding of the six-co-ordinated haem group, with His⁹⁶[F8 (eighth residue on helix F)] and His⁶⁴(E7) acting as the fifth/proximal and sixth/distal ligands respectively, in both the ferrous and ferric forms of NGB [3,4] (Scheme 1). The presence of a six-co-ordinated haem reduces the affinity of the protein for ligands, with respect to the expected values, in the absence of competitive co-ordination by the distal histidine residue [3].

Scheme 1. Cysteine residue oxidation states and co-ordination equilibria in metNGB.

Schematic representation of the low-spin/high-spin equilibria for metNGB_S-S (left-hand side) and metNGB_SH (right-hand side). The disposition of the side chains of the proximal and distal histidine residues [His⁹⁶(F8) and His⁶⁴(E7) respectively] is shown. The length of the vertical arrows indicates that, even if the co-ordination equilibrium is always shifted towards the six-co-ordinated species, in the case of metNGB_S-S, the amount of high-spin form is larger compared with metNGB_SH.

The function of NGB in the human brain remains unclear. In analogy with other globins, several possible physiological roles have been considered [5]. Indirect evidence suggests a role of NGB in neuronal O₂ homoeostasis, owing to the correlation between NGB expression and O₂ consumption [6]. The O₂ affinity of NGB depends on several factors, including the relative binding rates of O₂ with the haem sixth position and the competing distal histidine residue [3,7], and the redox state of the cell which controls the formation or cleavage of an internal disulfide bond between Cys⁴⁶ [CD7 (seventh residue on the inter-helix region between helices C and D)] and Cys⁵⁵(D5) [8,9]. The presence of the S–S bond could perturb the three-dimensional structure of the CD–D (inter-helix region between helices C and D, extending to the end of helix D) region and affect the location of the neighbouring E-helix, thus modulating the binding of the endogenous His⁶⁴(E7) ligand to the haem group [9,10]. As a consequence, in ferrous NGB, the distal histidine residue dissociation rate increases by a factor of 10 with respect to the protein form without the disulfide bond [9,11], leading to an effective increase in O₂ affinity by the same factor [12] (Scheme 1).

O₂ storage or diffusion to or from the haem group of NGB may be assisted by the presence of a large protein matrix cavity (approx. 120 ų, where 1 Å=0.1 nm), located between the haem distal site and the EF interhelical hinge and connected to the protein surface [10,13]. The cavity may be reshaped after ligand binding, which may allow trapping of harmful ROS (reactive oxygen species) or RNS (reactive nitrogen species) [14–17]. In particular, the increase in the concentration of NO up to the low micromolar range, as observed under ischaemic conditions [18], may be contrasted with its reaction with oxyNGB (oxygenated NGB), yielding metNGB (ferric-NGB) and NO₃⁻ through a haem-bound ONOO⁻ (peroxynitrite) intermediate [15]:

(1)

As a metNGB reductase system has not yet been identified in the brain [15], the in vivo formation of metNGB through this mechanism may be physiologically relevant. Also, metNGB reacts with NO [19] to generate the stable NGBFe^IINO species by reductive nitrosylation. This form of NGB is a good scavenger of ONOO⁻, which oxidizes the protein back to metNGB as follows [19]:

(2)

(3)

(4)

In contrast with Mb and Hb, the protein does not generate the ferryl form on reaction with H₂O₂ or ONOO⁻ [20]. Also, no reaction was observed on the addition of NO₂⁻ [19].

Therefore even if the Fe^III centre of metNGB appeared to be protected by the co-ordinated distal histidine residue from attack by oxidizing species, the competitive co-ordination of exogenous ligands could account for the activation of alternative reactions. In the present study, we report on the reactivity of human metNGB in the presence of NO₂⁻ and H₂O₂, both of which are increased in vivo under conditions of oxidative stress [21]. The activation of NO₂⁻/H₂O₂ by metNGB was assayed by the nitration of phenolic substrates [22–24] and an unexpectedly high reactivity was observed. Moreover, the two forms of the protein, metNGB_S-S (NGB with the internal disulfide bond) and metNGB_SH (NGB without the internal disulfide bond), are not equally as efficient in activating NO₂⁻ and H₂O₂. The issue of endogenous modification of metNGB, resulting from the reaction with NO₂⁻ and H₂O₂, at pathophysiological concentrations in the absence of external substrates was also addressed, since the identification of protein residues modified by ROS and RNS is essential for the understanding of the mechanisms involved in the development of various pathologies [25–29] and the postulated neuroprotective function of NGB. The endogenous targets of the reactive species generated by the protein are tyrosine and cysteine residues, and a more extensive pattern of modifications was observed for these residues than observed previously with hMb (human Mb) [24].

EXPERIMENTAL

Reagents

All buffer solutions were prepared using deionized Milli-Q water. ABTS [2,2′-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid)], DTT (dithiothreitol), GdnHCl (guanidine hydrochloride), 30% (v/v) H₂O₂ solution, HPA [3-(4-hydroxyphenyl)propionic acid], 4-hydroxybenzonitrile, 4-PDS (4,4′-dithiodipyridine), phenylacetic acid, sodium nitrite and trypsin were obtained from Sigma–Aldrich. All reagents were obtained at the best grade available. The concentration of H₂O₂ was controlled by monitoring the formation of the ABTS radical cation using a standard enzymatic method as published previously [30]. Recombinant human NGB was expressed and purified as reported previously [12]. In order to obtain a homogeneous sulfur-bridged form of NGB (i.e. NGB_S-S), the protein solution was dialysed overnight at 4 °C against 50 mM phosphate buffer, pH 7.5 (17 mM NaH₂PO₄ and 33 mM Na₂HPO₄). To reduce the intramolecular disulfide bond, forming NGB_SH, the NGB solution was dialysed at 37 °C against 1 mg/ml DTT dissolved in degassed 50 mM Tris/HCl, pH 7.5, and 0.5 mM EDTA under anaerobic conditions. After 30 min of dialysis, three exchanges of degassed 50 mM Tris/HCl, pH 7.5, and 0.5 mM EDTA were performed to eliminate DTT. Protein dialysis was carried out at 4 °C. In all of the experiments described below, when not explicitally stated, the proteins (NGB_S-S and NGB_SH) were used in their ferric form. All spectrophotometric measurements were performed on a Hewlett Packard HP 8452A diode array spectrophotometer.

Kinetic studies of phenol nitration catalysed by NGB_S-S and NGB_SH

The kinetic experiments were carried out in 200 mM phosphate buffer, pH 7.5 (68 mM NaH₂PO₄ and 132 mM Na₂HPO₄), using a quartz cuvette with a 1 cm-path-length, thermostated at 25.0±0.1 °C and equipped with a magnetic stirrer. The initial solution, containing protein and variable substrate and NO₂⁻ concentrations in a final volume of 1.6 ml, was obtained by mixing solutions of the reagents (at an appropriate concentration) into the buffer. The reaction was started by the addition of H₂O₂ and was followed by monitoring A₄₅₀ during the initial 10–15 s of the reaction. The conversion of the rate data from change in absorbance/s into [nitrophenol]/s was performed using the known molar absorption coefficient of HPA at 450 nm, ϵ=3350 M⁻¹·cm⁻¹ [22]. The kinetic parameters were obtained from fitting of the plots of experimental rates at different substrate/NO₂⁻ concentrations to the appropriate equations.

For each substrate, the change in the reaction rate in the presence of various reactant concentrations was studied through a series of steps. First, a suitable concentration of H₂O₂ was found which would maximize the reaction rate, but avoided the use of excess oxidant. Secondly, this H₂O₂ concentration was used to study the dependence of the reaction rate on the concentration of HPA, and, finally, the dependence of the reaction rate on the concentration of NO₂⁻ was examined. These steps were performed by following the iterative procedure described previously [23]. The concentration of NGB_S-S or NGB_SH used in these reactions was 1 μM, whereas the concentration of the other reactants for each of the steps is detailed as follows. Optimization of H₂O₂ concentration: using NGB_S-S, [HPA]=1 mM, [NO₂⁻]=0.3 M and [H₂O₂]=0.09–0.8 M; using NGB_SH, [HPA]=0.4 mM, [NO₂⁻]=1.0 M and [H₂O₂]=0.09–0.8 M. Dependence of the rate on varying HPA concentration: using NGB_S-S, [H₂O₂]=0.6 M, [NO₂⁻]=0.3 M and [HPA]=0.031–1.2 mM; using NGB_SH, [H₂O₂]=0.6 M, [NO₂⁻]=1.0 M and [HPA]=0.0096–0.96 mM. Dependence of the rate on varying the concentration of NO₂⁻: using NGB_S-S, [H₂O₂]=0.6 M, [HPA]=0.4 mM and [NO₂⁻]=0.0062–0.44 M; using NGB_SH, [H₂O₂]=0.6 M, [HPA]=0.3 mM and [NO₂⁻]=0.037–2.5 M.

The reaction rates observed in the absence of the protein (non-catalytic reaction) or without H₂O₂ were completely negligible.

Binding of NO₂⁻

In order to test the binding of NO₂⁻ to NGBs, a 1 cm-path-length quartz cuvette containing either 4.8 μM NGB_S-S or 5.2 μM NGB_SH in 200 mM phosphate buffer, pH 7.5, was incubated in increasing concentrations of NO₂⁻ in 200 mM phosphate buffer, pH 7.5 (final concentration from 0 to 0.42 M for NGB_S-S and from 0 to 0.64 M for NGB_S). The reaction was thermostatically controlled at 25.0±0.1 °C, and UV–visible spectra were recorded after each increase in NO₂⁻ concentration. Blank spectra were recorded in the same way, but in the absence of protein. After subtracting the corresponding blank from each spectrum, the resulting spectra were corrected for dilution and then transformed into difference spectra by subtracting the native protein spectrum. A plot was constructed demonstrating the difference between A₄₁₄ and A₄₃₄ (these wavelengths had the maximum variation in difference spectra) against the ligand concentration.

For the binding of NO₂⁻ to NGB_S-S, the constants K₁ and K₂ were obtained by interpolation of the absorbance data with the binding isotherm for low-affinity binding of two consecutive ligands:

where ΔA_∞1 and ΔA_∞2 represent the absorbance changes owing to the binding of one and two ligand molecules respectively, and [NO₂⁻]_tot is the total NO₂⁻ concentration, i.e. free plus bound NO₂⁻ concentrations. For the binding of NO₂⁻ to NGB_SH, the constant K_B was obtained by interpolation of the absorbance data with the binding isotherm for low-affinity binding of a single ligand:

where ΔA_∞ represents the absorbance change as a result of binding NO₂⁻, and [NO₂⁻]_tot is the total NO₂⁻ concentration.

HPLC analysis of the nitration products of 4-hydroxybenzonitrile and phenylacetic acid

The product mixtures derived from the nitration of 4-hydroxybenzonitrile and phenylacetic acid, promoted by the NGB/NO₂⁻/H₂O₂ system, were analysed by HPLC using a Jasco MD-1510 instrument with diode array detection equipped with a Supelcosil™ LC18 reverse-phase semipreparative column (5 μm; 250×10 mm). Elution was carried out using 0.1% TFA (trifluoroacetic acid) in distilled water (solvent A) and 0.1% TFA in acetonitrile (solvent B), with a flow rate of 5 ml/min. Elution started with the use of 100% solvent A for 5 min, followed by a linear gradient from 100% solvent A to 100% solvent B over 25 min. Spectrophotometric detection of the eluate was performed in the range of 200–600 nm.

The catalytic nitrations of 4-hydroxybenzonitrile and phenylacetic acid were studied under the following experimental conditions: [NGB_S-S]=1 μM, [substrate]=1 mM, [H₂O₂]=0.6 M and [NO₂⁻]=0.01, 0.05 or 0.15 M in 200 mM phosphate buffer, pH 7.5. The reaction mixtures were incubated for 10 min at room temperature (25 °C) and then were analysed by HPLC. The retention time of 4-hydroxybenzonitrile was 15.7 min, whereas the corresponding nitration product 4-hydroxy-3-nitrobenzonitrile was absent under all the conditions tested. Its retention time (17.4 min) was established using LPO (lactoperoxidase)-catalysed nitration, carried using the following conditions: [LPO]=80 nM, [phenol]=1 mM, [H₂O₂]=1 mM and [NO₂⁻]=0.01 M in 200 mM phosphate buffer, pH 7.5. The HPLC chromatograms of the mixtures obtained from the nitration of phenylacetic acid showed there was unreacted substrate present at 16.8 min, two main peaks at 16.0 and 16.2 min, and two minor peaks at 14.5 and 14.7 min (obtained at both low and high NO₂⁻ concentrations).

HPLC analysis showed that no modification of 4-hydroxybenzonitrile and phenylacetic acid occurred in the absence of the protein. The presence of H₂O₂ was also demonstrated to be essential for the modification.

Modification of NGB in the presence of NO₂⁻ and H₂O₂

Samples of modified NGB_S-S and NGB_SH proteins were prepared by the addition of sodium nitrite and H₂O₂ to 60 μM NGB_S-S in 50 mM phosphate buffer, pH 7.5, or 60 μM NGB_SH in 50 mM Tris/HCl, pH 7.5, and 0.5 mM EDTA. Four different concentrations of sodium nitrite and H₂O₂ were added to create a variety of reaction conditions as follows: (i) pathophysiological conditions [21] (p-NGB_S-S and p-NGB_SH) of 0.1 mM NO₂⁻ and 0.15 mM H₂O₂ (divided into five aliquots of 30 μM each); (ii) harsh conditions [h-NGB_S-S (modified NGB at high NO₂⁻ and H₂O₂ concentrations) and h-NGB_SH] of 0.1 M NO₂⁻ and 1.0 mM H₂O₂ (divided into five aliquots); (iii) modifications without NO₂⁻ [p′-NGB_SH and h′-NGB_SH (modified NGB_SH at high H₂O₂ concentrations)] of 0.15 mM H₂O₂ (divided into five ali-quots of 30 μM each) and 1.0 mM H₂O₂ (divided into five aliquots) respectively, and (iv) LPO-catalysed modifications [lpo-NGB_S-S (modified NGB in the presence of LPO, NO₂⁻ and H₂O₂) and lpo-NGB_SH] of 0.25 M NO₂⁻, 0.15 mM H₂O₂ (divided into five aliquots) and 80 nM LPO.

The proteins were allowed to react at 20 °C for 10 min. Excess NO₂⁻ and oxidant were removed by overnight dialysis against 20 mM phosphate buffer, pH 7.5 (6.8 mM NaH₂PO₄ and 13.2 mM Na₂HPO₄) for NGB_S-S derivatives, and against degassed 20 mM Tris/HCl, pH 7.5, and 0.2 mM EDTA for NGB_SH derivatives.

Analysis of protein fragments and haem modification

For analysis of protein fragments, approx. 0.5 mg of unmodified NGB_S-S and NGB_SH and the derivatives p-NGB_S-S, h-NGB_S-S, lpo-NGB_S-S, p-NGB_SH, h-NGB_SH, p′-NGB_SH, h′-NGB_SH and lpo-NGB_SH was transformed into the apo protein using the standard HCl/2-butanone method [31] and was subsequently hydrolysed by trypsin. Digestion was performed using 1:50 (w/w) trypsin at 37 °C for 3 h in 20 mM ammonium bicarbonate solution, pH 8.0. Prior to HPLC-MS/MS analysis, the samples were incubated with 5 mM DTT for 15 min at 37 °C to prevent coupling of the cysteine-containing peptides during digestion. The reduction of the disulfide bonds was necessary for the quantification of unreacted cysteine residues in NGB derivatives.

Haem modification was studied by direct HPLC-MS/MS analysis of NGB derivatives in 20 mM ammonium bicarbonate solution, pH 8.0, after acidification with HCl to pH∼1.

LC (liquid chromatography)-MS and LC-MS/MS data were obtained using an LCQ ADV MAX ion-trap mass spectrometer equipped with an ESI (electrospray ionization) source and controlled by Xcalibur software 1.3 (Thermo Finnigan). ESI experiments were carried out in positive-ion mode under the following constant instrumental conditions: source voltage of 5.0 kV, capillary voltage of 46 V, capillary temperature of 210 °C and tube lens voltage of 55 V. The system was run in automated LC-MS/MS mode using a Surveyor HPLC system (Thermo Finnigan) equipped with a BioBasic™ C₁₈ column (5 μm; 150×2.1 mm). The elution was performed using 0.1% methanoic acid in distilled water (solvent A) and 0.1% methanoic acid in acetonitrile (solvent B), at a flow rate of 0.2 ml/min. Elution started with 98% solvent A for 5 min, followed by a linear gradient from 98 to 55% solvent A in 65 min for the analysis of tryptic digests. A two-step linear gradient was used to analyse the non-digested protein solutions (98 to 50% solvent A in 5 min, followed a gradient from 50 to 20% solvent A in 40 min). MS/MS spectra obtained by CID (collision-induced dissociation) were performed with an isolation width of 2 m/z, the activation amplitude was approx. 35% of the ejection RF (radio frequency) amplitude of the instrument.

For the analysis of protein fragments derived from NGB derivatives, the mass spectrometer was set such that one full MS scan was followed by a zoomscan and a MS/MS scan on the most intense ion observed from the MS spectrum. To identify the modified residues, the acquired MS/MS spectra were automatically searched against a protein database for NGB, using the SEQUEST® algorithm incorporated into Bioworks 3.1 (Thermo Finnigan).

GdnHCl denaturation assay

The denaturation stability of NGB_SH, p-NGB_SH and h-NGB_SH was determined by monitoring the absorbance variation of the Soret band of the protein (approx. 7 μM) after addition of increasing amounts of 8 M GdnHCl solution (up to 4.5 M final concentration) in degassed 50 mM Tris/HCl, pH 6.0, and 0.5 mM EDTA. Data were corrected for dilution caused by the addition of GdnHCl. The GdnHCl concentration at 50% unfolding of the protein was evaluated from the curves of absorbance against denaturant concentration according to a standard method [32].

RESULTS

The two forms of NGB, with and without the disulfide bond

The purification of human NGB provides the protein in its ferric form, since oxyNGB was unstable in vitro and rapidly autoxidizes to metNGB (t_1/2=11 min) [7]. This was confirmed by the UV–visible spectrum characteristics of metNGB (Figure 1), as it was clearly distinguishable in the visible region from the spectrum of the Fe^II–O₂ form [for metNGB, λ_max=532 nm, with a shoulder at 554 nm (Figure 1); for oxyNGB, λ_max=543, 576 nm (results not shown)] [7,12]. The molar absorption coefficient of the Soret band of metNGB in the disulfide-bridged form (metNGB_S-S), obtained using the standard pyridine haemochrome assay [33] and used to determine protein concentration, was 1.29×10⁵ M⁻¹·cm⁻¹ at 414 nm (in 200 mM phosphate buffer, pH 7.5). The phosphate buffer used in our experiments is believed to promote the formation of disulfide bonds (Scheme 1) [12]. Actually, the number of accessible thiol groups per haem, as measured by the 4-PDS assay [12,34], was slightly lower than 1.0. This result is consistent with the presence of the single Cys¹²⁰(G19) in the reduced form and with the almost quantitative oxidation of the other two cysteines to a S–S bond [9]. These disulfide bonds are thought to play a significant role in the function of the protein [9,35].

Figure 1. UV–visible spectra of metNGB_S-S and NGB_SH.

UV–visible spectra of 6.7 μM metNGB_S-S (solid line) and 6.2 μM metNGB_SH (dashed line) in 200 mM phosphate buffer, pH 7.5. The wavelength maxima are indicated (arrows).

Treatment of NGB_S-S with DTT under anaerobic conditions leads to the partial reduction of the disulfide bridge, as NGB_SH was stable to oxidation when kept in degassed 50 mM Tris/HCl, pH 7.5, and 0.5 mM EDTA at 4 °C. Different experimental conditions (temperature and reaction time) were examined to maximize the reduction of the S–S bond, with 2.4 cysteines per haem titrated in the protein observed after reaction with DTT for 30 min at 37 °C. It should be noted that the 4-PDS assay underestimates the number of thiol groups, since the initial fast reduction of 4-PDS to 4-thiopyridone by the SH groups of cysteine residues overlaps with the subsequent slow formation of the latter product by reaction with some reducing species that can be present in solution; therefore, it can be assumed that, in NGB_SH, almost all cysteine residues are in the reduced form.

The UV–visible spectra of the ferric forms of NGB_S-S and NGB_SH are identical (λ_max=414 and 532 nm respectively, with a shoulder at 554 nm) and are characteristic of a six-co-ordinated ferric haem (Figure 1). Actually, even if the ratio between the high-spin and low-spin forms increased on formation of the disulfide bridge, the amount of the high-spin form would remain low (<5% at pH 7.5 [36] and <10% at pH 5 [37]). However, the only species detectable in the UV–visible spectra for both NGB_S-S and NGB_SH was the predominant six-co-ordinated species.

Lack of phenol oxidation by NGB

In the presence of H₂O₂, haem proteins such as Mb in their ferric form exhibit peroxidase-like activity towards phenolic substrates, oxidizing them to phenoxy radicals, which can give rise to dimer formation in solution [38,39]. The formation of these species can be followed spectrophotometrically through a characteristic absorption at a wavelength of approx. 300 nm. Moreover, a protein–ferryl intermediate (named compound II) can be detected in the reactions of both Mb and peroxidases after addition of the oxidant. This was characterized by a red-shift of the protein Soret band from 408 to 424 nm for Mb [40], 403 to 420 nm for horseradish peroxidase [41] and 412 to 422 nm for LPO [42]. In contrast, no spectral change was observed after mixing of NGB (1 μM) with H₂O₂ (increasing concentrations up to 100 mM) in the absence or presence of a phenol, such as HPA. Indeed, Herold et al. [19] also reported that, in contrast with Mb and Hb, NGB does not generate the ferryl form of the protein, since its reactivity was strongly limited by the co-ordinated distal histidine residue. In fact, when NGB was reacted with H₂O₂ and HPA, the UV–visual spectrum features that are characteristic of phenolic dimers do not develop, indicating that the reactivity of NGB with H₂O₂ was negligible.

Phenol nitration is catalysed by NGB_S-S and NGB_SH

Surprisingly, NGB activates NO₂⁻ in the presence of H₂O₂, and this results in the nitration of the phenolic substrate HPA into the position o-relative to the hydroxy group. An analogous catalytic activity was reported for peroxidases [21,22,43,44] and the haem proteins Mb [23,24,45,46] and Hb [46]. The investigation of the kinetics of phenol nitration catalysed by NGB (both the NGB_S-S and NGB_SH forms) was not simple, since the reaction rate depends on the concentration of all the reagents. In order to gain an insight into the mechanism of nitration by the metNGB/NO₂⁻/H₂O₂ system, it was necessary to investigate the reaction by varying the concentration of the reagents in a large range, far beyond the physiological concentrations likely to be encountered in vivo. Using the approximations and procedures followed previously for enzymatic nitration reactions [22], it was possible to obtain the kinetic parameters k_cat (the maximum turnover of the proteins), K_M^Ph-OH (the dissociation constant of HPA from the complex formed with the proteins) and k_cat/K_M^Ph-OH (used to calculate the dependence of the rate of reaction at low reagent concentrations on HPA) from the rate dependence on the HPA concentration, through the following equation:

(5)

and k_cat, K_M^nitrite (the dissociation constant of NO₂⁻ from the complex formed with the proteins) and k_cat/K_M^nitrite (used to calculate the dependence of the rate of reaction at low reagent concentrations on NO₂⁻) from the rate dependence on NO₂⁻ concentration, through the following equation:

(6)

The data collected are displayed in Table 1, and the corresponding data for the nitration of HPA as catalysed by hMb [24] are also shown.

Table 1. Kinetic data for HPA nitration catalysed by NGB and hMb.

Steady-state kinetic parameters for NGB- and hMb-dependent HPA nitration by NO₂⁻/H₂O₂ in 200 mM phosphate buffer, pH 7.5, at 25 °C.

Protein	KMPh-OH (mM)	kcat (s−1)	kcat/KMPh-OH (M−1·s−1)	KMnitrite (M)	kcat (s−1)	kcat/KMnitrite (M−1·s−1)
NGBS-S	0.078±0.008	0.177±0.004	2300±200	0.043±0.006	0.140±0.007	3.3±0.3
NGBSH	0.058±0.008	0.132±0.005	2300±200	0.22±0.03	0.139±0.005	0.63±0.06
hMb*	0.12±0.02	0.67±0.03	5600±900	0.24±0.04	1.03±0.08	4.3±0.5

*Data from [24].

It is worth noting that the overall catalytic efficiency of NGB (expressed by k_cat, k_cat/K_M^Ph-OH and k_cat/K_M^nitrite) was only slightly lower than that of Mb. This was surprising because the reactivity of the protein in the nitration reaction was expected to be negligible, owing to the presence of the six-co-ordinated haem group in NGB, as was observed for the phenol oxidation. Since the fraction of five-co-ordinated high-spin metNGB was less than 5% of total metNGB, its intrinsic reactivity must be very high, much greater than that of Mb. The interpretation of the kinetic parameters is complicated by the presence of the haem–distal histidine residue dissociation equilibrium, since the reported value of k_−H (rate constant for histidine residue release) of approx. 1 s⁻¹ [36]) was of the same order of magnitude as the NGB k_cat. It follows that the high-spin/low-spin equilibrium is neither slow nor fast, so that, during the kinetic studies, a steady-state equilibrium was not reached in the initial phase and the predominant form of the protein that was present in solution tended to change with time.

The most interesting information emerged from the catalytic activity studied as a function of NO₂⁻ concentration, since the disulfide-bridged NGB_S-S exhibited a higher affinity for NO₂⁻ with respect to the thiol form NGB_SH (K_M^nitrite=43 and 220 mM respectively). For ferrous NGB_SH, the distal histidine residue dissociation rate was nearly one order of magnitude smaller than that of NGB_S-S, thus leading to an effective decrease in O₂ affinity [9]. A similar effect of the S–S bond can be considered for the metNGBs on the iron affinity towards exogenous ligands such as NO₂⁻. Therefore the larger K_M^nitrite value observed for NGB_SH is in agreement with the requirement of NO₂⁻ co-ordination to the iron of NGB in order to promote nitrating activity.

The catalytic efficiency at low concentrations of NO₂⁻ (i.e. k_cat/K_M^nitrite) was similar for NGB_S-S and hMb, since the lower turnover rate of the former is compensated by its higher affinity for NO₂⁻.

Binding of NO₂⁻ to NGB_S-S and NGB_SH

Small but significant changes can be observed in the spectrum of NGB_S-S in solution after the addition of excess NO₂⁻. Difference spectra from optical titration clearly show both a smoothing and a slight shift towards higher energy of the Soret band during the first part of the binding process (NO₂⁻ concentration <50 mM), and a slight decrease in the intensity of the band in the second part of the titration (NO₂⁻ concentration >50 mM). The fitting of the changes in the Soret band as a function of NO₂⁻ concentration was consistent with low-affinity binding of two ligands, yielding the constants K₁=51±15 M⁻¹ and K₂=5±2 M⁻¹ (Figure 2). The former constant probably refers to the electrostatic interaction of an NO₂⁻ ion with the charged amino acid residues of the protein, and the latter constant to the co-ordination of an NO₂⁻ ion to the iron centre. As NO₂⁻ has to replace the distal histidine residue from the sixth co-ordination site, the affinity of NGB_S-S for NO₂⁻ results approx. one order of magnitude lower with respect to that previously obtained for hMb under the same conditions (K_B=76 M⁻¹) [24].

Figure 2. Binding of NO₂⁻ to NGB_S-S and NGB_SH.

The difference between A₄₁₄ and A₄₃₄ in the UV–visible spectra of NGB_S-S (Ngb_S-S) and NGB_SH (Ngb_SH) after addition of NO₂⁻ compared with the ligand concentration is shown. The absorbance data were fitted with the binding isotherm for low-affinity binding of two ligands for NGB_S-S, and a single ligand for NGB_SH.

The NO₂⁻-binding experiment performed with NGB_SH showed a different behaviour with respect to NGB_S-S, since only the initial spectral changes were similar for the two proteins. Indeed, the data could be fitted with a single binding isotherm, giving the association constant K_B=48±5 M⁻¹, which was very similar to the K₁ value obtained with NGB_S-S (Figure 2). By analogy, the process can be associated with the electrostatic interaction of NO₂⁻ to NGB_SH. It was impossible to obtain a reliable binding constant for the direct co-ordination of NO₂⁻ to the haem iron for NGB_SH, thus confirming the lower affinity of the thiol form of the protein for exogenous ligands with respect to the disulfide-bridged NGB, as was also observed from the kinetic studies of phenol nitration.

The interaction of NGB with NO₂⁻ was also investigated by Herold et al. [19], but, at the low NO₂⁻ concentration employed (<1 mM), the binding of NO₂⁻ was not observed. Since metNGB was active in promoting the nitration reaction at low concentrations of NO₂⁻ (pathophysiological values), a small and hardly detectable fraction of the haem iron must be accessible to the ligand.

Nitration of 4-hydroxybenzonitrile and phenylacetic acid

The substrates phenylacetic acid and 4-hydroxybenzonitrile were used as mechanistic probes for the nitration reaction catalysed by NGB/NO₂⁻/H₂O₂. The reaction could proceed through the formation of the same nitrating species involved in the reactions promoted by peroxidases and Mb, i.e. nitrogen dioxide (NO₂) or ONOO⁻, depending on the NO₂⁻ concentration [22–24]. Phenylacetic acid reacts with ONOO⁻, generating nitrated and/or hydroxylated compounds, but it is unreactive to NO₂ and thus it is a good probe for ONOO⁻. The reactivity of 4-hydroxybenzonitrile is different, since it reacts with NO₂ to generate 4-hydroxy-3-nitrobenzonitrile, whereas ONOO⁻ is a poor nitrating agent for this substrate [22] (Scheme 2).

Scheme 2. Reactivity of 4-hydroxybenzonitrile and phenylacetic acid with RNS.

Schematic representation of the reactivity of 4-hydroxybenzonitrile and phenylacetic acid with the nitrating agents NO₂ and ONOO⁻, and the catalytic system metNGB/NO₂⁻/H₂O₂.

4-Hydroxybenzonitrile was reacted with NGB and H₂O₂ at different NO₂⁻ concentrations (10–150 mM) and the reaction mixtures were analysed by HPLC. The nitrated product was absent under all of the conditions tested, indicating that NO₂ was not formed, or was formed in negligible amounts, by NGB under these conditions. In contrast, NGB/NO₂⁻/H₂O₂ induced the modification of phenylacetic acid, and by HPLC it was possible to separate four different products, regardless of the NO₂⁻ concentration used. These derivatives shared spectral features and retention times with the products obtained in the reaction of phenylacetic acid with ONOO⁻ [22]. These results suggest that the nitration catalysed by NGB in the presence of NO₂⁻ and H₂O₂ proceeds through the formation of a reactive species with the chemical properties of ONOO⁻ (Scheme 2).

Tandem MS/MS analysis of NGB_S-S and NGB_SH modified by H₂O₂ and NO₂⁻

To investigate the endogenous modifications undergone by NGB in the presence of NO₂⁻ and H₂O₂, NGB (both in the NGB_S-S and NGB_SH forms) was reacted with varying quantities of NO₂⁻ and H₂O₂ in the absence of an external substrate. Depending on the concentration of the reagents used, a series of protein derivatives was prepared under very mild conditions (similar to those that can occur under pathophysiological conditions [21]), and NGB_S-S and NGB_SH were reacted with 0.1 mM NO₂⁻ and 0.15 mM H₂O₂, obtaining p-NGB_S-S and p-NGB_SH respectively. Using less physiological and more harsh conditions, NGB_S-S and NGB_SH were reacted with 100 mM NO₂⁻ and 1 mM H₂O₂, obtaining h-NGB_S-S and h-NGB_SH respectively. The reactivity of the cysteine residues in the presence of H₂O₂ alone was studied by reacting NGB_SH with H₂O₂ at low (0.15 mM) and high (1 mM) oxidant concentrations, obtaining p′-NGB_SH and h′-NGB_SH respectively.

Moreover, generation of the NGB derivatives by NO₂⁻ and H₂O₂ was performed in the presence of LPO, as an external source of the nitrating agent. The protein derivatives obtained under these conditions were named as lpo-NGB_S-S and lpo-NGB_SH.

The UV–visible spectra of the p-NGB derivatives were indistinguishable from that of native metNGB, whereas the spectra for the h-NGB and lpo-NGB derivatives both exhibited a slight broadening of the Soret band and the appearance of a shoulder at approx. 440–450 nm was also observed. These spectral features were associated with haem nitration and reflected the colour change of the protein solutions from brown to greenish-brown, which accompanied the formation of h-NGB_S-S, h-NGB_SH, lpo-NGB_S-S and lpo-NGB_SH derivatives [24]. All of the protein derivatives were analysed by HPLC-ESI-MS/MS, with the aim of investigating the endogenous modifications and the results are shown in Tables 2 and 3. The potential targets of the oxidizing and nitrating species are, besides the haem prosthetic group, the protein tyrosine residues Tyr⁴⁴(CD3), Tyr⁸⁸ (F3), Tyr¹¹⁵(G14) and Tyr¹³⁷(H12), and the cysteine residues Cys⁴⁶(CD7), Cys⁵⁵(D5) and Cys¹²⁰(G19) in NGB_SH, with Cys¹²⁰(G19) being the only cysteine residue target in NGB_S-S (Figure 3).

Table 2. Haem and tyrosine residue modifications in NGB derivatives.

Nitration of haem and tyrosine residues in NGB derivatives obtained under the following conditions: for p-NGBs, [NGB]=60 μM, [NO₂⁻]=0.1 mM and [H₂O₂]=0.15 mM; for h-NGBs, [NGB]=60 μM, [NO₂⁻]=0.1 M and [H₂O₂]=1 mM; for lpo-NGBs, [LPO]=80 nM, [NGB]=60 μM, [NO₂]=0.25 M and [H₂O₂]=0.15 mM. All reactions were performed in 200 mM phosphate buffer, pH 7.5.

	Nitration (%)
Protein derivatives	Haem-NO2	Tyr44-NO2	Tyr88-NO2	Tyr115-NO2	Tyr137-NO2
p-NGBS-S	0	0	0.6	0	0
p-NGBSH	0	0	1	0	0
h-NGBS-S	38	50	11	65	38
h-NGBSH	14	11	7	36	14
lpo-NGBS-S	25	5	1.3	3.5	15
lpo-NGBSH	27	2.3	0.9	7	11

Table 3. Modification of cysteine residues in NGB derivatives.

Oxidation of cysteine residues in NGB derivatives obtained by protein modification under the following conditions: for p-NGBs, [NGB]=60 μM, [NO₂⁻]=0.1 mM and [H₂O₂]=0.15 mM; for h-NGBs, [NGB]=60 μM, [NO₂⁻]=0.1 M and [H₂O₂]=1 mM; for lpo-NGBs, [LPO]=80 nM, [NGB]=60 μM, [NO₂⁻]=0.25 M and [H₂O₂]=0.15 mM; for p′-NGB_SH, [NGB]=60 μM and [H₂O₂]=0.15 mM; for h′-NGBs, [NGB]=60 μM and [H₂O₂]=1 mM. All reactions were performed in 200 mM phosphate buffer, pH 7.5.

	Oxidation (%)
Protein derivatives	Cys46-SO2H	Cys46-SO3H	Cys55-SO2H	Cys55-SO3H	Cys120-SO2H	Cys120-SO3H
NGBS-S	*	*	*	*	1.7	1.5
NGBSH	0.3	0.9	0.6	1.7	2	2
p-NGBS-S	*	*	*	*	2	2
p-NGBSH	2.3	4.5	2.5	7	10	9
p′-NGBSH	1	2	0.5	2	2	1.5
h-NGBS-S	*	*	*	*	5	8
h-NGBSH	21	14	4	14	18	16
h′-NGBSH	0.9	4	0.7	4	5	6
lpo-NGBS-S	*	*	*	*	6	7
lpo-NGBSH	22	8	3	7	10	5

*Cys⁴⁶ and Cys⁵⁵ are involved in the S–S bond.

Figure 3. Structure of NGB.

Structure of the C46G/C55S/C120S mutant of NGB. The atomic co-ordinates were obtained from [10] and the Protein Data Bank (accession number 1OJ6). The disposition of the side chains of the tyrosine residues (Tyr⁴⁴, Tyr⁸⁸, Tyr¹¹⁵ and Tyr¹³⁷), and the cysteine residues (Cys⁴⁶, Cys⁵⁵, and Cys¹²⁰) present in the wild-type protein are shown.

The modification of the prosthetic group can be detected by direct HPLC-MS/MS analysis of the acidified protein solution, since the haem group was released from the protein. The modified haemin was detected only in the NGB derivatives obtained after reaction with high levels of NO₂⁻ and H₂O₂, i.e. h-NGB_S-S and h-NGB_SH. In particular, ions with m/z 616 and 661 were identified, corresponding to free haemin and haemin modified with a NO₂ group [23,24] respectively. The modification probably occurs at one of the haem vinyl groups, in analogy with the haem nitration observed after treatment of Mb with a large excess of NO₂⁻ at pH 5.5 [47]. Integration of the peaks in the EIC (extracted ion current) chromatograms indicated that haemin nitration occurred with a 38% yield in h-NGB_S-S, a 14% yield in h-NGB_SH, a 25% yield in lpo-NGB_S-S and a 27% yield in lpo-NGB_SH (Table 2).

The modifications at the cysteine and tyrosine residues were characterized on polypeptide fragments resulting from tryptic digestion of the apoNGB derivatives. The HPLC-ESI-MS/MS analysis showed the presence of many modified peptide fragments (for the molecular masses and m/z values as bicharged ions, see Supplementary Table S1 at http://www.BiochemJ.org/bj/407/bj4070089add.htm). The analysis of the MS/MS spectra using the SEQUEST® algorithm allowed the assignment of the modifications to the nitration of the tyrosine residues to 3-nitrotyrosine [23] and the oxidation of the cysteine residues to sulfinic (R-SO₂H) and sulfonic (R-SO₃H) acids. As an example of the MS results, the MS/MS spectrum of the peptide from amino acid residue 68 to 94, obtained by the analysis of h-NGB_S-S (nitrated at Tyr⁸⁸), is shown in Figure 4.

Figure 4. CID spectrum of the modified amino acid 68–94 peptide.

MS/MS spectrum of the m/z 1477.6 peak (mass of 2953.3 Da) assigned to the amino acid 68–94 peptide in a double-charged state and with Tyr⁸⁸ modified to 3-nitrotyrosine. The assignment of the y and b ion series is shown. Above the spectrum, the sequence of the 68–94 peptide, with the modified residue in bold, and the summary of the y and b ions found in the spectrum is shown.

The percentage of residue modifications is shown in Tables 2 and 3, and was obtained by comparing the areas of the peaks in the EIC chromatograms that corresponded to the derivatized peptides with those of the corresponding peptides in the starting proteins. The greater efficiency of NGB_S-S, in promoting self-nitration of both the haem group and the tyrosine residues by NO₂⁻ and H₂O₂, compared with NGB_SH was clearly apparent from the amount of endogenous modifications observed for h-NGB_S-S and h-NGB_SH (Table 2). This is in agreement with the kinetic parameters for phenol nitration (Table 1), since NGB_S-S exhibits a significantly higher catalytic efficiency in terms of k_cat/K_M^nitrite compared with NGB_SH.

A complete picture of the cysteine residue derivative formation is provided by the modification of NGB_SH under different conditions, resulting in the formation of the h-, p-, h′- and p′-NGB_SH derivatives (Table 3). All of the cysteine residues were oxidized to sulfinic and sulfonic acids and the amount of modification induced by NO₂⁻ and H₂O₂ increased with the concentration of the reagents (compare p-NGB_SH with h-NGB_SH), and was significantly lower in the absence of NO₂⁻ (h′-NGB_SH compared with p′-NGB_SH), indicating that the NGB/NO₂⁻/H₂O₂ system generates distinctive oxidant species.

The mechanism of protein modification was explored by inducing the formation of the NGB derivatives in the presence of an external catalyst. The enzyme LPO was chosen, since it reacts with H₂O₂ and NO₂⁻ to generate nitrating species with a rate two orders of magnitude larger when compared with NGB [22]. Thus when NGB reacted with NO₂⁻ at low H₂O₂ concentration and in the presence of LPO, the reactive species were generated by the enzyme, making NGB self-promoted derivative formation negligible. This was confirmed by the similarity in the extent of nitration (both at the haem group and the tyrosine residues) obtained for the LPO-promoted modification of NGB_S-S and NGB_SH (Table 2).

From the analysis of the derivatives lpo-NGB_S-S and lpo-NGB_SH, it emerged that the IPO-promoted NGB tyrosine residue nitration and cysteine residue oxidation occurred to a lower extent compared with the modifications which are self-promoted by NGB. In contrast, the haem group was the favoured modification site for both lpo-NGBs (Tables 2 and 3). These results indicated that the prosthetic group was the most accessible target for the LPO-generated reactive species, whereas the NGB-generated reactive species exhibited a preference for intramolecular, rather than intermolecular, reactivity.

GdnHCl denaturation assay

The unfolding midpoint [D₀] in the presence of GdnHCl for NGB_SH, p-NGB_SH and h-NGB_SH was estimated to be in the range of 4.1–4.2 M, as obtained from the curves of absorbance against denaturant concentration. These values can be compared with those obtained under the same conditions for sperm whale Mb ([D₀]=1.46 M [32]), and horse heart Mb ([D₀]=1.32 M) (E. Monzani, S. Nicolis, R. Roncone, M. Barbieri, A. Granata and L. Casella, unpublished work). Therefore NGB exhibited a considerably higher stability to GdnHCl than Mbs. Moreover, the derivatization of the protein residues (nitration of tyrosine residues and oxidation of cysteine residues) in the h-NGB_SH derivative had very little effect on the stability of the protein. Treatment of NGB with pathophysiological NO₂⁻ and H₂O₂ concentrations, yielding p-NGB_SH, induced completely negligible effects on protein stability to GdnHCl denaturation. A similar negligible effect towards denaturation was observed for horse heart Mb after nitration of tyrosine residues in the presence of NO₂⁻ and H₂O₂ ([D₀]=1.35 M compared with 1.32 M for native Mb). A more detailed characterization of the unfolding properties of NGB and its derivatives will be carried out separately.

DISCUSSION

In vivo NGB species

Since NGB possesses a bis-histidine six-co-ordinated haem group, the reactivity of the protein towards exogenous ligands is regulated by the dissociation rate of the distal histidine residue. Nevertheless, NGB in its ferrous state has significant reactivity towards small ligands, such as NO, O₂ and CO (carbon monoxide). The competition between these external molecules and the histidine residue for the sixth co-ordination site of the Fe^II centre emerges from the kinetic and thermodynamic constants previously reported [3,4,48]. It follows that, in spite of the high intrinsic affinity of NGB for both O₂ and CO, the binding of these ligands is suggested to be slow in vivo [7]. Furthermore, the influence of the internal disulfide bridge on the affinity of ferrous NGB_S-S for molecular O₂ has been clearly demonstrated [9].

In the reducing conditions that exist within living cells, NGB would be present in the ferrous NGB_SH form, saturated with O₂ at approx. 12% [12]. Although the ferric form studied in the present paper may not be the most physiologically relevant form of the protein, the presence of H₂O₂ and NO₂⁻ would rapidly convert ferrous or oxyNGB into the ferric form. Furthermore, metNGB is the product of rapid NO scavenging by oxyNGB, so, in the absence of a metNGB reductase, a system would accumulate metNGB (according to reaction 1) [15].

MetNGB promoted nitration of phenolic compounds

Currently, the only data concerning the behaviour of metNGB with H₂O₂ indicated an absence of reactivity [19]. The results reported in the present study show an interesting pseudo-enzymatic activity of metNGB in the presence of H₂O₂ and NO₂⁻. Even if the mechanism of the NGB reaction could be considered to be similar to that of peroxidases, this is in contrast with the absence of activity of metNGB in the oxidation of phenolic compounds observed in the presence of H₂O₂ alone. In fact, under these conditions, metNGB does not generate the ferryl species compounds I and II, which are the characteristic intermediates of the catalytic cycle of peroxidases. Nevertheless, in analogy with peroxidases, metNGB promotes the nitration of phenolic substrates by activation of NO₂⁻ and H₂O₂. This result is not in contradiction with the absence of the reaction observed between NGB and H₂O₂, since, besides the peroxidase-like mechanism of nitration through NO₂, an alternative mechanism for the nitration of phenolic substrates emerged in our previous studies on peroxidases and Mb in the presence of NO₂⁻ and H₂O₂ [22,23]. This mechanism involved the initial binding of NO₂⁻ to the haem iron, followed by the reaction of H₂O₂ with the co-ordinated NO₂⁻, to produce an iron-bound ONOO⁻ [NGBFe^III-N(O)OO], which acts as the nitrating active species. In the case of peroxidases and Mbs, the ONOO⁻ pathway becomes important only at relatively high NO₂⁻ concentrations [22–24], whereas with NGB it is demonstrated that this is the only mechanism active in the nitration of exogenous and endogenous phenolic groups (i.e. tyrosine residues) even in the pathophysiological concentration range of NO₂⁻ and H₂O₂ (Scheme 3).

Scheme 3. Mechanism of metNGB promoted nitration of phenolic compounds.

Schematic representation of the activation of NO₂⁻ and H₂O₂ by the high-spin metNGBs (metNGB_S-S or metNGB_SH) through the ONOO⁻ pathway, involving the initial binding of NO₂⁻ to the haem iron, followed by the reaction with H₂O₂, to produce an iron-bound ONOO⁻ (NGBFe^III-N(O)OO) nitrating active species. Nitration of phenol occurs at the o-position relative to its hydroxy group. The upper part of the scheme shows the absence of reactivity of metNGB with H₂O₂.

The ONOO⁻ pathway

The kinetic parameters for HPA nitration reported in Table 1 show the importance of NO₂⁻ in discriminating the reactivity of metNGB_S-S and metNGB_SH. As expected from the influence of the internal S–S bond on the binding rate of exogenous ligands to the iron centre, the kinetic parameters for the two forms of the protein and their dependence on NO₂⁻ concentration are significantly different, supporting the hypothesis of direct co-ordination of NO₂⁻ to the Fe^III centre to promote catalysis. In particular, metNGB_S-S exhibits larger values, both for NO₂⁻ affinity [extrapolated from kinetic data (1/K_M^nitrite)] and for reactivity at low NO₂⁻ concentrations (k_cat/K_M^nitrite) with respect to metNGB_SH, indicating that the interaction of NO₂⁻ with the catalytic site regulates the efficiency of nitration at low concentrations of the reagent, in conditions that can occur in vivo. The higher affinity of NO₂⁻ for metNGB_S-S compared with metNGB_SH was confirmed by NO₂⁻-binding studies (Figure 2).

The ONOO⁻ pathway proposed for the nitrating activity of metNGB was examined by the observation of the reaction with 4-hydroxybenzonitrile and phenylacetic acid, which are mechanistic probes of NO₂ and ONOO⁻ respectively. The lack of reactivity of hydroxybenzonitrile on one hand, and the presence of several products deriving from nitration and/or hydroxylation of phenylacetic acid on the other hand, is in agreement with the formation of an active species with the chemical properties of ONOO⁻ by metNGB (Scheme 2). It follows that the formation of a protein-bound ONOO⁻ intermediate is a reasonable step for nitration catalysed by metNGB in the presence of NO₂⁻ and H₂O₂, even at rather low concentrations of the reagents.

NGB endogenous modifications

The generation and activity of the oxidant species formed by metNGB was examined by identification of the pattern of protein modifications obtained using LPO as an external catalyst, in comparison with metNGB-promoted endogenous modifications (Tables 2 and 3). At the high NO₂⁻ concentration (250 mM) used for the preparation of lpo-NGB_S-S and lpo-NGB_SH derivatives, LPO-promoted nitration proceeded through the formation of the ONOO⁻ active species [22]. The differences in the pattern of residues modified by the LPO-generated ONOO⁻ in comparison with those self-induced by NGB indicated that, in the latter case, ONOO⁻ reacts intramolecularly. Moreover, the ONOO⁻ reactive species are probably responsible for the oxidation of cysteine residues to form the corresponding sulfinic and sulfonic acids. Not only can it promote this type of modifications, as observed in the oxidations of BSA and hMb [24,49], but the high reactivity of this species also accounts for the lower extent of derivatization observed in the h′-NGB_SH and p′-NGB_SH derivatives (obtained by reaction with H₂O₂ only) in comparison with the h-NGB_SH and p-NGB_SH derivatives respectively. The less physiological conditions used in the preparation of the h-NGB_SH and h-NGB_S-S derivatives, which forced the extent of endogenous modifications, allowed the investigation of the competitive reactivity of the protein residues. Both the presence of three reactive cysteine residues (which may scavenge RNS) in NGB_SH, and its poorer catalytic efficiency in the nitration of phenolic compounds compared with NGB_S-S, may account for the lower degree of tyrosine residue nitration in h-NGB_SH.

From the point of view of the physiological relevance of NGB endogenous modifications, the analysis of p-NGB_SH was of particular interest, since this form was obtained from NGB_SH (presumed to be the predominant form in vivo [12]) with low concentrations of NO₂⁻ and H₂O₂. Among the protein residues, cysteine residues had the highest reactivity towards the ONOO⁻ active species, and their reaction prevented nitration of tyrosine residues and the haem prosthetic group. It is possible to calculate that approx. 36% of H₂O₂-oxidizing equivalents introduced were used for oxidation of cysteine residues to sulfinic and sulfonic acids. Oxidation of cysteine residues to sulfinic and sulfonic acid derivatives has been observed [50,51], together with other types of oxidations that can be reversed by reaction with biological thiols [52]. It is worth noting that, among the tyrosine residues, the only one that undergoes nitration in both p-NGB_S-S and p-NGB_SH is Tyr⁸⁸, the phenolic hydroxy group of which points toward the large NGB cavity connected to the haem distal site [10]. This may indicate that the ONOO⁻ reactive species endogenously generated at the haem distal site can easily diffuse through the cavity and nitrate Tyr⁸⁸.

Does human NGB act only as a scavenger?

In the globins, cysteine residues often play specific roles, and this is modulated by the formation of intramolecular or intermolecular disulfide bonds [9]. In the case of NGB, the oxidation state of Cys⁴⁶ and Cys⁵⁵ is relevant because it is linked to the peculiar six-co-ordination of the haem. The present study confirms the importance of the internal S–S bond in controlling the reactivity of metNGB towards NO₂⁻ and H₂O₂. It also shows that the protein can be involved in other activities. NGB is thought to act as a scavenger of toxic species generated in vivo under conditions of oxidative stress. But the ability of the metNGB form generated by this activity to activate NO₂⁻ and H₂O₂ means that, besides acting as a scavenger, NGB can be the source of RNS, thus affecting its physiological activity. A significant fraction of the nitrating and oxidizing species (approx. 36% in the conditions used to prepare p-NGB_SH) does not escape from the protein, but is consumed in the self-modification of NGB. The easy oxidation of cysteine residues to sulfinic and sulfonic acids may be one way used by NGB to scavenge part of the generated RNS. To gain a better understanding of the physiological relevance of these reactions, it will be important to analyse the state of endogenous modification of samples of the protein isolated after in vivo exposure to oxidative stress conditions.