Introduction

Peroxynitrite, generated from the diffusion-limited reaction of ·NO with O2·− [1], is limited from the availability of ·NO and O2·−. In particular, the reactions of Hb(II)-O2 and Mb(II)-O2 with ·NO and of Hb(II)-NO and Mb(II)-NO with O2 limit the ·NO disposal [2,3,4,5], leading to the formation of the harmless NO3− and of the ferric heme-protein derivative (i.e., Mb(III) and Hb(III), respectively) via the transient formation of the heme-Fe-bound peroxynitrite [3,4,5,6]. Then, Mb(III) and Hb(III) are converted to Mb(II) and Hb(II) by the cytochrome b5/cytochrome b5 reductase system [7,8,9]. In parallel, the availability of O2·− is modulated by superoxide dismutase, which leads to the formation of H2O2 and O2 [10]; then, H2O2 is removed by catalase [11].

Peroxynitrite displays a relevant biological role oxidizing proteins, nucleic acids, and membrane lipids [12,13,14,15,16,17]. Among others, peroxynitrite impairs O2 uptake, transport and delivery by ferrous hemoglobin (Hb(II)-O2) and myoglobin (Mb(II)-O2) oxidizing the heme-Fe(II)-O2 atom via the transient formation of ferryl-Hb (Hb(IV)=O) and ferryl-Mb (Mb(IV)=O) [6]. Then, Hb(IV)=O and Mb(IV)=O convert ·NO and NO2− to nitrite and nitrate, respectively [18, 19]. Moreover, peroxynitrite induces the oxidation of the heme-Fe(II) atom of nitrosylated and carbonylated Hb (i.e., Hb(II)-NO and Hb(II)-CO) and Mb (i.e., Mb(II)-NO and Mb(II)-CO), leading to the release of these diatomic ligands [20,21,22]. In turn, Hb(III) and Mb(III) catalyze the isomerization of peroxynitrite to NO3− displaying cytoprotective effects, such as the prevention of protein nitration [3, 13, 14].

Peroxynitrite scavenging by Hb(III) and Mb(III) is impaired by cyanide [3]. Here, we report the inhibition of the Mb(III)-mediated peroxynitrite detoxification also by ·NO, at pH 7.2 and 20.0 °C. However, unlike for cyanide, this process is not due to a simple competition between the two ligands, since even NO-bound Mb(III) (i.e., Mb(III)-NO) is able to isomerize peroxynitrite to nitrate, though with a reduced rate; this feature appears to be peculiar of Mb(III), which does not undergo reductive nitrosylation at neutral pH [23], whereas Hb(III) is rapidly converted to Hb(II)-NO at pH > 6 [23].

Materials

Recombinant ferric sperm whale Mb was obtained from Sigma-Aldrich (St. Louis, MO, USA). The Mb(III) concentration was determined spectrophotometrically at 409.5 (ε = 1.57 × 105 M−1 cm−1) [24]. Peroxynitrite was purchased from Cayman Chemical (Ann Arbor, Michigan, USA). The concentration of peroxynitrite was determined spectrophotometrically at 302 nm (ε = 1.705 × 103 M−1 cm−1) [14]. Gaseous ·NO (from Linde Caracciolossigeno S.r.l., Roma, Italy) was purified under anaerobic conditions by flowing through a glass column packed with NaOH pellets and then by passage through a trap containing 20 mL of 5.0 M NaOH solution to remove trace impurities. The NO pressure was 760.0 mmHg [24]. The stock ·NO solution was prepared anaerobically by keeping in a closed vessel the degassed 5.0 × 10–2 M phosphate buffer solution (pH 7.2) under ·NO at P = 760.0 mm Hg (T = 20.0 °C). The solubility of ·NO in the aqueous buffered solution is 2.05 × 10–3 M, at P = 760.0 mm Hg and T = 20.0 °C [25]. The concentration of ·NO was determined spectrophotometrically (between 380 and 450 nm), in the presence of dithionite (5.0 × 10–3 M), by the titration of ferrous sperm whale Mb (Mb(II); 7.5 × 10–6 M). The ·NO:Mb(II) molar ratio ranged between 0.12 and 3.6. No gaseous phase was present [26]. All the other chemicals were purchased from Merck KGaA (Darmstadt, Germany). All chemicals were of analytical grade and were used without further purification.

Methods

·NO binding to Mb(III) was investigated at pH 7.2 (5.0 × 10–2 M phosphate buffer) and 20.0 °C, and analyzed in the framework of Scheme 1 [23].

Scheme 1
scheme 1

·NO binding to Mb(III)

Full size image

Values of the apparent first-order rate constant for ·NO binding to Mb(III) (i.e., ·NOk) were obtained by rapid-mixing the heme-protein solution (final concentration ranging between 3.1 × 10–6 M and 3.4 × 10–6 M) with the ·NO solution (final concentration, 2.0 × 10–5 M to 1.0 × 10–3 M). The ·NO:Mb(III) molar ratio ranged between 5.9 and 320. No gaseous phase was present. Kinetics was monitored by single-wavelength rapid-mixing stopped-flow spectroscopy between 380 and 450 nm [23].

Values of ·NOk for ·NO binding to Mb(III) were obtained according to Eqs. 1 and 2 [23]:

$$ \left[ {{\text{Mb}}\left( {{\text{III}}} \right)} \right]_{{\text{t}}} = \, \left[ {{\text{Mb}}\left( {{\text{III}}} \right)} \right]_{{\text{i}}} \times {\text{ e}}^{{{-}^{{ \cdot {\text{NO}}}} k \times t}} $$
(1)
$$ \left[ {{\text{Mb}}\left( {{\text{III}}} \right)} \right]_{{\text{t}}} = \, \left[ {{\text{Mb}}\left( {{\text{III}}} \right)} \right]_{{\text{i}}} \times \, \left( {{1 } - {\text{ e}}^{{{-} \, ^{{ \cdot {\text{NO}}}} k \times t}} } \right) $$
(2)

depending on the observation wavelength. The amplitude of the time courses for ·NO binding to Mb(III) was normalized to that observed at 409 nm.

The value of the apparent second-order rate constant for ·NO binding to Mb(III) (i.e., ·NOkon) and of the first-order rate constant for Mb(III)-NO denitrosylation (i.e., ·NOkoff) were obtained according to Eq. 3 [23]:

$$^{{ \cdot {\text{NO}}}} k =^{{ \cdot {\text{NO}}}} k_{{{\text{on}}}} \times \, \left[ {{}^{ \cdot }{\text{NO}}} \right] \, +^{{ \cdot {\text{NO}}}} k_{{{\text{off}}}} $$
(3)

The value of the apparent dissociation equilibrium constant for ·NO binding to Mb(III) (i.e., ·NOK = ·NOkoff/·NOkon) was obtained by mixing the heme-protein solution (final concentration, 3.8 × 10–6 M) with the ·NO solution (final concentration, 4.2 × 10–6 M to 1.0 × 10–3 M). The ·NO:Mb(III) molar ratio ranged between 1.1 and 264. No gaseous phase was present. Thermodynamics was monitored spectrophotometrically between 380 and 450 nm [23].

The value of ·NOK was determined from the dependence of the molar fraction of the ligand-bound ferric heme-protein (i.e., ·NOY) on the ligand concentration (i.e., [·NO]) according to Eq. 4 [23]:

$$^{{ \cdot {\text{NO}}}} Y = {{\left[ {{}^{ \cdot }{\text{NO}}} \right]^{n} } \mathord{\left/ {\vphantom {{\left[ {{}^{ \cdot }{\text{NO}}} \right]^{n} } {\left( {\left[ {{}^{ \cdot }{\text{NO}}} \right]^{n} +^{{ \cdot {\text{NO}}}} K^{n} } \right)}}} \right. \kern-\nulldelimiterspace} {\left( {\left[ {{}^{ \cdot }{\text{NO}}} \right]^{n} +^{{ \cdot {\text{NO}}}} K^{n} } \right)}} $$
(4)

where n is the Hill coefficient.

In the absence and presence of ·NO (final concentration, 2.5 × 10–5 M to 2.5 × 10–4 M), values of the apparent pseudo-first-order rate constant at pH 7.2 (5.0 × 10–2 M phosphate buffer) and 20.0 °C for Mb(III)-mediated peroxynitrite isomerization (i.e., Pk) were determined by rapid mixing the ferric heme-protein solutions (final concentration ranging between 3.2 × 10–6 M and 3.0 × 10–5 M) with the peroxynitrite solution (final concentration, 2.0 × 10–4 M). The peroxynitrite:Mb(III) molar ratio ranged between 6.7 and 62.5. No gaseous phase was present. Kinetics was monitored spectrophotometrically at 302 nm [1, 3, 4, 14, 27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42].

Values of Pk for Mb(III)-mediated peroxynitrite isomerization were obtained, in the absence and presence of ·NO (final concentration, 2.5 × 10–5 M to 1.0 × 10–3 M), according to Eq. 5 [1, 3, 4, 14, 27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42]:

$$ \left[ {{\text{peroxynitrite}}} \right]_{{\text{t}}} = \, \left[ {{\text{peroxynitrite}}} \right]_{{\text{i}}} \times {\text{ e}}^{{{-} \, {}^{{\text{P}}}k \times t}} $$
(5)

Values of the apparent second-order rate constant for peroxynitrite isomerization by Mb(III) and Mb(III)-NO (i.e., Pkonobs) as well as of the apparent first-order rate constant for the spontaneous decay of peroxynitrite (i.e., Pk0) were obtained from the dependence of Pk on the total Mb concentration (i.e., [Mb]tot = [Mb(III)] + [Mb(III)-NO]), according to Eq. 6 [1, 3, 4, 14, 27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42]:

$$ {}^{{\text{P}}}k = {}^{{\text{P}}}k_{{{\text{on}}}}^{{{\text{obs}}}} \times \, \left[ {{\text{Mb}}} \right]_{{\text{tot}}} + {}^{{^{{\text{P}}} }}k_{0} $$
(6)

The dependence of Pkonobs on the ·NO concentration was analyzed in the framework of Scheme 2 that envisages the possibility that peroxynitrite isomerization is undertaken not only by Mb(III) but also by Mb(III)-NO unless Pkon* = 0.

Scheme 2
scheme 2

Mb(III)-mediated peroxynitrite isomerization in the absence and presence of ·NO. Pkon and Pkon* are the second order rate constants for peroxynitrite isomerization by Mb(III) and Mb(III)-NO, respectively, Pk0 is the first-order rate constant for peroxynitrite isomerization in the absence of Mb(III), ·NOkon is the second order rate constant for ·NO association to Mb(III), ·NOkoff is the first-order rate constant for ·NO dissociation from Mb(III)-NO, and ·NOK (= ·NOkoff/·NOkon) is the equilibrium dissociation constant for ·NO binding to Mb(III)

Full size image

On the basis of Scheme 2, values of the different parameters have been obtained according to the following Eq. 7, under the assumption that Pkon and Pkon* are the rate-limiting steps of the whole process reported in Scheme 2 [3, 4, 30,31,32,33,34,35,36]:

$$^{{\text{P}}} k_{{{\text{on}}}}^{{{\text{obs}}}} = {{^{{\text{P}}} k_{{{\text{on}}}} \times^{{ \cdot {\text{NO}}}}K} \mathord{\left/ {\vphantom {{^{{\text{P}}} k_{{{\text{on}}}} \times^{{ \cdot {\text{NO}}}}K} {{{\left( {^{{ \cdot {\text{NO}}}} K + \, \left[ {^{ \cdot } {\text{NO}}} \right]} \right) \, +^{{\text{P}}}{k_{{{\text{on}}}}}^{*} \, \times \, \left[ {^{ \cdot } {\text{NO}}} \right] \, } \mathord{\left/ {\vphantom {{\left( {^{{ \cdot {\text{NO}}}} K + \, \left[ {^{ \cdot } {\text{NO}}} \right]} \right) \, +^{{\text{P}}}{k_{{{\text{on}}}}}^{*} \, \times \, \left[ {^{ \cdot } {\text{NO}}} \right] \, } {\left( {^{{ \cdot {\text{NO}}}} K + \, \left[ {^{ \cdot } {\text{NO}}} \right]} \right)}}} \right. \kern-\nulldelimiterspace} {\left( {^{{ \cdot {\text{NO}}}} K + \, \left[ {^{ \cdot } {\text{NO}}} \right]} \right)}} \, }}} \right. \kern-\nulldelimiterspace} {{{\left( {^{{ \cdot {\text{NO}}}} K + \, \left[ {^{ \cdot } {\text{NO}}} \right]} \right) \, +^{{\text{P}}}{k_{{{\text{on}}}}}^{*} \, \times \, \left[ {^{ \cdot } {\text{NO}}} \right] \, } \mathord{\left/ {\vphantom {{\left( {^{{ \cdot {\text{NO}}}} K + \, \left[ {^{ \cdot } {\text{NO}}} \right]} \right) \, +^{{\text{P}}}{k_{{{\text{on}}}}}^{*} \, \times \, \left[ {^{ \cdot } {\text{NO}}} \right] \, } {\left( {^{{ \cdot {\text{NO}}}} K + \, \left[ {^{ \cdot } {\text{NO}}} \right]} \right)}}} \right. \kern-\nulldelimiterspace} {\left( {^{{ \cdot {\text{NO}}}} K + \, \left[ {^{ \cdot } {\text{NO}}} \right]} \right)}} \, }} $$
(7)

Although from resulting parameters (see below) it comes out that actually a significant fraction of peroxynitrite is isomerized through Pk0 (see Scheme 2), the amount of peroxynitrite reacting with Mb(III) and Mb(III)-NO is sufficiently high to give a significant and reproducible signal, allowing a thorough analysis of these pathways. On the basis of Scheme 2 and Eq. 7 it comes out that if Pkon* = 0, ·NO behaves as a purely competitive ligand, which must be dissociated from Mb(III) in order it to isomerize peroxynitrite; conversely, if Pkon* ≠ 0, ·NO behaves as a non-competitive modulator of the peroxynitrite isomerization, which can then be accomplished to a different rate by both Mb(III) and Mb(III)-NO.

The results are given as mean values of at least five experiments plus or minus the corresponding standard deviation. All data were analyzed using the GraphPad Prism program, version 5.03 (GraphPad Software, La Jolla, CA, USA).

All density functional theory (DFT) calculations were performed using the Gaussian 16 program package [43]. In these calculations, the meta-GGA functional TPSSh was employed in combination with the def2-TZVP basis set for all atoms [44, 45]. A superfine integration grid was used throughout. Geometry optimizations were followed by frequency calculations to confirm that the structures correspond to local energy minima on the potential energy surface. Following previous studies, spin-unrestricted calculations using a singlet configuration were employed [46]. Bulk solvent effects (water) were included using the integral equation formalism of the polarized continuum model [47].

Results and discussion

At pH < 8, Mb(III) binds reversibly ·NO, the reductive nitrosylation of the metal center being negligible [23]. Indeed, pumping off ·NO from the Mb(III)-NO solution leads to the formation of Mb(III). Also the reaction of peroxynitrite with ·NO is negligible, the value of the second order rate constant being 5 × 10−2 M−1 s−1 at most [48].

Over the whole wavelength range explored (i.e., between 380 and 450 nm), the time course of ·NO binding to Mb(III) follows a single exponential process for more than 95 ± 6% of its course (Fig. 1a) according to Eqs. 1 and 2. Values of the apparent pseudo-first-order rate constant for ·NO binding to Mb(III) (i.e., ·NOk), determined according to Eqs. 1 and 2 (Fig. 1a), depend linearly on the ·NO concentration (Fig. 1b). The analysis of data (Fig. 1b) according to Eq. 3 allowed the determination of values of ·NOkon (representing the slope of the straight line) and ·NOkoff (corresponding to the y intercept of the straight line). Values of ·NOkon for Mb(III) nitrosylation and of ·NOkoff for ·NO dissociation from Mb(III)-NO are (6.8 ± 0.7) × 104 M−1 s−1 and 3.1 ± 0.3 s−1, respectively.

Fig. 1
figure 1

·NO binding to Mb(III), at pH 7.2 and 20.0 °C. a Averaged time courses of ·NO binding to Mb(III). Data analysis according to Eq. 1 allowed us to determine the following values of ·NOkobs: 9.6 s−1 (trace a), 1.6 × 101 (trace b), and 3.1 × 101 s−1 (trace c). The ·NO concentration was 1.0 × 10–4 M (trace a), 2.0 × 10–4 M (trace b), and 4.0 × 10–4 M (trace c). The Mb(III) concentration was 3.2 × 10–6 M. b Dependence of ·NOkobs on the ·NO concentration. Kinetics were analyzed according to Eq. 3 with ·NOkon = (6.8 ± 0.7) × 104 M−1 s−1 and ·NOkoff = 3.1 ± 0.3 s−1. c Dependence of the molar faction of Mb(III)-NO (i.e., ·NOY) on the ·NO concentration. Data analysis according to Eq. 4 allowed us to determine the following value of ·NOK = (3.8 ± 0.5) × 10−5 M. The Mb(III) concentration was 3.8 × 10–6 M. The ·NO concentration refers to that of the free ligand. Where not shown, the standard deviation is smaller than the symbol

Full size image

Values of ·NOY for ·NO binding to Mb(III) (Fig. 1c) increase with the ·NO concentration tending to level off at ·NOY ≫ ·NOK (see Scheme 1 and Eq. 4). The analysis of data (Fig. 1c) according to Eq. 4 allowed the determination of the ·NOK value for Mb(III) nitrosylation corresponding to (3.8 ± 0.5) × 10−5 M. According to Scheme 1: (i) the value of the Hill coefficient n (see Eq. 4) is 0.99 ± 0.2, and (ii) the value of ·NOK [= (3.8 ± 0.5) × 10−5 M] is in excellent agreement with that of the ·NOkoff/·NOkon ratio (= (4.6 ± 0.7) × 10−5 M). Moreover, the values of ·NOkon (= 6.8 ± 0.7 × 104 M−1 s−1) and ·NOK (= 3.8 ± 0.5) × 10−5 M) reasonably agree with those reported in the literature (i.e., ·NOkon = 7.0 × 104 M−1 s−1 and ·NOK = 7.7 × 10−5 M, respectively), considering the different experimental conditions [23].

Under all the experimental conditions, the time course of peroxynitrite isomerization was fitted to a single exponential process for more than 94 ± 6% of its course (Fig. 2a, b) according to Eq. 5. In agreement with literature [31,32,33,34,35,36, 38, 39, 41, 42, 49], this suggests that: (i) no stable intermediate species (see Scheme 2) accumulate(s) in the course of peroxynitrite isomerization, and (ii) the formation of the transient Fe(III)-OONO− species represents the rate-limiting step in catalysis.

Fig. 2
figure 2

Effect of Mb(III) and Mb(III)-NO on kinetics of peroxynitrite isomerization, at pH 7.2 and 20.0 °C. a Averaged time courses of Mb(III)-mediated isomerization of peroxynitrite in the absence of NO. Data analysis according to Eq. 5 allowed us to determine the following values of Pkobs: 2.8 × 10−1 s−1 (trace a), 4.8 × 10−1 s−1 (trace b), and 8.5 × 10−1 s−1 (trace c). The Mb(III) concentration was 0.0 M (trace a), 1.0 × 10–5 M (trace b), and 3.0 × 10–5 M (trace c). The peroxynitrite concentration was 2.0 × 10–4 M. b Averaged time courses of the Mb(III)-mediated isomerization of peroxynitrite in the presence of NO. Data analysis according to Eq. 5 allowed us to determine the following values of Pkobs: 2.6 × 10−1 s−1 (trace a), 3.3 × 10−1 s−1 (trace b), and 4.5 × 10−1 s−1 (trace c). The Mb(III) concentration was 0.0 M (trace a), 1.0 × 10–5 M (trace b), and 3.0 × 10–5 M (trace c). The peroxynitrite concentration was 2.0 × 10–4 M. The ·NO concentration was 1.0 × 10–4 M. c Dependence of Pkobs on the Mb(III) concentration in the absence (open circles) and presence of 7.0 × 10–5 M (filled circles) and 2.5 × 10–4 M (filled diamonds) ·NO. The continuous lines were obtained according to Eq. 6 with the following parameters: open circles: Pkonobs = (1.9 ± 0.2) × 104 M–1 s–1 and Pk0 = (2.8 ± 0.3) × 10−1 s–1; filled circles: Pkonobs = (1.1 ± 0.1) × 104 M–1 s–1 and Pk0 = (2.8 ± 0.3) × 10−1 s–1; filled diamonds: Pkonobs = (5.4 ± 0.5) × 103 M–1 s–1 and Pk0 = (2.8 ± 0.3) × 10−1 s–1. d Dependence of Pkonobs for Mb(III)-catalyzed isomerization of peroxynitrite on the ·NO concentration. Continuous and dashed lines were obtained according to Eq. 7 with values of parameters given in Table 1. e Dependence of Pk0 for the spontaneous isomerization of peroxynitrite on the ·NO concentration. The average value of Pk0 obtained in the presence of ·NO (= (2.8 ± 0.3) × 10–1 s−1; filled circles) corresponds to that obtained in the absence of ·NO (= (2.8 ± 0.3) × 10–1 s−1; open circle on the ordinate). Where not shown, the standard deviation is smaller than the symbol

Full size image
Table 1 Values of kinetic and thermodynamic parameters (Pkon, ·NOK, and Pkon*) derived from the analysis of data shown in Fig. 2d according to Eq. 7
Full size table

In the absence and presence of ·NO, values of the apparent pseudo-first-order rate constant for Mb(III)-mediated peroxynitrite scavenging (i.e., Pk), determined according to Eq. 5 (Fig. 2a, b), increase linearly with the heme-protein concentration (i.e., Fig. 2c). The analysis of the data shown in Fig. 2c, according to Eq. 6, allowed us to determine the values of the second-order rate constant for peroxynitrite isomerization by Mb(III) (i.e., Pkonobs, corresponding to the slope of the linear plots) and of the first-order rate constant for the spontaneous isomerization of peroxynitrite (i.e., Pk0, corresponding to the y intercept of the straight line).

As shown in Fig. 2c, d, the apparent second-order rate constant for peroxynitrite isomerization (i.e., Pkonobs) decreases upon increasing the ·NO concentration. In fact, the Pkonobs value for Mb(III)-catalyzed isomerization of peroxynitrite decreases from (1.9 ± 0.2) × 104 M−1 s−1 in the absence of ·NO (i.e., Pkon; see Scheme 2 and Eq. 7) to (3.8 ± 0.5) × 103 M−1 s−1 at the ·NO concentration of 2.5 × 10−4 M. In contrast, the values of Pk0 are unaffected by ·NO, being (2.8 ± 0.3) × 10–1 s−1 in the absence and (2.8 ± 0.3) × 10–1 s−1 in the presence of ·NO (Fig. 2e). The value of Pkon obtained in the absence of ·NO and the values of Pk0 determined in the absence and presence of ·NO (Fig. 2c, d, e) are in excellent agreement with those reported in the literature (Pkon = 1.2 × 104 M−1 s−1 and Pk0 = (3.0 ± 0.5) × 10–1 s−1, respectively) [1, 3, 30, 32, 33, 36, 39, 41, 42]. The pivotal role of the heme-Fe(III) atom in peroxynitrite isomerization is outlined from the fact that the value of the apparent pseudo-first-order rate constant for peroxynitrite scavenging obtained in the presence of apoMb (Pk = 2.7 × 10−1 s−1) [3] corresponds to that obtained in the absence of Mb(III) (Pk0 = 2.8 × 10−1 s−1) (present study and Ref [3].). Moreover, the occurrence in Mb(III) of a ligand (e.g., nitrate) binding site that enhances the heme-Fe(III)-mediated peroxynitrite detoxification must be discarded since the time course of peroxynitrite isomerization follows a simple behavior (Fig. 2a, b) for more than 94% of its course.

Since with the exception of Pkon* we know all parameters in Scheme 2 and Eq. 7, we determined this value from the dependence of Pkonobs on the ·NO concentration, as reported in Fig. 2d. From this analysis it clearly emerges that Pkon* ≠ 0, ruling out the possibility that ·NO works as a competitive inhibitor. Therefore, also the Mb(III)-NO species is able to facilitate the isomerization of peroxynitrite, though with a reduced rate, corresponding to Pkon* = (2.8 ± 0.3) × 103 M‒1 s‒1. This result is very interesting, since it drastically differs from what observed for azide, cyanide, and isoniazide, which behave as purely competitive inhibitors of peroxynitrite isomerization by ferric horse heart Mb [3], human heme-albumin [30], horse heart carboxymethylated cytochrome c [33], Mycobacterium tuberculosis truncated HbN [36], Pseudoalteromonas haloplanktis TAC125 truncated HbO [37], Campylobacter jejuni truncated HbP [40], Methanosarcina acetivorans protoglobin [38], human haptoglobin:hemoglobin complex [41], human nitrobindin [42], and microperoxidase-11 [39], envisaging a peculiar behavior for the modulation of peroxynitrite isomerization in the case of ·NO.

The mixed competitive-non-competitive mechanism of ·NO (see Fig. 2d) implies that peroxynitrite isomerization is facilitated also when the heme-Fe(III) atom is NO-bound. This occurrence can be explained by assuming that either (i) peroxynitrite is an outer sphere ligand, which can be isomerized to NO3‒ even without binding directly to the ferric metal center, or (ii) that ·NO and peroxynitrite bind to the heme-Fe(III) atom on the opposite axial coordination position in a trans geometrical arrangement, or else (iii) the ferric metal assumes a seven-coordination geometry.

The possibility that peroxynitrite is an outer sphere ligand is reminiscent to that of dithionite bound to Hb(II) at the interface of two adjacent tetramers in the crystal lattice [50]. However, this hypothesis appears unlikely in the light of the observation that in cytochrome c only the cleavage of the axial heme-Fe(III)-Met80 bond allows peroxynitrite detoxification [32, 33].

A more interesting and realistic event is the occurrence of an opposite axial coordination, as it has been shown to take place in the case of ferrous NO-bound forms of guanylate cyclase [51], Alcaligenes xylosoxidans cytochrome c’ [52], and of cardiolipin-bound cytochrome c [53, 54], but it has never been reported in the case of NO-bound ferric heme-proteins. However, it is known that: (i) the strength of the proximal bond to the heme-Fe atom is lower for the ferric than for the ferrous form [55, 56], and (ii) ·NO binding brings about a dramatic reduction of the strength of the proximal His-Fe bond [57]. Since in ferric hemoproteins the pKa of the proximal His-Fe bond (= 4.7) [55] is about 1.7 pH units higher than in ferrous hemoproteins (= 3.0 ± 0.3) [56] and ·NO binding brings about an increase of pKa in ferrous NO-bound hemoproteins (= 4.8 ± 0.2) [57], it is reasonable to expect that in Mb(III)-NO the pKa could be about 6.4 ± 0.4. Of note, pKa values of ~ 6.9 and ~ 6.5 have been reported for azide-bound and ligand-free Dermochelys coriacea and sperm whale Mb(III) [58], even though they were not explicitly attributed to the cleavage of the proximal Fe-His bond. This being eventually the case, it is likely that, under conditions of the present study, a significant fraction (10–40%) of the Mb(III)-NO might be penta-coordinated with only one axial ligand, thus being able to coordinate peroxynitrite as a sixth axial ligand. Moreover, we cannot rule out the possibility of a simultaneous binding of ·NO and peroxynitrite, even in the presence of the Fe-His proximal bond, bringing about the formation of a seven-coordinate species, as reported for phthalocyaninatoiron(II), which is able to bind transiently two pyridine and one carbon monoxide molecule [59].

Of note, since the X-ray irradiation facilitates the reduction of heme-Fe(III) to heme-Fe(II) via water radiolysis, including the formation of OH− [60, 61], the structure of the Mb(III)-NO complex cannot be determined. In fact, OH− induces the fast conversion of Mb(III)-NO to Mb(II)-NO [23]. In any way, azide, fluoride, formate, and thiocyanate that bind selectively to the heme-Fe(III) atom of Mb weaken significantly the proximal HisF8-Fe(III) bond [62,63,64,65].

Therefore, DFT calculations were carried out to gain information on the role that the heme-Fe(III) group has in peroxynitrite isomerization (see computational details above). Thus, the heme group has been modelled as a porphynate dianion coordinated to Fe(III) and ·NO, with the sixth coordination position being occupied by either a 4-methyl-1H-imidazolate (MI−) or a peroxynitrite anion (ONOO−), leaving peroxynitrous acid (ONOOH) or 4-methyl-1H-imidazol (HMI) in the second-sphere. The calculated structures of the two models, denoted as [FeP(NO)(MI)]·ONOOH and [FeP(NO)(OONO)]·HMI, are shown in Fig. 3. The calculated Fe–N distances involving the ·NO ligand, which give values of 1.642 and 1.637 Å, are in good agreement with the values observed for Fe(III) porphyrinate complexes binding different ligands at the sixth coordination position. As expected, the Fe–N–O angle (> 177º) shows little deviations from linearity. The calculated Fe–N distances to the porphyrinate donor atoms (2.009–2.011 Å) are also in excellent agreement with experimental data obtained from X-ray diffraction studies [66]. The Fe–N distance to the donor atom of the MI− ligand (1.967 Å) and the ν(N–O) stretching frequency (1973 cm−1) are also in excellent agreement with the experimental values (1.973 Å and 1920 cm−1, respectively). The peroxynitrite ligand coordinates to Fe(III) through the terminal peroxy O atom and adopts a cis conformation, in agreement with previous studies on related systems [67].

Fig. 3
figure 3

Optimized geometries of the FeP(NO)(OONO)]·HMI and [FeP(NO)(MI)]·ONOOH complexes obtained with DFT calculations (TPSSh/def2-TZVP)

Full size image

The calculated relative energies of the [FeP(NO)(MI)]·ONOOH and [FeP(NO)(OONO)]·HMI systems favor peroxynitrite coordination by 0.9 kcal mol−1. The inclusion of thermal corrections and zero-point energy corrections does not change this trend, affording a free energy of 3.3 kcal mol−1. Thus, these calculations: (i) support the hypothesis that peroxynitrite isomerization is facilitated upon binding to the opposite axial coordination position of the heme-Fe(III) atom with respect to the ·NO ligand, and (ii) disregard the occurrence of the seven-coordination geometry of the heme-Fe(III) atom (Fig. 3).

Conclusions

Over decades, Mb(II) has been reported to act not only as an O2 reservoir but also to facilitate O2 transport from the cell periphery to the mitochondria [68]. Recently, ferrous oxygenated Mb (Mb(II)-O2) has been reported to act as an intracellular scavenger of ·NO in red muscle, thereby protecting cellular respiration [8, 9, 69]. Here, ·NO-linked inhibition of the Mb(III)-based detoxification of peroxynitrite is reported; this expands to a wider extent the multiple roles of ·NO in modulating heme-protein actions. In this respect, it is very important to outline that peroxynitrite has been reported to react with ferrous nitrosylated human Hb [70], horse heart Mb [70], human Ngb [71], soybean (Glycine max) leghemoglobin [72], and rabbit hemopexin [73]. Peroxynitrite induces the oxidation of the ferrous nitrosylated metal center (i.e., heme-Fe(II)-NO) leading to the formation of the ferric nitrosylated species (i.e., Fe(III)-NO) and eventually the release of NO under conditions where the heme-Fe(II):NO stoichiometry is approx. 1:1. Therefore, since also in the Fe(II)-NO the proximal Fe-His bond is known to be fairly weak, the simultaneous coordination of NO and peroxynitrite might be envisaged as a likely transient species and present data for Mb(III)-NO and peroxynitrite underlie a more general mechanism of reciprocal modulation by the two ligands. As a whole, the reversible binding of ·NO to Mb(III) introduces a new modulatory role of ·NO, which does not impair peroxynitrite detoxification, but regulates the rate and the extent of this process. Therefore, such a feature envisages a delicate balancing between peroxynitrite and ·NO, which is modulated through the relative amount of Mb(III) and Mb(III)-NO.