The Balancing of Peroxynitrite Detoxification between Ferric Heme-Proteins and CO2: The Case of Zebrafish Nitrobindin

Nitrobindins (Nbs) are all-β-barrel heme proteins and are present in prokaryotes and eukaryotes. Although their function(s) is still obscure, Nbs trap NO and inactivate peroxynitrite. Here, the kinetics of peroxynitrite scavenging by ferric Danio rerio Nb (Dr-Nb(III)) in the absence and presence of CO2 is reported. The Dr-Nb(III)-catalyzed scavenging of peroxynitrite is facilitated by a low pH, indicating that the heme protein interacts preferentially with peroxynitrous acid, leading to the formation of nitrate (~91%) and nitrite (~9%). The physiological levels of CO2 dramatically facilitate the spontaneous decay of peroxynitrite, overwhelming the scavenging activity of Dr-Nb(III). The effect of Dr-Nb(III) on the peroxynitrite-induced nitration of L-tyrosine was also investigated. Dr-Nb(III) inhibits the peroxynitrite-mediated nitration of free L-tyrosine, while, in the presence of CO2, Dr-Nb(III) does not impair nitro-L-tyrosine formation. The comparative analysis of the present results with data reported in the literature indicates that, to act as efficient peroxynitrite scavengers in vivo, i.e., in the presence of physiological levels of CO2, the ferric heme protein concentration must be higher than 10−4 M. Thus, only the circulating ferric hemoglobin levels appear to be high enough to efficiently compete with CO2/HCO3− in peroxynitrite inactivation. The present results are of the utmost importance for tissues, like the eye retina in fish, where blood circulation is critical for adaptation to diving conditions.

Tyrosine nitration through the peroxynitrite pathway (see Scheme 1) is operative in most fish, including zebrafish (Danio rerio), under stressed conditions [47,48]. Further, the effect of the enrichment of CO2 in the atmosphere brings about the raising of CO2 levels also in the oceans [49]. Although the increased levels of CO2 create only limited behavioral effects, they may significantly affect the body response to external insults. In addition, the increased concentration of CO2 in the blood is of particular relevance in fish, for which retinal circulation is of crucial importance due to the high hydrostatic pressure of the environment where they live [50,51].
Here, the kinetics of ONOO -/ONOOH (hereafter peroxynitrite) inactivation by all-βbarrel ferric Danio rerio nitrobindin (Dr-Nb(III)), in the absence and presence of CO2, are reported and analyzed in parallel with those of all-α-helical globins. The present results open a question on the competition between the protecting activity of heme proteins (which inactivate the damaging reactions of peroxynitrite on L-tyrosine nitration) and the non-protecting action of CO2/HCO3 − , which indeed depends on the levels of ferric hemeproteins. Thus, the levels (and the scavenging activity) of most heme proteins might be too low to act as peroxynitrite scavengers in vivo; only circulating ferric hemoglobin (Hb(III)) levels appear to be high enough to efficiently compete with CO2/HCO3 − /CO3 2-in peroxynitrite inactivation. Moreover, the reaction of ONOO − with CO 2 /HCO 3 − /CO 3 2− redirects the ONOO − / ONOOH-based oxidation of aromatic and aliphatic amino acid residues, facilitating the nitration of tyrosine and tryptophan and limiting the oxidation of methionine and cysteine [5,[9][10][11][12][13][14][15]20,21,[42][43][44][45]. Remarkably, • NO 2 and CO 3 •− radicals are much stronger oxidant species than their precursors • NO, O 2 •− , and ONOO − /ONOOH [9][10][11][12][13]20,21,46]. Tyrosine nitration through the peroxynitrite pathway (see Scheme 1) is operative in most fish, including zebrafish (Danio rerio), under stressed conditions [47,48]. Further, the effect of the enrichment of CO 2 in the atmosphere brings about the raising of CO 2 levels also in the oceans [49]. Although the increased levels of CO 2 create only limited behavioral effects, they may significantly affect the body response to external insults. In addition, the increased concentration of CO 2 in the blood is of particular relevance in fish, for which retinal circulation is of crucial importance due to the high hydrostatic pressure of the environment where they live [50,51].
Here, the kinetics of ONOO − /ONOOH (hereafter peroxynitrite) inactivation by allβ-barrel ferric Danio rerio nitrobindin (Dr-Nb(III)), in the absence and presence of CO 2 , are reported and analyzed in parallel with those of all-α-helical globins. The present results open a question on the competition between the protecting activity of heme proteins (which inactivate the damaging reactions of peroxynitrite on L-tyrosine nitration) and the non-protecting action of CO 2 /HCO 3 − , which indeed depends on the levels of ferric heme-proteins. Thus, the levels (and the scavenging activity) of most heme proteins might be too low to act as peroxynitrite scavengers in vivo; only circulating ferric hemoglobin (Hb(III)) levels appear to be high enough to efficiently compete with CO 2 /HCO 3 − /CO 3 2-in peroxynitrite inactivation.

Methods
In the absence and presence of CO 2 , peroxynitrite isomerization by Dr-Nb(III) and apo-Dr-Nb was investigated at pH 5.8, 7.0, and 8.5 (5.0 × 10 −2 M Bis-Tris propane buffer) and 22.0 • C, under anaerobic conditions. In the presence of CO 2 , peroxynitrite isomerization by Dr-Nb(III) and apo-Dr-Nb was investigated by adding NaHCO 3 (final concentration, 5.0 × 10 −1 M) to the Dr-Nb(III) and apo-Dr-Nb solutions; this NaHCO 3 concentration corresponds to different concentrations of CO 2 , HCO 3 − , and CO 3 2depending on pH. After the addition of NaHCO 3 , the Dr-Nb(III) and apo-Dr-Nb solutions were allowed to equilibrate for at least 5 min; then, the pH was readjusted to the desired pH if needed [25,27,28,30,36,39].
The percentage of NO 3 − and NO 2 − obtained from peroxynitrite isomerization was determined spectrophotometrically at 543 nm by using the Griess reagent and vanadium(III) chloride (VCl 3 ) to catalyze the conversion of NO 3 − to NO 2 − , according to literature [25,30,36,39,56]. Kinetic data were analyzed using the GraphPad Prism program, version 5.03 (Graph-Pad Software, San Diego, CA, USA). The results are given as the mean values of at least four experiments plus and minus the standard deviation. Data were analyzed using GraphPad Prism 6 (GraphPad Software Inc., San Diego, CA, USA) and were expressed as mean values ± standard deviation (SD). Statistical analysis was performed using either the Student's t-test or the one-way ANOVA comparison test (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001).

Results and Discussion
In the absence and presence of CO 2 , the absorbance at 302 nm decreases upon mixing the Dr-Nb(III), apo-Dr-Nb, and peroxynitrite solutions over the whole pH range explored.
Under all the experimental conditions, the time course of peroxynitrite isomerization in the absence and presence of Dr-Nb(III), apo-Dr-Nb, and CO 2 was fitted to a singleexponential decay for more than 93% of its course, according to Equation (1) (Figures 1-3, panels A and B). The values of the pseudo-first-order rate constant for peroxynitrite isomerization catalyzed by Dr-Nb(III) (i.e., k obs ) increase with the heme protein concentration (Figures 1-3, panels C and D). Previous experiments have shown that, once bound to a heme protein (likely as Fe(III)-ONOOH), peroxynitrite isomerizes to NO 3 − at a rate of 70 s −1 [13]. Therefore, since the values of k obs range between 1 × 10 −1 s −1 and 8 s −1 (Figures 1-3, panel C), they indicate that, under our conditions, the formation of the Dr-Nb(III)-OONOH species represents the rate-limiting step of the catalytic process. Notably, the Dr-Nb(III)-OONOH conversion to Dr-Nb(III) and NO 3 − is faster by at least 10-fold than the formation of the Dr-Nb(III)-OONOH complex.
Under all the experimental conditions, the time course of peroxynitrite isomerization in the absence and presence of Dr-Nb(III), apo-Dr-Nb, and CO2 was fitted to a singleexponential decay for more than 93% of its course, according to Equation (1) (Figures 1-3, panels A and B). The values of the pseudo-first-order rate constant for peroxynitrite isomerization catalyzed by Dr-Nb(III) (i.e., kobs) increase with the heme protein concentration (Figures 1-3, panels C and D). Previous experiments have shown that, once bound to a heme protein (likely as Fe(III)-ONOOH), peroxynitrite isomerizes to NO3 − at a rate of ~70 s −1 [13]. Therefore, since the values of kobs range between 1 × 10 −1 s −1 and 8 s −1 (Figures 1-3, panel C), they indicate that, under our conditions, the formation of the Dr-Nb(III)-OONOH species represents the rate-limiting step of the catalytic process. Notably, the Dr-Nb(III)-OONOH conversion to Dr-Nb(III) and NO3 − is faster by at least 10-fold than the formation of the Dr-Nb(III)-OONOH complex.      The analysis of the data shown in Figures 1-3 (panels C and D), according to Equation (2), allowed us to determine the values of kon and k0, corresponding to the slope and the yintercept of the linear plots, respectively (Table 1). Moreover, the values of k0 were measured in the absence of Dr-Nb(III) via the rapid mixing of the peroxynitrite solution with the appropriate Bis-Tris propane buffer solution (Figures 1-3, panels E and F). Both in the absence and presence of CO2, the values of k0 obtained by the different methods match well with each other (Table 1) Table 2). The analysis of the data shown in Figures 1-3 (panels C and D), according to Equation (2), allowed us to determine the values of k on and k 0 , corresponding to the slope and the y-intercept of the linear plots, respectively (Table 1). Moreover, the values of k 0 were measured in the absence of Dr-Nb(III) via the rapid mixing of the peroxynitrite solution with the appropriate Bis-Tris propane buffer solution (Figures 1-3, panels E and F). Both in the absence and presence of CO 2 , the values of k 0 obtained by the different methods match well with each other (Table 1) and agree well with those previously reported [12,13,25,27,28,30,33,34,36,39,54] (see Table 2).     As shown in Table 1, the values of kon for the interaction between Dr-Nb(III) and peroxynitrite were unaffected by CO2/HCO3 − /CO3 2− , suggesting that CO2/HCO3 − /CO3 2− does not alter the binding properties of Dr-Nb(III). The values of kon for peroxynitrite isomerization via all-α-helical globins and all-β-barrel nitrobindins ranged between 1.2 × 10 4 M −1 s −1 for Hs-Hb(III) [25] and 6.9 × 10 4 M −1 s −1 for Mt-Nb(III) [40] ( Table 2), suggesting that the very different structural organization [40,[57][58][59][60][61] is not at the root of the different rate of peroxynitrite isomerization. However, the coordination of the heme-Fe(III) atom, the in-or out-of-plane position of the metal with respect to the pyrrole nitrogen atoms of the porphyrin, the ligand accessibility of the heme-Fe(III) atom, and its Lewis acidity may tune the kinetics of the related peroxynitrite decomposition [36].

Heme-Protein
To outline the role of the metal center, the values of kobs have been determined as a function of (i) apo-Dr-Nb concentration (at a fixed peroxynitrite concentration) and (ii) peroxynitrite concentration (at fixed Dr-Nb(III) concentration). Apo-Dr-Nb does not induce the isomerization of peroxynitrite; in fact, the values of kobs and k0 match each other in the presence of apo-Dr-Nb (Figures 1-3, panels E-H), as reported for apo-Efc-Mb, apo-Hs-Hb, and apo-Hs-Nb [25,39]. As shown in Figure 4, the values of kobs slightly decrease with an increasing peroxynitrite concentration, reflecting either the slow isomerization process of peroxynitrite-peroxynitrous acid dimers or their slow dissociation preceding the Dr-Nb(III)-catalyzed isomerization of peroxynitrite [30,36].     As shown in Table 1, the values of kon for the interaction between Dr-Nb(III) and peroxynitrite were unaffected by CO2/HCO3 − /CO3 2− , suggesting that CO2/HCO3 − /CO3 2− does not alter the binding properties of Dr-Nb(III). The values of kon for peroxynitrite isomerization via all-α-helical globins and all-β-barrel nitrobindins ranged between 1.2 × 10 4 M −1 s −1 for Hs-Hb(III) [25] and 6.9 × 10 4 M −1 s −1 for Mt-Nb(III) [40] ( Table 2), suggesting that the very different structural organization [40,[57][58][59][60][61] is not at the root of the different rate of peroxynitrite isomerization. However, the coordination of the heme-Fe(III) atom, the in-or out-of-plane position of the metal with respect to the pyrrole nitrogen atoms of the porphyrin, the ligand accessibility of the heme-Fe(III) atom, and its Lewis acidity may tune the kinetics of the related peroxynitrite decomposition [36].

Heme-Protein
To outline the role of the metal center, the values of kobs have been determined as a function of (i) apo-Dr-Nb concentration (at a fixed peroxynitrite concentration) and (ii) peroxynitrite concentration (at fixed Dr-Nb(III) concentration). Apo-Dr-Nb does not induce the isomerization of peroxynitrite; in fact, the values of kobs and k0 match each other in the presence of apo-Dr-Nb (Figures 1-3, panels E-H), as reported for apo-Efc-Mb, apo-Hs-Hb, and apo-Hs-Nb [25,39]. As shown in Figure 4, the values of kobs slightly decrease with an increasing peroxynitrite concentration, reflecting either the slow isomerization process of peroxynitrite-peroxynitrous acid dimers or their slow dissociation preceding the Dr-Nb(III)-catalyzed isomerization of peroxynitrite [30,36].

Heme-Protein
To outline the role of the metal center, the values of kobs have been determined as a function of (i) apo-Dr-Nb concentration (at a fixed peroxynitrite concentration) and (ii) peroxynitrite concentration (at fixed Dr-Nb(III) concentration). Apo-Dr-Nb does not induce the isomerization of peroxynitrite; in fact, the values of kobs and k0 match each other in the presence of apo-Dr-Nb (Figures 1-3, panels E-H), as reported for apo-Efc-Mb, apo-      [55]. b pH 7.0 and 20.0 °C. From [34]. c pH 7.0 and 20.0 °C. From [35]. d pH 7.3 and 25.0 °C. From [36]. e pH 7.0 and 20.0 °C. From [25]. f pH 7.5 and 20.0 °C. From [27]. g pH 7.2 and 25.0 °C. From [40]. h pH 7.0 and 22.0 °C. Present study. i pH 7.1 and 25.0 °C. From [39]. n.d., not determined.

Heme-Protein
As shown in Table 1, the values of kon for the interaction between Dr-Nb(III) and peroxynitrite were unaffected by CO2/HCO3 − /CO3 2− , suggesting that CO2/HCO3 − /CO3 2− does not alter the binding properties of Dr-Nb(III). The values of kon for peroxynitrite isomerization via all-α-helical globins and all-β-barrel nitrobindins ranged between 1.2 × 10 4 M −1 s −1 for Hs-Hb(III) [25] and 6.9 × 10 4 M −1 s −1 for Mt-Nb(III) [40] ( Table 2), suggesting that the very different structural organization [40,[57][58][59][60][61] is not at the root of the different rate of peroxynitrite isomerization. However, the coordination of the heme-Fe(III) atom, the in-or out-of-plane position of the metal with respect to the pyrrole nitrogen atoms of the porphyrin, the ligand accessibility of the heme-Fe(III) atom, and its Lewis acidity may tune the kinetics of the related peroxynitrite decomposition [36].
To outline the role of the metal center, the values of kobs have been determined as a function of (i) apo-Dr-Nb concentration (at a fixed peroxynitrite concentration) and (ii) peroxynitrite concentration (at fixed Dr-Nb(III) concentration). Apo-Dr-Nb does not induce the isomerization of peroxynitrite; in fact, the values of kobs and k0 match each other in the presence of apo-Dr-Nb (Figures 1-3, panels E-H), as reported for apo-Efc-Mb, apo-Hs-Hb, and apo-Hs-Nb [25,39]. As shown in Figure 4, the values of kobs slightly decrease with an increasing peroxynitrite concentration, reflecting either the slow isomerization process of peroxynitrite-peroxynitrous acid dimers or their slow dissociation preceding the Dr-Nb(III)-catalyzed isomerization of peroxynitrite [30,36].  [55]. b pH 7.0 and 20.0 • C. From [34]. c pH 7.0 and 20.0 • C. From [35]. d pH 7.3 and 25.0 • C. From [36]. e pH 7.0 and 20.0 • C. From [25]. f pH 7.5 and 20.0 • C. From [27]. g pH 7.2 and 25.0 • C. From [40]. h pH 7.0 and 22.0 • C. Present study. i pH 7.1 and 25.0 • C. From [39]. n.d., not determined.
As shown in Table 1, the values of k on for the interaction between Dr-Nb(III) and peroxynitrite were unaffected by CO 2 /HCO 3 − /CO 3 2− , suggesting that CO 2 /HCO 3 − /CO 3 2− does not alter the binding properties of Dr-Nb(III). The values of k on for peroxynitrite isomerization via all-α-helical globins and all-β-barrel nitrobindins ranged between 1.2 × 10 4 M −1 s −1 for Hs-Hb(III) [25] and 6.9 × 10 4 M −1 s −1 for Mt-Nb(III) [40] ( Table 2), suggesting that the very different structural organization [40,[57][58][59][60][61] is not at the root of the different rate of peroxynitrite isomerization. However, the coordination of the heme-Fe(III) atom, the inor out-of-plane position of the metal with respect to the pyrrole nitrogen atoms of the porphyrin, the ligand accessibility of the heme-Fe(III) atom, and its Lewis acidity may tune the kinetics of the related peroxynitrite decomposition [36].
To outline the role of the metal center, the values of k obs have been determined as a function of (i) apo-Dr-Nb concentration (at a fixed peroxynitrite concentration) and (ii) peroxynitrite concentration (at fixed Dr-Nb(III) concentration). Apo-Dr-Nb does not induce the isomerization of peroxynitrite; in fact, the values of k obs and k 0 match each other in the presence of apo-Dr-Nb (Figures 1-3, panels E-H), as reported for apo-Efc-Mb, apo-Hs-Hb, and apo-Hs-Nb [25,39]. As shown in Figure 4, the values of k obs slightly decrease with an increasing peroxynitrite concentration, reflecting either the slow isomerization process of peroxynitrite-peroxynitrous acid dimers or their slow dissociation preceding the Dr-Nb(III)-catalyzed isomerization of peroxynitrite [30,36].
To clarify the differential roles of various ferric heme proteins and CO2/HCO3 -/CO3 2levels in peroxynitrite detoxification, the overall effect of pH and CO2/HCO3 -/CO3 2-on the observed rate of peroxynitrite isomerization was investigated. Thus, the protonation equilibria of OONO -/HOONO and CO2/HCO3 -/CO3 2-were considered. The effect of CO2 on peroxynitrite isomerization in the absence of heme proteins (i.e., kobs) can be described by Equation (3):  To clarify the differential roles of various ferric heme proteins and CO 2 /HCO 3 − /CO 3 2− levels in peroxynitrite detoxification, the overall effect of pH and CO 2 /HCO 3 − /CO 3 2− on the observed rate of peroxynitrite isomerization was investigated. Thus, the protonation equilibria of OONO − /HOONO and CO 2 /HCO 3 − /CO 3 2− were considered. The effect of CO 2 on peroxynitrite isomerization in the absence of heme proteins (i.e., k obs ) can be described by Equation (3):  showing that HOONO decays to NO 3 − faster than OONO − , with a pK a ≈ 6.8 [13,36]. On the other hand, as already outlined before [9,13,22,27,28,30,36,39], when CO 2 /HCO 3 − /CO 3 2− is present (i.e., [L] = 5.0 × 10 −1 M), the rate of peroxynitrite degradation becomes much faster and increases with a rising pH (see Table 1 and [36]). Therefore, employing Equation (3) and knowing the values of K c1 and K c2 (see above), the pH dependence of peroxynitrite degradation ( Figure 5, panel B) gives information on the different values of k ci and k ci H (with i = 1, 2, 3). The inspection of Table 3 allows for the following considerations: (i) k c1 , k c2 H , and k c3 H are not playing any role, likely because either one or both reactants are too scarcely populated over the pH range investigated for the pseudo-firstorder rate constant to have a detectable value, which can then be considered as ≈0 s −1 ; (ii) At pH ≤ 6.2, the peroxynitrite degradation rate is mostly characterized by the reaction between HOONO and CO 2 , corresponding to k c1 H = 8.0 M −1 s −1 ); (iii) For 6.2 < pH < 8.2, the peroxynitrite degradation rate is mostly characterized by the reaction between OONO − and HCO 3 − , corresponding to k c2 = 6.2 × 10 1 M −1 s −1 ; (iv) For pH ≥ 8.2, the peroxynitrite degradation rate is mostly characterized by the reaction between OONO − and CO 3 2− , corresponding to k c3 = 1.8 × 10 3 M −1 s −1 .
The pH-dependence of peroxynitrite degradation in the absence of CO2/HCO3 − /CO3 2-(i.e., [L] = 0, see Equation (3)), allowed us to determine the values of k0 (= 3.0 × 10 -2 s -1 ), k0 H (= 8.3 × 10 -1 s -1 ) and KP (= 10 -6.8 ) ( Figure 5, panel A), showing that HOONO decays to NO3 − faster than OONO − , with a pKa ≈ 6.8 [13,36]. On the other hand, as already outlined before [9,13,22,27,28,30,36,39], when CO2/HCO3 -/CO3 2− is present (i.e., [L] = 5.0 × 10 −1 M), the rate of peroxynitrite degradation becomes much faster and increases with a rising pH (see Table 1 and [36]). Therefore, employing Equation (3) and knowing the values of Kc1 and Kc2 (see above), the pH dependence of peroxynitrite degradation ( Figure 5, panel B) gives information on the different values of kci and kci H (with i = 1, 2, 3). The inspection of Table  3 allows for the following considerations: (i) kc1, kc2 H , and kc3 H are not playing any role, likely because either one or both reactants are too scarcely populated over the pH range investigated for the pseudo-first-order rate constant to have a detectable value, which can then be considered as ≈ 0 s -1 ; (ii) At pH ≤ 6.2, the peroxynitrite degradation rate is mostly characterized by the reaction between HOONO and CO2, corresponding to kc1 H = 8.0 M -1 s -1 ); (iii) For 6.2 < pH < 8.2, the peroxynitrite degradation rate is mostly characterized by the reaction between OONOand HCO3 -, corresponding to kc2 = 6.2×10 1 M -1 s -1 ; (iv) For pH ≥ 8.2, the peroxynitrite degradation rate is mostly characterized by the reaction between OONOand CO3 2-, corresponding to kc3 = 1.8×10 3 M -1 s -1 .  Table 3. In (A,B), the open circles correspond to experimental data obtained from the literature [36] on the peroxynitrite isomerization rate constants as a function of pH under the conditions described above. The continuous lines were obtained according to Equation (3) by the non-linear least-squares fitting of data. In (C), the open circles correspond to values of kobs in the presence of 3.0 × 10 −5 M Dr-Nb(III) in the absence of CO2/HCO3 − /CO3 2− at different pH values (present study). Data shown in (C) indicate that, at pH 5.8, the efficiency of the heme protein (at the indicated concentration) is higher than that of the CO2/HCO3 − /CO3 2− system, while at pH 8.5, the efficiency is lower.

Parameter
Value k0 (s −1 )  Table 3. In (A,B), the open circles correspond to experimental data obtained from the literature [36] on the peroxynitrite isomerization rate constants as a function of pH under the conditions described above. The continuous lines were obtained according to Equation (3) by the non-linear least-squares fitting of data. In (C), the open circles correspond to values of k obs in the presence of 3.0 × 10 −5 M Dr-Nb(III) in the absence of CO 2 /HCO 3 − /CO 3 2− at different pH values (present study). Data shown in (C) indicate that, at pH 5.8, the efficiency of the heme protein (at the indicated concentration) is higher than that of the CO 2 /HCO 3 − /CO 3 2− system, while at pH 8.5, the efficiency is lower. Table 3. Values of the parameters fitting data with Equation (3).  (Figure 5, panel C). The pH-dependent mechanism of peroxynitrite degradation can be described by Scheme 2, which reports the different pathways for peroxynitrite degradation. Obviously, the value of kobs and its pH dependence depend on the CO2/HCO3 -/CO3 2levels (i.e., [L]), envisaging that the relative levels of the ferric heme protein and CO2/HCO3 -/CO3 2-are crucial in defining the respective role for peroxynitrite detoxification. Of note, data reported in Figures 1-3 (panel D) refer to [CO2/HCO3 -/CO3 2-] = 5.0 × 10 -1 M, while in the bloodstream, the physiological levels of CO2/HCO3 -/CO3 2-(~3.0 × 10 -2 M) are about 10-20 fold lower. Therefore, under these conditions, kobs decreases significantly even in the presence of CO2/HCO3 -/CO3 2-, as indicated by Equation (3) ( Figure 5, panel C). The pH-dependent mechanism of peroxynitrite degradation can be described by Scheme 2, which reports the different pathways for peroxynitrite degradation.  Figure 5C). For alkaline pH values, Dr-Nb(III) levels higher than 3.0 × 10 −5 M would be required to play a relevant role in peroxynitrite detoxification (see Figure 5, panel C).

Parameter Value
The relative importance of the heme proteins and CO2/HCO3 − /CO3 2− catalyzing the peroxynitrite isomerization is crucial since, according to the literature [9,12,25,30,36,62,63], ferric heme proteins prevent L-tyrosine nitration, which, instead, occurs either in the presence of apo-Dr-Nb and/or of CO2/HCO3 − /CO3 2− (Figure 6) (see also Scheme 1). In fact, the relative yield of NO3 − and NO2 − , obtained from peroxynitrite isomerization catalyzed by Dr-Nb(III), ranged between 89 and 92%, and between 7 and 12%, respectively. However, in the absence of Dr-Nb(III) and/or in the presence of apo-Dr-Nb and/or CO2/HCO3 − /CO3 2− , the values of the relative yield of NO3 − and NO2 − ranged between 69 and 74%, and between 7 and 12%, respectively ( Table 4). The great relevance of these observations is confirmed by the fact that, under stressed conditions, zebrafish triggers a defense mechanism through nitrosative stress and peroxynitrite production for which its degradation may be, on one side, enhanced by the elevated levels of CO2 [48,49], but, on the other side, inhibited by the consequent lowering of the pH level. In any event, the correlation between the peroxynitrite degradation by ferric heme proteins and CO2 levels appears to be relevant within in vivo models of zebrafish, and it may have important consequences for the O2 supply to poorly oxygenated tissues, such as the retina [50,51].

Conclusions
Ferric heme proteins and the CO2/HCO3 − /CO3 2− system are the major players in peroxynitrite detoxification, for which efficiency depends on both the concentration of these two actors and the pH level. CO3 2− is much more effective at peroxynitrite inactivation, as compared to CO2, with HCO3 − displaying an intermediate activity. Of note, the CO3 2− levels were much lower than those of the other two components in the system under physiological conditions. Although ferric heme proteins are intrinsically much more effective than the CO2/HCO3 − /CO3 2− system, their levels are usually significantly lower than CO2/HCO3 − /CO3 2− . Since over the physiological pH range the rate of peroxynitrite detoxification by CO2/HCO3 − /CO3 2− (i.e., ~3.0 × 10 −2 M) is ~1.5 s −1 , values of kobs (= kon × [heme-Fe(III)]), for peroxynitrite scavenging by ferric heme-proteins must be larger than 2 s −1 . Overall, only the level of the circulating Hs-Hb(III) (~2.0 × 10 −4 M;

Conclusions
Ferric heme proteins and the CO 2 /HCO 3 − /CO 3 2− system are the major players in peroxynitrite detoxification, for which efficiency depends on both the concentration of these two actors and the pH level. CO 3 2− is much more effective at peroxynitrite inactivation, as compared to CO 2 , with HCO 3 − displaying an intermediate activity. Of note, the CO 3 2− levels were much lower than those of the other two components in the system under physiological conditions. Although ferric heme proteins are intrinsically much more effective than the CO 2 /HCO 3 − /CO 3 2− system, their levels are usually significantly lower than CO 2 /HCO 3 − /CO 3 2− . Since over the physiological pH range the rate of peroxynitrite detoxification by CO 2 /HCO 3 − /CO 3 2− (i.e.,~3.0 × 10 −2 M) is~1.5 s −1 , values of k obs (= k on × [heme-Fe(III)]), for peroxynitrite scavenging by ferric heme-proteins must be larger than 2 s −1 . Overall, only the level of the circulating Hs-Hb(III) (~2.0 × 10 −4 M; corresponding to about 2-3% of the total Hs-Hb in red cells) appears to be sufficient to play a relevant role in the detoxification by peroxynitrite under physiological conditions. Furthermore, it must be remarked that acidification, which often occurs in the bloodstream of fish when they dive to depth [64][65][66] and can even be magnified by increased levels of CO 2 [49], decreases the effect of the CO 2 /HCO 3 − /CO 3 2− system, rendering the role of ferric heme proteins even more crucial. This is especially important in the ocular sys-tem and in the protection of the retina against oxidative stress linked to peroxynitrite [3,50]. Therefore, under specific environmental conditions (typical of diving fish), Dr-Nb(III) indeed may play a relevant physiological role in peroxynitrite scavenging from poorly oxygenated tissues, such as the retina.