Mycobacterial and Human Ferrous Nitrobindins: Spectroscopic and Reactivity Properties

Structural and functional properties of ferrous Mycobacterium tuberculosis (Mt-Nb) and human (Hs-Nb) nitrobindins (Nbs) were investigated. At pH 7.0 and 25.0 °C, the unliganded Fe(II) species is penta-coordinated and unlike most other hemoproteins no pH-dependence of its coordination was detected over the pH range between 2.2 and 7.0. Further, despite a very open distal side of the heme pocket (as also indicated by the vanishingly small geminate recombination of CO for both Nbs), which exposes the heme pocket to the bulk solvent, their reactivity toward ligands, such as CO and NO, is significantly slower than in most hemoproteins, envisaging either a proximal barrier for ligand binding and/or crowding of H2O molecules in the distal side of the heme pocket which impairs ligand binding to the heme Fe-atom. On the other hand, liganded species display already at pH 7.0 and 25 °C a severe weakening (in the case of CO) and a cleavage (in the case of NO) of the proximal Fe-His bond, suggesting that the ligand-linked movement of the Fe(II) atom onto the heme plane brings about a marked lengthening of the proximal Fe-imidazole bond, eventually leading to its rupture. This structural evidence is accompanied by a marked enhancement of both ligands dissociation rate constants. As a whole, these data clearly indicate that structural–functional relationships in Nbs strongly differ from what observed in mammalian and truncated hemoproteins, suggesting that Nbs play a functional role clearly distinct from other eukaryotic and prokaryotic hemoproteins.


Introduction
Globins are a superfamily of evolutionary conserved heme-proteins that bind, sense, and transport diatomic gases [1][2][3][4][5][6][7][8][9]. Most of these proteins (e.g., hemoglobin (Hb) and myoglobin (Mb)) are characterized by eight α-helical segments shaped around the heme In the low-frequency region of the spectrum, obtained with the 441.6 nm excitation line, an intense band was observed at 213 cm −1 that decreases upon excitation with the 413.1 nm in both Mt-Nb(II) and Hs-Nb(II) ( Figure 1B). This band was assigned to the ν(Fe-His) stretching mode since it is expected to give rise to a strong band in five-coordinate In the low-frequency region of the spectrum, obtained with the 441.6 nm excitation line, an intense band was observed at 213 cm −1 that decreases upon excitation with the 413.1 nm in both Mt-Nb(II) and Hs-Nb(II) ( Figure 1B). This band was assigned to the ν(Fe-His) stretching mode since it is expected to give rise to a strong band in five-coordinate high spin ferrous heme proteins upon excitation in the Soret band [32]. The ν(Fe-His) stretching mode frequency spans from about 200 cm −1 (neutral proximal His) to 250 cm −1 (deprotonation of N δ , as in the heme-containing peroxidases). Its frequency, very sensitive to the protein matrix, is an optimum probe of the proximal cavity structure [33]. In ferrous Ec-Mb(II) and Pc-Mb(II), where the N δ proton is H-bonded to a neutral backbone carbonyl group, the ν(Fe-His) stretch was found at 220 cm −1 [34,35]. Likewise, in the Nbs(II), where the N δ proton is H-bonded to a neutral backbone ND1 of Lys26 (Mt-Nb) or with the O atom of Thr29 (Hs-Nb), the frequency of the ν(Fe-His) band is at 213 cm -1 , indicating a weaker interaction than in Mb ( Figure 1B).

UV-Vis and RR Spectroscopic Properties of Mt-Nb(II)-CO and Hs-Nb(II)-CO
CO was found very informative to examine the distal cavity of heme-proteins [36]. In fact, the back-donation from the Fe dπ to the CO π* orbitals depends on polar interactions. A very important role is played by H-bonds between the bound CO and the distal protein residues. A strong H-bond favors back-donation, with a strengthening of the Fe-C bond and a correspondingly weakening of the CO bond [37].
Within this context, a linear correlation with a negative slope between the frequencies of the ν(Fe-C) and ν(CO) stretching modes was found for a large class of carbonylated heme-proteins and heme-model compounds containing imidazole as the fifth heme-Fe(II) ligand ( Figure 1C) [38]. The ν(Fe-C)/ν(CO) position along the correlation line reflects the type and strength of distal polar interactions [36]. Wild type Ec-Mb(II)-CO and Pc-Mb(II)-CO is characterized by moderate back-bonding induced by weak H-bonding from the distal His residue (Ec-Mb(II)-CO: 509 and 1944 cm −1 ; Pc-Mb(II)-CO: 508 and 1944 cm −1 ). When the distal His residue is replaced by non-polar residues (e.g., in the His64Val mutant of the Pc-Mb), the ν(Fe-C)/ν(CO) point slides down the line (488 and 1966 cm −1 ), reflecting the expected decrease in back-bonding [37]. Variations in the donor strength of the trans-ligand also affect the frequencies. In fact, CO complexes with a weak or absent proximal ligand are located above the histidine line [36]. Hence, the upper dashed line in Figure 1C represents either five-coordinate heme-Fe(II)-CO complexes with no trans-ligand or six-coordinate heme-Fe(II)-CO adducts with weak trans-ligands [36,39].
Upon CO binding, Mt-Nb(II) and Hs-Nb(II) give rise to a six-coordinate low spin complex between pH 6.0 and 10.2 with UV-Vis absorption bands at 419, 537, and 568 nm ( Figure 1A). The RR modes of the Mt-Nb(II)-CO complex were identified by an isotopic shift at 510 cm −1 ν(Fe-C) and 1958 cm −1 ν(CO) ( Figure 1C). This latter value is very close to that obtained for the At-Nb(Fe(II)-CO complex by FTIR [29]. The RR modes of Hs-Nb(II)-CO show a similar ν(Fe-C) mode at 509 cm −1 , but the intense fluorescence in the 1900 to 2000 cm −1 region observed in the Hs-Nb(II)-13 CO sample does not allow us to identify the ν( 13 CO) mode. The ν(CO) mode was tentatively assigned to the band observed at 1950 cm −1 ( Figure 1C). The ν(Fe-C)/ν(CO) position for both Nbs appears displaced above the solid His line, the effect being more pronounced for Mt-Nb(II)-CO than for Hs-Nb(II)-CO, moving toward frequencies typical of CO complexes with no or weakly-bound trans-ligand ( Figure 1C). Therefore, this behavior might reflect a weaker proximal Fe-His bond in the two carbonylated Nb(II) with respect to mammalian Mbs.

UV-Vis and EPR Spectroscopic Properties of Mt-Nb(II)-NO and Hs-Nb(II)-NO
Extensive studies support the view that the UV-Vis and EPR spectroscopy of ferrous nitrosylated heme-proteins and heme-model compounds are indicative of the strength of the proximal His-Fe(II) bond and in turn of the ferrous metal center reactivity [1, [40][41][42][43][44][45][46][47] Absorption spectra of the Fe(II)-NO derivative of Ec-Mb, Mt-Nb, and Hs-Nb are reported in Figure 2A. The difference between these heme-proteins is strikingly remarkable with a blue-shift of Mt-Nb(II)-NO and Hs-Nb(II)-NO, associated to a marked decrease of the extinction coefficient, both features suggesting a weakening of the heme-Fe-His proximal bond [48].  In the EPR spectra, the experimental spectrum is showed in red while the simulated spectra in black. Mirroring what already observed for absorption spectra of Figure 2A, a dramatic difference between the nitrosylated hemoproteins clearly comes out by the EPR spectra performed at 110 K ( Figure 2B-D). Thus, heme proteins in a histidine-Fe(II)-NO conformation are characterized by a temperature-dependent EPR spectrum composed of a combination of two paramagnetic species whose relative composition depends on the temperature, such that (a) at high temperatures (>150 K) the cw-EPR spectra are dominated by an axial species (denoted state A, from axial) (called SysA below), and (b) at low temperature (< 150 K) a rhombic species prevails (species R, from rhombic) (called SysR below). When the histidine-iron bond is elongated or broken, the effect of the histidine nitrogen on the EPR spectrum is lost. Consequently, the EPR spectrum is only split by the NO nitrogen and resolved into three sharp lines with a hyperfine splitting constant of 17 G (called Sys5C below) [1, [40][41][42][43][44][45][46][47].
In Figure 2B-D, the experimental EPR spectra are compared with their simulations, employing different percentages of the three forms (i.e., SysA, SysR and Sys5C), which are reported in the figure caption. On the basis of this simulation, the EPR signal of Ec-Mb(II)-NO ( Figure 2B), detected at pH 7.0, displays a full (~100%) rhombic shape with some resolution of the superhyperfine structure in the g z region of the spectrum characteristic of a hexa-coordinated form [49]. Conversely, in the case of Hs-Nb(II)-NO and Mt-Nb(II)-NO, indeed a three-line pattern in the high magnetic field region of EPR spectra was detected at pH 7.0 ( Figure 2C,D), even though important spectroscopic differences occur between the two Nb(II)-NO. Thus, in the case of Hs-Nb(II)-NO we observe at pH 7.0 the predominance of the species Sys5C (~88%), characterized by the three-line hyperfine structure ( Figure 2C), clearly indicating that most molecules display the five-coordination of the heme-Fe(II)-NO species as the result of cleavage of the proximal His-Fe bond. On the other hand, the EPR spectrum of Mt-Nb(II)-NO exhibits a mixture of different forms, with only 55% attributable to the species Sys5C, as from simulations ( Figure 2D). Therefore, like for CO-bound (see above), and even to a higher extent, in NO-bound Nb(II) the evidence for a weak proximal bond emerges in a clear cut fashion, suggesting that upon distal ligand binding a dramatic strain is exerted on the proximal Fe-His bond, eventually leading to the cleavage of the Fe-His proximal bond in a large percentage of molecules. The time course of CO binding to Mt-Nb(II) and Hs-Nb(II) by rapid-mixing technique is strictly monophasic (>95%) ( Figure 3A,B) and wavelength-independent. The amplitude of the exponentials is dependent on the CO concentration under all the experimental conditions, since CO concentration is similar to the value of the dissociation equilibrium constant K (CO) (i.e., [CO] does not fully saturate Mt-Nb(II) and Hs-Nb(II)) ( Figure 3A,B). Values of k obs(CO) are independent of the heme-protein concentration (Table S1) and increase linearly with the CO concentration over the whole CO concentration range explored (between 2.0 × 10 -5 M and 2.0 × 10 -4 M). In Figure 3A,B are reported some kinetic progress curves for CO binding to Hs-Nb(II) ( Figure 3A) and Mt-Nb(II) ( Figure 3B); the analysis of data, shown in Figure 3C according to Equation (2), allowed to determine values of k on(CO) and k off(CO) for (de)carbonylation of Mt-Nb(II)(-CO) and Hs-Nb(II)(-CO). Moreover, values of k off(CO) for decarbonylation of Mt-Nb(II)-CO and Hs-Nb(II)-CO were obtained by CO displacement with NO ( Figure 4). The time course of CO displacement from Mt-Nb(II)-CO and Hs-Nb(II)-CO by NO, investigated by rapid-mixing technique, is strictly monophasic (>93%), as indicated by the distribution of residuals ( Figure 4).
The low reactivity of CO for Mt-Nb(II) and Hs-Nb(II) (Figure 3), as compared to that for Pc-Mb(II) [15] (Table 1), may be ascribed either to (i) proximal effects, possibly related to a higher activation free energy for the in-plane motion of the Fe-His proximal bond [56] and/or to (ii) distal effects, due to either the steric hindrance exerted by the heme distal residues (His85 in Mt-Nb and Thr91 in Hs-Nb), altering the Fe(II)-C-O angle [57] and/or crowding of H2O molecules in the vicinity of the heme because of the exposure of the distal side to the bulk solvent. The unusually high CO dissociation rate constant from Mt-Nb(II)-CO ( Figure 4) indeed may reflect a weakening of the heme-Fe-His proximal bond for the carbonylated species, as suggested by the unusual ν(Fe-C)/ν(CO) position ( Figure 1C). Such a feature was also observed in other hemoproteins, such as soluble guanylate cyclase, cytochrome c, and sensor proteins (e.g., FixL) [58,59], accompanied by an increase of the CO dissociation rate constant, as observed for heme model compounds [60].
An additional piece of information may come from the pH dependence of CO binding to both Hs-Nb and Mt-Nb, which does not show any enhancement over the 2.2-7.0 pH range ( Figure 5A). This behavior is drastically different from what observed in most of the other hemoproteins, wherefore at pH < 5.0 a relevant increase of the CO binding rate constant is observed with variable pKa values [56,[61][62][63][64][65][66][67]. This feature, which was attributed to the cleavage (or severe weakening) of the heme-Fe-His proximal bond in the unliganded form, as demonstrated by the spectroscopic features, is characterized by the blue-shift of the absorption spectrum in the Soret region and the appearance of two peaks at 525 and 565 nm [56,61,62]. However, in the case of Nbs the pH independence of the CO Values of k on(CO) for CO binding to Mt-Nb(II) and Hs-Nb(II) are 2-to 4-fold slower, respectively, than that reported for At-Nb(II) carbonylation (=2.3 × 10 5 M −1 s −1 ) [29], and even slower (i.e., 5-and 10-folds, respectively) than that of Pc-Mb(II) [56] (Table 1). The unusually high values of k off(CO) for CO dissociation from Mt-Nb(II)-CO and Hs-Nb(II)-CO, as derived from linear plots of k obs(CO) versus [CO] ( Figure 3C), are closely similar with those directly measured following CO displacement by NO ( Figure 4). The values of k off(CO) are~140-fold higher than that reported for At-Nb(II)-CO decarbonylation [29] and~500-fold higher than those observed in mammalian Mbs (e.g., Pc-Mb(II)) [15]. The resulting values of the dissociation equilibrium constant for CO binding to Mt-Nb(II) and Hs-Nb(II) (i.e., k off(CO) /k on(CO) ) are very high, being 6.3 × 10 -5 M and 3.8 × 10 −5 M, respectively. These values are about 200-fold and 1000-fold higher than those of At-Nb(II) (2.2 × 10 −7 M) [29] and mammalian Mbs (e.g., Ec-Mb(II), 5.7 × 10 −8 M; and Pc-Mb(II), 3.7 × 10 −8 M) [15,52].
The low reactivity of CO for Mt-Nb(II) and Hs-Nb(II) (Figure 3), as compared to that for Pc-Mb(II) [15] (Table 1), may be ascribed either to (i) proximal effects, possibly related to a higher activation free energy for the in-plane motion of the Fe-His proximal bond [56] and/or to (ii) distal effects, due to either the steric hindrance exerted by the heme distal residues (His85 in Mt-Nb and Thr91 in Hs-Nb), altering the Fe(II)-C-O angle [57] and/or crowding of H 2 O molecules in the vicinity of the heme because of the exposure of the distal side to the bulk solvent. The unusually high CO dissociation rate constant from Mt-Nb(II)-CO ( Figure 4) indeed may reflect a weakening of the heme-Fe-His proximal bond for the carbonylated species, as suggested by the unusual ν(Fe-C)/ν(CO) position ( Figure 1C). Such a feature was also observed in other hemoproteins, such as soluble guanylate cyclase, cytochrome c, and sensor proteins (e.g., FixL) [58,59], accompanied by an increase of the CO dissociation rate constant, as observed for heme model compounds [60].
An additional piece of information may come from the pH dependence of CO binding to both Hs-Nb and Mt-Nb, which does not show any enhancement over the 2.2-7.0 pH range ( Figure 5A). This behavior is drastically different from what observed in most of the other hemoproteins, wherefore at pH < 5.0 a relevant increase of the CO binding rate constant is observed with variable pK a values [56,[61][62][63][64][65][66][67]. This feature, which was attributed to the cleavage (or severe weakening) of the heme-Fe-His proximal bond in the unliganded form, as demonstrated by the spectroscopic features, is characterized by the blue-shift of the absorption spectrum in the Soret region and the appearance of two peaks at 525 and 565 nm [56,61,62]. However, in the case of Nbs the pH independence of the CO binding rate constants ( Figure 5A) is mirrored by an absorption spectrum of the deoxygenated form which remains unchanged for 1 s (keeping the same features as at pH 7.0) even at pH 2.2 ( Figure 5B) before decaying for denaturation (data not shown). This clearly indicates that the heme-Fe-His proximal bond remains unaltered even at this low pH value, likely reflecting a highly compact proximal side of the heme pocket in the unliganded form of Hs-Nb(II) and Mt-Nb(II), which dramatically lowers the pK a of the proximal bond. binding rate constants ( Figure 5A) is mirrored by an absorption spectrum of the deoxygenated form which remains unchanged for 1 s (keeping the same features as at pH 7.0) even at pH 2.2 ( Figure 5B) before decaying for denaturation (data not shown). This clearly indicates that the heme-Fe-His proximal bond remains unaltered even at this low pH value, likely reflecting a highly compact proximal side of the heme pocket in the unliganded form of Hs-Nb(II) and Mt-Nb(II), which dramatically lowers the pKa of the proximal bond.

Rebinding Kinetics
The progress curves of CO rebinding to Mt-Nb(II) and Hs-Nb(II) are characterized by spectral changes reflecting: (i) geminate CO rebinding, (ii), bimolecular carbonylation process and (iii) conformational changes.
The negligible or low CO geminate rebinding for Mt-Nb(II and Hs-Nb(II)), respectively, (Figures 6 and 7) may either arise from the low heme reactivity or from the easy escape of the ligand from the distal pocket. It turns out that in both Nbs φgem ≤ 0.05, indeed suggesting that kBC(CO) >> kBA(CO) (Equation (3)); in this respect, this outcome is consistent with the structural evidence of a remarkable heme exposure to the solvent and the absence of relevant distal structural constraints, thus leading to a very low energy barrier for the escape to the solvent of the photolyzed CO.

Rebinding Kinetics
The progress curves of CO rebinding to Mt-Nb(II) and Hs-Nb(II) are characterized by spectral changes reflecting: (i) geminate CO rebinding, (ii), bimolecular carbonylation process and (iii) conformational changes.
The negligible or low CO geminate rebinding for Mt-Nb(II and Hs-Nb(II)), respectively, (Figures 6 and 7) may either arise from the low heme reactivity or from the easy escape of the ligand from the distal pocket. It turns out that in both Nbs ϕ gem ≤ 0.05, indeed suggesting that k BC(CO) >> k BA(CO) (Equation (3)); in this respect, this outcome is consistent with the structural evidence of a remarkable heme exposure to the solvent and the absence of relevant distal structural constraints, thus leading to a very low energy barrier for the escape to the solvent of the photolyzed CO. multiple exponential behavior of the CO rebinding, which occurs with a continuum of reactivity-changing species. Figure 6B shows that the amplitude of the slow decay systematically increases at lower CO concentrations, a fact that is expected for such transitions. Moreover, the slower bimolecular rate, observed in flash photolysis, is similar to the one measured in the stopped-flow experiments, which may be taken as a further hint towards the identification of the slow phase as a relaxed deoxy structure which is functionally distinct from the liganded state [69]. Most of the absorption change for CO rebinding kinetics to Mt-Nb(II) ( Figure 6A) is due to bimolecular rebinding, which is best described by a sum of two exponential decay functions, even though the two kinetic phases display similar amplitudes and only a two-fold difference for observed rate constants ( Figure 6B). Interestingly, values of the bimolecular rates constants at pH 7.4 and 20.0 • C (i.e., k 1 on(CO) = (1.6 ± 0.01) × 10 5 M -1 s -1 and k 2 on(CO) = (0.83 ± 0.01) × 10 5 M -1 s -1 ) are only slightly (2-to 3-fold) faster than what observed by stopped-flow (Table 1). This difference, though quite small, indeed might suggest that immediately after the CO detachment the heme is in a somewhat faster conformation, relaxing over the ms time regime (that is a time interval overlapping with CO rebinding) to a structural arrangement characterized by a slightly higher (bỹ 3 kJ/mole) energy barrier for CO binding. This occurrence seems supported by the evidence ( Figure 6A) that the progress curves of CO rebinding to Mt-Nb(II) show a small increase in the signal over the microsecond time scale, which is likely due to a protein conformational change following the photodissociation of the bound ligand [50]. This signal is independent of the CO concentration and is weakly temperature-dependent. To highlight this signal, we collected the absorbance change at 421.5 nm, which is an isosbestic point of the spectral difference between carbonmonoxy-and deoxy-species. This signal (on a × 4 scale) is compared with the one measured at 436 nm in Figure 6C. The time course shows the typical shape for a time-extended conformational change observed in many heme-proteins [68] and it can be described with a stretched exponential decay with the time constant of 770 µs and a stretching exponent of 0.24, followed by exponential relaxation with a lifetime identical to the long-lived decay detected at 436 nm, corresponding to CO rebinding. Therefore, the overlapping of the heme relaxation time with CO rebinding time is responsible for the non-exponential behavior of the conformational transition and the apparent multiple exponential behavior of the CO rebinding, which occurs with a continuum of reactivity-changing species. Figure 6B shows that the amplitude of the slow decay systematically increases at lower CO concentrations, a fact that is expected for such transitions. Moreover, the slower bimolecular rate, observed in flash photolysis, is similar to the one measured in the stopped-flow experiments, which may be taken as a further hint towards the identification of the slow phase as a relaxed deoxy structure which is functionally distinct from the liganded state [69]. From the temperature dependence of the bimolecular rebinding rate linear Eyring plots of kon(CO) can be obtained between 10.0 °C and 40.0 °C, allowing to determine for both Nbs values of the activation enthalpy and entropy ( Table 2). The amplitude of each phase was not influenced by the temperature (Figures 6B and 7B), while the CO concentration seemed to have a small systematic effect consistent with the hypothesis that the slow phase is populated after a structural relaxation. Interestingly, for the slower bimolecular CO rebinding process all activation parameters (i.e., ΔG2 ‡ , ΔS2 ‡ and ΔH2 ‡ , Table 2) are closely similar between Hs-Nb(II) and Mt-Nb(II), clearly indicating that their structural arrangement is essentially the same after the conformational change following the CO detachment. On the other hand, a striking difference can be observed between the two Nbs before this structural transition (Table 2); thus, in the faster process, observed in Hs-Nb(II), the lower free energy barrier is fully attributable to a much lower activation entropy, which is essentially 0, as compared to the very negative value observed in Mt-Nb(II) ( The time course of Mt-Nb(II) and Hs-Nb(II) nitrosylation is strictly monophasic (>91%) (Figure 8A,B) and wavelength-independent. The amplitude of the exponentials is independent of the NO concentration under all the experimental conditions since the NO Somewhat different behavior is observed for CO recombination to Hs-Nb(II), wherefore appreciable geminate recombination is detected (amounting to~5 % of the total amplitude of the rebinding process) with a r gem = 9 × 10 7 s −1 at 25 • C. According to Equation (3) it suggests that for Hs-Nb(II) k BA(CO) ≈ r gem × ϕ gem ≈ 4.5 × 10 6 s −1 and k BC(CO) ≈ r gem -k BA(CO) ≈ 8.5 × 10 7 s −1 . Furthermore, the bimolecular rebinding is more markedly biphasic in Hs-Nb(II) than in Mt-Nb(II) (Figures 6A and 7A), mostly because of a minor faster phase (corresponding to~15% of the absorption change due to the bimolecular process) with k f on(CO) = (3.9 ± 0.1) × 10 6 M -1 s -1 at 25.0 • C ( Figure 7A). On the other hand, the second-order rate constant of the slower process (i.e., k s on(CO) = (1.7 ± 0.1) × 10 5 M -1 s -1 , corresponding to~85% of the absorption change due to the bimolecular process) is closely similar to what observed by stopped-flow (Table 1 and Figures 3A and 7A). As reported for Mt-Nb(II), the CO rebinding kinetics shows a small rise in the micro-seconds timescale, possibly reflecting a conformational relaxation following photolysis. Unlike the case of Mt-Nb(II), it was not possible to identify a wavelength at which the structural relaxation could be clearly observed. However, the larger amplitude of the slow phase and its rate very close to that observed by stopped-flow indeed suggest that the relaxation is likely faster than in the case of Mt-Nb(II) and it gets closer to completion during the time scale of the flash photolysis experiment. However, the larger extent of geminate recombination phase and a faster bimolecular recombination process also indicates that in Hs-Nb(II) after CO detachment the heme is in a higher reactivity structural arrangement, as indicated by the 7 kJ/mole lower free energy barrier than in Mt-Nb(II) for CO binding to the faster process (Table 2). Table 2. Activation enthalpies (∆H ‡ ) and entropies (∆S ‡ )"parameters for bimolecular CO binding rates in Mt-Nb(II) and Hs-Nb(II).

Mt-Nb(II)
Hs-Nb(II) From the temperature dependence of the bimolecular rebinding rate linear Eyring plots of k on(CO) can be obtained between 10.0 • C and 40.0 • C, allowing to determine for both Nbs values of the activation enthalpy and entropy ( Table 2). The amplitude of each phase was not influenced by the temperature (Figures 6B and 7B), while the CO concentration seemed to have a small systematic effect consistent with the hypothesis that the slow phase is populated after a structural relaxation. Interestingly, for the slower bimolecular CO rebinding process all activation parameters (i.e., ∆G 2 ‡ , ∆S 2 ‡ and ∆H 2 ‡ , Table 2) are closely similar between Hs-Nb(II) and Mt-Nb(II), clearly indicating that their structural arrangement is essentially the same after the conformational change following the CO detachment. On the other hand, a striking difference can be observed between the two Nbs before this structural transition (Table 2); thus, in the faster process, observed in Hs-Nb(II), the lower free energy barrier is fully attributable to a much lower activation entropy, which is essentially 0, as compared to the very negative value observed in Mt-Nb(II) ( Table 2). The time course of Mt-Nb(II) and Hs-Nb(II) nitrosylation is strictly monophasic (>91%) (Figure 8A,B) and wavelength-independent. The amplitude of the exponentials is independent of the NO concentration under all the experimental conditions since the NO concentration was larger by at least three orders of magnitude than the dissociation equilibrium constant K (NO) (i.e., [NO] was largely sufficient to saturate both Mt-Nb(II) and Hs-Nb(II)) ( Figure 8A,B). The values of k obs(NO) are independent of the heme-protein concentration and increase linearly with the NO concentration over the whole gaseous ligand concentration range explored (between 1.5 × 10 -5 M and 1.0 × 10 -4 M). The analysis of data, shown in Figure 8C according to Equation (4), allowed to determine only values of k on(NO) for the nitrosylation of Mt-Nb(II) and Hs-Nb(II) (1.7 × 10 6 M −1 s −1 and 9.3 × 10 5 M −1 s −1 , respectively). In fact, the intercept of the straight lines with the y axis is close to zero. Therefore, the values of k off(NO) for Mt-Nb(II)-NO and Hs-Nb(II)-NO denitrosylation (6.8 × 10 −2 s −1 and 2.1 × 10 −2 s −1 , respectively) were obtained by NO displacement with CO. The time course of NO displacement from Mt-Nb(II)-NO and Hs-Nb(II)-NO by CO (i.e., of Mt-Nb(II)-CO and Hs-Nb(II)-CO formation) is strictly monophasic (>96%) (Figure 9).

Rebinding Kinetics
After nanosecond laser photolysis of NO, the time course of Hs-Nb(II) nitrosylation displays: (i) a geminate rebinding phase, corresponding to about 35% of the total recombination absorption change (φgem = 0.35) with an apparent lifetime of 10 ns (rgem ≈ 6.9 × 10 7 s −1 , Equation (3), and (ii) a bimolecular biphasic phase, characterized by a faster process (~8% of the total recombination absorption change with kon(NO) = 1.5 × 10 7 M −1 s −1 ) and a slower one (~57% of the total recombination absorption change with kon(NO) = 8.5 × 10 5 M −1 s −1 ) (Figure 10). The higher geminate recombination underlies a quite fast recombination rate kBA(NO) (≈ 2.3 × 10 7 s −1 ), about 6 times faster than that for CO, while the escape rate constant turns out to be fairly similar for CO and NO, reflecting the substantially similar size of the two ligands. The faster value of the second-order rate constant for NO binding is closely similar to what observed for Pc-Mb (Table 1 and [53]) and about 6-fold slower than that reported for At-Nb(II) [29]. On the other hand, the slower value of the second-order rate constant for NO binding to Hs-Nb(II) species is about 20-fold and 100-fold slower than what reported for Pc-Mb and At-Nb(II), respectively (Table 1).  Values of k on(NO) for NO binding to Mt-Nb(II) and Hs-Nb(II) (1.7 × 10 6 M −1 s −1 and 9.3 × 10 5 M −1 s −1 , respectively) are 50-to 100-fold slower, respectively, than that reported for At-Nb(II) nitrosylation (=8.1 × 10 7 M −1 s −1 ) [29], and 10-and 20-folds, respectively, of that observed for Pc-Mb(II) (2.2 × 10 7 M −1 s −1 ) [53] (Table 1). The unusually high values of k off(NO) for NO dissociation from Mt-Nb(II)-NO and Hs-Nb(II)-NO (6.8 × 10 −2 s −1 and 2.1 × 10 −2 s −1 , respectively) ( Figure 9) are similar to that for NO dissociation from At-Nb(II)-NO (~8 × 10 −2 s −1 ) [29] (Table 1). The values of k off(NO) for the denitrosylation of Mt-Nb(II)-NO, and Hs-Nb(II)-NO, as well as for At-Nb(II)-NO [29], are 200-to 800fold faster than those reported for the denitrosylation of mammalian Mbs (e.g., Pc-Mb(II); k on(NO) = 1.2 × 10 −4 s −1 ) [54] (Table 1). Lastly, the affinity of NO (i.e., K = k off(NO) /k on(NO) ) for the fast-reacting form of Hs-Nb(II) (=1.4 × 10 −9 M), calculated using the k on(NO) value determined by laser photolysis, is similar to that of At-Nb(II) (~1 × 10 −9 M), obtained with the same approach [29]. On the other hand, the NO affinity for the slow-reacting form of Hs-Nb(II) (=2.5 × 10 −8 M) calculated using the k on(NO) value determined by laser photolysis agrees with those of Mt-Nb(II) and Hs-Nb(II) (i.e., k off(NO) /k on(NO) = 4.0 × 10 −8 M and 2.3 × 10 −8 M, respectively) calculated with k on(NO) values determined by rapid mixing technique (Table 1). Of note, the affinity of NO for Nbs(II) is lower than that of mammalian Mb(II), displaying values of k off(NO) /k on(NO) for Nb(II) nitrosylation which are 200-to 10,000-fold higher than that of Pc-Mb(II) (=5.5 × 10 −12 M) [53,54] (Table 1).

Rebinding Kinetics
After nanosecond laser photolysis of NO, the time course of Hs-Nb(II) nitrosylation displays: (i) a geminate rebinding phase, corresponding to about 35% of the total recombination absorption change (ϕ gem = 0.35) with an apparent lifetime of 10 ns (r gem ≈ 6.9 × 10 7 s −1 , Equation (3), and (ii) a bimolecular biphasic phase, characterized by a faster process (~8% of the total recombination absorption change with k on(NO) = 1.5 × 10 7 M −1 s −1 ) and a slower one (~57% of the total recombination absorption change with k on(NO) = 8.5 × 10 5 M −1 s −1 ) ( Figure 10). The higher geminate recombination underlies a quite fast recombination rate k BA(NO) (≈ 2.3 × 10 7 s −1 ), about 6 times faster than that for CO, while the escape rate constant turns out to be fairly similar for CO and NO, reflecting the substantially similar size of the two ligands. The faster value of the second-order rate constant for NO binding is closely similar to what observed for Pc-Mb (Table 1 and [53]) and about 6-fold slower than that reported for At-Nb(II) [29]. On the other hand, the slower value of the second-order rate constant for NO binding to Hs-Nb(II) species is about 20-fold and 100-fold slower than what reported for Pc-Mb and At-Nb(II), respectively (Table 1). turns out to be fairly similar for CO and NO, reflecting the substantially similar size of the two ligands. The faster value of the second-order rate constant for NO binding is closely similar to what observed for Pc-Mb (Table 1 and [53]) and about 6-fold slower than that reported for At-Nb(II) [29]. On the other hand, the slower value of the second-order rate constant for NO binding to Hs-Nb(II) species is about 20-fold and 100-fold slower than what reported for Pc-Mb and At-Nb(II), respectively (Table 1).

Discussion
The highly solvent exposed heme-Fe-atom is at the root of the fast auto-oxidation rate of Mt-Nb(II), At-Nb(II) and Hs-Nb(II), which is similar to that of Rp-NPs and 10 4-10 5 times higher than that of mammalian globins [22,29,30]. This impairs oxygenation, carbonylation and nitrosylation of Nb(II)s under non-reducing conditions. Nonetheless, the kinetic and thermodynamic behavior of Nb(II) forms can be investigated under appropriate con-

Discussion
The highly solvent exposed heme-Fe-atom is at the root of the fast auto-oxidation rate of Mt-Nb(II), At-Nb(II) and Hs-Nb(II), which is similar to that of Rp-NPs and 10 4-10 5 times higher than that of mammalian globins [22,29,30]. This impairs oxygenation, carbonylation and nitrosylation of Nb(II)s under non-reducing conditions. Nonetheless, the kinetic and thermodynamic behavior of Nb(II) forms can be investigated under appropriate conditions which prevent autoxidation to significantly affect the investigation.
Remarkably, the easy access to the heme pocket of Nbs [26,29,30] does not lead to fast ligand binding rate constants (Tables 1 and 2), indicating that the easier access pathway does not affect to a meaningful extent the activation free energy of ligand binding to the heme-Fe(II) atom of Nbs. Actually, the much slower CO binding rate constants for Hs-Nb and Mt-Nb (Table 1), which display an activation free energy (∆G ‡ = 43.7 kJ/mol for Hs-Nb and 45.0 kJ/mol for Mt-Nb) much higher than that of Ec-Mb (∆G ‡ = 39.2 kJ/mol) and Pc-Mb (∆G ‡ = 39.7 kJ/mol), might stem from crowding of H 2 O molecules in the distal side of the heme pocket, as observed from X-ray structures of Fe(III) Nbs [26], which would raise the free energy barrier for ligand binding to the heme's Fe atom. However, an additional contribution might arise from a higher energy barrier for the in-plane movement of the unliganded heme-Fe-His proximal bond to bind CO [56,[61][62][63][64][65][66][67]. The strain, exerted on the proximal His-Fe(II) bond by this movement, may be due to the clustering of amino acid side chains in the proximal side of the heme pocket, which might be also responsible for the resistance of the proximal bond even at very low pH values ( Figure 5). As a matter of fact, an inspection of the available protein three-dimensional structures of Mt-Nb(III) and Hs-Nb(III) supports this hypothesis [26], suggesting that the increased clustering of residues in the proximal Nb heme pocket, relative to Mb, can be related to the different surrounding secondary structures (β-strands in Nbs versus α-helices in Mbs), which imply different residue spacing and structural arrangement around the heme group. In particular, the Mt-Nb three-dimensional structure [26] shows that five amino acid residues (i.e., Ile30, Phe33, Tyr35, Met145, and Leu156) directly contact the porphyrin ring on the proximal side through van der Waals interactions (distance ≤ 4.0 Å). Such a scheme of contacts is also observed in At-Nb [29] and Hs-Nb [30]. On the other hand, on the proximal side of Pc-Mb [70] and Ec-Mb [71] only three amino acid residues (i.e., Leu89, His97, and Ile99) are in contact with the heme group. This strain, imposed by the protein structure, would also explain the very weak proximal His-Fe(II) bond in (i) the CO-bound form, as indicated by the resonance Raman frequencies of the ν(Fe-C) and ν(CO) modes ( Figure 1C), and in (ii) the NO-bound form, as indicated by the absorption and EPR spectroscopy ( Figure 2). This effect is then mirrored by the functional behavior of liganded forms, wherefore much faster CO dissociation rate constants (Figure 4 and Table 1) as well as faster NO dissociation rate constants ( Figure 9 and Table 1) are observed, being in keeping with a severe weakening (or even a cleavage) of the proximal Fe-His bond in the liganded species. This peculiar structural arrangement of the liganded forms of both Nbs finds further support in the relatively slow relaxation (overlapping with the bimolecular recombination process) toward the equilibrium reformation of the Fe-His bond in the unliganded species, which brings about a multi-exponential rebinding both for CO (Figures 6 and 7 and Table 1) and for NO ( Figure 10 and Table 1), not observed by stopped-flow (Figures 3 and 8).
On the other hand, some difference can be observed between Hs-Nb(II) and Mt-Nb(II), wherefore the Fe-His bond looks weaker in Hs-Nb(II) than in Mt-Nb(II), as suggested by EPR spectroscopy for the NO-bound forms, since the Fe-His bond is completely missing already at pH 7.0 in Hs-Nb(II)-NO ( Figure 2C) while in Mt-Nb(II)-NO an equilibrium between a penta-coordinated species and a rhombic one is observed ( Figure 2D). Additionally, in the case of the CO-bound form, some difference can be detected between the two Nbs, which shows up in a slightly lower ν(CO) frequency for Hs-Nb(II)-CO (i.e., 1950 cm −1 ) with respect to Mt-Nb(II)-CO (i.e., 1958 cm −1 )( Figure 1C), possibly reflecting a different interaction of the ligand with residues of the distal heme pocket. It might be also respon-sible for the larger geminate recombination, observed after photolysis of Hs-Nb(II)-CO (Figure 7), envisaging the possibility of a higher barrier for the ligand escape with respect to Mt-Nb(II)-CO, where only a negligible geminate rebinding is observed ( Figure 6). No relevant difference instead can be detected between the two Nbs for the Fe-His bound in the CO-bound forms, as indicated by the closely similar fast CO dissociation rate constant (Figure 4) and the similar rate for the relaxation to the equilibrium unliganded conformation after laser photolysis (Figures 6 and 7).
In conclusion, the results here presented show that ferrous Nbs display a significantly reduced reactivity toward exogenous ligands, such as CO and NO, likely due to both (i) H 2 O crowding in the distal side of the heme pocket and (ii) a very high barrier for the concerted movement of the proximal Fe-His bond toward the heme plane upon ligand binding. Such a proximal strain brings about also a severe weakening of the Fe-His proximal bond in the liganded forms, thus leading to a markedly accelerated dissociation rate constants for both CO and NO. Indeed, all the Nb three-dimensional structures determined so far indicate a weakening of the Fe-His-proximal bond, that is 0.10-0.17 Å longer than that observed in Pc-Mb [70] and Ec-Mb [71]; such bond length differences are meaningful given the high resolution (ranging from 1.79 Å to 1.36 Å) of the threedimensional structures analyzed. The drastically different regulation of ligand-linked conformational changes in Nbs, as compared to other monomeric hemoproteins (such as mammalian Mbs), is in keeping with the likely different physiological role exerted by this new class of hemoproteins [26].
Gaseous 12 CO and 13 CO for Resonance Raman (RR) measurements were purchased from Rivoira (Milan, Italy) and FluoroChem (Hadfield, UK), respectively. Gaseous CO for laser flash photolysis and rapid-mixing stopped-flow kinetics was purchased from Linde AG (Höllriegelskreuth, Germany). The CO solution was prepared by keeping in a closed vessel the 5.0 × 10 −2 M phosphate buffer solution (pH = 7.0) under CO at p = 760.0 mm Hg anaerobically (T = 20.0 • C). The solubility of CO in the aqueous buffered solution is 1.03 × 10 −3 M at p = 760.0 mm Hg and T = 20.0 • C [51]. NO solutions for UV-Vis and EPR spectroscopy were prepared by dissolving in a phosphate buffer solution (pH = 7.0, T = 20.0 • C) sodium dithionite and sodium nitrite (Approx. 1 × 10 -2 M). Gaseous NO for rapid-mixing stopped-flow kinetics was purchased from Merck KGA (Darmstadt, Germany). NO was purified by flowing through a glass column packed with NaOH pellets and then by passage through a trapping solution, containing 20 mL of 5.0 M NaOH, to remove traces impurities; the NO pressure was 760.0 mmHg [72]. The All chemicals were of analytical or reagent grade and were used without further purification unless stated. The RR spectra were recorded using a 5 mm NMR tube by excitation with the 413.1 nm line of a Kr + laser (Coherent, Innova 300 • C; Coherent, Santa Clara, CA, USA) and with the 441.6 nm line of a He-Cd laser (Kimmon IK4121R-G; Kimmon Koha Co. LTD, Tokyo, Japan). Backscattered light from a slowly rotating NMR tube was collected and focused into a triple spectrometer (consisting of two Acton Research SpectraPro 2300i and a SpectraPro 2500i in the final stage with a grating of 3600 or 1800 grooves/mm; Princeton Instruments, Trenton, NJ, USA), which works in the subtractive mode, equipped with a liquid nitrogen-cooled CCD detector. Spectral resolution, calculated theoretically based on the optical properties of the spectrometer, was of 1.2 cm −1 and spectral dispersion of 0.4 cm −1 /pixel, and 4 cm −1 and spectral dispersion 1.2 cm −1 /pixel, for the 3600 and 1800 grating, respectively. This latter grating was used to measure the RR spectra of the CO complexes in the 1800-2300 cm −1 region with the 441.6 nm excitation. These spectra were obtained with a cylindrical lens to minimize ligand photolysis since it focuses the laser light into a line instead of a point. The RR spectra were calibrated using as standards carbon tetrachloride, indene, and n-pentane, to an accuracy of 1 cm −1 for intense isolated bands.
To improve the signal-to-noise ratio, several spectra were accumulated and summed only if no spectral differences were noted. All spectra were baseline corrected. The UV-Vis spectra were measured both prior to and after RR measurements to ensure that no degradation occurred under the experimental conditions that were used. The RR spectra were recorded using the experimental set-up as previously reported [73] . Within a few seconds after mixing, the solutions were frozen with liquid N 2 and the EPR spectrum was recorded [43]. EPR measurements were carried out on a Bruker ELEXSYS E500 spectrometer (Bruker Bruker BioSpin GmbH, Germany) operating in continuous wave at X band and equipped with a high sensitivity SHQ cavity. A temperature of 110 K was achieved by a nitrogen Bruker VT system. Spectra were recorded using a microwave power of 20 mW and a modulation amplitude of 0.5 mT. Simulation of the EPR spectra was performed with Easypsin v. 5.2.28 [74] [56]. Kinetics of CO binding to Mt-Nb(II) and Hs-Nb(II) was recorded over the 380-450 nm wavelength range.
CO binding to Mt-Nb(II) and Hs-Nb(II) was analyzed in the framework of Scheme 1. Progress kinetic curves at selected wavelengths were analyzed according to Equation (1): where ODobs is the observed optical density at a selected wavelength and at a given time interval, OD0 is the optical density at t = 0, n is the number of exponentials, ΔODi is the optical density change associated to the exponential i, ki is the pseudo-first-order rate constant of the exponential i (i.e., kobs(CO)) and t is the time. Since data collection occurs on a logarithmic scale, experimental points in the first second represent the absolute majority (about 70 out of 100 points) of total collected ones.
Values of the second-order rate constant for Mt-Nb(II)-CO and Hs-Nb(II)-CO formation (i.e., kon(CO)) and of the first-order rate constant for CO dissociation from Mt-Nb(II)-CO and Hs-Nb(II)-CO (i.e., koff(CO)) were obtained from the dependence of the pseudo-firstorder rate constant for Mt-Nb(II) and Hs-Nb(II) carbonylation (i.e., kobs(CO)) on the ligand concentration (i.e., [CO]) according to Equation (2): The values of the first-order rate constant for CO dissociation from Mt-Nb(II)-CO and Hs-Nb(II)-CO (i.e., for CO replacement by NO; koff(CO)) were also determined by mixing the Progress kinetic curves at selected wavelengths were analyzed according to Equation (1): where OD obs is the observed optical density at a selected wavelength and at a given time interval, OD 0 is the optical density at t = 0, n is the number of exponentials, ∆OD i is the optical density change associated to the exponential i, k i is the pseudo-first-order rate constant of the exponential i (i.e., k obs(CO) ) and t is the time. Since data collection occurs on a logarithmic scale, experimental points in the first second represent the absolute majority (about 70 out of 100 points) of total collected ones.
Values of the second-order rate constant for Mt-Nb(II)-CO and Hs-Nb(II)-CO formation (i.e., k on(CO) ) and of the first-order rate constant for CO dissociation from Mt-Nb(II)-CO and Hs-Nb(II)-CO (i.e., k off(CO) ) were obtained from the dependence of the pseudo-firstorder rate constant for Mt-Nb(II) and Hs-Nb(II) carbonylation (i.e., k obs(CO) ) on the ligand concentration (i.e., [CO]) according to Equation (2): The values of the first-order rate constant for CO dissociation from Mt-Nb(II)-CO and Hs-Nb(II)-CO (i.e., for CO replacement by NO; k off(CO) ) were also determined by mixing the  [75]. Kinetics of CO dissociation from Mt-Nb(II)-CO and Hs-Nb(II)-CO was recorded at 421 nm.
The conversion of Mt-Nb(II)-CO and Hs-Nb(II)-CO to Mt-Nb(II)-NO and Hs-Nb(II)-NO, respectively, was analyzed in the framework of Scheme 2 [75]: concentration (i.e., [CO]) according to Equation (2): The values of the first-order rate constant for CO dissociation from Mt-Nb(II)-CO and Hs-Nb(II)-CO (i.e., for CO replacement by NO; koff(CO)) were also determined by mixing the The values of koff(CO) were determined from data analysis, according to Equation (1). The over 100-fold excess of NO over CO guarantees that the reaction proceeds rightward because kon(NO) × [NO] >> kon(CO) × [CO] [75].
The pH-dependence of CO binding kinetics to ferrous Nbs was carried out by mixing in the stopped-flow apparatus the reduced hemoprotein solution (in 1.0 × 10 −3 M phos-  [75]. This type of reaction proceeds according to Scheme 6. The deoxygenation of Mt-Nb(II)-O2 and Hs-Nb(II)-O2 process was followed at 431 nm. Kinetics of Mt-Nb(II)-O2 and Hs-Nb(II)-O2 deoxygenation were analyzed according to Equation (1).
Rapid-mixing experiments were carried out employing an SX18.MV stopped-flow apparatus (Applied Photophysics, Salisbury, UK) equipped with a diode array for spectra acquisition over a 1 ms time range; the light path of the observation chamber was 10 mm.  [75]. This type of reaction proceeds according to Scheme 6. The deoxygenation of Mt-Nb(II)-O2 and Hs-Nb(II)-O2 process was followed at 431 nm. Kinetics of Mt-Nb(II)-O2 and Hs-Nb(II)-O2 deoxygenation were analyzed according to Equation (1).
Rapid-mixing experiments were carried out employing an SX18.MV stopped-flow apparatus (Applied Photophysics, Salisbury, UK) equipped with a diode array for spectra acquisition over a 1 ms time range; the light path of the observation chamber was 10 mm.