A functional nitric oxide reductase model

Abstract

A functional heme/nonheme nitric oxide reductase (NOR) model is presented. The fully reduced diiron compound reacts with two equivalents of NO leading to the formation of one equivalent of N₂O and the bis-ferric product. NO binds to both heme Fe and nonheme Fe complexes forming individual ferrous nitrosyl species. The mixed-valence species with an oxidized heme and a reduced nonheme Fe_B does not show NO reduction activity. These results are consistent with a so-called “trans” mechanism for the reduction of NO by bacterial NOR.

Nitric oxide reductase (NOR) is a membrane-bound enzyme that catalyzes the 2e⁻ reduction of nitric oxide (NO) to nitrous oxide (N₂O), an obligatory step involved in the sequential reduction of nitrate to dinitrogen known as bacterial denitrification. The active site of NOR consists of a monohistidine ligated five-coordinate heme and a trisimidazole ligated nonheme Fe_B. This structure strongly resembles the active site of oxygen reduction enzyme-cytochrome c oxidase (CcO), which possesses a heme-a₃/Cu_B center (Fig. 1) (1, 2). Essentially, the distal metal Cu_B in CcO is replaced by a nonheme Fe metal in NOR; NOR and CcO are thought to be distant relatives.

Fig. 1.

Schematic representation of the bimetallic active sites of NOR and CcO.

The dinuclear iron active site in NOR was confirmed a decade ago by spectroscopic studies (3). Presumably, two NO molecules are turned over to give one molecule of N₂O and one molecule of H₂O at the diiron center with the consumption of two electrons and two protons. Although many enzyme studies of NOR have been focused on the intermediate trapping and elucidation of the reaction mechanism (4–15), the details of the catalytic cycle are still unresolved because of the lack of structural information and uncertainty regarding short-lived intermediates.

In contrast to enzyme studies, synthetic biomimetic model complexes provide a straightforward and controlled method to understand how this chemical transformation proceeds at the enzyme active site. However, only a few synthetic models have been developed that mimic the active site of NOR; moreover, these compounds either lack a proximal imidazole ligand (16, 17) or use pyridine as a replacement for the histidine ligands (18–20). No functional NOR models have been reported to date. Our CcO model complexes have proved to be functionally active for oxygen reduction reaction with minimal reactive oxygen species (ROS) formation (21–24). These appear to be promising NOR model candidates if the distal Cu metal is replaced by an iron because the resulting diiron compound has almost all of the key components in NOR: a heme Fe with a proximal imidazole ligand and a trisimidazole ligated nonheme Fe center.

In this report, we disclose the first synthetic functional NOR model LFe^II/Fe^II (Fig. 2), which reacts with two equivalents of NO to give one equivalent of N₂O and the bis-ferric product. We have shown that NO binds to both heme Fe and Fe_B to form a possible bis-nitrosyl intermediate; subsequently, the two bound NO molecules are reduced to N₂O by electrons from both Fe centers, leaving both heme Fe and Fe_B in an uncoupled ferric state. N₂O has been quantitatively identified by using an enzyme, nitrous oxide reductase (N₂OR) that reduces N₂O to N₂. The bis-ferric product was characterized by both FTIR and EPR: FTIR reveals a heme ferric nitrosyl band at 1,924 cm⁻¹ that shifts to 1,887 cm⁻¹ when ¹⁵NO is used, whereas the EPR spectrum manifests a low-spin ferric signal that is assigned to the nonheme Fe_B because a ferric heme nitrosyl would be EPR silent. The NO adducts of both heme Fe and Fe_B were obtained separately from the reaction of NO with LFe^II and LZn^II/Fe^II. These NO adducts were characterized with a series of spectroscopic methods including UV-vis, EPR, resonance Raman, FTIR, and mass spectroscopy. The reaction of NO with the mixed-valence compound LFe^III/Fe^II was also investigated. Our data show the formation of a mixture of LFe^II-NO/Fe^III and LFe^III-NO/Fe^II-NO species, but no N₂O was detected. These results are consistent with a so-called “trans” mechanism for the reduction of NO to N₂O by NOR.

Fig. 2.

Representation of the model complexes of ligand L: LFe^II (no M_B); LFe^II/Fe^II (M = M_B = Fe²⁺); LZn^II/Fe^II (M = Zn²⁺, M_B = Fe²⁺); LFe^III/Fe^II (M = Fe³⁺, M_B = Fe²⁺); LFe^III/Fe^III (M = Fe³⁺, M_B = Fe³⁺) (metals have triflate counterions).

Results

Syntheses and Characterization of Dinuclear Complexes.

The reduced diiron complex LFe^II/Fe^II is readily synthesized by reaction of LFe^II complex (25) with 1 equivalent of Fe(OTf)₂(MeCN)₂ in THF at room temperature under a N₂ atmosphere. The UV-vis spectrum shows a fast and clean formation of a dinuclear product with shifts for both the Soret (426–424 nm) and Q bands (535–530 nm) and a slightly diminished intensity in both bands (Fig. 3, solid to dotted line). The dinuclear compound LZn^II/Fe^II was synthesized by using a similar method from reaction of LZn^II with Fe(OTf)₂(MeCN)₂ [UV-vis spectra are in supporting information (SI) Fig. S1]. The mixed-valence compound LFe^III/Fe^II was obtained by oxidation of LFe^II/Fe^II with one equivalent of ferrocenium tetrafluoroborate. The bis-ferric LFe^III/Fe^III was generated by oxidation of LFe^II/Fe^II with either two equivalents of ferrocenium tetrafluoroborate or with dioxygen. The structures of these dinuclear compounds were confirmed by electrospray mass spectroscopy.

Fig. 3.

UV-vis spectra of LFe^II (solid line), LFe^II/Fe^II (dotted line), and LFe^II/Fe^II+NO (dashed line).

Reaction of NO with LFe^II, LZn^II/Fe^II and LFe^II/Fe^II.

LFe^II + NO.

Addition of purified NO to a THF solution of LFe^II leads to formation of its mono-nitrosyl derivative LFe-NO. This reaction is associated with a decreased intensity of the Soret band at 427 nm and a Q band shift from 535 to 550 nm (Fig. 4). The EPR spectrum shows a S = 1/2 signal (Fig. 8, solid line) typical of a six-coordinate ferrous NO complex akin to those reported for the active site in CcO (26–30). The FTIR spectrum shows a new band at 1,630 cm⁻¹ for the LFe-NO complex that shifts to 1,600 cm⁻¹ with ¹⁵NO (Fig. 5, dotted and dashed lines). The resonance Raman spectrum obtained by excitation with a 425-nm laser shows an Fe-N stretch at 581 cm⁻¹ that shifts to 545 cm⁻¹ with ¹⁵NO substitution. An Fe-N_Imidazole stretch is also observed at 238 cm⁻¹, providing further evidence for the presence of a six-coordinate iron nitrosyl (Fig. 6).

Fig. 4.

UV-vis spectra of LFe^II (solid line) and LFe^II+NO (dotted line).

Fig. 8.

EPR data of the LFe^II-NO (solid line), LZn^II/Fe^II-NO (dotted line), and LFe^II/Fe^II+NO (dotted-to-dashed line) complexes.

Fig. 5.

IR spectra of LFe^II (solid line) and its ¹⁴NO (dotted line) and ¹⁵NO (dashed line) derivatives.

Fig. 6.

Resonance Raman data on LFe-¹⁴NO (solid line) and LFe-¹⁵NO (dotted line).

LZn^II/Fe^II+NO.

Addition of NO to LZn^II/Fe^II in THF results in slight blue shifts (0.5 nm for the Soret band and 1 nm for the Q band) in the UV-vis spectra (Fig. S1). The FTIR of the ¹⁴NO adduct of LZn^II/Fe^II exhibits a new vibration frequency at 1,810 cm⁻¹ (Fig. 7, dotted line) that shifts to 1,774 cm⁻¹ with ¹⁵NO substitution (Fig. 7, dashed line). These features are not observed in the LZn^II/Fe^II (Fig. 7, solid line). The Fe-NO vibrations could not be identified with resonance Raman because this region was obscured by porphyrin bands. The EPR spectrum reveals a well characterized S = 3/2 nonheme Fe-NO signal at g = 4.0 (Fig. 8, dashed line). Note that the small perturbation of the Zn-porphyrin Soret implies that the NO probably binds in the pocket but does not bridge the two metals. Binding of NO to Fe_B leads to LZn^II/Fe^II-NO species. Both the LFe^II-NO and LZn^II/Fe^II-NO species are stable in the solid state for a prolonged period when exposed in air. These mononitrosyl complexes were also characterized by high-resolution electrospray mass spectroscopy (Figs. S2–S5).

Fig. 7.

IR spectra of LZn^II/Fe^II (solid line) and its ¹⁴NO (dotted line) and ¹⁵NO (dashed line) derivatives.

LFe^II/Fe^II+NO.

Addition of NO to the THF solution of LFe^II/Fe^II at room temperature leads to rapid changes in the UV-vis spectrum. The Soret band shifts from 424 to 423 nm, and the Q band shifts from 530 to 550 nm (Fig. 3, dotted to dashed line). FTIR spectra obtained from solid samples of the end product exhibit a new band at 1,924 cm⁻¹ that is absent in LFe^II/Fe^II (Fig. 9, dotted and solid lines). This band is shifted to 1,887 cm⁻¹ when ¹⁵NO is used for the sample preparation (Fig. 9, dashed line). The 37-cm^{−1 15}N-isotope shift is consistent with literature data for heme ferric nitrosyls (18, 31, 32). The resonance Raman data on this complex show a Fe-NO stretch at 610 cm⁻¹ (Fig. 10, solid line) that shifts to 598 cm⁻¹ upon ¹⁵NO substitution (Fig. 10, dotted line). The Fe-NO bending vibration of this species is observed at 589 cm⁻¹ that shifts to 580 cm⁻¹ upon ¹⁵NO substitution. These values are characteristic of a ferric heme NO species. The EPR data of the end product show an S = 1/2 signal (g = 2.07, 2.02, 1.96) that is not perturbed by ¹⁵NO substitution (Fig. 8, dotted line). These data and spin integration indicate that a single low-spin Fe^III is present in the product. The above spectroscopic data are consistent with the formation of LFe^III-NO/Fe^III-OH (Scheme 1); the vibrational features i.e., Fe-N = 610 cm⁻¹ in the Raman and N-O = 1,924 cm⁻¹ in the FTIR are derived from a ferric heme nitrosyl that is diamagnetic (i.e., EPR silent), and the S = 1/2 EPR signal is derived from the nonheme ferric center in the distal pocket. The assignment of this end product is further buttressed by comparable FTIR, resonance Raman, and EPR (g = 2.07, 2.00, 1.95) obtained by addition of NO to LFe^III/Fe^III, followed by addition of an equivalent of sodium methoxide (comparable to OH) (Fig. S6).^†

Fig. 9.

FTIR spectra of LFe^II/Fe^II (solid line), LFe^II/Fe^II +¹⁴NO (dotted line), and LFe^II/Fe^II +¹⁵NO (dashed line)

Fig. 10.

Resonance Raman of the reaction product of LFe^II/Fe^II+NO with ¹⁴NO (solid line) and ¹⁵NO (dotted line).

Scheme 1.

Reactions of mono- and bis-iron complexes with NO. (A) LFe^II + NO. (B) LZn^II/Fe^II + NO. (C) LFe^III/Fe^II + NO. (D) LFe^II/Fe^II + NO.

We have used a nitrous oxide reductase enzyme (N₂OR) that reduces N₂O to N₂ to identify the formation of N₂O in this reaction. Samples with 1 mM LFe^II/Fe^II on addition of 3 mM NO show specific activities of 43 ± 5, whereas the background activity with the same amount of NO gas is ≈10 ± 2. This activity reflects an N₂O concentration of 1 mM in solution, which, in this case, implies a quantitative yield. This, in addition to the vibrational and EPR data presented above, indicates that the synthetic model complex LFe^II/Fe^II reduces two molecules of NO to N₂O (Scheme 1).

Reaction of NO with the Mixed-Valence Compound LFe^III/Fe^II.

Addition of NO to the mixed-valence compound LFe^III/Fe^II shifts the Soret from 418 to 424 nm and the Q band from 530 to 546 nm (Fig. 11). The FTIR of the solid samples from the reaction mixture reveals a band at 1,924 cm⁻¹, which indicates the formation of a ferric heme nitrosyl species, and another weak band at 1,812 cm⁻¹, which is indicative of a ferrous nonheme nitrosyl species (Fig. S7). EPR data from the NO adduct with LFe^III/Fe^II indicates a mixture of a heme Fe-NO (signal at g = 2.0), a nonheme Fe-NO species (signal at g = 4.0), and a high-spin Fe^III species (at g = 6) (Table 1). This demonstrates that the product of LFe^III/Fe^II+NO is a mixture of LFe^III-NO/Fe^II-NO and LFe^II-NO/Fe^III. The ferric heme nitrosyl is EPR silent, but an N-O stretch is observed in the IR. This implies that NO binding to the LFe^III/Fe^II complex leads to some electron transfer from the ferrous nonheme to the ferric heme center, resulting in an equilibrium of three iron nitrosyl species: ferric heme nitrosyl, ferrous heme nitrosyl, and ferrous nonheme nitrosyl. In any case, these mixed-valence species do not form N₂O.

Fig. 11.

The optical spectra of LFe^III/Fe^II (solid line) and LFe^III/Fe^II + NO (dotted line)

Table 1.

Summary of spectroscopic features of LFe^II-NO, LZn^II/Fe^II-NO, LFe^II/Fe^II+NO, and LFe^III/Fe^III-OMe + NO

Spectroscopy method	LFeII-NO	LZnII/FeII-NO	LFeII/FeII+NO	LFeIII/FeIII+NO	LFeIII/FeII+NO
IR 14NO/ 15NO (cm−1)	1,630/1,600	1,810/1,774	1,924/1,887	1,924/1,885	1,924, 1,812 (14NO)
EPR	S = 1/2, g = 2.08, 2.02, 1.97 14NAy=22 cm−115NAy=31 cm−1	S = 3/2 g = 4.0	S = 1/2 g = 2.07, 2.02, 1.96	S = 1/2, g = 6 after reaction of NaOMe (1 eq): g = 2.07, 2.00, 1.95	g = 2.0 g = 4.0 g = 6

Discussion

The Synthetic Heme/Nonheme Diiron Compound Is a Functional NOR Model.

In this study, a synthetic model compound is reported that integrates the essential features proposed for a bacterial NOR active site; namely, a heme active site with a covalently attached imidazole ligand and a nonheme site coordinated via three imidazole ligands. This model possesses the functional NO reductase activity i.e., two molecules of NO are reduced to one molecule of N₂O at the fully reduced diiron center, leaving a di-ferric compound. The yield of N₂O is nearly quantitative within the error of the enzyme assay. A putative diferric complex formed initially after reduction of NO to N₂O reacts further with NO, forming a Fe³⁺-NO/Fe³⁺-OH species as indicated by EPR, FTIR, and resonance Raman spectroscopy methods. The ν(N-O) of our ferric heme nitrosyl is ≈20 cm⁻¹ higher than the ν(N-O) of the ferric heme nitrosyl of NOR (1,904 cm⁻¹) (33) and cytochrome cbb₃ oxidase (1,903 cm⁻¹) (34) but similar to the neutral Met Mb-NO (1,921 cm⁻¹) (31, 35). The EPR spectrum of the reaction product reveals a low-spin ferric signal, which is assigned to a ferric nonheme because a ferric heme nitrosyl would be EPR silent (S = 0). A bis-ferric compound LFe^III-NO/Fe^III-OMe prepared by the reaction of LFe^III/Fe^III with NO and one equivalent of NaOMe exhibits a very similar low-spin ferric signal in the EPR. This strongly suggests that an OH ligand is bound to the ferric nonheme center (Table 1).

Mechanism of NO Reduction.

The molecular mechanism of the NO reduction by NOR is still under debate. Two mechanisms have been proposed for binding and reduction of NO at the active site: A “trans” mechanism invokes binding of two NO molecules to the heme-iron and nonheme iron separately at the fully reduced active site, whereas a “cis” mechanism suggests binding of both NO molecules to only one iron metal (typically Fe_B). Recent quick-freezing EPR (15) indicates the formation of both heme-iron nitrosyl and nonheme-iron nitrosyl species, supporting the “trans” mechanism. Meanwhile, the mixed-valent state (heme-Fe^III/Fe_B^II) because of the low midpoint potential of heme b₃ (heme b₃, E_m = 60 mV; Fe_B, E_m = 320 mV), was suggested to be the active form of the enzyme (10). Therefore, it had been proposed that the binding and reduction of NO occurs exclusively on Fe_B, leaving the heme Fe uninvolved during the catalytic process (10, 11). In addition, it has also been suggested that two molecules of NO bind consecutively at the heme Fe³⁺ site to form a hyponitrite intermediate, which is thought to decay, giving the ferric heme, N₂O, and H₂O. A six-coordinate heme Fe³⁺-NO species of cytochrome cbb₃ oxidase and a hyponitrite species in the Fe-Cu dinuclear center were detected by FTIR and the resonance Raman spectroscopy method, respectively (34, 36).

We have shown that NO can readily bind to heme Fe and nonheme Fe separately to form stable heme Fe-NO and nonheme Fe-NO species. This indicates that the enzymatic reaction probably proceeds by a “trans” mechanism. Two equivalents of NO react with the diiron center and form individual heme and nonheme nitrosyl species. Then, the two adjacent bound nitrosyls undergo reductive coupling to form N₂O (possibly through a Fe-N(O)-N(O)-Fe intermediate), leaving both the heme Fe and the nonheme Fe in an oxidized ferric state. It is not clear if a μ-oxo diiron compound is formed as an intermediate during this reaction, because excess NO may bind to the heme Fe and cause the rupture of Fe-O-Fe bond, leading to the observed final LFe^III-NO/Fe^III-OH compound. The reaction product of LFe^II/Fe^II with two equivalents of NO did not show a ferric heme nitrosyl signal in the FTIR, and a small ion peak at m/z 1,491 amu corresponding to a [LFe-O-Fe(MeCN)₂]⁺ was detected by electrospray mass spectroscopy.

It has been postulated that NO binding to the heme iron of NOR or CcO causes bond cleavage between heme Fe and the proximal imidazole producing a five-coordinate heme nitrosyl complex (5, 8). However, we show that binding NO to LFe^II results in a stable six-coordinate heme nitrosyl in which the proximal imidazole is still coordinated to the heme iron. Such a six-coordinate ferrous nitrosyl does not appear to be a “dead-end” species as previously claimed (10). Instead, the bound nitrosyl can undergo N-N coupling with the adjacent nonheme nitrosyl to produce N₂O. Praneeth et al. (37) demonstrated that a six-coordinate iron(II) porphyrin NO adduct with a proximal imidazole ligand has a distinctive Fe^IINO^· character relative to a five-coordinate Fe^II-NO compound (37). Moreover, the enhanced radical character of the heme nitrosyl should be advantageous for the central (radical) N-N coupling step in the “trans” mechanism. On the other hand, a less reactive five-coordinate Fe^II-NO would be a dead-end species, as demonstrated by a recent synthetic diiron model based on a five-coordinate heme nitrosyl, which does not show any NO reductase activity (18).

The Mixed-Valence Form of NOR Is Not Active for the Reduction of NO to N₂O.

It has been suggested that NO activation occurs with a mixed-valence form of the NO reductase with an oxidized heme b₃ and a reduced nonheme Fe_B (10). Based on this, it was proposed that the binding of two molecules of NO happens exclusively on either heme Fe or Fe_B, leaving the other metal essentially a witness. However, our studies demonstrated that reaction of a mixed-valence compound LFe^III/Fe^II with NO leads to a mixture of two species: LFe^III-NO/Fe^II-NO and LFe^II-NO/Fe^III. No N₂O was detected from either of these species, suggesting that a so-called “cis” mechanism is unlikely in the NO reduction by NOR.

Conclusions

We have described a functional heme/nonheme nitric oxide reductase model LFe^II/Fe^II that can reduce NO to N₂O stoichiometrically, leading to a bis-ferric product. NO binds to the heme Fe of LFe^II, producing a stable six-coordinate heme Fe-NO complex, whereas binding of NO to a nonheme Fe of LZn^II/Fe^II leads to a nonheme Fe-NO species. These results suggest that the reaction of LFe^II/Fe^II with NO follows a “trans” mechanism: Two molecules of NO bind to heme Fe and nonheme Fe separately, forming a heme Fe-NO and a nonheme Fe-NO species; then the two adjacent nitrosyls undergo reductive coupling, producing N₂O and the di-ferric product. The mixed-valence form of our model compound LFe^III/Fe^II does not show any NO reduction activity; instead, stable nitrosylated species were formed. Experiments focusing on the reaction intermediates to further clarify the reaction mechanism await completion.

Materials and Methods

All reagents were obtained from commercial suppliers and used without further purification unless otherwise indicated. Fe(OTf)₂(MeCN)₂ was prepared according to literature procedures (38). Heme compound LFe^II was synthesized as reported (25). All air- and moisture-sensitive reactions were carried out under a nitrogen atmosphere in oven-dried glassware. Acetonitrile, tetrahydrofuran, and dichloromethane were purified and dried by passing reagent-grade solvent through a series of two activated alumina columns under nitrogen atmosphere. These solvents were further deoxygenated by bubbling with nitrogen for 30 min in a nitrogen glove box. DMF was distilled over molecular sieves and properly deoxygenated. Nitric oxide (NO) was obtained from Matheson Gas Products or generated by adding saturated NaNO₂ solution into a sulfuric acid solution (98% sulfuric acid /water = 3:1). It is purified by passage through a series of two thoroughly degassed 3.0 M KOH solutions and water. The saturated NO solution in acetonitrile or THF was made by bubbling purified NO gas through deoxygenated solvent in a gas-tight vial for 15 min. All reactions with NO (including making EPR and UV-vis samples) were carried out by injecting saturated NO solution into the sample solution in a glovebag purged and filled with nitrogen. The ¹⁵NO (99%) was purified by passing through a column packed with dry KOH powder under N₂.

Infrared spectra were obtained on a Mattson Galaxy 4030 FT-IR spectrometer. Solid samples were prepared by dissolving a sample in solution in a glovebox, spotting on a KBr or NaCl palate, allowing the solvent to evaporate, and then covering it by another palate and sealing the sides with parafilm. The palates containing the sample were sealed in a container and brought to the IR spectrometer for measurement. Room-temperature UV-vis spectra were recorded with a HP8452 diode array spectrophotometer. Mass spectra were obtained from the Stanford Mass Spectrometry Laboratory. The air-sensitive sample solutions were prepared in a glovebox and sealed in gas-tight vials. They were brought to the spectrometer and injected into the instrument immediately before the measurement.

EPR spectra were obtained by using a Bruker EMX spectrometer, ER 041 XG microwave bridge, and ER 4102ST cavity. All X band samples were run at 77 K in a liquid nitrogen finger dewar. A Cu standard (1.0 mM CuSO₄·5H₂O with 2 mM HCl and 2 M NaClO₄) was used for spin quantitation of the EPR spectra.

Resonance Raman (rR) spectra were obtained by using a Princeton Instruments ST-135 back-illuminated CCD detector on a Spex 1877 CP triple monochromator with 1,200, 1,800, and 2,400 grooves per millimeter holographic spectrograph gratings. Excitation was provided by a Dye Laser (Stilbene 599; Coherent) that was energized by a Coherent Innova Sabre 25/7 Ar⁺ CW ion laser. The laser line 425 nm (≈10 mW) was used for excitation. The spectral resolution was <2 cm⁻¹. Sample concentrations were ≈1 mM in Fe. The samples were either cooled to 77 K in a quartz liquid nitrogen finger dewar (Wilmad) and hand spun to minimize sample decomposition during scan collection or cooled to 190–196 K by using a flow of liquid-N₂-cooled He gas in a spinner setup.

In evaluating for the possibility of generation of N₂O in the reaction mixture, the reaction of LFe^II/Fe^II with purified NO was performed in dichloromethane, and a buffer solution was used to extract this organic layer and then used for the activity assay. Simultaneously, PnN₂OR enzyme was incubated in an excess of an anaerobic solution of methyl viologen and dithionite in Tris buffer (pH ≈7.3), in the glovebox (required to activate the enzyme) (39). Activity of the enzyme was determined spectrophotometrically, after the oxidation of dithionite reduced methyl viologen at 600 nm by using a standard protocol under anaerobic conditions (40, 41). The activity initiated by adding 20 μl of the 100-μl buffer solution used to extract the reaction mixture in CH₂Cl₂ was 43 μmol of N₂O reduced min⁻¹mg⁻¹ of N₂OR. As a control, activity measured by initiating the reaction with NO-saturated buffer solution was 10 units. A N₂O concentration-vs.-activity curve shows that 43 units of activity correspond to an N₂O concentration of 1 mM.