Lewis Acid Activation of the Ferrous Heme–NO Fragment Towards the N–N Coupling Reaction with NO to Generate N₂O

Abstract

Bacterial NO reductase (bacNOR) enzymes utilize a heme-nonheme active site to couple two NO molecules to N₂O. We show that BF₃ coordination to the nitrosyl O-atom in (OEP)Fe(NO) activates it towards N–N bond formation with NO to generate N₂O. ¹⁵N-isotopic labeling reveals a reversible nitrosyl exchange reaction and follow-up N–O bond cleavage in the N₂O formation step. Other Lewis acids (B(C₆F₅)₃ and K⁺) also promote the NO coupling reaction with (OEP)Fe(NO). These results, complemented by DFT calculations, provide experimental support for the cis:b₃ mechanism in bacNOR.

Graphical Abstract

The conversion of nitric oxide (NO) to the greenhouse gas nitrous oxide (N₂O) is an important component of the global N cycle.¹ Elucidating the N–N bond forming reaction between two NO molecules to generate N₂O remains one of the high impact research goals in NO bioinorganic chemistry. Bacterial NO reductases (bacNOR) utilize a di-Fe heme/non-heme active site (heme b₃:non-heme Fe_B) to couple two NO molecules to generate N₂O and the μ-oxo (heme)Fe–O–Fe_B that is reused in the reaction cycle.^2–3 X-ray crystal structures of bacNOR enzymes are available,^2–4 and provide a framework for the elucidation of the NO coupling reaction.

Three mechanisms for the NO coupling reaction catalyzed by bacNOR have been proposed (Figure 1), and all involve the formation of a new N–N bond and a hyponitrite intermediate. In the cis:b₃ mechanism, the second (external) NO molecule attacks the heme-nitrosyl N atom to form the hyponitrite intermediate. In the trans mechanism, both active site Fe atoms first bind one NO each prior to N–N bond formation. In the cis:Fe_B mechanism, both NO molecules bind at the non-heme Fe_B site prior to the coupling reaction.

Figure 1.

The three common proposed mechanisms for NO coupling by bacterial NO reductases.

Detailed computational investigations on the bacNOR active site by Blomberg are in support of the cis:b₃ pathway, where the heme-NO moiety is stabilized by an O atom interaction with the Fe_B center.^5–6 Definitive experimental evidence for this cis:b₃ mechanism is, however, lacking. Spectroscopic studies on heme models have suggested the trans mechanism as a viable mechanism for NO coupling by bacNOR.⁷ The operative mechanism for NO reduction by the di-Fe bacNOR enzyme is, however, still under active debate ^{3, 8–9}

Importantly, outstanding work by Lu^{8, 10} and Karlin¹¹ on engineered myoglobins and on synthetic dinitrosyl (por)Fe(NO)₂/Cu^I/acid systems, respectively, have clearly established the requirement for the second non-heme metal to enable N₂O formation. Given the current high interest in elucidating the mechanism(s) of NO reduction to generate N₂O, it is imperative that molecular-level experimental insight into factors that control the initial N–N bond formation be obtained.

It is interesting to note that two of the three proposed mechanisms for NO reduction by bacNOR (the cis:b₃ and trans mechanisms) involve the ferrous (por)Fe(NO) fragment, whose participation is central to both these mechanisms. Curiously, however, the (por)Fe(NO) compounds in isolation are unreactive towards external NO for N₂O generation.^11–14 We thus sought to probe experimental factors that could activate the isolated (por)Fe(NO) fragment towards N–N bond formation with external NO to result in the generation of N₂O. In this article, we demonstrate experimentally, complemented by DFT calculations, that Lewis acid addition to a synthetic (por)Fe(NO) compound enables activation of the FeNO moiety for engagement in a three-way synergistic reaction with external NO for the critical N–N bond formation step to generate N₂O.

Addition of 2.5 equiv. of the Lewis acid BF₃•OEt₂ to an anaerobic CH₂Cl₂ solution of (OEP)Fe(NO) (υ_NO 1664 cm⁻¹) at 0 °C results in a ~32–35% decrease in υ_NO intensity and the appearance of a new IR band at 1425 cm⁻¹ (sh) that shifts to 1406 cm⁻¹ when (OEP)Fe(¹⁵NO) is used (Figure 2). It is intriguing to note that the large υ_NO shift of −241 cm⁻¹ that occurs when the Lewis acid is added to the {FeNO}⁷ precursor (OEP)Fe(NO) to form the (OEP)Fe(NO) •BF₃ adduct is similar to those observed for the 1-electron reduced {FeNO}⁸ anions (Table 1) whose υ_NOs are also of weak-to-medium intensities.^15–18 This large υ_NO shift in (OEP)Fe(NO) •BF₃ is suggestive of a significant increase in backdonation of electron density from the Fe center to the nitrosyl ligand as a result of Lewis acid coordination to the terminal nitrosyl O atom.^19–21 A smaller yet still significant negative Δυ_NO shift of ~50 cm⁻¹ has been documented for the engineered mononitrosyl di-Fe myoglobin mutant, Fe^II-Fe_BMb(NO).²²

Figure 2.

IR spectra showing the reduction in intensity of the υ_NO band of the starting (OEP)Fe(NO) (υ_NO 1664 cm⁻¹) in CH₂Cl₂ upon reaction with BF₃•Et₂O. Inset: IR spectrum of the product mixture (as a NaCl plate) when (OEP)Fe(NO) (solid line) is reacted with BF₃•OEt₂. The corresponding IR spectrum for the (OEP)Fe(¹⁵NO) reaction is shown as a dashed line. The new bands are obscured in the reaction solvent (CH₂Cl₂) mixture.

Table 1.

NO stretching frequencies (cm⁻¹) for (OEP)Fe(NO) •BF₃ and anionic {FeNO}⁸ iron porphyrins.

system	υNO	ΔυNOa	ref
(OEP)Fe(NO) •BF3	1425	−241	t.w.
[(OEP)Fe(NO)]−	1445	−228	15
[(TPP)Fe(NO)]−	1496	−185	16
[(TFPPBr8)Fe(NO)]−	~1450	−165	17
[(3,5-Me-BAFP)Fe(NO)]−	~1466	−218	18
[(To-F2PP)Fe(NO)]−	1473	−214	18

^aΔυ_NO with respect to the neutral {FeNO}⁷ precursors.

We then proceeded to evaluate the effect of the Lewis acid addition on the reaction of (OEP)Fe(NO) with external NO. When the ¹⁵N-labeled (OEP)Fe(¹⁵NO) is reacted with ¹⁴NO in the absence of Lewis acid, a nitrosyl exchange reaction occurs (eq 1) as determined by IR spectroscopy, consistent with an earlier report.²³ Further, and consistent with previous results for related (por)Fe(NO) compounds,^{11, 13–14} we find that (OEP)Fe(NO) does not react with NO to generate N₂O.¹²

$$(\text{OEP})\text{Fe}({}^{15}\mathrm{N}\mathrm{O})+\text{NO}\rightleftharpoons (\text{OEP})\text{Fe}(\text{NO})+{}^{15}\mathrm{N}\mathrm{O}$$ (1)

When (OEP)Fe(NO) is reacted with NO in the presence of BF₃•OEt₂, however, we observed the ready formation of the N–N coupled product N₂O (υ_N2O 2237/2212 cm⁻¹; Figure 3A top) in 55–63% yields as determined by IR spectroscopy.²⁴ When ¹⁵N-labeled (OEP)Fe(¹⁵NO) is reacted with unlabeled ¹⁴NO in the presence of BF₃•Et₂O, the mixed labeled ¹⁵N¹⁴NO and ¹⁴N¹⁵NO products form in addition to ¹⁴N₂O and ¹⁵N₂O (Figure 3A bottom),²⁵ due to the exchange process in eq. 1. In this reaction, [(OEP)Fe(NO)]BF₃OH (υ_NO 1839 cm⁻¹) is isolated as the byproduct in 59–64% unoptimized isolated yield.²⁶

Figure 3.

(A) Top: IR spectrum of the headspace from the reaction of (OEP)Fe(NO) with BF₃•OEt₂ and NO, showing the formation of N₂O (2237/2212 cm⁻¹). Bottom: Headspace IR spectrum when (OEP)Fe(¹⁵NO) is reacted with BF₃•Et₂O and unlabeled ¹⁴NO, to give the mixed labeled ¹⁵N¹⁴NO (~2212/2194 cm⁻¹) and ¹⁴N¹⁵NO (~2194/2170 cm⁻¹) products as well as ¹⁴N₂O and ¹⁵N₂O (~2167/2145 cm⁻¹). These products also arise when (OEP)Fe(¹⁴NO) is reacted with ¹⁵NO. (B) Corresponding spectra when [K(2.2.2)]OSiMe₃ was used instead of BF₃•Et₂O. NO₂ (υ_NO2 ~1618 cm⁻¹) was not formed in these reactions, but appeared over time when leakage of the gas IR cell occurred if excess NO was present. The isotopic N₂O products were identified by their characteristic IR spectra.²⁵

The requirement for BF₃ in the reaction suggests that the Lewis acid plays an important role in activating (OEP)Fe(NO) towards reactivity with NO, most likely via the pathway shown in Figure 4.

Figure 4.

Proposed pathway for the N–N coupling reaction enabled by the Lewis acid BF₃.

We employed density functional theory calculations using an unsubstituted porphine macrocycle (P; see SI for details) to probe the proposed involvement of 1-NO•BF₃ and 1-(NO)₂•BF₃ in the N–N coupling reaction depicted in Figure 4. The optimized structures are shown in Figure 5 (right), and are compared with those for the non-BF₃ containing analogs (left).

Figure 5.

Calculated geometries (in Å and deg) and selected NPA charges (in e units) of the optimized precursor (P)Fe(NO) (1-NO) and products from its reactions with NO and/or BF₃. Bond distances are in italics-underline, and atomic charges in bold font.

The interaction of BF₃ with the terminal O atom in 1-NO•BF₃ results in a lengthening of both the Fe–NO (from 1.718 Å to 1.911 Å) and FeN–O (1.157 Å to 1.225 Å) bonds. Interestingly, the NO moiety in 1-NO•BF₃ becomes partially anionic (−0.239e), compared with a partially cationic charge (+0.131e) in the precursor 1-NO. This helps explain the significant lowering of the experimental υ_NO by −241 cm⁻¹ (Table 1) upon BF₃ coordination to the FeNO moiety in (OEP)Fe(NO) •BF₃ with substantial Fe^III and nitroxyl character (Tables S1 and S2). We note that a related Δυ_NO of −228 cm⁻¹ was observed for the 1-electron reduced {FeNO}⁸ anion [(OEP)Fe(NO)]⁻ (as its K⁺(2.2.2) salt), whose N–O bond length increased by ~0.02–0.04 Å from that of the neutral precursor.¹⁵ A similar nitroxyl character was proposed for the protein model Fe^II-Fe_BMb(NO).²² Importantly, the spin density on the nitrosyl N atom of 1-NO•BF₃ is −1.108e, which is substantially increased from that of its precursor 1-NO (+0.401e) (Table S2), rendering it susceptible to attack by external NO.

As mentioned above, (OEP)Fe(NO) in the absence of the Lewis acid BF₃ does not engage in the N–N coupling reaction with external NO to generate N₂O. Likely structures of 1-(NO)₂ (bottom left of Figure 5) were probed. The reaction ΔG of 11.90 kcal/mol (Table S4) for NO coupling essentially shows no catalytic role of the (P)Fe unit alone for NO coupling.^27–28

Inclusion of BF₃ in the reaction of (OEP)Fe(NO) with external NO leads to the successful generation of N₂O which necessitates the consideration of the N–N bond-forming step. The optimized structure of 1-(NO)₂•BF₃ is shown in Figure 5 (bottom right). In the presence of BF₃, the N–N coupling reaction has a large decrease in ΔG from ~12 kcal/mol (without BF₃ present) to −0.54 kcal/mol (with BF₃ present). The reductions in other energy terms are also pronounced and result in −13.53, −10.60, and −11.62 for reaction ΔE, zero-point-energy-corrected ΔE, and ΔH, respectively. The data clearly show a strong enthalpic driving force to help overcome the big entropy drop due to NO coupling. Both the Fe–N and (Fe)N–O(B) distances become longer in 1-(NO)₂•BF₃ (by 0.284 and 0.122 Å, respectively) than those in the precursor 1-NO•BF₃, thus weakening these interactions upon binding of the second NO. The effect of BF₃ addition on the nature of the NO-coupled intermediate is even more striking when one considers that without BF₃ (i.e., in 1-(NO)₂), both the proximal NO (−0.338e charge; directly bound to Fe) and distal NO (−0.257e charge) groups are negatively charged, which hinders NO coupling due to Coulombic repulsion and explains the lack of NO coupling determined experimentally. In contrast, however, when BF₃ is attached to the proximal O atom in 1-(NO)₂•BF₃, it induces a large negative charge of −0.539e to this proximal NO moiety, whereas the distal NO moiety changes to be positively charged (+0.146e), which results in a strong favorable Coulombic attraction to facilitate the NO coupling reaction as observed experimentally. The resulting short N–N bond distance of 1.240 Å in 1-(NO)₂•BF₃ is indicative of a strong N–N bond typical of that determined in stable hyponitrite [O–N=N–O]-salts.^29–30 Additional results in the SI show that BF₃ alone does not enable NO coupling (nor does NO by itself),³¹ so it is the synergistic effect of both iron porphyrin and Lewis acid that is critical for NO coupling, which may help understand the biological NOR mechanism in systems where both heme and nearby Lewis acid are available.

Our results provide experimental support for the cis:b₃ mechanism for the NO-coupling reaction, where it is the original heme nitrosyl FeN–O(Lewis acid) bond in the 1-(NO)₂•BF₃ intermediate that, is cleaved to give N₂O. The observed N₂O may derive from two isomers of the cis-hyponitrite intermediate, with the second (O-coordinated) form (Figure 6; middle) having a slightly lower energy (by ~1 kcal/mol; Table S1 footnote) but a clearly longer proximal N–O bond (by 0.081 Å). This latter intermediate can explain both the facile regioselective N–O cleavage and formation of the (OEP)Fe(μ-OH)B(C₆F₅)₃ product (isolated in ~55% yield when B(C₆F₅)₃ was used as the Lewis acid) whose X-ray structure was obtained. Such a cis-hyponitrite isomerization has recently been proposed for bacNOR based on DFT calculations.⁶ Investigations are on-going to determine the detailed energetics of all possible pathways for these processes, and will be the subject of a follow-up paper.

Figure 6.

Proposed pathway for the formation of N₂O.

In summary, our experimental results, supported by DFT calculations, demonstrate for the first time that Lewis acid activation of a synthetic monoheme-NO model, via coordination of the Lewis acid to the terminal nitrosyl O-atom, renders the FeNO moiety susceptible to attack by external NO to generate a new N–N bond en route to N₂O formation. Importantly, we show that it is the synergistic and not the separate effects of (por)Fe and Lewis acid BF₃ that is critical for the NO coupling reaction. Our preliminary extensions to other Lewis acids such as [K(2.2.2)]OSiMe₃ (Figure 3B) and B(C₆F₅)₃ (Figure S2) show that they also enable the stoichiometric {heme-¹⁵NO + ¹⁴NO} coupling reaction. Indeed, our results suggest that a redox-active (second) metal is not necessarily required for the first but critical stoichiometric N–N bond formation step for NO coupling with the (por)Fe(NO) fragment. However, in the bacNOR catalysis, a second redox center in the active site would be required to complete the reaction to generate both N₂O and water, and be catalytically competent while maintaining strict control of this biological heme-NO activation. These results lend the first chemical support for the proposed cis-b₃ mechanism as a viable, but not necessarily exclusive, pathway for NO coupling by the bacterial NO reductase enzymes and related systems where both heme and nearby metal ions/Lewis acids are in close proximity.