Not limited to Iron: A Cobalt Heme–NO Model Facilitates N–N Coupling with External NO in the Presence of a Lewis Acid to Generate N₂O

Abstract

Some bacterial heme proteins catalyze the coupling of two NO molecules to generate N₂O. We previously reported that a heme Fe–NO model engages in this N–N bond-forming reaction with NO. We now demonstrate that (OEP)Co^II(NO) similarly reacts with 1 equiv. of NO in the presence of the Lewis acids BX₃ (X = F, C₆F₅) to generate N₂O. DFT calculations support retention of the Co^II oxidation state for the experimentally observed adduct (OEP)Co^II(NO•BF₃), the presumed hyponitrite intermediate (P•⁺)Co^II(ONNO•BX₃), and the porphyrin π-radical cation by-product of this reaction, and that the π-radical cation formation likely occurs at the hyponitrite stage. In contrast, the Fe analogue undergoes a ferrous-to-ferric oxidation state conversion during this reaction. Our work shows, for the first time, that cobalt hemes are chemically competent to engage in the NO-to-N₂O conversion reaction.

Graphical Abstract

The biological conversion of NO to N₂O is important to the global N-cycle. We show that a cobalt heme model couples NO in the presence of a Lewis acid. DFT calculations provide insight into the role of the Lewis acid in activating the bound NO towards this N–N bond forming reaction.

Introduction

Cobalt-containing proteins are prevalent in biology,^[1] and in these proteins, the Co center may be part of a porphyrin-like prosthetic group. While much attention has been focused on vitamin B₁₂ containing a corrin ring, other naturally occurring Co proteins have been reported whose prosthetic groups resemble porphyrins rather than corrins. For example, a Co porphyrin-like enzyme, a fatty aldehyde decarboxylase, has been isolated from the colonal alga Botyrococcus braunii.^[2] Other Co porphyrins have been isolated from the gram-negative anaerobic sulfate reducing bacterium Desulfovibrio gigas^[3] and from D. desulfuricans.^[4] However, Battersby and Sheng have shown, using UV-vis spectroscopy, that the latter cobalt porphyrins are better described as Co^III-isobacteriochlorins.^[5]

Importantly, substitution of Co for Fe in heme proteins and in heme model compounds has provided useful information regarding the function of natural heme proteins, especially in their reactions with the signaling agent and vasodilator nitric oxide (NO). For example, substitution of Co for Fe in the reconstituted bona fide NO receptor soluble guanylyl cylase (sGC) results in a higher activation of the enzyme by NO, consistent with the ready formation of the five-coordinate (por)Co(NO)-containing adduct with proximal His dissociation from the metal center.^[6–7] NO adducts of cobalt-reconstituted heme proteins such as Mb and Hb have also been reported,^[8–9] as has the NO adduct of vitamin B₁₂.^[10–11]

There is intense interest in the N–N bond-forming reaction where two NO molecules couple, in the presence of a biological heme cofactor, to generate the greenhouse gas N₂O, as this reaction is a critical component of, and of fundamental importance to, the global N-cycle. Bacterial NO reductase (bacNOR) enzymes utilize a bimetallic heme/non-heme active site to perform this 2NO→N₂O reaction.^[12–13] We recently reported that the five-coordinate heme model (OEP)Fe(NO) engages in an N–N coupling reaction with external NO in the presence of a boron-containing Lewis acid (BX₃; X = F or C₆F₅) to generate N₂O.^[14] In this reaction, the BX₃ served to polarize the ferrous nitrosyl moiety to possess substantial ferric nitroxyl character (eq. 1), thus rendering it susceptible to attack by another NO molecule. In effect, this observation lent chemical

$$\underset{\begin{array}{l}\text{ferrous}\\ \text{nitrosyl}\end{array}}{{\text{Fe}}^{\text{II}}-\text{NO}}\text{\hspace{0.17em}+\hspace{0.17em}LA}\to \underset{\begin{array}{l}\text{ferric\hspace{0.17em}nitroxyl}\\ \text{anion\hspace{0.17em}character}\end{array}}{\left\{{\text{Fe}}^{\text{III}}-{\text{NO}}^{-}\cdot \cdot \cdot \cdot \text{LA}\right\}}$$ (1)

support for the viability of the proposed bacNOR cis:b₃ mechanism for the N–N bond formation step of N₂O generation,^[15–17] where the non-heme Fe in the di-Fe heme/non-heme active site could, in principle, serve as an appropriately positioned Lewis acid to “activate” the heme-NO fragment towards attack by a second NO molecule. We then sought to probe whether the related (por)Co(NO) fragment is chemically competent to engage in this N–N coupling reaction with external NO to generate N₂O, or whether this reaction was unique to heme Fe. In this communication, we show that N₂O is indeed generated from the reaction of (OEP)Co(NO) with external NO in the presence of the Lewis acid BX₃.

Results and Discussion

Reaction of (OEP)Co(NO) in CH₂Cl₂ at 0 °C with NO (in the absence of an added Lewis acid) for 1 h, with subsequent warming of the reaction mixture to room temperature for an additional 2 h period, does not result in N₂O formation as judged by IR spectroscopy of the reaction headspace (not shown). The analogous reaction with labeled ¹⁵NO (in excess), however, reveals that an NO exchange reaction occurs (eq 2 and Fig. S1),

$$\left(\text{OEP}\right)\text{Co}\left(\text{NO}\right)+\text{xs.\hspace{0.17em}}{}^{15}\text{N}\text{O}\to \left(\text{OEP}\right)\text{Co}\left({}^{15}\text{N}\text{O}\right)+\text{NO}$$ (2)

similar to that previously observed with the Fe analog,^[14] to give (OEP)Co(¹⁵NO) (υ_NO 1645 cm⁻¹ (KBr), Δυ_NO −30 cm⁻¹; Fig. S1).^[18]

We then examined the effect of the Lewis acid on the Co–NO moiety in the absence of added NO. Addition of ~2 equiv. of BF₃•OEt₂ to (OEP)Co(NO) in CH₂Cl₂ (υ_NO 1672 cm⁻¹) at −45 °C for 30 min resulted in the appearance of a new low-intensity band in the IR spectrum at 1632 cm⁻¹ (i.e., Δυ_NO −40 cm⁻¹; calcd. Δυ_NO −28 cm⁻¹, see later) in addition to the starting 1672 cm⁻¹ band; this new band shifted to 1607 cm⁻¹ when (OEP)Co(¹⁵NO) (υ15_NO 1641 cm⁻¹) was used in the reaction (Fig. 1, right). We assign this new band at 1632 cm⁻¹ to the υ_NO of the (OEP)Co(NO•BF₃) adduct.

Figure 1.

IR spectra from the reaction of (OEP)Co(NO) with BF₃•OEt₂ in CH₂Cl₂ at −45 °C. (Left) IR spectra of the starting (OEP)Co(NO) (υ_NO 1672 cm⁻¹) and (OEP)Co(¹⁵NO) (υ_NO 1641 cm⁻¹). (Right) IR spectra after addition of BF₃•OEt₂ to (OEP)Co(NO) (solid line trace) and (OEP)Co(¹⁵NO) (dashed line trace); the assigned υ_NO bands for the (OEP)Co(^14/15NO•BF₃) products are underlined. The low-intensity band at ~1700 cm⁻¹ is near-coincident with a similar weak band in our BF₃•Et₂O/CH₂Cl₂ control solution (Fig. S2b) and in the final product after addition of external NO (Fig. S3).

Density functional theory (DFT) calculations were employed to probe possible spin state interaction patterns for the starting neutral (P)Co(NO) compound (1-NO; P = unsubstituted porphinato dianion) and other species mentioned in this work (see Tables 1 and S2).^[19] In a typical (P)Co(NO) system, valence electrons can be distributed between the metal center (e.g., Co²⁺ or Co³⁺), the porphyrin (e.g., P^2– or P^– (π cation)) or the NO ligand (NO, NO⁺, or NO^–) (Fig. S14). The allowable combinations of these options, including additional options for spin-state variations, were studied for such systems in Table 1 (see SI for details). For 1-NO (S = 0), both open-shell and closed shell singlet states with, respectively, Co²⁺ and Co³⁺ were studied and successfully optimized. 1-NO is best described as a (P^2–)Co^II(NO) species (N-II-1 in Table 1) with antiferromagnetic coupling between Co²⁺(S = ½) and •NO (S = −½), consistent with an earlier formulation.^[20] The alternate N-III-1 state with Co^III (S = 0) and NO^– (S = 0) is calculated to be 6.02 kcal/mol higher in energy.

Table 1.

Selected Fragment Spin States Studied, and the Spin Densities and Relative Electronic Energies of the Optimized Geometries^[a]

system[b]	Spin	State[c]	Co	P	axial (NO)n	ραβCo[d]	ραβP[d]	ραβNO[d]	ραβN’O’[d]	ραβBF3[d]	ΔESCF[e]
1-NO	0	N-II-1	Co2+ (S = ½)	P2− (S = 0)	NO (S = −½)	0.895	−0.064	−0.831			0.00
	0	N-III-1	Co3+ (S = 0)	P2− (S = 0)	NO− (S = 0)	0.000	0.000	0.000			6.02
1-NO•BF3	0	N-II-1	Co2+ (S = ½)	P2− (S = 0)	NO (S = −½)	0.876	−0.067	−0.802		−0.067	0.00
	0	N-III-1	Co3+ (S = 0)	P2− (S = 0)	NO− (S = 0)	0.000	0.000	0.000		0.000	4.38
2-(NO)2	½	NN-II-4	Co2+ (S = ½)	P2− (S = 0)	(NO)2 (S = 0)	0.899	–0.111	0.056	0.156		0.00
	½	NN-III-1	Co3+ (S = 0)	P2− (S = 0)	(NO)2− (S = ½)	−0.179	0.075	0.722	0.381		0.05
	½	NN-III-2	Co3+ (S = 1)	P2− (S = 0)	(NO)2− (S = −½)	1.878	–0.071	–0.430	−0.378		11.25
	3/2	NN-II-9	Co2+ (S = 3/2)	P− (S = ½)	(NO)2− (S = −½)	2.768	1.147	−0.561	−0.354		12.80
	3/2	NN-III-3	Co3+ (S = 1)	P2– (S = 0)	(NO)2− (S = ½)	1.808	−0.038	0.824	0.406		9.25
2-(NO)2•BF3	½	NN-II-6	Co2+ (S = ½)	P− (S = −½)	(NO)2− (S = ½)	0.984	−1.005	0.041	0.980	0.001	0.00
	½	NN-II-7	Co2+ (S = −½)	P– (S = ½)	(NO)2− (S = ½)	−0.956	1.000	0.000	0.948	0.008	0.36
	½	NN-III-1	Co3+ (S = 0)	P2− (S = 0)	(NO)2− (S = ½)	–0.010	0.035	0.036	0.932	0.006	1.84
	½	NN-III-2	Co3+ (S = 1)	P2− (S = 0)	(NO)2− (S = −½)	1.964	−0.087	0.039	−0.911	−0.006	1.51
	3/2	NN-III-3	Co3+ (S = 1)	P2− (S = 0)	(NO)2− (S = ½)	1.964	−0.051	0.095	0.990	0.002	0.69
	3/2	NN-II-9	Co2+ (S = 3/2)	P− (S = ½)	(NO)2− (S = −½)	2.780	1.163	−0.018	−0.923	−0.003	1.25
[1-NO]+	½	C-II-1	Co2+ (S = −½)	P− (S = ½)	NO (S = ½)	−0.900	1.056	0.844			0.00
	½	C-III-1	Co3+ (S = 0)	P− (S = ½)	NO− (S = 0)	0.014	0.989	−0.003			7.09

^[a]See SI for full listing of spin states studied

^[b]2-(NO)₂ = (P)Co(ONN’O’), and 2-(NO)₂•BF₃ = (P)Co(ONN’O’).BF₃, where the distal (i.e., added/second) NO group is labeled N’O’.

^[c]A single “N” represents a mono-NO system; “NN” represents NO-coupling products, and “C” represents the cation. The “II” and “III” refer to the Co oxidation states.

^[d]in e units.

^[e]kcal/mol.

As BF₃ is diamagnetic, the overall spin state of the 1-NO•BF₃ adduct is still singlet. However, we investigated additional fragment spin states with both Co²⁺ and Co³⁺ as well (Tables 1 and S4) to examine if there could be any differences in electronic state due to interaction with BF₃. The calculated lowest energy formulation of the 1-NO•BF₃ adduct is the singlet (P^2–)Co^II(NO•BF₃) retaining the N-II-1 spin state feature of 1-NO with antiferromagnetic coupling between Co²⁺ and NO; the alternate Co³⁺-based N-III-1 formulation is 4.38 kcal/mol higher in energy. The associated Δυ_NO of −28 cm⁻¹ for the lowest energy N-II-1 state of 1-NO•BF₃ correlates well with the experimental value of −40 cm⁻¹ for the (OEP)Co(NO•BF₃) adduct. The calculated long O…B distance of 2.726 Å in 1-NO•BF₃ and near-planarity of the BF₃ moiety (top right of Fig. 2) point to a weak CoNO…BF₃ interaction in this adduct. Importantly, no N₂O formation was detected in the headspace of the “(OEP)Co(NO)/BF₃” reaction in the absence of added NO (Fig. 3, top).

Figure 2.

Calculated geometries (in Å and deg) and selected NPA charges (in e units) of the optimized precursor (P)Co(NO) (1-NO) and products from its reactions with NO and/or BF₃. Bond distances are in italics-underline, and atomic charges are in regular font. The atomic charges and spin densities (in e units) for the Co and P units are shown below the structures.

Figure 3.

Gas phase infrared spectra of the headspace from the reaction of: (Top) (OEP)Co(NO) with BF₃•OEt₂ showing no N₂O formation. (Middle) (OEP)Co(NO) with NO in the presence of BF₃•OEt₂ showing the formation of N₂O (2237/2212 cm⁻¹). (Bottom) BF₃•OEt₂ with NO showing no N₂O formation. No NO₂ gas (υ_NO2 ~ 1618 cm⁻¹) formed in these reactions in the absence of air.

Interestingly, however, when (OEP)Co(NO) was reacted with NO (~1 equiv) at 0 °C in the presence of BF₃•OEt₂ (~2 equiv.) for 1 hr, and then at room temperature for an additional 2 hr, the gas-phase IR spectrum of the headspace of the reaction mixture revealed the formation of N₂O in ~12% yield, together with unreacted NO (middle of Fig. 3).^[21–22] Sampling the headspace of the reaction after 1 hr at 0 °C (or −45 °C) revealed N₂O formation in lower (3–5%) yields. The related reaction of ¹⁵N-labeled (OEP)Co(¹⁵NO) with NO in the presence of BF₃•OEt₂ generated an isotopic mixture of ¹⁴N₂O, ¹⁴N¹⁵NO, ¹⁵N¹⁴NO and ¹⁵N₂O (Fig. S4),^[23] consistent with the exchange reaction shown in eq 2. The use of excess NO (and excess BF₃•OEt₂) did not result in an increased yield of N₂O. As we reported previously,^[14] BF₃•OEt₂ alone does not couple two molecules of NO to give N₂O under these reaction conditions (bottom of Fig. 3), nor does (OEP)Co(NO) by itself as noted earlier.

The observed N₂O generation necessitates consideration of the crucial N–N bond formation step, which we probed by DFT methods. We first considered the BF₃-free NO-coupled hypothetical product (P)Co(ONNO) (i.e., 2-(NO)₂, bottom left of Fig. 2). As (P)Co(NO) has S = 0 and NO has S = ½, the product with total spin state of ½ was studied first, with various fragment spin state options and with both Co²⁺ and Co³⁺ centers (Tables 1 and S5). The recent theoretical study of Fe-based N–N coupling showed a spin state change from low-spin to intermediate-spin, hence we also investigated the overall intermediate-spin state for 2-(NO)₂. The lowest energy state (see more detailed discussion of the studied spin states in the Supporting Information) was determined to be NN-II-4 with the radical feature concentrated on the Co center, with Co²⁺ (S = ½), P^2– (S = 0), and ONNO (S = 0), in essence retaining the N-II-1 spin feature of 1-NO. Although this NN-II-4 state is only slightly more favorable (by 0.05 kcal/mol) than NN-III-1 at the electronic level, its Gibbs free energy is more stable by 2.17 kcal/mol. Importantly, the calculated long N…N distance of 1.832 Å in the NN-II-4 state of 2-(NO)₂ (Fig. 2 and Table S8) is suggestive of a weak interaction and is consistent with our observation that no N₂O was formed under our experimental conditions.

Addition of NO (S = ½) to the bound NO in 1-NO•BF₃ (S = 0) likely yields the initial hyponitrite-like adduct 2-(NO)₂•BF₃, and the coupling products with total spin S = ½ were investigated first. Of the several spin-state combinations considered for this adduct (Tables 1 and S5), the Co²⁺-based states were determined to be more favorable than the Co³⁺-based states and the states with total spin of S = 3/2 are of higher energies than the most favorable S = ½ state (see more detailed discussion of the studied spin states in the Supporting Information), which are similar to the case of BF₃-free 2-(NO)₂ but with relatively smaller energy differences. These results suggest an important role of BF₃ on NO coupling as described below. As a result, the lowest energy state of 2-(NO)₂.BF₃ changes to the NN-II-6 state possessing Co²⁺ (S = ½), P^– (S = −½), and (NO)₂^– (S = −½) characters, based on the calculated spin densities (Table 1). The calculated structure and selected data for 2-(NO)₂•BF₃ in the NN-II-6 state are shown at the bottom right of Fig. 2. The data are reflective of the formation of both a porphine π-radical cation and a hyponitrite (NO)₂^– radical bonded to a formally Co²⁺ center. In this 2-(NO)₂•BF₃ structure, the O…BF₃ distance is significantly reduced to 1.503 Å from 2.726 Å in 1-NO•BF₃, and the ON–NO interaction becomes stronger with a bond distance of 1.237 Å (from 1.832 Å in the BF₃-free 2-(NO)₂). Further, in 2-(NO)₂, the proximal NO (i.e., directly bonded to Co) charge of −0.013e and the distal NO charge of +0.045e indicates a weak attraction for NO coupling. However, in the case of 2-(NO)₂•BF₃, the Lewis acid induces much larger opposite charges for the proximal NO (–0.547e) and distal NO (+0.124e) moieties, significantly enhancing their attraction to facilitate N–N bond formation. These calculated features in 2-(NO)₂•BF₃ vs. 2-(NO)₂ clearly demonstrate the role of the Lewis acid BF₃ in facilitating the experimental NO coupling reaction with (OEP)Co(NO).

The analogous reaction of (OEP)Co(NO) with 1 equiv. of NO in the presence of the related Lewis acid B(C₆F₅)₃ (2.2 equiv.) also resulted in the generation of N₂O in ~10% yield (Fig. S5). The use of excess NO and B(C₆F₅)₃ (10 equiv.) in this reaction resulted in a higher yield of N₂O (~22%). The use of excess NO in this reaction allowed us, fortunately, to isolate the cobalt-containing product in ~30% yield which we identified to be the porphyrin π-radical cation [(OEP•⁺)Co^II(NO)]-containing derivative by its IR spectrum with a new band at 1715 cm⁻¹ assigned to υ_NO (Δυ_NO +45 cm⁻¹; calcd. Δυ_NO +58 cm⁻¹) and a new band at 1581 cm⁻¹ assigned to the π-radical cation macrocyle (Fig. S6).^[24–27]

The identity of the Co-containing product was further characterized by X-ray crystallography as [(OEP•⁺)Co(NO)][HO(B(C₆F₅)₃)₂] whose crystal structure is shown in Fig. 4 (right) and Fig. S8A–B. The ∠Co–N–O angle of 121.3(2)° for the major disordered nitrosyl O-atom (71% occupancy) is similar to that of the neutral (OEP)Co(NO) precursor (at 122.70(8)°),^[28] consistent with the positive charge being remote from the Co^IINO moiety and the low Δυ_NO of +45 cm⁻¹ observed in the π-radical cation. Its identity as a π-radical cation is further substantiated by the following observations: (i) the porphyrin macrocycle is significantly distorted from planarity (Fig. S8D) when compared with its neutral (OEP)Co(NO)^[28] precursor, (ii) the N–Cα and N–Cm bond lengths within the 16-membered porphyrin core display alternating short-long distances (Fig. S8E) characteristic of some (but not all) porphyrin π-radical cations,^[29] and (iii) adjacent porphyrin macrocycles form an almost completely overlapping π-π dimer (right of Fig. 4) with a mean plane separation (M.P.S.) and lateral shift of 3.16 Å and 0.21 Å, respectively. These latter values are characteristic of “classical” porphyrin π-radical cations.^[30] Further, a twist angle (∠(por)N–Co…Co–N(por) torsion angle of 34°, Fig. S8C) of the π-π interacting dimers is observed. Selected comparisons between the structural data of the (OEP•⁺)Co(NO) π-radical cation product and those of the neutral (OEP)Co(NO) are summarized in Figure 4 and Table S1.

Figure 4.

(Left) X-ray crystal structures of (left) (OEP)Co^II(NO) and (right) the cation of [(OEP•+)Co^II(NO)][HO(B(C₆F₅)₃)₂] (CCDC 1939777). The bottom panels show the edge-on views of the π-π interactions and associated mean plane separations (M.P.S.) and lateral shifts (L.S.).

Our DFT results for [(P)Co(NO)]⁺ ([1-NO]⁺; bottom of Table 1) substantiated the relative stability of the Co²⁺ π-radical cation product over its Co³⁺-based analogues.^{[20, 31–34]} The structural model used for [1-NO]⁺ was based on the X-ray structure, with several possible spin state interaction patterns examined (Tables 1 and S2). For this [1-NO]⁺ (S = ½) product, both low-spin Co²⁺ and Co³⁺ initial states were studied. We determined that the π-radical cation state C-II-1 with antiferromagnetic coupling between Co²⁺ and NO was ~7 kcal/mol lower in energy than the Co³⁺-based C-III-1 valence isomer. We thus conclude that the Co-containing product from the “(OEP)Co(NO)+NO+BX₃” coupling reaction was indeed the π-radical cation.

The observation that (OEP)Co(NO) couples with external NO in the presence of BF₃ to generate N₂O necessitates a comparison with the previously reported Fe system.^{[14, 35]} In this regard, there are several interesting points to note. First, the small Δ_NO shift of −40 cm⁻¹ observed in (OEP)Co(NO•BF₃) (c.f., −247 cm⁻¹ in the related (OEP)Fe(NO•BF₃)) is consistent with the longer calculated NO…BF₃ distance of 2.726 Å in 1-NO•BF₃ vs. 1.675 Å in the Fe analog (left of Fig. 5). This weak interaction allows the (P)Co(NO) unit to essentially retain its initial Co^II state; the calculated Co charge increases by only +0.138e upon BF₃ addition. In contrast, the relatively strong “NO-BF₃” interaction in the Fe analog induces electron density transfer to the NO unit to give it ferric-nitroxyl (i.e., Fe^III–NO^–) character (the calculated Fe charge increases by +0.650e upon BF₃ addition). Second, interactions of external NO with the (P)M(NO•BF₃) (M = Co, Fe) adducts result in the presumed formation of hyponitrite (NO)₂^– radical anion intermediates, with formal oxidation of the “(P)M” units as shown in the middle of Fig. 5. In the (P)Co case, the calculated lowest energy structure is the one in which the electron was removed from the porphine macrocycle while retaining the low-spin Co^II (S = ½) state, whereas for Fe, the metal was oxidized to an intermediate-spin Fe^III state. In 2-(NO)₂•BF₃, the proximal (i.e., the NO group directly bonded to Co) and distal (i.e., the added NO group) have calculated opposite charges of −0.547e and +0.124e, respectively (the Fe values are, respectively, −0.539e and +0.146e), significantly enhancing their attraction to facilitate the N–N bond formation step. Third, the isolation of the final products, when excess NO was used, displaying formal porphyrin oxidation for Co (i.e., (OEP•⁺)Co^II(NO)) vs. formal metal oxidation for Fe (i.e., [(OEP)Fe^III(NO)]⁺) (shown on the right of Fig. 5) is consistent with the calculated lowest energy structures for their respective 2-(NO)₂•BF₃ intermediates (middle of Fig. 5). Finally, the determination that the weak interaction in (OEP)Co(NO•BF₃) was sufficient to enable reaction with external NO is consistent with our observation that both (OEP)Co(NO) and (OEP)Fe(NO) in THF at 0 °C are also activated by the weak Lewis acid [K(2.2.2)]OTf (2.5 equiv)^{[14, 36]} towards N–N coupling with excess NO to generate N₂O in detectable but trace (~2–4%) yields (Figs. S9–10). Our observed non-redox active Lewis acid promoted NO coupling reactions involving (OEP)Co(NO) and (OEP)Fe(NO) add to the range of non-redox-active Lewis acid promoted reactions including those of metal-oxo complexes.^[37–39]

Figure 5.

Comparison of key features in the NO coupling reaction for Co (top) and Fe (bottom) starting from the respective (P)M^II(NO) (M = Co, Fe) precursors.

Conclusion

In summary, we demonstrate for the first time that a cobalt-containing heme model, (OEP)Co(NO), is chemically competent to couple with external NO in a synergistic interaction with the Lewis acids BX₃ (X = F, C₆F₅) to generate the greenhouse gas N₂O. Our results thus show that this 2NO→N₂O reaction is not limited to Fe heme models, although the observed N₂O yields (~12%) from the Co reaction with ~1 equiv. of NO were lower than those from the analogous Fe reaction (~16%). We speculate that the relatively easier polarization in (OEP)Fe(NO) vs. (OEP)Co(NO) by BF₃ (eq. 1), coupled with trends in their respective calculated 2-(NO)₂•BF₃ structures, contributes to a more favorable hyponitrite (NO)₂^– radical anion intermediate^[40–45] and subsequent N₂O formation for Fe vs Co. Our results show that a relatively weak (P)M(NO•LA) interaction can activate the NO coupling reaction, and lends support for a possible similar role of the non-heme Fe site in bacNOR. Interestingly, a recent report describes a 1-electron pathway by a heme/non-heme engineered Mb (e.g., Zn^II-Fe_BMb1) that was sufficient for NO coupling and N₂O generation.^[46] A more in-depth analysis of the reaction pathways, including assessment of transition states, is needed to elucidate the complete mechanism of N₂O formation from this experimentally observed NO coupling reaction by Co and Fe heme models facilitated by Lewis acids. We are pursuing such studies and we will report the results in a future paper.