Oxidation of a [CU₂S] Complex by N₂O and CO₂: Insights into a Role of Tetranuclearity in the Cu_z Site of Nitrous Oxide Reductase

Abstract

Oxidation of a [Cu₂(μ-S)] complex by N₂O or CO₂ generated a [Cu₂(μ-SO₄)] product. In the presence of a sulfur trap, a [Cu₂(μ-O)] species also formed from N₂O. A [Cu₂(μ-CS₃)] species derived from CS₂ modeled initial reaction intermediates. These observations indicate that one role of tetranuclearity in the Cu_z catalytic site of nitrous oxide reductase is to protect the crucial S^2-ligand from oxidation.

Because nitrous oxide (N₂O) is harmful as a greenhouse gas and ozone layer destroyer,^1,2 it is critical to understand the mechanisms by which nature regulates its atmospheric concentrations. Notably, during the terminal step of bacterial denitrification, N₂O undergoes two-electron reduction catalyzed by nitrous oxide reductase (N₂OR), generating benign N₂ + H₂O as products.³ The active site for catalytic N₂O reduction in N₂OR is a tetranuclear copper cluster, Cu_z.^4-6 The kinetically competent form of Cu_z that participates in enzymatic catalysis has a [Cu₄(μ₄-S)] structural core,⁷ which activates N₂O upon reduction to the 4Cu^I state.⁸ Computational studies indicate that the reaction coordinate for N-O cleavage involves μ-1,3 binding between two of the Cu centers, labeled Cu_I and Cu_IV (Scheme 1), with crucial assistance from hydrogen bond donation by a protonated lysine residue in the secondary coordination sphere.⁹ Thus, only two Cu centers, Cu_I and Cu_IV, participate directly in N₂O binding and reductive N-O cleavage. The other two Cu centers, Cu_II and Cu_III, do not interact directly with the N₂O substrate according to this model. Furthermore, within the synthetic copper-sulfur model literature, it is clear that dicopper species are capable of mediating two-electron reduction of N₂O,^10-12 and thus that tetranuclearity is not a requirement for N₂O reduction with copper-sulfur clusters.¹³ However, it should be noted that none of the synthetic copper- sulphur clusters with activity towards N₂O feature an unprotected S^2-ligand.

Scheme 1.

N₂O reduction at Cu_z.

An intriguing question, then, is what are the roles of the “spectator” Cu_II and Cu_III centers in the function of N₂OR. In other words, why is the catalytic site of N₂OR tetranuclear as opposed to dinuclear if only two Cu centers are required for the two-electron reduction of N₂O? Two possible roles for Cu_II and Cu_III have been proposed in the literature. First, the two electrons that ultimately transfer from Cu_z to N₂O are delocalized over all four Cu centers via the covalent μ₄-S^2-bridge, which serves to lower the energetic barrier for electron transfer.⁹ Second, a computational study indicates that a putative [Cu₂(μ-S)] core would be susceptible to deactivation by protonation of the S² ligand under catalytic conditions.¹⁴ Thus, the conventional wisdom is that while the Cu_II and Cu_III sites do not interact directly with substrate, they serve to facilitate electron transfer and to protect against protonation.

Studying the small-molecule activation chemistry of a model [Cu₂(μ-S)] complex could add further insight into possible roles of tetranuclearity in Cu_z. An excellent candidate for such studies is the complex [(IPr*)Cu]₂(μ-S) (1), which was published posthumously by Hillhouse in 2015.¹⁵ Reactivity studies of 1 with organic substrates established the nucleophilic character of its bridging S^2-ligand, but no reactivity studies with small molecules such as N₂O were reported. In this study, we have examined the small-molecule activation chemistry of 1 towards N₂O and its isoelectronic analogue, CO₂. Our results lead us to propose that additional roles of tetranuclearity in Cu_Z that have not been appreciated before might include protection of the S^2-ligand against oxidation by N₂O and against expulsion from the active site during N₂O reduction.

Exposing 1 to N₂O (1 atm, room temperature) resulted in a mixture of six different compounds, according to ¹H NMR analysis (Scheme 1a). The major product of the reaction was found to be [(IPr*)Cu]₂(μ-SO₄) (2), which exhibits a diagnostic ¹H NMR resonance for the IPr* para-methyl group at 1.75 ppm (Figure 1). The identity of 2 was confirmed by generating it independently from (IPr*)CuCl + Ag₂SO₄, from observing the [M+H]⁺ ion by ESI-MS, and by X-ray crystallography^‡ (Figure 2a). A minor product was found to be [(IPr*)Cu]₂(μ-O) (3) when the reaction was continued to >36 h; this species has a diagnostic ¹H NMR resonance at 1.78 ppm. The identity of 3 is tentatively proposed by noting that it was generated independently from the dehydration of (IPr*)CuOH^‡ by heating over molecular sieves, and from observing the [M+H]⁺ ion by ESI-MS. We have been unable to grow X-ray quality crystals of 3 even after repeated attempts. A third species, yet unidentified, was determined to be an intermediate that converts to 2 upon further reaction with N₂O. This conclusion was reached by observing that this species, which has a diagnostic ¹H NMR resonance at 1.74 ppm (Figure 1), converted to 2 when the reaction was continued beyond 24 h to 60 h under N₂O. In contrast, no further conversion of this intermediate to 2 was evident even at 5 d when the N₂O atmosphere was replaced with N₂ at the 24-h time point. The other three components of the product mixture (1.82, 1.69, 1.68 ppm in Figure 1) have not been identified but formed in only small quantities. These three products also formed when 1 was exposed to air, but no formation of 2 or 3 from either air or O₂ was evident.

Figure 1.

The poro-methyl region of the ¹H NMR spectrum from reaction of 1 and N₂O (1 atm) in the absence of PPh₃

Figure 2.

Solid-state structures of (a) [(IPr*Cu)₂]([μ-SO₄) (2) and (b) [(IPr*)Cu)₂](μ-CS₃) (4) determined by X-ray crystallography. The IPr* ligands are shown as wireframes, the inorganic cores are shown as 50%-probability thermal ellipsoids, hydrogens are omitted, and only one of two molecules from the asymmetric unit of 2 is shown. Selected bond distances (Å) and angles (°) for 2: Cu(3)-O(5), 1.937(11); Cu(4)-O(6), 1.883(10); S(2)-O(5), 1.444(12); S(2)-O(6), 1.479(12); S(2)- O(7), 1.454(14); S(2)-O(8), 1.488(14). For 4: Cu(1)-S(1), 2.1627(17); Cu(1)-S(2), 2.905(2); S(1)-C(70), 1.677(4); S(2)-C(70), 1.708(7); S(1)-C(70)-S(2), 119.4(2); S(1)- C(70)-S(1#), 121.1(4).

Interestingly, when N₂O was added to a mixture of 1 and PPh₃, the products 2 and 3 formed in roughly equimolar amounts products rather than 2 being the major product (Scheme 2b). The formation of Ph3P=S was detected by ³¹P NMR analysis, indicating that the byproduct of 1 + N₂O converting to 3 is likely elemental sulfur. It is important to note that no reaction was observed between 1 and PPh₃ under N₂, and that Ph₃P=S was not observed when PPh₃ was added to the product mixture after 24 h of exposing 1 to N₂O as opposed to having PPh₃ present from the beginning.

Scheme 2.

Small molecule activation experiments with [(IPr*)Cu]₂(μ-S) (1) and {[(IPr)Cu]₂(μ₃-S)}{BF₄} (5).

Exposing 1 to CO₂ (1 atm, room temperature) provided similar results to the N₂O reaction (Scheme 2a). Once again, the major product was 2. The same three unidentified byproducts formed in small quantities, although no obvious intermediate was observed. An additional minor product was the (IPr*)-CO₂ adduct, whose identity was established by independently exposing the free IPr* carbene to CO₂ (1 atm, room temperature) as is well known for related carbenes.¹⁶ Unlike in the N₂O case, the presence of PPh₃ in the reactant mixture did not impact the nature of the product distribution, and neither compound 3 nor Ph₃P=S were observed from CO₂ under any conditions we examined.

To gain insight into the initial interaction between 1 and N₂O/CO₂, we examined the reactivity of 1 with CS₂ as a model substrate. A rapid reaction was observed between 1 and CS₂, generating pink-colored [(IPr*)Cu]₂(μ-CS₃) (4) quantitatively (Scheme 2c). The identity of 4 was confirmed by X-ray crystallography^‡ (Figure 2b) and by observation of the [M+H]⁺ ion by ESI-MS. Using the structure of 4 as a starting point, we were able to optimize the structure of a DFT model, [(IMe)Cu]₂(μ-CS₃) (4′), whose [Cu₂CS₃] core closely resembles the experimentally determined structure of 4 (IMe = N,N′-dimethylimidazol-2-ylidene). Then, using the 4′ structure as a starting point, we located energy minima for the DFT models [(IMe)Cu]₂(μ-SCO₂) (5′) and [(IMe)Cu]₂(μ-SN₂O) (6′), which are shown in Figure 3. Unlike 4′, the 5′ and 6′ structures exhibit unsymmetrical bridging within the dicopper core. In both cases, one Cu center is engaging in Lewis acid activation of an oxygen atom from the small molecule, with multiple bond character being evident from the optimized C-O/N-N bond distances distal to this Lewis acid activation. Loss of CO and N₂ from 5′ and 6′, respectively, would generate [(IMe)Cu]₂(μ-SO) (7′), which we also were able to optimize (Figure 3). The optimized structure of 7′ places the sulfur and oxygen centers clearly within bonding distance (1.80 Å) and features an unsymmetrical, puckered [Cu₂(μ-SO)] core. We propose that 7′ is a reasonable approximation for an early intermediate that ultimately is either exhaustively oxidized by the small-molecule oxidant to generate 2 or trapped by PPh₃ to generate 3.

Figure 3.

DFT-optimized structures (BVP86/LANL2TZ(f)/6-311+(d)) of dicopper complexes relevant to the small-molecule activation chemistry of [(IPr*)Cu]₂(μ-S). Bond distance labels are in units of Å.

Based on the experimental and computational observations outlined above, our preliminary mechanistic proposal is outlined in Scheme 3. Initial insertion of the small-molecule heteroallene into a Cu-S bond of 1 produces intermediate A, which in turn evolves N₂ or CO to generate intermediate B. From here, presumably there are two distinct pathways. One involves further oxidation of B by three sequential equivalents of the heteroallene oxidant to produce 2. The other involves reversible expulsion of elemental sulfur to produce 3. In the case of N₂O, the formation of 3 can be accelerated by trapping the elemental sulfur with PPh₃. In the case of CO₂, the formation of 2 is sufficiently rapid that 3 does not form even in the presence of PPh₃. Further studies will be required to understand the conversion of A to B and to elucidate how B is further oxidized by N₂O and CO₂.

Scheme 3.

Hypothetical mechanism.

These studies reveal the chemistry of a [Cu₂(μ₂-S)] core towards small molecules including N₂O. If one were to imagine the catalytic site of N₂OR having such a dinuclear core instead of a tetranuclear [Cu₄(μ₄-S)] core, productive catalysis would be hampered or prevented by oxidation and/or expulsion of the S^2-ligand. Thus, in addition to participating in electron delocalization and preventing protonolysis, another role of the two “spectator” Cu centers in Cu_z may be to prevent these unproductive side reactions of the bridging S^2-ligand from competing with productive N₂O reductase reactivity. In this context, it is intriguing to note that a “protected” tricopper analogue of 1, {[(IPr)Cu]₃(μ₃-S)}+ (5),¹⁷ was found to be inert towards N₂O and CO₂ under conditions where 1 undergoes the reactions described in this manuscript (Scheme 2d).