Coordination chemistry of the Cu_Z site in nitrous oxide reductase and its synthetic mimics

Abstract

Atmospheric nitrous oxide (N₂O) has garnered significant attention recently due to its dual roles as an ozone depletion agent and a potent greenhouse gas. Anthropogenic N₂O emissions occur primarily through agricultural disruption of nitrogen homeostasis causing N₂O to build up in the atmosphere. The enzyme responsible for N₂O fixation within the geochemical nitrogen cycle is nitrous oxide reductase (N₂OR), which catalyzes 2H⁺/2e⁻ reduction of N₂O to N₂ and H₂O at a tetranuclear active site, Cu_Z. In this review, the coordination chemistry of Cu_Z is reviewed. Recent advances in the understanding of biological Cu_Z coordination chemistry is discussed, as are significant breakthroughs in synthetic modeling of Cu_Z that have emerged in recent years. The latter topic includes both structurally faithful, synthetic [Cu₄(µ₄-S)] clusters that are able to reduce N₂O, as well as dicopper motifs that shed light on reaction pathways available to the critical Cu_I-Cu_IV cluster edge of Cu_Z. Collectively, these advances in metalloenzyme studies and synthetic model systems provide meaningful knowledge about the physiologically relevant coordination chemistry of Cu_Z but also open new questions that will pose challenges in the near future.

1. Introduction

1.1. Nitrous oxide

Nitrous oxide (N₂O) is one of the oxidized forms of nitrogen (N₂) present in the global nitrogen cycle [1,2], in which the atmospheric nitrogen is accumulated in terrestrial systems through biological nitrogen fixation and later released back to the atmosphere as N₂ by bacterial denitrification pathways [1,3,4]. Anthropogenic production of N₂O disturbs nitrogen homeostasis, resulting in accumulation of N₂O in the atmosphere. According to the ‘‘Inventory of greenhouse gas emissions and sinks, 1990–2018” by the U.S. Environmental Protection Agency, the global atmospheric concentration of N₂O has increased from 270 parts per billion (ppb) to 331 ppb (by 23%) since 1750, reaching a concentration that had not been exceeded during the last 800,000 years [5]. The increasing atmospheric N₂O concentration is an emerging threat as N₂O is one of the main contributors towards both global warming and ozone layer depletion [1,5–7]. Certainly, N₂O is not the most abundant greenhouse gas (Fig. 1); nonetheless, its global warming potential is 298-fold higher than that of CO₂ due in part to its long atmospheric half-life of 121 years [5]. The anthropogenic production of N₂O in the United States primarily occurs by agricultural soil management, stationary combustion, manure management, fuel combustion, and adipic acid production (Fig. 2) [5,6].

Fig. 1.

2018 greenhouse gas emissions [% by million metric tons of CO₂ equivalents] in the United States. Figure was taken from the ‘‘Inventory of U.S. Greenhouse Gas Emissions and Sinks, 1990–2018” by U.S. EPA [5].

Fig. 2.

2018 Sources of N₂O emissions (in million metric tons of CO₂ equivalents) in the United States. Figure was taken from the ‘‘Inventory of U.S. Greenhouse Gas Emissions and Sinks, 1990–2018” by U.S. EPA [5].

From 1990 to 2018, the total emissions of CO₂ have increased by 300.9 million metric tons of CO₂ equivalents (5.9%) and that of CH₄ have decreased by 139.9 million metric tons of CO₂ equivalents (18.1%) [5]. Despite fluctuations, the N₂O emissions have remained constant [5]. The impact of N₂O on the ecosystem due to higher global warming potential, atmospheric lifetime, and ozone layer depletion raises awareness in the scientific community to understand its metabolic pathways.

1.2. N₂O metabolism

Atmospheric N₂O is primarily removed by the photolytic action of sunlight in the stratosphere [5], while terrestrial N₂O is metabolized through microbial denitrification in soils and fresh and marine waters [1]. Under anoxic conditions, these denitrification organisms consume the oxidized forms of N₂ in place of molecular dioxygen (O₂) for essential metabolic pathways including anaerobic respiration and ATP synthesis [8–11]. To date only three enzymes – nitrogenase, multicopper oxidase, and nitrous oxide reductase (N₂OR) – have been identified to metabolize N₂O, with the latter being predominant biologically. Interestingly, these metalloproteins neither share the same catalytic site nor metal composition. Nitrogenase, a metalloenzyme found in the nitrogen fixation pathway, has a FeMo cofactor that primarily produces NH₃ using N₂ as its substrate [12]. Studies show that N₂O acts as an competitive inhibitor; in fact, nitrogenase can utilize N₂O as a source of its N₂ substrate [13–16]. The multicopper oxidase from archaeon Pyrobaculum aerophilum produced in E. coli remains the only example of that enzyme that utilizes N₂O in place of its primary substrate, O₂, to function as a metallo-oxidase for Fe²⁺ and Cu⁺ [17]. Its catalytic site is composed of a trinuclear copper cluster supported by histidine groups, but the absence of sulfur and the geometric orientation may disfavor the coordination and activation of N₂O as compared to N₂OR [17]. To date, the products resulting from N₂O activation by multicopper oxidase under physiological conditions are not known. N₂OR is the most efficient and physiologically relevant enzyme for removing terrestrial N₂O. N₂OR is one of the enzymes participating in the microbial denitrification process that involves consecutive reductions of nitrate $\left(N{O}_3^{-}\right)$ by a sequence of enzymes in which the N₂O is converted into inert N₂ and H₂O by N₂OR at the last step (Scheme 1) [8]. The 2H⁺/2e⁻ reduction of N₂O is thermodynamically favorable (ΔG° = ‒339.5 kJ/mol) but requires to be catalytically driven by N₂OR due to a high activation energy barrier (250 kJ/mol) coupled to a spin forbidden process [9,10].

Scheme 1.

(A) The sequence of bacterial denitrification highlighting the final step of converting N₂O to N₂. The enzyme involved in each catalytic step is given above the arrow. The denitrification sequence was adapted from reference [9]. (B) Reduction of N₂O by N₂OR and its thermochemical parameters.

2. Nitrous oxide reductase (N2OR)

2.1. N₂OR overview

N₂OR is a homodimeric metalloenzyme with a molecular weight of ~ 120 kDa, and each monomer carries six copper atoms that are located in two distinct metal-containing protein domains [18]. The electron transfer site (Cu_A) at the C-terminal domain holds two copper atoms, while the active site (Cu_Z) at the N-terminal domain contains four copper atoms [9,19]. The dimer is arranged in head-to-tail fashion such that the Cu_A site of one monomer lies ~ 10 Å from the Cu_Z site of the other monomer, while the two sites within the same monomer are approximately 40 Å apart (Fig. 3) [4,9]. It is likely that the copper sites within the same monomer are far too away for efficient electron transfer, but Cu_A and Cu_Z from different subunits are in proximity for cooperative substrate (N₂O) reduction [4].

Fig. 3.

Ribbon diagram of N₂OR dimer from Marinobacter hydrocarbonoclasticus at 2.4-Å resolution (PDB ID 1QNI). Two monomers are shown in plum and tan colors. Ribbon segments covering the metal domains have been removed for clarity. Image was created using UCSF Chimera.

In all N₂OR crystal structures to date, additional electron densities corresponding to Ca²⁺/K⁺ are seen close to the metal sites [4]. However, studies of non-metalated N₂OR suggest that Ca²⁺ was incorporated into the protein after the in vivo metalation events, implying that Ca²⁺/K⁺ ions are not required for dimer formation but presumably help the protein’s structural integrity and rigidity by keeping the two monomers together [21].

2.2. Cu_A electron transfer site

Cu_A is a binuclear copper site (Fig. 4) located at the C-terminus of N₂OR, and its existence was first discovered by Kroneck and co-workers using spectroscopic techniques including EPR, EXAFS and UV–Vis [22]. Multifrequency EPR studies of the oxidized form of Cu_A demonstrated a direct Cu…Cu interaction similar to that characterized for cytochrome c oxidase, corresponding to a mixed-valent [Cu^1.5Cu^1.5] dicopper site [23–25]. The X-band EPR spectrum of the oxidized form of Cu_A exhibits a 7-line hyperfine (intensity ratios 1:2:3:4:3:2:1) pattern with g_|| = 2.18, g⊥ = 2.13 and A_|| = 3.38 mT due to the unpaired electron being delocalized over the two metal centers equally (I_Cu = 3/2), while the reduced form [Cu^I-Cu^I] is EPR silent as no unpaired electrons are present [8,9]. EXAFS studies initially suggested that the Cu_A is a Cu₂S₂ cluster with 2.43-Å Cu···Cu and 2.2-Å Cu–S distances [26]. Later, these geometric parameters of Cu_A were confirmed with the X-ray structure of N₂OR from Marinobacter hydrocarbonoclasticus, at 2.4-Å resolution [20]. The visible spectrum of the oxidized form of Cu_A displays absorption maxima at 480 and 525–540 nm, both associated with S (Cys) → Cu charge transfer, and at 800 nm for intervalence charge transfer of the mixed valent site [8,9]. Overall, Cu_A has been well characterized using crystallographic and spectroscopic techniques for its function as an electron shuttle for N₂O reduction at the Cu_Z catalytic site.

Fig. 4.

The dicopper electron transfer site (Cu_A) of N₂OR from Marinobacter hydrocarbonoclasticus, at 2.4-Å resolution (PDB ID 1QNI) with selected bond distances shown [20]. His-526 and His-569 are coordinated to Cu ions while Cys-565 and Cys-561 residues are bridging between Cu atoms through their S atoms. Image was created using UCSF Chimera.

2.3. N₂OR catalytic site

The structure and the identity of the tetranuclear copper sulfide active site of N₂OR remained obscure for several years until it was characterized using X-ray crystallography and spectroscopic techniques starting in the late 1990′s [27–29]. To date, two forms of the active site have been biologically isolated and are often denoted as Cu_Z (Cu₄S₂) and Cu_Z* (Cu₄S), differing in their compositional, structural, spectroscopic, and kinetic features (Fig. 5) [8]. It is challenging to isolate N₂OR samples with purely Cu_Z or Cu_Z*, making enzymological studies challenging. Generally, anaerobic conditions favor Cu_Z while aerobic isolation favors Cu_Z*, in which the µ-sulfide along the Cu_I–Cu_IV cluster edge has been replaced with a solvent-derived ligand [8,9]. In both forms, the active site resembles a distorted tetrahedron supported by 7 conserved histidine residues, with each copper except Cu_IV bound to 2 histidine groups. For both forms, the average Cu–S bond distance remains ~ 2.3 Å, but with one Cu atom (Cu₄ in Fig. 5A and Cu₂ in Fig. 5B) located significantly away (>3.3 Å) from the rest of the Cu atoms. [11,27,30].

Fig. 5.

The two forms of the active site of N₂OR isolated under different conditions. (A) Cu_Z* isolated aerobically from Pseudomonas stutzeri (PDB ID 3SBP) at 2.1 Å resolution. (B) Cu_Z isolated anaerobically from Paracoccus denitrificans (PDB ID 1FWX) at 2.16 Å resolution. Image was created using UCSF Chimera.

2.4. Cu_Z, Cu_Z* and their redox forms

Cu_Z is typically isolated as its oxidized [2Cu^I:2Cu^II] state with either the oxidized (2Cu^1.5+) or reduced (2Cu^I) form of Cu_A [4,8,9,31]. By convention, the oxidized form of Cu_Z termed as a ‘‘2-hole” state, as two Cu²⁺ (d⁹) ions bring two electron holes: from now onward, the oxidized form of Cu_Z will be called ‘‘2-hole Cu_Z”. Selective reduction of Cu_A by sodium ascorbate allows the isolation and spectroscopic characterization of 2-hole Cu_Z with the reduced form of Cu_A (2Cu^I) [32]. The UV–Vis spectrum of 2-hole Cu_Z exhibits a single absorption band at ~ 550 nm, assigned as S (3p) → Cu(3d) charge transfer as further supported by resonance Raman spectroscopy and DFT calculations [33].

Dithionite reduces both Cu_A and Cu_Z sites, and the reduced form of Cu_Z features a [3Cu^I:1Cu^II] electronic state. Accordingly, the reduced form of Cu_Z is called ‘‘1-hole Cu_Z”, and it becomes EPR active with the electron hole being delocalized over the four Cu atoms [33]. Multifrequency EPR spectroscopy of 1-hole Cu_Z indicates parameters of g_|| = 2.152, g⊥ = 2.042, A_|| = 56 10⁻⁴ cm^—1, and A⊥ = 20 10⁻⁴ cm^—1 simulated with the unpaired electron delocalized equally over three of the four Cu centers [33]. It is assumed that these three Cu centers are Cu_I, Cu_II, and Cu_IV, which are coplanar with the two S atoms in the crystal structure. The 1-hole Cu_Z has a characteristic optical band at ~ 670 nm, which is attributed to overlapping S(3p) → Cu(3d) charge transfer, His → Cu(3d) charge transfer, and Cu d → d transitions [9]. Spectroscopic studies suggest that upon reduction to the 1-hole state, the µ₂-S²⁻ ligand of Cu_Z undergoes protonation to become a µ₂-SH⁻ ligand [33]. The reduction potential of the 2-hole/1-hole Cu_Z redox pair has been measured to be + 65 ± 5 mV vs. SHE at pH 7.6 [34]. To date, no 3-hole [1Cu^I:3Cu^II], 4-hole [4Cu^II] or fully reduced [4Cu^I] forms of Cu_Z have been observed. In fact, the attempted prolonged incubation with reduced methyl viologens to reach the fully reduced [4Cu^I] form of Cu_Z was unsuccessful [32].

Cu_Z* can also be isolated with either oxidized or reduced form of Cu_A. The primary differences of geometry and coordination between Cu_Z* and Cu_Z arise from the solvent-derived molecule coordinated across the Cu_I–Cu_IV cluster edge. The observed spectroscopic features [33,35,36] have been best fitted to an occupancy of a hydroxyl (OH^–) ligand across the Cu_I–Cu_IV edge that is slightly closer to Cu_I (2.00 Å) than to Cu_IV (2.09 Å) [37,38]. Oxidized Cu_Z* possesses a [3Cu^I:1Cu^II] electronic state and is accordingly called ‘‘1-hole Cu_Z*”. Multifrequency EPR spectroscopy of 1-hole Cu_Z* indicates parameters of g_|| = 2.160, g⊥ = 2.043, A_|| = 61 × 10⁻⁴ and 23 × 10⁻⁴ cm⁻¹, and A⊥ = 25 × 10⁻⁴ and 20 × 10⁻⁴ cm⁻¹, simulated with the unpaired electron delocalized unequally over two of the four Cu centers in a ~ 5:2 ratio [33]. The visible spectrum of 1-hole Cu_Z* consists of a strong absorption band at ~ 640 nm that is attributed to overlapping S(3p) → Cu(3d) charge transfer, His → Cu(3d) charge transfer, and Cu d → d transitions [29,37,39]. Literature reports indicate that the absorption band at ~ 550 nm in 2-hole Cu_Z may originate from the µ₂-S₂ atom, as it is absent in the visible spectrum of 1-hole Cu_Z* [4]. However, that conclusion is debatable as the 1-hole Cu_Z does not feature the absorption band at ~ 550 nm. Unlike 2-hole Cu_Z, the 1-hole Cu_Z* is resistant to reduction. In fact, it can only be reduced to the spectroscopically silent, fully reduced state (4Cu^I) upon prolonged incubation (3–5 h) with large excess of reduced viologen [32]. The reduction potential of the 1-hole/fully-reduced Cu_Z* redox pair has not been reported, as electrochemical reduction of 1-hole Cu_Z* is not achievable [34,40]. However, it should be noted that reduced violegens are generally considered to be too reducing to fall within a reasonable physiological reduction potential window.

Apart from the different redox states of N₂OR mentioned above (Table 1), no other redox forms of the active site have been observed to date. However, some interesting derivatives of those forms have been characterized using spectroscopic and crystallographic techniques and will be discussed in the following sections.

Table 1.

Isolable redox forms of N₂OR with their purification method, active site composition, spin state and visible spectrum absorption.

Redox form of N2OR		Source	Catalytic site composition	CuZ Spin state	CuZ visible spectrum
CuA	CuZ
[2Cu1.5]	[2CuI:2CuII]	Anerobic purification	CuZ (4Cu:2S)	S = 0	N/Aa
[2CuI]	[2CuI:2CuII]	Anerobic purification and ascorbate reduction		S = 0	550 nm
[2CuI]	[3CuI:1CuII]	Anerobic purification and dithionite reduction		S = 1/2	670 nm
[2Cu1.5]	[3CuI:1CuII]	Aerobic purification	CuZ* (4Cu:1S)	S = 1/2	N/Aa
[2CuI]	[3CuI:1CuII]	Aerobic purification and ascorbate reduction		S = 1/2	640 nm
[2CuI]	[4CuI]	Aerobic purification and prolonged methyl viologen reduction		S = 0	Silent

^aSpectral features of Cu_A and Cu_Z are convoluted.

3. Substrate/inhibitor interactions with the N₂OR active site

3.1. Overview

Substrate binding studies of N₂OR are typically challenging due to N₂O being a weak ligand with poor σ-donor and π-acceptor ability [9]. Literature assignment of the Cu_I–Cu_IV (Fig. 5B) edge of Cu_Z in the 4Cu:1S form as the substrate binding site of N₂OR has largely depended on spectroscopic measurements [41] combined with DFT modeling [35]. Additionally, this proposal is strongly supported by crystallograpic studies of N₂OR inhibition by iodide (I⁻) and the N₂O-pressurized crystal structure of N₂OR [30,42]. It is also fundamentally supported by the fact that the Cu_IV atom is coordinatively unsaturated by virtue of only being supported by one histidine residue, thus providing a potential coordination site for the substrate (N₂O); the other three Cu atoms are supported by two histidine residues each. Furthermore, one could imagine the labile solvent derived molecule across the Cu_I–Cu_IV edge being replaced by N₂O in the beginning of a substrate activation process.

3.2. Interaction of substrate (N₂O) with Cu_Z

The study reported by Einsle and co-workers remains as the only report of a N₂OR crystal structure in which the substrate (N₂O) is occupied in proximity to the proposed substrate binding site [42]. The anoxic isolation of N₂OR from P. stutzeri contained the oxidized form of Cu_A [2Cu^1.5] and the 2-hole form of Cu_Z (4Cu:2S, [2Cu^I:2Cu^II]). The X-ray diffraction data collected from N₂O-pressurized N₂OR crystals contained extra electron density that was adequately modeled to be a linear N₂O molecule [42]. It is important to note that the substrate was not located within covalent bonding distance to any metal atom. Nonetheless, weak interactions of N₂O and Cu_Z were deduced from EPR and UV–Vis data [42]. The authors proposed that the substrate is carried through the hydrophobic channels from the protein surface to the cluster, positioning it in a tight binding pocket created by Phe-621, His-626 and Met-627 residues [42]. Upon N₂O activation and reduction, the apolar N₂ product then escapes through the hydrophobic channels while the produced H₂O molecule is retained in the distal water-filled cavity adjacent to the binding pocket. However, no subsequent reactivity or redox chemistry was observed with this substrate-incorporated N₂OR sample, and the only significant structural difference compared to the substrate-free N₂OR structure is that the His-583 histidine residue is coordinated to the Cu_A site, which otherwise is rotated ~ 130⁰ away from Cu_A and participating in hydrogen bonding with backbone residues. This phenomenon suggests that the electron transfer event between Cu_A and Cu_Z only occurs upon flipping and coordination of His-538 to Cu_A, which is apparently triggered by introduction of the N₂O molecule within the hydrophobic channel [42].

3.3. Inhibitor (I⁻) interaction with Cu_Z

The N₂OR inhibition study reported by Hasnain and co-worker is particularly important not only because it represented the first inhibitor bound N₂OR crystal structure but also because it strongly supported the proposed substrate binding site for N₂OR [30]. The crystallographic characterization of anaerobically isolated N₂OR from Achromobacter cycloclastes (AcN₂OR, PDB ID 2IWF) at 1.86-Å resolution revealed ligation of two oxygen atoms (H₂O/OH^–) to Cu_I and Cu_IV atoms separately (Fig. 6A), which is different from conventional Cu_Z or Cu_Z* forms. However, the visible and EPR spectra confirmed the active site to be a 1-hole Cu_Z* (4Cu:1S, [3Cu^I:1-Cu^II]) form [30]. Incubation of AcN₂OR with the known N₂OR inhibitor NaI over a prolonged period resulted reduction of the Cu_A center as evidenced by the complete loss of spectroscopic features associated to Cu_A [30]. However, the iodide-inhibited Cu_Z* active site remained unchanged in terms of redox state, as evidenced by characteristic 650 nm absorption band in visible spectrum, even after the attempted oxidation by K₃Fe(CN)₆ [30].

Fig. 6.

(A) The active site of native N₂OR from Achromobacter cycloclastes (AcN₂OR, PDB ID 2IWF) at 1.86 Å showing the Oxy₁ and Oxy₂ atoms coordinated to Cu_I and Cu_IV respectively. (B) inhibitor (I⁻) occupies the proposed substrate binding site (PDB ID 2IWK, at 1.7 Å), causing the active site to lose its redox function. Helix-red, sheet-yellow and loop-green. Images were created using PyMol educational version.

In the native AcN₂OR, the Cu_I and Cu_IV atoms are coordinated by solvent-derived Oxy₁ and Oxy₂ atoms at 2.2 and 2.5 Å respectively (Fig. 6A). The distance between two oxygen atoms is 2.3 Å, which could accommodate a bent N₂O molecule (where Cu_IV is bound to the oxygen and Cu_I to the nitrogen atoms of N₂O), as proposed by DFT calculations [30]. The active site loses its redox activity upon iodide binding (Fig. 6B, PDB ID 2IWK). Overall, the proposed catalytic mechanism involves N₂O binding across the Cu_I–Cu_IV edge, and upon reduction and release of N₂ the remaining oxygen atom may coordinate between Cu_I and Cu_IV like iodide does in the inhibitor bound complex (Fig. 6B) [30]. However, such a complex resembles a dormant resting state of the catalytic site, as will be discussed further in Section 4.3.

4. The mechanism of N₂O reduction by N₂OR

4.1. Dependence of catalytic activity of N₂OR

The redox activity and catalytic features of N₂OR have been seen to vary depending on the isolation procedure, the mode of activation (reducing agent), the pH, and the redox state of the active site [4,8,9]. Typically, the specific activity is reported as the µmol of N₂O reduced/min/mg of N₂OR and in most cases is determined by indirect spectrometric assay using viologen dyes [43,44]. Alternatively, a direct chromatographic determination has been used to measure the N₂O consumption or N₂ production [43,45].

The idea that N₂OR requires external activation is supported by the fact that crude N₂OR extracts display specific activities ranging from 48 to 72 µmol of N₂O reduced/min/mg of N₂OR that drop down to 1–10 µmol of N₂O reduced/min/mg of N₂OR upon either aerobic or anaerobic purification [45,46]. The latter samples are believed to be in an unready state of the enzyme [47], though other possible in vitro mechanisms cannot be ruled out. It is a valid argument that the in vitro spectroscopic assays deviate from the physiological conditions, as the reduction potentials of the electron donors used (methyl viologen, 450 mV vs. SHE at pH 7.0; benzyl viologen, 374 mV vs. SHE at pH 7.0) [48] are too low to be physiologically relevant to the native bacteria. However, studies using physiologically relevant electron donors have also given higher specific activity for the crude cell extracts than for purified N₂OR [49–51]. Moreover, the purified N₂OR from W. succinogenes shows high specific activity (160 µmol of N₂O reduced/min/mg of N₂OR), implying that it does not require an external activation [52]. The next section will discuss the redox forms of copper sites (Cu_A and Cu_Z) found in the activated and/or non-activated N₂OR.

4.2. Catalytic activity of different Cu_Z forms

So far, 4 different redox forms of Cu_Z have been discussed, with either the reduced or oxidized form of Cu_A: 2-hole Cu_Z, 1-hole Cu_Z, 1-hole Cu_Z*, and fully reduced Cu_Z* (Table 1). Of these, 2-hole Cu_Z and 1-hole Cu_Z* do not react with N₂O even after Cu_A is reduced, as evidenced by lack of changes to spectroscopic features in the presence of N₂O [32]. The 1-hole Cu_Z coupled to reduced Cu_A has been found to slowly react with N₂O, oxidizing back to 2-hole Cu_Z and oxidized Cu_A, completing a two-electron process [32]. However, the corresponding turnover number (k = 0.6 h^—1) is too low to be physiologically relevant. On the other hand, the fully reduced Cu_Z*, resulting from prolonged incubation of 1-hole Cu_Z* with methyl viologen (a non-physiologically relevant reductant), reacts with N₂O as evidenced by reappearance of spectroscopic features corresponding to 1-hole Cu_Z* and oxidized Cu_A, as well as GC–MS detection of ¹⁵N₂ production upon using ¹⁵N-labeled N₂O [53]. The turnover frequency for the catalytic reaction reached 320 s^—1, but the reductive activation of 1-hole Cu_Z* is too slow (1.2 × 10⁻³ s^—1) for this to be considered as physiologically relevant [32,41,53].

An interesting intermediate named Cu_Z° has been observed within the first two minutes of the stoichiometric reaction between fully reduced Cu_Z* and N₂O [32,40]. This intermediate is characterized by an absorption band at ~ 680 nm (ε = 2000 M^—1 cm^—1) and has a turnover rate of 200 s^—1, followed by a decay rate of ~ 5 × 10⁻³ s^—1 with reappearance of characteristic absorption bands for 1-hole Cu_Z* [32,40]. Both the formation and decay rates are compatible to reported steady state kinetic assays for enzymatic N₂O reduction, thus giving Cu_Z° a potential physiological relevance [32,40]. The observed spectral features of Cu_Z° are best explained by a DFT model in which the Cu_IV atom is coordinated to a terminal OH^– ligand at a distance of 1.93 Å, with further stabilization coming from hydrogen bonding between the hydroxyl ligand and a protonated lysine residue (Lys397) that is interacting with negatively charged glutamate residue (Glu435) [38]. The EPR parameters for Cu_Z° were determined to be g_|| = 2.177, g⊥ = 2.05, and A_|| = 42 × 10⁻⁴ cm^—1, simulated with the unpaired electron being delocalized equally over two of the four Cu centers, presumably Cu_I and Cu_IV [38].

4.3. Latest proposed mechanism for N₂O reduction by N₂OR

As concluded in the previous section 4.2, the only form of Cu_Z that is catalytically competent and physiologically relevant is the fully reduced (4Cu^I) form of Cu_Z*. Putting all of the previous information together with several computational modeling studies produced over the years by Solomon [10,54,55], an updated mechanism has been proposed for the in vitro reaction of N₂OR activated by reduced methyl viologen with equimolar N₂O [38]. The slow reductive activation of N₂OR from an off-cycle resting state, 1-hole Cu_Z*, results in a reduced Cu_A and fully reduced Cu_Z* (Scheme 2, intermediate 1), which lies on the active cycle.

Scheme 2.

The mechanism of in vitro reduction of N₂O by N₂OR at its Cu_Z* center, showing the intermediates and the proton coupled electron transfer events from Cu_A center. The catalytically competent cycle is shown using solid arrows while the dashes indicate the slow alternative pathway in the absence of reductants. The conserved amino acid residues are labeled according to M. hydrocarbonoclasticus N₂OR mature primary sequence. The scheme was adapted from reference [9].

N₂O coordinates across the Cu_I–Cu_IV edge in bent (139⁰) µ−1,3 fashion with N and O termini coordinated to Cu_I and Cu_IV respectively (Scheme 2, intermediate 2) [38,41,56]. The intermediate 2 is further stabilized by H-bonding between the coordinated oxygen atom of N₂O and a nearby protonated lysine residue (Lys397) [38]. The strong backbonding from the Cu cluster into the π* system of N₂O makes the Cu_I–N and Cu_IV–O bonds stronger than the N–N and N–O bonds, which in turn makes the inert N₂O susceptible for reduction [39,57]. Next, N₂ is liberated upon a two-electron transfer from the cluster to N₂O that induces N–O and Cu_I–N bond cleavage, resulting in a hypothetical 2-hole Cu_Z*. This presumed, short-lived intermediate has never been observed as the N₂ liberation is accompanied by subsequent protonation and single electron transfer from Cu_A, resulting in Cu_Z° (Scheme 2, intermediate 3), evidenced by spectroscopic and DFT studies [4]. A second proton-coupled electron transfer event converts the Cu_Z° back to the fully reduced state, completing the catalytic cycle [38]. In the absence of sufficient reductants, Cu_Z° slowly decays (5 × 10⁻³ s^—1, pH 7.6) to its resting state, 1-hole Cu_Z* with the OH^– ligand bridging along the Cu_I–Cu_IV edge rather than remaining terminally attached to Cu_IV [40]. Furthermore, the intramolecular reduction rate of Cu_Z° by sodium ascorbate has been found to be ~ 10⁴ times faster than that of 1-hole Cu_Z*, supporting the biological competence and physiological relevance of Cu_Z° and the off-cycle nature of the 1-hole Cu_Z* resting state.[38] The catalytic activity of N₂OR has been found to be sensitive to the pH of the medium, even though the precise mechanistic influence remains to be fully understood [40]. Disturbance in the H-bonding network at different pH values may alter the geometric orientation around the substrate binding site and could also affect the protonation state of the interacting lysine residue that facilitates reduction of Cu_Z° by increasing its reduction potential and preventing decay to the off-cycle resting state [4,37,38,40].

There remain open questions about the role of 4Cu:2S Cu_Z in physiological N₂O reduction. As described above, this form of the active site is favored when protein purification and crystallization is conducted under anaerobic conditions, which better approximate the anaerobic microorganisms involved in bacterial denitrification. However, neither redox form of the 4Cu:2S Cu_Z is known to be catalytically competent thus far. Given the observation that the second µ₂-S²⁻ ligand along the Cu_I-Cu_IV edge undergoes protonation upon reduction [33], it is tempting to consider whether there is a mechanism by which the 4Cu:2S Cu_Z can convert to the 4Cu:1S Cu_Z. However, there is no evidence at this time to indicate such conversion, although loss of µ₂-sulfur atoms from other biological clusters such as the FeMo cofactor of nitrogenase has been observed recently [58–60], and instability of a dicopper µ₂-sulfido unit under N₂O conditions is suggested by synthetic model studies (see Section 5.3).

5. Structural and/or functional model complexes of Cu_Z

5.1. Overview

The content discussed in Section 4 summarizes the catalytically competent Cu_Z forms and the proposed intermediates of the catalytic cycle. Protein samples containing the fully-reduced [4Cu^I] and off-cycle [3Cu^I:1Cu^II] Cu_Z* have been isolated, but crucial catalytic intermediates 2 and 3 (Scheme 2) have been studied only using spectroscopic and DFT methods, as the isolation of these short-lived species is challenging. Synthetic model complexes supported by appropriate ligands could provide an alternative way to isolate and study such intermediates, lending insight into the chemistry of these unusual inorganic functional groups in the absence of the protein matrix. In this section, the reported structural and/or functional copper-sulfur model complexes of Cu_Z will be briefly discussed for their relevance to the structure and activity of N₂OR. Several notable breakthroughs have come about in this area since other reviews focused on synthetic models of Cu_Z were published [61]. Fundamental knowledge gained from synthetic model studies can be used not only to guide hypotheses about N₂OR itself, but also to inspire catalyst designs for applications in environmental remediation as well as chemical synthesis using N₂O as an oxidant [62].

5.2. Cu-S clusters supported by nitrogen-rich ligands

Nitrogen rich ligands like ethylenediamines (EDAs) and diketimines (nacnacs) have been utilized to mimic the histidine environment supporting the native Cu_Z site [63]. In many cases, these bidentate ligands chelate single Cu ions, resulting a mononuclear copper complexes that are then reacted with appropriate sulfur-atom donors to assemble copper-sulfur clusters. However, this self-assembly strategy is sensitive to nature of copper salt, the sulfur source, and the reaction conditions, leading to a wide range of di-, tri-, and multinuclear copper-sulfur clusters reported in the literature (Fig. 7) [63–65]. These complexes do not resemble the composition and/or connectivity of the Cu_Z active site, and observation of activity towards N₂O reduction has been elusive because of the presence of coordinatively saturated copper centers in many cases. Attempts have also been made to stabilize multicopper µ-S²⁻ complexes resulting in multicopper complexes with no sulfur incorporation [66–68].

Fig. 7.

Selected copper-sulfur complexes supported by nitrogen-rich ligands.

In selected cases, well-defined complexes with a single sulfur-atom bridge and containing coordinatively unsaturated Cu centers were accessible by avoiding construction through self-assembly. The complex L₃Cu₃S (1) synthesized by Murray and coworkers (Fig. 7) represents the first coordinatively unsaturated copper-sulfur cluster supported by nitrogen rich ligands and was accessed using a cage-like cyclophane ligand that templates the cluster geometry and composition [69]. Interestingly, 1 possesses a reversible one-electron redox event at E_1/2 = −1.44 V vs Fc⁺/Fc (~−0.8 V vs SHE [70]) that is assigned to a [Cu₃S]^3+/2+ redox couple. The one-electron reduced complex was afforded by the chemical reduction of 1 with decamethylcobaltocene, but no N₂O reduction activity was reported with either redox form [69]. Subsequently, Hillhouse and coworkers devised an elegant protection/deprotection strategy to control the sequential introduction of copper and sulfur atoms in assembling the $C{u}_3^I\left({\mu }_3-S\right)$ complex 2 supported by N-heterocyclic carbene (NHC) ligands, which resemble histidine residues despite not being N-donor ligands (Fig. 7) [71]. This complex was isolated in the 3Cu^I redox state, and no redox chemistry or N₂O reactivity was reported.

It should be noted in this section that Tolman and coworkers published the first multicopper–sulfur complex able to reduce N₂O to N₂ using neutral nitrogen donor ligands [72]. This result will be discussed further in Section 5.4.

5.3. Structural models of Cu_Z that contain Cu₄(µ₄-S) cores

Regarding structural model complexes of Cu_Z, the first Cu₄(µ4-S) complex was reported in 1993 by Yam and co-workers, actually pre-dating the first structural study of N₂OR [73]. The phosphine-supported [Cu₄(µ₄-S)(dppm)₄][PF₆]₂ (3) (dppm = bis(diphenylphosphino)methane) (Scheme 3) complex and related derivatives were studied for their photochemical properties [74]. Complex 3 contains four Cu^I ions bridged by a µ₄-S²⁻ ion, resembling the connectivity and composition of Cu_Z, yet neither displays any reversible redox features nor reactivity towards N₂O. Following Yam’s synthetic procedure, our group synthesized the complex [Cu₄(µ₄-S)(dppa)₄][(PF₆)₂] (dppa = bis(diphenylphosphino)amine) (4) with potential hydrogen-bonding amine groups [75] (Scheme 3).

Scheme 3.

Synthesis of structural models [Cu₄(µ₄-S)(dppm)₄][(PF₆)₂] (3) and [Cu₄(µ₄-S)(dppa)₄][(PF₆)₂] (4).

Indeed, complex 4 was found to have hydrogen bonding interactions with solvent molecules as well as the PF⁻ counterions, both in solution and the solid state. When comparing the solid-state structures of 3 and 4, it is evident that the hydrogen bonding interactions cause a distortion of the Cu₄(µ₄-S) core in complex 4. Fortuitously, the core of complex 4 resembles the connectivity, composition, and geometry of Cu_Z, as evidenced by similar Cu-Cu distances, Cu-S distances, and geometry index $\left({{\tau }^{\prime }}_4\right)$ at the µ₄-S²⁻ ligand to those of the enzyme active site (Fig. 8). In particular, the structures of the Cu_Z* active site and the core of 4 are both unsymmetrical, with one Cu atom (Cu_I) positioned further away from the other three. On the other hand, 3 is symmetrical with a nearly idealized S₄ axis of symmetry passing through the sulfur atom.

Fig. 8.

Comparison of the Cu-Cu distances (shown in units of Å) of (left) the active site (Cu_Z*) of Paracoccus denitrificans (PDB ID 1FWX), (middle) complex 3, and (right) complex 4. Both solution and solid-state structures of 4 indicate hydrogen bonding between the NH group shown in red and solvent or anion residues, which are absent in complex 3 and presumably induce the loss of symmetry.

Initial attempts to discover any N₂O reactivity of complex 4 were not successful, though reactions were observed with an isoelectronic analogue, $N_3^{-}$, and with the N OR inhibitor, I⁻ [75]. In the presence of $N_3^{-}$, complex 4 produces a mixture of [Cu₃(µ₃-S) (dppa)₃](PF₆) and [Cu₃(µ₃- N₃)₂(dppa)₃](PF₆). In the presence of I⁻, complex 4 produces an interesting tricopper complex [Cu₃(µ₃-S)(µ₃-I)(dppa)₃] with mixed µ₃-S²⁻/µ₃-I⁻ bridging, as well as [Cu₃(µ₃-I)₂(dppa)₃](PF₆). Later, we found that 4 could reduce N₂O to N₂ under very specific conditions (Scheme 4) [76]. In the presence of CoCp₂ (2 eq) as an external electron donor, the reaction between 4 and N₂O in MeOH was found to produce a stoichiometric amount of N₂ as confirmed by quantitative headspace GC–MS analysis. Production of H₂O was confirmed by a near-IR spectroscopy assay. A dicopper complex containing two deprotonated dppa ligands was isolated, and the [CoCp₂]PF₆ was also recovered upon the completion of the reaction. Most importantly, a series control experiments have shown that all four components (4, N₂O, H-bond accepting solvent, and CoCp₂) must be present for the reaction to proceed [76].

Scheme 4.

Reaction between complex 4 and N₂O in MeOH in the presence of CoCp₂ (2 eq) as an external electron donor. The scheme was adapted from reference [76].

DFT analysis of the frontier orbitals of complex 4 revealed localized electron density along a Cu-Cu edge (Cu_II-Cu_III in Fig. 8) within its unsymmetrical tetracopper core, whereas the same analysis of 3 indicated even delocalization of electron density throughout the symmetrical tetracopper unit. Thus, we hypothesize that N₂O activation occurs along the electron-rich dicopper cluster edge in 4 that is better able to π-backbond to N₂O compared to inert 3, and critically the ‘‘loading” of electron density along that dicopper edge is induced by hydrogen bonding to the MeOH reaction solvent and the resulting structural distortion. Attempts to probe this hypothesis further using both experimental and computational studies are ongoing in our laboratory. This reaction mimics several aspects of the Cu_Z catalytic site of N₂OR: activity in the 4Cu^I:1S²⁻ redox state, use of a second-sphere proton donor, and reactivity dependence on both primary and secondary sphere effects.

As discussed in Section 5.2, both anionic and/or neutral nitrogen donor ligands have been utilized in synthesizing Cu_Z models, yet none of them reproduce both the structure and function of Cu_Z. We extended the scope of Yam’s synthetic route by utilizing anionic formamidinate ligands (NCN^–) in place of neutral phosphine ligands [77]. The resulting Cu₄S complex 5 was isolated in a 2-hole (2Cu^I:2Cu^II) redox state through use of elemental sulfur as a sulfur source and redox reagent to oxidize the appropriate 2Cu^I precursor (Scheme 5). Representing the first open-shell [Cu₄(µ₄-S)] synthetic compound, complex 5 has an intense purple color corresponding to an absorption at 560 nm (ε = 14000 M^—1cm^—1) with a shoulder at 470 nm, compared to 550 nm for the 4Cu:2S form of Cu_Z in its 2-hole state (Table 1). Although 5 was initially assigned as having a low-lying paramagnetic excited state [77], rigorously purified samples of 5 are strictly diamagnetic across a wide range of temperatures [78].

Scheme 5.

Synthesis of [Cu₄(µ₄-S)(NCN)₄]ⁿ⁻ complexes 5 (n = 0) and 6 (n = 1). Scheme was adapted from reference [68].

Cyclic voltammetry indicated a reversible 1-electron reduction of 5 to the 1-hole (3Cu^I:1Cu^II) redox state at ‒1.28 V vs. Fc⁺/Fc (~−0.7 V vs. SHE) and an irreversible reduction at ~ −2.36 V vs. Fc⁺/Fc (~−1.7 V vs. SHE) to the fully reduced (4Cu^I) state, the latter of which is presumably unstable towards loss of S²⁻ on the electrochemical timescale [77]. Chemical reduction of 5 with [FeCp (CO)₂]⁻, whose potential falls between the reversible and irreversible redox couples of 5, indeed provided 1-hole complex 6 (Scheme 5) [78]. Complex 6 displays an intense blue-purple color with minimal shift in the optical absorption energy (565 nm) but loss of the shoulder feature and a dramatic decrease in extinction coefficient (ε = 8600 M^—1cm^—1). This lack of shifting in the major absorption is distinct from Cu_Z, whose absorption profile is redox state-dependent (Table 1). The most comparable states to 6 of Cu_Z are the 4Cu:1S 1-hole resting state, Cu_z*, and active state, Cu_Z°, which have absorptions at 640 and 680 nm, respectively. The shifts in absorption energies when comparing 6 to 1-hole states of Cu_Z likely reflect differences in electronic properties of neutral histidines vs. anionic formamidinates. The complex hyperfine splitting patterns found in X- and Q-band EPR spectra of 6 were simulated with the single unpaired electron being delocalized equally over the four Cu atoms.

Anionic complex 6 was found to react with N₂O at low temperature (−78 °C), quantitatively producing neutral complex 5. Use of isotopically labeled ¹⁵N₂O confirmed formation of ¹⁵N₂ [78]. Quantitative GC–MS headspace measurements indicated evolution of 0.5 equivalents of N₂ per molecule of complex 6 [79]. Although no oxygen-containing product was detected from the crude reaction mixtures, the product was assumed to be K₂O. Consistent with this proposal, addition of electrophiles Me₃SiCl or PhC(O)Cl to the final product mixture produced (Me₃Si)₂O and [PhC(O)]₂O, respectively [78]. Collecting this information together, a balanced stoichiometric reaction shown in Scheme 6 (top) is evident. The reduction of N₂O to N₂ + O^2– can be considered the aprotic analogue of the biological reaction catalyzed by N₂OR during bacterial denitrification. Consistent with this reaction being a 2-electron redox reaction but complex 6 only being able to mediate reversible 1-electron chemistry, the balanced reaction clearly indicates the consumption of 2 equivalents of [Cu₄S] cluster per molecule of N₂O. Our working hypothesis is that one equivalent of the cluster is responsible for N₂O activation, while the second equivalent acts as a sacrificial reductant given that no external reductant was added. The final copper-containing product of this reaction, complex 5, can be re-reduced to 6, thus completing a closed synthetic cycle for N₂O reduction. However, we have thus far been unable to accomplish N₂O activation and cluster reduction in the same reaction pot.

Scheme 6.

N₂O reduction by a redox-active synthetic [Cu₄(µ₄-S)] model system.

A mechanism for this reaction was proposed on the basis of DFT modeling (Scheme 6, bottom) [79]. First, N₂O binds to 1-hole complex 6. The π-accepting nature of N₂O withdraws electron density from the [Cu₄S] core, making electron transfer from a sacrificial molecule of 6 to the [6·N₂O] adduct spontaneously favorable. The resulting intermediate A mimics the N₂O-bound form of the active 4Cu^I state of Cu_Z (intermediate 2 in Scheme 2). However, a surprising finding in this model study involved the N₂O binding mode. Rather than binding across a dicopper cluster edge as proposed for Cu_Z, the most energetically favorable binding mode was found to be along a Cu-S cluster edge, implying cooperative metal/ligand cooperativity in N₂O activation. Although this binding mode has not been considered for Cu_Z before, one must be cautious about extrapolating results from the model system due to the differences in sulfur geometries (and therefore electronic structures): the sulfur atom in 6 is the apex of a square pyramid, whereas the sulfur atom in Cu_Z is the body center of a distorted tetrahedron. In any case, presumed π-backdonation from the electron-rich 4Cu^I core of A into the N₂O π* manifold induces N–O cleavage and loss of N₂, producing 2-hole intermediate B with a terminal oxo ligand. This intermediate resembles an aprotic form of the meta-stable Cu_Z° state characterized for N₂OR [38]. Indeed, intermediate B was calculated to be unstable relative to rearrangement to intermediate C, in which the oxo ligand occupies a bridging position across a dicopper cluster edge, akin to the off-cycle state Cu_Z*. Given that the final product of the reaction was found experimentally to be O^2–, oxide could dissociate from either intermediate B or C to produce final product 5. An interesting feature of the calculated mechanism is that the coordination number at sulfur is variable by virtue of a Cu atom dissociating from the sulfur center when N₂O is bound, then reassociating when N₂ leaves, then dissociating again when oxide shifts from terminal to bridging.

The involvement of the sulfur atom bridge in N₂O activation and N–O bond cleavage raises the question of sulfur participation in the redox-active molecular orbitals (RAMOs) of complexes 5 and 6. Indeed, a combination of Cu K-edge and S K-edge XAS spectroscopy combined with high-level DFT analysis is consistent with this proposal and indicates a high degree of Cu/S covalency [79]. Notably, the XAS spectroscopic signatures of these model complexes closely match those measured for Cu_Z from Achromobacter cyclastes [54] and Paracoccus denitrificans [10], indicating the fidelity of the model complexes in terms of electronic structure. Experimentally, the RAMOs of complexes 5 and 6 were found to have 20.5 ± 0.1% and 21.1 ± 0.5% S 3p character, respectively. Computational modeling indicated 20.6% and 20.5% S 3p character for the RAMOs of 5 and 6, respectively, compared to 12.7% and 14.5% average Cu 3d character, respectively. Again, these conclusions closely match the prevailing view of Cu_Z, whose RAMO is proposed to have 15– 22% S 3p character [10]. The combined spectroscopic interrogation of 5 and 6 represents the only case where a [Cu₄(µ₄-S)] redox pair could be analyzed side-by-side under identical conditions, therefore adding significantly to the Cu_Z literature.

5.4. Model studies of the reactive dicopper edge site of Cu_Z

Synthetic dicopper species relevant to Cu_Z are motivated by the proposed N₂O binding to the Cu_I-Cu_IV edge of the cluster in its 4Cu:1S form [38], the observed inhibition along that same edge by I⁻ [30], and the observation of an additional µ₂-S²⁻ ligand along that edge in the 4Cu:2S form [42].

Predating the N₂O reactivity of complexes 4 and 6, the first synthetic copper-sulfur complex reported to react with N₂O was studied by Tolman and coworkers [72]. The reaction of [LCu(CH₃CN)]X (L = 1,4,7-trimethyltriazacyclononane, $X=O_{3}SC{F}_3^{-}$ or $Sb{F}_6^{-}$) with Na₂S₂ at room temperature yielded tricopper complex [(LCu)₃(µ-S₂)][X]₂ (7), which was further found to equilibrate with meta-stable dicopper intermediate [(LCu)₂(µ₂-S₂)][X]₂ (8) in solution (Scheme 7). Exposure of 7 to N₂O at low temperature formed N₂ and [L₂Cu₂(OH)₂]X, although the exact stoichiometry of the reaction is unclear. The mechanism for N₂O reduction, proposed based on DFT modeling, involves pre-equilibrium formation of 8, followed by N₂O coordination between the two copper atoms in a µ−1,1 fashion through the oxygen atom. This N₂O-ligated species was found to be a transition state rather than a stable intermediate along the calculated reaction coordinate. The proposal was further supported by the fact that the N₂O reactivity of intermediate 7 was suppressed by excess [LCu(CH₃CN)]X, which presumably disfavors formation of 8. Neither 7 nor 8 is a strict structural mimic of Cu_Z, lacking any S²⁻ ligand and failing to assemble into a tetranuclear cluster. However, the dinuclear N₂O activation can be thought of as modeling the chemistry of the critical Cu_I-Cu_IV cluster edge of Cu_Z.

Scheme 7.

First functional model complex of Cu_Z reported by Tolman and co-workers.

Torelli and coworkers followed an initial report of a mixed-valent dicopper(I,II) m-thiolate complex with saturated coordination spheres [80] by reporting unsaturated derivative 9, which was found to be reactive towards N₂O [81] (Scheme 8). Complex 9 is a delocalized mixed-valent system, with unpaired spin density being delocalized roughly equally throughout the covalent Cu₂(µ-SR) core according to DFT calculations: Cu¹, 27%; Cu², 27%; S, 23%. An optical band assigned to intervalence charge transfer was identified, and EPR spectroscopy was found to be consistent with the delocalized assignment. Although 9 was isolated only in the 1Cu^I:1Cu^II formal redox state, cyclic voltammetry indicated quasi-reversible access to 2Cu^I and 2Cu^II states at ‒0.11 and + 0.42 V (vs. Fc⁺/Fc), respectively. In addition to featuring coordinatively unsaturated metal centers, the metals in 9 each feature labile reactive sites occupied by triflate and water ligands, respectively, thought to be key for enabling N₂O activation. Exposure of 9 to N₂O resulted in evolution of N₂ confirmed by GC–MS analysis, and the copper-containing product crystallized from the product mixture was found to be dicopper(II) µ-hydroxide complex 10 (Scheme 8). A stoichiometry of 0.5 N₂O consumed per molecule of 9 was determined from titration experiments, and an intermediate assigned as an N₂O-bound complex was observed by UV–vis and ¹⁹F NMR spectroscopy and implied by cyclic voltammetry experiments under N₂O atmosphere. Based on computational modeling, this intermediate was tentatively assigned as having a terminally bonded N₂O ligand in place of the H₂O residue in 9. Similar to our proposed mechanism for complex 6 (Scheme 6), the authors propose a mechanism in which one equivalent of 9 serves to activate N₂O, while a second equivalent acts as a sacrificial reductant (and, in this case, also a proton donor). Although the sulfur ligand in 9 is a thiolate rather than a sulfide, nonetheless this system can be thought of as accurately modeling the chemistry of the Cu_I-Cu_IV cluster edge in Cu_Z. Here, the equivalent of dicopper complex 9 responsible for N₂O activation plays the role of the Cu_I-Cu_IV edge, while the equivalent responsible for electron transfer reproduces a possible electron-relay role of Cu_II and Cu_III. The presence of a labile site at one of the Cu centers of 9 mimics the Cu_IV center in Cu_Z, which is bound by a single histidine and features a solvent-derived ligand like the H₂O ligand in 9. A follow-up study from the same group exhibited enhanced N₂O reduction rates obtained through manipulating solvent coordination effects [82].

Scheme 8.

Dicopper µ-thiolate system studied by Torelli.

The only well-characterized dicopper µ-sulfide complex, [(IPr*) Cu]₂(µ-S) (11), was reported posthumously by Hillhouse and coworkers [83]. In that report, the authors devised synthetic procedures for 11, structurally characterized this dicopper(I) complex, and highlighted the potent nucleophilicity of the bridging sulfur atom using reactivity studies with alkyl halide electrophiles. Our group decided to probe the N₂O reactivity of 11 [84]. Exposure of 11 to N₂O atmosphere resulted in slow conversion to dicopper(I) µ-sulfate complex 12, in which the copper centers remained unchanged, but the bridging sulfur atom had been oxidized exhaustively from S²⁻ to $S{O}_4^{2-}$ (Scheme 9). Although no reaction was observed between 11 and PPh₃, exposure of the 11/PPh₃ mixture to N₂O instead produced a roughly equimolar mixture of 12 and a new product tentatively assigned as [(IPr*)Cu]₂(µ-O) (13). An equivalent of S = PPh₃ was produced per molecule 13, with the remaining PPh₃ left unreacted. Here, once again, the copper centers remained unchanged while the bridging sulfur had undergone partial oxidation from S²⁻ to S⁰, the latter being trapped in situ by PPh₃. On the basis of these observations as well as DFT modeling, we proposed that initial N₂O activation results in formation of a [(IPr*)Cu]₂(µ-SO) intermediate that equilibrates between µ-thioperoxide and µ-sulfide/µ-oxide forms. The µ-thioperoxide form presumably undergoes further N₂O oxidation, ultimately leading to 12. On the other hand, the mixed [(IPr*)Cu](µ-S)(µ-O) intermediate could expel elemental sulfur, producing 13. The position of this equilibrium, and thus the 12:13 ratio, would be impacted by the presence of PPh₃ that is able to trap elemental sulfur and drive the equilibrium towards 13. We proposed two lessons to be learned from this study. First of all, regarding the 4Cu:2S form of Cu_Z that features a µ₂-S²⁻ ligand along the Cu_I-Cu_IV edge, we consider it unlikely that the µ₂-S²⁻ ligand would survive N₂O exposure upon cluster reduction, thus providing a possible pathway for converting the 4Cu:2S form to the corresponding 4Cu:1S form through expulsion of sulfur (either by oxidation to sulfate or to S⁰). Second, this study highlights an underappreciated role of the ‘‘spectator” Cu_II and Cu_III centers: anti-oxidation protection of the µ₄-S²⁻ ligand. To highlight the latter point, we noted that tricopper analogue 2, whose bridging sulfur atom is ‘‘protected” by a third Cu^I, did not react with N₂O under the same conditions.

Scheme 9.

N₂O reactivity of Hillhouse’s Cu₂S cluster.

6. Conclusion

This review served to highlight recent advances in crystallographic, enzymological, spectroscopic, and computational interrogations of the Cu_Z site of N₂OR, as well as significant progress that has been made in synthetic model chemistry during the past few years by several groups. Understanding the coordination chemistry and intimate chemical details of Cu_Z aids in rationalizing effects on ecological N₂O fixation of factors such as pH or iodide concentration, which could even be manipulated purposely to favor or disfavor N₂ evolution using soil additives.

In terms of the Cu_Z itself, several mysteries remain, particularly with regard to the different forms, 4Cu:2S vs. 4Cu:1S, of Cu_Z that have been characterized. What is the physiological role of the 4Cu:2S form of Cu_Z, which was characterized through anaerobic isolation that is more relevant to the anaerobic microbes performing N₂O fixation than aerobic procedures? What is the N₂O binding mode, if any, for this form of Cu_Z given the presence of a µ₂-S²⁻ ligand along the critical Cu_I-Cu_IV cluster edge? Is there a mechanism by which proton/electron transfers or substrate interactions convert the 4Cu:2S Cu_Z to the 4Cu:1S form under physiological conditions, as suggested by synthetic model studies described in Section 5? Could the µ₂-S²⁻ ligand somehow be labile to the reaction medium, as has recently been proposed for a ‘‘belt” µ₂-S²⁻ ligand in the FeMo-cofactor of nitrogenase [58–60], to open the reactive site? What is the physiological role of the resting Cu_Z* state of the 4Cu:1S form of the enzyme, and which factors allow N₂OR to avoid this off-cycle intermediate?

In terms of synthetic modeling chemistry, a major void in the literature consists of constructing 4Cu:2S models that could serve to probe some of the above questions. Additionally, given that the native Cu_Z active site mediates a two-electron redox reaction, another prominent goal of synthetic modeling efforts should be to identify a single system capable of stabilizing all three catalytically relevant redox states: 4Cu^I, 3Cu^I:1Cu^II, and 2Cu^I:2Cu^II. Currently for synthetic [Cu₄(µ₄-S)] clusters, soft phosphine donors are known to stabilize the reduced 4Cu^I state but not the more oxidized open-shell states, which require hard nitrogen donors. A mixed hard/soft donor set may enable stabilization of all three redox states in a single system, thus allowing for more extensive spectroscopic and reactivity studies than are currently available. Lastly, a major challenge for synthetic systems to overcome is that of overpotential, as all of the available electrochemical data for biomimetic complexes indicates that they operate at significantly more cathodic potentials than Cu_Z itself.