Nitrosomonas europaea cytochrome P460 is a direct link between nitrification and nitrous oxide emission

Significance

Nitrous oxide (N₂O) is a potent ozone-depleting greenhouse gas. This work identifies a means by which N₂O is generated during nitrification, or biological ammonia oxidation. Fertilizer use in agriculture stimulates nitrification, thus increasing the volume of N₂O emissions worldwide. The results presented herein will inform models and strategies toward optimized, sustainable agriculture. Moreover, these results highlight a rare example of biological N–N bond formation.

Abstract

Ammonia oxidizing bacteria (AOB) are major contributors to the emission of nitrous oxide (N₂O). It has been proposed that N₂O is produced by reduction of NO. Here, we report that the enzyme cytochrome (cyt) P460 from the AOB Nitrosomonas europaea converts hydroxylamine (NH₂OH) quantitatively to N₂O under anaerobic conditions. Previous literature reported that this enzyme oxidizes NH₂OH to nitrite ($\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$) under aerobic conditions. Although we observe $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ formation under aerobic conditions, its concentration is not stoichiometric with the NH₂OH concentration. By contrast, under anaerobic conditions, the enzyme uses 4 oxidizing equivalents (eq) to convert 2 eq of NH₂OH to N₂O. Enzyme kinetics coupled to UV/visible absorption and electron paramagnetic resonance (EPR) spectroscopies support a mechanism in which an Fe^III–NH₂OH adduct of cyt P460 is oxidized to an {FeNO}⁶ unit. This species subsequently undergoes nucleophilic attack by a second equivalent of NH₂OH, forming the N–N bond of N₂O during a bimolecular, rate-determining step. We propose that $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ results when nitric oxide (NO) dissociates from the {FeNO}⁶ intermediate and reacts with dioxygen. Thus, $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ is not a direct product of cyt P460 activity. We hypothesize that the cyt P460 oxidation of NH₂OH contributes to NO and N₂O emissions from nitrifying microorganisms.

Nitrous oxide (N₂O) participates in ozone-layer depletion and possesses a global warming potential nearly 300-fold greater than carbon dioxide (1). Atmospheric N₂O concentrations have increased ∼120% since the preindustrial era, largely due to the widespread use of fertilizers required to produce sustenance for humans and livestock. N₂O is a byproduct of the microbial metabolism of fertilizer components, including ammonia (NH₃) and nitrate ($\mathrm{N}{\mathrm{O}}_{\mathrm{3}}^{\mathrm{-}}$); consequently, agricultural soils account for an estimated 60–75% of global N₂O emissions. The metabolic pathway by which microorganisms oxidize NH₃, nitrification, occurs in two phases, both of which are mediated by autotrophic microorganisms. In the first, NH₃-oxidizing bacteria (AOB) or archaea (AOA) oxidize NH₃ to nitrite (${\text{NO}}_2^{\mathrm{-}}$). In the second, ${\text{NO}}_2^{\mathrm{-}}$ is subsequently oxidized to ${\text{NO}}_3^{\mathrm{-}}$ by ${\text{NO}}_2^{\mathrm{-}}$-oxidizing bacteria. NH₃-oxidizing microbes contribute substantially to global N₂O emissions, whereas ${\text{NO}}_2^{\mathrm{-}}$-oxidizing bacteria produce negligible N₂O (2, 3). AOB are proposed to emit N₂O either as a byproduct of the nitrification pathway or as a product of the nitrifier denitrification pathway (i.e., the reduction of ${\text{NO}}_2^{\mathrm{-}}$) (4–6).

Nitrification of NH₃ to ${\text{NO}}_2^{\mathrm{-}}$ occurs in two steps (7, 8). The first step is catalyzed by NH₃ monooxygenase, which uses copper (Cu) and dioxygen (O₂) to hydroxylate NH₃ to hydroxylamine (NH₂OH) (9). In AOB, the second step is thought to be the four-electron oxidation of NH₂OH to ${\text{NO}}_2^{\mathrm{-}}$ by NH₂OH oxidoreductase (HAO). HAO is a multiheme enzyme with eight c-type hemes per subunit: seven are electron transfer cofactors, and the eighth is the so-called P460 active site that contains a unique tyrosine cross-link to the heme ring. The enzyme (or enzymes) that AOA uses to oxidize NH₂OH is currently unknown (10).

AOB possess machinery for nitrifier denitrification that reduces $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ to N₂O via a nitric oxide (NO) intermediate (11, 12). The archetypal AOB Nitrosomonas europaea possesses genes for a Cu-containing nitrite reductase (NirK) and a membrane-bound, heme-containing NO reductase (NorB). NirK reduces $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ by one electron to NO, whereas NorB catalyzes the two-electron reduction of 2 eq of NO to N₂O. Nitrifier denitrification is thought to lead to the increased production of N₂O and NO by AOB in microaerophilic or anaerobic conditions (13). However, this pathway does not account for the total AOB N₂O emission under aerobic conditions.

Under aerobic conditions, N₂O emission from AOB is proposed to result from the incomplete oxidation of NH₂OH to either nitroxyl (HNO) or NO. Two eq of HNO rapidly react to form N₂O (14), whereas NO is reduced by an NO reductase (12). In support of these hypotheses, both NO and N₂O have been observed to form during steady-state turnover of purified HAO under aerobic conditions (15). Although NorB could facilitate this N₂O production, a NorB knockout strain also produces N₂O at atmospheric O₂ concentration, consistent with the presence of an alternate N₂O-producing pathway (16).

In this study, we demonstrate that there indeed exists a direct enzymatic pathway from NH₂OH to N₂O, and that this pathway is mediated by cytochrome (cyt) P460, a constitutively expressed (17), soluble, periplasmic metalloenzyme originally isolated from N. europaea (18). N. europaea cyt P460 is a 36-kDa homodimeric protein in which each subunit bears a mono-His c-type heme with an N–C cross-link from the 13′ mesocarbon to the amine of Lys70 (19) (Fig. 1A). Such cross-links alter porphyrin π-conjugation. Moreover, the heme P460 macrocycle exhibits significant ruffling, a common distortion mode for c-type hemes (20, 21). P460 hemes are named for the 460-nm Soret band observed in their ferrous state. The P460 center found in N. europaea HAO differs from cyt P460: it is doubly cross-linked by Tyr491 at the 5′ mesocarbon and an adjacent pyrrole α-carbon (Fig. 1B). Despite lacking homology to HAO (19, 22), the presence of a P460 cofactor in cyt P460 has implicated this enzyme in NH₂OH oxidation (23–25).

Fig. 1.

Ferric P460 cofactors in cyt P460 [A, Protein Data Bank (PDB) ID code 2JE2] and HAO (B, PDB ID code 4N4N).

Cyt P460 was previously reported to oxidize NH₂OH to $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ (26), implicating all P460 centers in mediating four-electron oxidation of NH₂OH. However, an octaheme HAO-like protein from the anaerobic NH₃-oxidizing microorganism Kuenenia stuttgartiensis was shown to oxidize NH₂OH to NO instead of $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ (27). An X-ray crystal structure of the K. stuttgartiensis HAO-like enzyme features a tyrosine cross-linked active site and heme distribution that was identical to N. europaea HAO. The K. stuttgartiensis enzyme was proposed to yield NO due to the absence of a nearby tyrosine residue that is present in N. europaea HAO. The N. europaea cyt P460 active site also lacks the nearby tyrosine residue and is far more solvent-exposed (19). Herein, we show that N₂O is the enzymatic product of anaerobic NH₂OH oxidation by N. europaea cyt P460. We demonstrate that this pathway proceeds through an NH₂OH-bound form that is oxidized by three electrons to an Fe–NO species that then reacts with NH₂OH to form N₂O in the rate-determining step.

Results and Discussion

Exclusive Conversion of NH₂OH to N₂O by Cyt P460 Under Anaerobic Conditions.

Consistent with previous work on the Methylococcus capsulatus cyt P460 (26), under aerobic conditions, cyt P460 reacts with NH₂OH in the presence of the oxidant phenazine methosulfate (PMS) to form $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$, as detected by Griess diazotization assays (SI Appendix, Fig. S1A). However, the maximum stoichiometry achieved was 0.7 mol of $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ per mole of NH₂OH (SI Appendix, Table ST1). Gas chromatography (GC) analysis reveals that the remainder of the NH₂OH is converted to N₂O (SI Appendix, Fig. S1A).

Under anaerobic conditions, no $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ is formed when cyt P460 is treated with NH₂OH and PMS. GC analysis (SI Appendix, Fig. S1B) reveals that in the presence of an oxidant ([Ru(NH₃)₆]Cl₃, horse heart cyt c, or PMS), cyt P460 catalyzes the exclusive formation of N₂O from NH₂OH. In the absence of either enzyme or oxidant, only trace N₂O is formed (SI Appendix, Fig. S1). By monitoring the amount of N₂O produced under various NH₂OH or oxidant concentrations (Fig. 2A), a stoichiometry of 2 NH₂OH eq and 4 oxidizing eq producing 1 eq of N₂O was established (Eq. 1):

$$2\hspace{0.17em}\mathrm{N}{\mathrm{H}}_{\mathrm{2}}\mathrm{OH}\to {\mathrm{N}}_{\mathrm{2}}\mathrm{O}\mathrm{+}{\mathrm{H}}_{\mathrm{2}}\mathrm{O}\mathrm{+}\mathrm{4}\hspace{0.17em}{\mathrm{e}}^{\mathrm{-}}+4\hspace{0.17em}{\mathrm{H}}^{\mathrm{+}}.$$ [1]

Fig. 2.

(A) Stoichiometry of N₂O production by cyt P460 determined with GC. Data points are averages of triplicate trials with 5 μM ferric cyt P460 in anaerobic 50 mM Hepes, pH 8.0, at 25 °C overnight. Error bars represent 1 SD of three trials. For the red triangles, the concentration of PMS is held at 1 mM, whereas the NH₂OH concentration is varied; for the blue circles, the NH₂OH concentration is held at 1 mM, whereas PMS concentration varies. (B) Steady-state NH₂OH oxidase activity plot for cyt P460. The assay conditions were 1 μM cyt P460, 6 μM PMS, and 100 μM DCPIP with various NH₂OH concentrations in anaerobic 50 mM Hepes, pH 8.0, at 25 °C. Each data point is the average of three trials, with error bars representing one SD.

Steady-State Kinetics.

Steady-state activity assays of cyt P460 were performed under anaerobic conditions using a 2,6-dichlorophenolindophenol (DCPIP)/PMS coupled assay (23) and an N₂O-selective electrode. By monitoring the decay of the DCPIP absorbance at 605 nm (A₆₀₅), it was determined that 2 eq of DCPIP was reduced per 1 eq of N₂O produced under steady-state conditions (SI Appendix, Fig. S2). Because DCPIP is a two-electron oxidant, this stoichiometry is consistent with Eq. 1. The DCPIP/PMS assay therefore provides a convenient means of measuring steady-state N₂O production by cyt P460.

Steady-state activities of cyt P460 have been reported previously (23, 26), but to our knowledge, no studies have presented steady-state activity plots. Fig. 2B shows the steady-state activity plot exhibiting nonsaturating, linear behavior from 0.05 to 20 mM NH₂OH that spans turnover frequencies of 0.3–80 μM DCPIP consumed min⁻¹⋅µM⋅enzyme⁻¹. Although this nonsaturating behavior precludes determination of k_cat or K_m, the slope of the linear region suggests a k_cat/K_m of 5,000 M⁻¹⋅min⁻¹ (28). However, the characterization of pathway intermediates suggests a multistep reaction mechanism that is inconsistent with classical Michaelis–Menten kinetics (vide infra).

Characterization of Cyt P460 Fe^III–NH₂OH.

Recombinant expression and purification of cyt P460 was previously achieved by Elmore et al. (29). This method yields a British racing green protein with UV/visible absorption and electron paramagnetic resonance (EPR) spectra consistent with those reported for the enzyme isolated from N. europaea (18, 23, 30). The absorption spectrum of the as-isolated cyt P460 has a Soret band at 440 nm with a shoulder at 414 nm and Q-band maxima at 570 nm and 627 nm (Fig. 3). The corresponding EPR spectrum (Fig. 4A) is characteristic of an S = ⁵/₂ Fe^III, with g-values of 6.57, 5.09, and 1.97 (^E/_D = 0.03). The crystal structure of the as-isolated cyt P460 (19, 29) (Fig. 1A) shows phosphate ligated to the Fe center. These crystallization conditions used 2.4 M phosphate buffer; we expect that H₂O will occupy this site under our experimental conditions, and therefore assign the as-isolated protein as an Fe^III–OH₂ heme center.

Fig. 3.

UV/visible absorption spectra of Fe^III–OH₂ cyt P460 (black), Fe^III–NH₂OH cyt P460 (blue), and {FeNO}⁶ cyt P460 generated via treatment with PROLI-NONOate (red line) or oxidation of Fe^III–NH₂OH (black dashed line). (Inset) Magnification of the Q-bands.

Fig. 4.

EPR spectra of species on the proposed cyt P460 NH₂OH oxidase pathway. Cyt P460 at 170 μM (A) was treated with 100 mM NH₂OH (B), with 2 mM NH₂OH and 2 mM DCPIP (C), or with 45 mM NH₂OH and 2 mM DCPIP (D) and incubated for 10 min. Black traces are spectra simulated with the parameters listed in SI Appendix, Table ST2. Spectra were collected at 10 K and 633 μW or at 20 K and 63 μW. A 5% impurity of an {FeNO}⁷ species is indicated by a single asterisk. An Mn²⁺ EPR signal is indicated by double asterisks. dχ″/dH, derivative of magnetic susceptibility vs. magnetic induction.

The addition of NH₂OH to cyt P460 produces a species with a UV/visible absorption spectrum distinct from Fe^III–OH₂ (Fig. 3). Within the time of manual mixing, the 414-nm shoulder of the Fe^III–OH₂ spectrum disappears concomitant with a shift in the Soret band to 445 nm and broadening of the Q-bands. The simplest interpretation is that NH₂OH substitutes for H₂O at the Fe site (i.e., Fe^III–NH₂OH). Cyt P460 was titrated with NH₂OH, and the resulting series of spectra (SI Appendix, Fig. S3A) show an isosbestic point at 438 nm, which suggests a one-step conversion from Fe^III–OH₂ to the putative Fe^III–NH₂OH. To determine the NH₂OH dissociation constant [$K_{\text{d}\left({\text{NH}}_2\text{OH}\right)}$], the absorption at 414 nm (A₄₁₄) was plotted against NH₂OH concentration. Fitting the data to a hyperbolic binding curve resulted in a $K_{\text{d}\left(N{H}_{2}OH\right)}$ of 9 ± 1 mM (SI Appendix, Fig. S3A, Inset).

Treatment of cyt P460 with 100 mM NH₂OH results in the disappearance of the S = ⁵/₂ Fe^III–OH₂ signal and the appearance of two rhombic S = ¹/₂ EPR signals (Fig. 4B). The first signal has g-values of 2.75, 2.28, and 1.54, and is consistent with a low-spin Fe^III. The EPR spectra of cyt P460 treated with increasing concentrations of NH₂OH (SI Appendix, Fig. S3B) corroborate the $K_{\text{d}\left({\text{NH}}_2\text{OH}\right)}$ value determined with UV-vis spectrometry. The conversion from a high-spin to low-spin Fe^III is consistent with the binding of NH₂OH to the Fe center; thus, we assign this S = ¹/₂ species as a rare example of a stable heme Fe^III–NH₂OH complex (31).

The second rhombic S = ¹/₂ signal has g-values of 2.10, 2.02, and 2.01, and ¹⁴N hyperfine coupling values of 50 MHz, 57 MHz, and 45 MHz. This second signal represents less than 5% of the total spin as determined by spin quantitation and is consistent with a five-coordinate heme Fe^II–NO species (32), or {FeNO}⁷ in Enemark–Feltham notation (33). This {FeNO}⁷ species can be independently generated by treating cyt P460 with the HNO donor, disodium diazen-1-ium-1,2,2 triolate (Na₂N₂O₃, Angeli’s salt). This species is stable in the presence of NH₂OH or oxidant (SI Appendix, Fig. S4). The low yield observed in the EPR spectrum and the lack of reactivity suggest that the {FeNO}⁷ is an off-path product and does not contribute to the productive N₂O-generating cyt P460 pathway. At present, we are unsure why this oxidized species appears in the absence of O₂, oxidant, or NO. One possibility is that the samples are exposed to a small amount of O₂ during freezing, which would oxidize Fe^III–NH₂OH to {FeNO}⁷.

Characterization of an Intermediate in NH₂OH Oxidation.

The addition of the oxidant [Ru(NH₃)₆]Cl₃ under anaerobic conditions results in formation of a new species. Over the course of the reaction (ca. 20 min), there are two distinct phases: an accumulation phase and a decay phase. In the accumulation phase, a new species appears within 2 min. The absorption spectrum of the new species exhibits an intense and sharp Soret band centered at 455 nm and Q-bands at 554 nm, 603 nm, and 652 nm (SI Appendix, Fig. S5A). Several isosbestic points are observed during the spectral time course, indicating a single-step conversion from Fe^III–NH₂OH to the new species. In the decay phase, the new species is converted back to Fe^III–NH₂OH. The spectral time course of this phase also features several isosbestic points (SI Appendix, Fig. S5B). These data strongly suggest the accumulation and decay of an intermediate on the cyt P460 pathway, hereafter referred to as the 455-nm intermediate owing to its Soret band maximum.

In the presence of DCPIP, the formation of the 455-nm intermediate is complete at the time of mixing; the intermediate persists for 1 min, until DCPIP is completely consumed, at which point the A₄₅₅ slowly decreases (SI Appendix, Fig. S6). Subsequent addition of 10 mM NH₂OH hastened the decay of the 455-nm intermediate (SI Appendix, Fig. S6C). The spectral time course of the decay phase features an isosbestic point at 445 nm, consistent with the direct conversion of the 455-nm intermediate to a single species. Comparison of the final decay product spectrum to the decay product of as-isolated cyt P460 mixed with 10 mM NH₂OH confirms that the decay product is Fe^III–NH₂OH (SI Appendix, Fig. S6B).

The persistence of the 455-nm intermediate in the presence of excess oxidant suggests that, under turnover conditions, the decay of the 455-nm intermediate is the rate-determining step. The increase in decay rate at higher NH₂OH concentrations implies a bimolecular reaction between the 455-nm intermediate and NH₂OH. To confirm this relationship, the formation and decay of the 455-nm intermediate at various NH₂OH concentrations were monitored (SI Appendix, Fig. S8A). The sum of two exponentials (SI Appendix, Eq. S1) was fit to the A₄₅₅ traces, providing observed rate constants (k_obs) for both the formation [k_obs(1)] and decay [k_obs(2)] of the 455-nm intermediate. At higher NH₂OH concentrations, k_obs(1) was too fast to fit accurately, but, qualitatively, it increases with increasing NH₂OH concentration. The k_obs(2) parameter showed a linear dependence on NH₂OH, consistent with a bimolecular reaction of NH₂OH and the 455-nm intermediate. A linear fit to a plot of k_obs versus NH₂OH concentration provided a second-order rate constant of 0.07 mM⁻¹⋅min⁻¹ (SI Appendix, Fig. S8B).

The 455-nm intermediate was trapped for EPR characterization by freezing the reaction of 170 μM cyt P460 with 2 mM NH₂OH and 2 mM DCPIP within 1 min. The EPR spectrum of the trapped 455-nm intermediate lacks the signals associated with Fe^III–OH₂ and Fe^III–NH₂OH, suggesting that the protein is completely converted to the 455-nm intermediate or to another EPR-silent species (Fig. 4C). The only EPR signal observed is an S = ¹/₂ signal with hyperfine structure that can be attributed to Mn²⁺ contamination in the protein sample. The lack of an Fe-based EPR signal implies that the 455-nm intermediate is either a diamagnetic or non-Kramer’s (integer spin greater than 0) species.

The decay product was prepared by treating 150 μM cyt P460 with 45 mM NH₂OH and 2 mM DCPIP, followed by incubation for 10 min at room temperature. The EPR spectrum of this sample shows signals for Fe^III–NH₂OH and {FeNO}⁷ (Fig. 4D). The combined spins account for the total Fe concentration in the sample, indicating the quantitative conversion of the 455-nm intermediate to either Fe^III–NH₂OH (80 μM) or {FeNO}⁷ (70 μM), following the depletion of oxidant.

Identification of the 455-nm Intermediate as an {FeNO}⁶.

The need for oxidant to form the 455-nm intermediate suggests that it is an oxidized form of Fe^III–NH₂OH. Bari and coworkers (34) have postulated several oxidized species in their proposed mechanism of NH₂OH oxidation by HAO, including an {FeNO}⁷, an {FeNO}⁶, and a ferrous-nitrous acid (Fe^II–ONOH) species. Other possibilities include a ferric hydroxylamine radical (Fe^III–^•NH₂OH), which has been proposed as an intermediate in the enzymatic pathway of P450 NO reductase (35, 36) and Fe^II–HNO, which has been characterized in only one biological system, the myoglobin–HNO complex (37, 38). Finally, because N₂O is the product of NH₂OH oxidation by cyt P460 under our experimental conditions, an Fe–N₂O complex is also possible. We propose that our experimental data are most consistent with assignment of the 455-nm intermediate as an {FeNO}⁶ species (vide infra).

The number of oxidizing equivalents required to convert Fe^III–NH₂OH to the 455-nm intermediate was determined by treating cyt P460 with substoichiometric NH₂OH and excess cyt c (SI Appendix, Fig. S9). The reduction of cyt c was monitored by the increase in absorption at 550 nm (ε₅₅₀ = 19,600 M⁻¹⋅cm⁻¹ (39). These experiments show that 3 electrons are required to oxidize Fe^III–NH₂OH to the 455-nm intermediate, suggesting an {FeNO}⁶ or Fe^II–ONOH species (33). Consistent with this observation, treatment of Fe^III–OH₂ with NO supplied either as the gas or via the NO donor PROLI-NONOate [1-(hydroxy-NNO-azoxy)-l-proline] produces an EPR-silent species with an absorption spectrum identical to the absorption spectrum of the 455-nm intermediate (Fig. 3). The lack of an EPR signal precludes assignment of the 455-nm intermediate as an {FeNO}⁷, because these species typically exhibit S = ¹/₂ EPR signals (32). The 455-nm intermediate generated in this way persists for ca. 1 h and is stable to excess DCPIP (SI Appendix, Fig. S10). This lack of reactivity is inconsistent with the hypothesis of Bari and coworkers (34) that the hydrolysis of an {Fe–NO}⁶ generates an Fe^II–ONOH adduct that is poised for proton-coupled, one-electron oxidation to ${\text{NO}}_2^{\mathrm{-}}$ because cyt P460 does not produce ${\text{NO}}_2^{\mathrm{-}}$ under anaerobic conditions. Taken together, these data support the assignment of the 455-nm intermediate as an {FeNO}⁶.

The stability of the {FeNO}⁶ intermediate in the absence of NH₂OH and the NH₂OH dependence of its decay suggests that {FeNO}⁶ reacts with NH₂OH to form N₂O. A so-called shunted {FeNO}⁶ species was prepared by adding 2 eq of NO supplied via PROLI-NONOate to cyt P460. The addition of 2 mM NH₂OH results in the decay of {FeNO}⁶ to Fe^III–NH₂OH (SI Appendix, Fig. S11). Consumption of the shunted {FeNO}⁶ exhibits the same NH₂OH concentration dependence, 0.07 ± .01 mM⁻¹⋅min⁻¹, as the {FeNO}⁶ that accumulates under turnover conditions.

To confirm the production of N₂O in the above reaction, cyt P460 was treated with varying concentrations of NO and excess NH₂OH. The amount to N₂O produced was monitored by GC/mass spectrometry (MS). There is a clear 1:1 stoichiometry of N₂O produced versus NO added (SI Appendix, Fig. S12A). This result is consistent with NO binding to cyt P460 to form the {FeNO}⁶ intermediate and subsequent reaction with NH₂OH to form N₂O. The same experiment was performed with NO and isotopically labeled ¹⁵NH₂OH. The mass shift from 44 atomic mass units (amu) to 45 amu (SI Appendix, Fig. S12B) in the presence of cyt P460 clearly demonstrates that NO is coupled to NH₂OH via cyt P460. The reaction of NO with NH₂OH in the absence of cyt P460 does not result in N₂O production.

Mechanism for N₂O Formation from NH₂OH by Cyt P460.

Given the above results, we propose the mechanism described in Fig. 5 for the NH₂OH oxidase activity of cyt P460. The catalytic cycle initiates from the S = ⁵/₂ Fe^III–OH₂, which binds NH₂OH to form an S = ¹/₂ Fe^III–NH₂OH species that is stable in the absence of oxidant. In the presence of oxidant, Fe^III–NH₂OH is rapidly oxidized by 3 electrons to the EPR-silent {FeNO}⁶, which undergoes nucleophilic attack by a second equivalent of NH₂OH to yield N₂O and H₂O. We propose that the Fe-containing product of this reaction is an Fe^II species that is rapidly oxidized and converted to the starting Fe^III–OH₂ species.

Fig. 5.

Proposed Cyt P460 NH₂OH oxidase mechanism. RDS, rate-determining step.

Several other intermediates can be envisioned in the conversion of Fe^III–OH₂ to {FeNO}⁶. One-electron (K₃[Fe(CN)₆], Cu^II-azurin, and [Ru(NH₃)₆]Cl₃) and two-electron (DCPIP and PMS) oxidants can access {FeNO}⁶, which suggests that rapid, subsequent one-electron oxidation steps occur in the conversion of Fe^III–NH₂OH to {FeNO}⁶. Although {FeNO}⁷ appeared under certain conditions (i.e., after complete consumption of oxidant), it is not a catalytically competent intermediate. The EPR spectrum of this species is consistent with a five-coordinate {FeNO}⁷ (32); however, a six-coordinate {FeNO}⁷ could be on the reaction pathway. Other possible one-electron oxidized intermediates include Fe^III–^•NH₂OH, Fe^III–^•NHOH, or {FeN(H)O}^8″ (40). To date, we have found no evidence to exclude any of these possibilities, and the characterization of these possible intermediates is the subject of future investigations in our laboratory.

The rate-determining step of the anaerobic oxidation of NH₂OH by cyt P460 is the bimolecular reaction of NH₂OH with {FeNO}⁶ that results in N₂O formation. The intensities of the UV/visible absorption spectral features of Fe^III–OH₂ treated with NO are identical to the intensities observed for the {FeNO}⁶ intermediate formed using NH₂OH and oxidant (Fig. 3). Combined with the lack of Fe-based EPR signals in these samples, these data suggest that under both conditions, all Fe sites are quantitatively converted to the {FeNO}⁶ species. This quantitative accumulation is consistent with the nucleophilic attack by NH₂OH being the rate-determining step of the catalytic cycle. There is precedent for NH₂OH reacting with [Fe(CN)₅(NO)]²⁻, a classic example of an {FeNO}⁶, to form N₂O with stoichiometry matching the stoichiometry above for cyt P460 (41).

We propose that NH₂OH reacts with {FeNO}⁶ to form an Fe^II species, which is rapidly oxidized to return the enzyme to Fe^III–OH₂. This proposal is based on the precedent that [Fe(CN)₅NO]²⁻ reacts with NH₂OH to form [Fe(CN)₅(OH₂)]³⁻. However, the characteristic 463-nm UV/visible absorption peak of Fe^II cyt P460 was never observed in our experiments as an intermediate or an end product. The Fe^II cyt P460 is expected to react rapidly with other species present during turnover (oxidant, NH₂OH, or NO), thereby precluding its observation (vide infra). The reaction of Fe^II cyt P460 with NO produces the same inactive {FeNO}⁷ species observed when Fe^III–OH₂ is treated with HNO (SI Appendix, Fig. S4). Fe^II cyt P460 reacts with NH₂OH to form Fe^III–OH₂ via an intermediate with a UV/visible absorption feature at 663 nm that has not been observed in any of our other experiments (SI Appendix, Fig. S13). This 663-nm intermediate forms and disappears within 3 s; assignment of this 663-nm intermediate will require rapid-mixing techniques. Critically, the reaction of Fe^II cyt P460 with NH₂OH does not inactivate the enzyme.

Stopped-flow UV/visible absorption spectroscopy provided insight into why inactivation of the enzyme by NO is avoided during turnover (SI Appendix, Fig. S14). Fe^II cyt P460 is quantitatively oxidized to Fe^III–OH₂ by [Ru(NH₃)₆]Cl₃ within the time of mixing (<3 ms), placing a lower limit on k_obs of 1,100 s⁻¹. The reactions with NH₂OH and NO exhibit k_obs of 1.3 s⁻¹ and 2.8 s⁻¹, respectively. These results show that oxidation of Fe^II cyt P460 to Fe^III–OH₂ is much faster than the reaction with NH₂OH or NO, thereby avoiding the production of inactive {FeNO}⁷. After the oxidant is completely consumed, Fe^II cyt P460 will react with either NH₂OH or NO. The reaction of Fe^II cyt P460 with NH₂OH forms Fe^III–OH₂ or, in the case of high NH₂OH concentration, Fe^III–NH₂OH, whereas the reaction of Fe^II cyt P460 with NO results in the formation of inactive {FeNO}⁷. Therefore, these reactions with Fe^II cyt P460 likely account for the products observed in the EPR spectrum of the {FeNO}⁶ intermediate decay sample (Fig. 5D).

The King–Altman method (42) was used to derive a steady-state equation based on the minimal mechanism shown in Fig. 5. In this simplified model, NH₂OH binding is treated as reversible, with association (k₁) and dissociation (k₋₁) rate constants. Both the Fe^III–NH₂OH oxidation (k₂) and the subsequent reaction of {FeNO}⁶ with NH₂OH (k₃) are assumed to be irreversible. The resulting model (Eqs. 2–4) indicates that k₂[ox] influences both k_cat and K_m:

$$\frac{velocity}{{[\text{cyt}\hspace{0.17em}\text{P}460]}_0}=\frac{k_2\left[\text{ox}\right]{\left[{\text{NH}}_2\text{OH}\right]}_0}{K_{\text{m}}+{\left[{\text{NH}}_2\text{OH}\right]}_0},$$ [2]

$$K_{\text{m}}=\frac{k_{-1}{k}_3+k_2\left[\text{ox}\right](k_1+\hspace{0.17em}{k}_3)\hspace{0.17em}}{k_1{k}_3},$$ [3]

$$k_{\text{cat}}=k_2\left[\text{ox}\right].$$ [4]

There are two limiting regimes of the derived steady-state model in which either k₋₁k₃ or k₂[ox](k₁ + k₃) dominates the numerator of the K_m term. In the former, K_m should resemble $K_{\text{d}\left(N{H}_{2}OH\right)}$, which was determined to be 9 mM. The steady-state activity plot lacks any clear curvature despite the inclusion of activity measurements in up to 20 mM NH₂OH. This lack of curvature suggests that the second regime, which is highly dependent on k₂[ox], contributes substantially under the assay conditions, thereby increasing the observed K_m. Currently, we have an estimate of only k₃ (0.07 mM^–1⋅min^–1). Given the assumptions made, we attribute the absence of saturation in the steady-state activity plot to a large K_m, which suggests a large second-order rate constant for Fe^III–NH₂OH oxidation.

N. europaea tolerates at least 10 mM NH₂OH without showing inhibited O₂ uptake (43). Under the steady-state conditions studied, we estimate the rate of N₂O production at 10 mM NH₂OH to be 23 μM N₂O min⁻¹. The properties of the oxidant heavily influence k₂ and, by extension, the rate at which N₂O is produced by cyt P460. Thus, lacking the identity of the native electron transfer partner, we cannot verify that this rate is the physiological rate.

The observation that NO binding accesses a shunt in the cyt P460 catalytic pathway provides evidence for the hypothesis that cyt P460 contributes to NO detoxification in the cell (11). Furthermore, if NO binding is reversible, there may be an alternate cyt P460 NH₂OH oxidation pathway that results in NO as the product. At low NH₂OH concentrations, the bimolecular reaction with {FeNO}⁶ should be slow and NO dissociation may outpace N₂O formation. This alternate NO-forming pathway could also be responsible for the observation of $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ formation under aerobic conditions, as NO reacts with O₂ in aqueous solution to form $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$. To test this hypothesis, {FeNO}⁶ was generated by treating Fe^III–OH₂ with 1 eq of NO generated from PROLI-NONOate. Exposure of the resulting {FeNO}⁶ species to O₂ results in the return to Fe^III–OH₂ and generation of ${\text{NO}}_2^{\mathrm{-}}$ (SI Appendix, Fig. S15). We therefore propose that the ${\text{NO}}_2^{\mathrm{-}}$ observed as an aerobic product is not directly formed by cyt P460, but rather is a byproduct resulting from NO dissociation from the {FeNO}⁶ intermediate. A detailed kinetic analysis will determine the partitioning between the ${\text{NO}}_2^{\mathrm{-}}$- and N₂O-forming pathways.

Cell-free extracts of N. europaea were previously shown to oxidize NH₂OH to N₂O and NO without formation of NO₂⁻ (44). Furthermore, purified HAO was shown to react with NH₂OH and PMS or DCPIP under aerobic conditions to form a mixture of products, including NO, N₂O, $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$, and $\mathrm{N}{\mathrm{O}}_{\mathrm{3}}^{\mathrm{-}}$ (15). Additionally, a stable {FeNO}⁶ species on the HAO P460 cofactor was observed after Fe^III–OH₂ was allowed to react with NO in the absence of NH₂OH (25). Given this evidence, our results, and the recent report that an HAO-like protein oxidizes NH₂OH to NO (27), we suggest that the biochemistry of HAO be revisited to determine if $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ is indeed its terminal, enzymatic product.

Outlook: Environmental Consequences.

AOB are major contributors to N₂O emissions from wastewater treatment plants (WWTPs), at which N. europaea is the dominant AOB species (45). There are two proposed methods for N₂O emission from AOB: The first is as a product in the nitrifier denitrification pathway, and the second is as a byproduct in incomplete NH₂OH oxidation. The results of our study demonstrate that the constitutively expressed cyt P460 is a direct link between NH₂OH oxidation and the emission of N₂O from N. europaea, thus establishing an alternative oxidative pathway to N₂O. We have established through GC and kinetic analysis a strict stoichiometry of 2 eq of NH₂OH and 4 oxidizing equivalents to produce 1 eq of N₂O. Due to the reactivity of NH₂OH with biological electrophiles, N. europaea likely has a detoxifying role at high NH₂OH concentrations. Previously, cyt P460 was thought to oxidize NH₂OH and NO to $\mathrm{N}{\mathrm{O}}_{\mathrm{2}}^{\mathrm{-}}$ (46). However, we suggest herein an alternative role in NH₂OH detoxification through the production of N₂O (vide supra) (11).

We have demonstrated the production of N₂O from NH₂OH oxidation by cyt P460 under anaerobic conditions, thereby establishing a direct enzymatic link between nitrification and N₂O formation via NH₂OH. The identification of the source of N₂O emission helps to explain the production of N₂O in WWTPs under conditions of low dissolved O₂ concentrations and high NH₃ concentrations, conditions in which the nitrifier denitrification pathway would not dominate (47). The influx of high concentrations of NH₃ in WWTPs causes an increase in NH₃ oxidation rates, which, in turn, increases intracellular NH₂OH concentration (4). Studies using activated sludge from WWTPs with high concentrations of NH₃ have shown that transitioning from an aerobic environment to an anoxic environment increases the amount of N₂O released. Isotopic labeling studies show that the N₂O produced under these conditions originates from the NH₂OH oxidation pathway rather than the nitrifier denitrification pathway (48). Our study pinpoints direct N₂O production via an enzymatic, anaerobic NH₂OH oxidation mechanism. Identifying the chemical source of this emission should aid in the design and operation of WWTPs with curtailed N₂O emission.

Experimental Procedures

Materials.

Cyt P460 was prepared as previously described (29). PROLI-NONOate and Na₂N₂O₃ (Angeli’s salt) were purchased from Cayman Chemicals. DCPIP was purchased from Alfa Aesar, PMS from Bean Town Chemical, and NH₂OH⋅HCl from TCI Chemicals.

Spectroscopy.

X-band (9.40-GHz) EPR spectra were obtained using a Bruker Elexsys-II spectrometer equipped with a liquid He cryostat maintained at 10 or 20 K. Temperatures and microwave powers are listed in the figure legends. UV/visible absorption spectra were obtained using a Cary 60 UV/visible absorption spectrometer.

N₂O Quantification.

All reactions were prepared and sealed in 5-mL headspace GC vials (Wheaton). The final N₂O concentration was analyzed with GC (Agilent), GC/MS (GC-MATE II; JEOL), or an N₂O microsensor housed within a septum-piercing needle (Unisense). For GC experiments, the headspace was measured with GC analysis using a Supel-Q PLOT (30 mm × 0.32 mm) or an RT Q-bond column. For N₂O microsensor measurement, the needle probe was inserted through the septum and into the solution. Calibration standards for all experiments were made either by diluting an N₂O-saturated solution into water or by decomposing Na₂N₂O₃ in Hepes buffer, pH 8.0, in a sealed headspace vial.