Bacterial NOR (nitric oxide reductase) is a major source of the powerful greenhouse gas N2O. NorBC from Paracoccus denitrificans is a heterodimeric multi-haem transmembrane complex. The active site, in NorB, comprises high-spin haem b3 in close proximity with non-haem iron, FeB. In oxidized NorBC, the active site is EPR-silent owing to exchange coupling between FeIII haem b3 and FeBIII (both S=5/2). On the basis of resonance Raman studies [Moënne-Loccoz, Richter, Huang, Wasser, Ghiladi, Karlin and de Vries (2000) J. Am. Chem. Soc. 122, 9344–9345], it has been assumed that the coupling is mediated by an oxo-bridge and subsequent studies have been interpreted on the basis of this model. In the present study we report a VFVT (variable-field variable-temperature) MCD (magnetic circular dichroism) study that determines an isotropic value of J=−1.7 cm−1 for the coupling. This is two orders of magnitude smaller than that encountered for oxo-bridged diferric systems, thus ruling out this configuration. Instead, it is proposed that weak coupling is mediated by a conserved glutamate residue.

INTRODUCTION

Bacterial denitrification is a fundamental component of the global nitrogen cycle and involves the reduction of nitrate (NO3) to dinitrogen (N2) in four steps:

 
formula

Use of nitrate fertilizer stimulates soil denitrification and the emission into the atmosphere of nitrous oxide (N2O), a greenhouse gas 300 times more potent than CO2, with a half-life of >100 years [1]. N2O has now become the most serious ozone-depleting emission [2]. NOR (nitric oxide reductase), a haem-containing enzyme in the cytoplasmic membrane of denitrifying bacteria [3] that is responsible for the majority of N2O production, catalyses the one-electron reduction of nitric oxide:

 
formula
(1)

NOR is also used by the opportunistic pathogen Pseudomonas aeruginosa, in the lung of cystic fibrosis patients, to detoxify macrophage-produced NO [4]. A suggestion, on the basis of primary sequence analyses [5], that NOR is structurally related to HCOs (haem-copper oxidases) was supported by spectroscopic studies [610] and by the crystal structure of NorBC from Ps. aeruginosa [11]. NorBC contains two low-spin haems and a haem/non-haem iron active site, arranged within two protein subunits as illustrated in Figure 1. FeB takes the place of the copper ion found at the active site of HCOs. As with HCOs, the active site of oxidized NOR is EPR-silent and it is assumed that haem b3 and FeB are spin-coupled via a bridging ligand (X in Figure 1) [6,8,12,13].

Subunit arrangement and metal cofactor content of NorBC in the 595 form

Figure 1
Subunit arrangement and metal cofactor content of NorBC in the 595 form

All iron atoms are in the FeIII oxidation state.

Figure 1
Subunit arrangement and metal cofactor content of NorBC in the 595 form

All iron atoms are in the FeIII oxidation state.

A bridging oxygen ligand (H2O, OH or O2−) was identified in the Pseudomonas NorBC structure, which also supports suggestions that a conserved glutamate residue could provide an additional ligand to FeB [8,13,14]. However, absorption spectra of the crystals showed that substantial reduction occurs upon exposure to the X-ray beam [11]: the structure provides an invaluable framework to which other data can be referenced, but is not that of fully oxidized NorBC. Spectroscopic studies are therefore essential to understand how metal cofactor configuration alters with redox state. FeIII haem b3 in oxidized NorBC from Paracoccus denitrificans gives rise to an electronic CT (charge-transfer) transition characteristic of high-spin FeIII haem, but at the unusually short wavelength of 595 nm [6,8]. Resonance Raman showed that haem b3 is five-co-ordinate: since lack of EPR signals necessitates distal ligand bridging to FeBIII, it was concluded that the proximal histidine residue is dissociated from haem b3. Sensitivity of Raman bands to isotopic oxygen substitution was taken to indicate that X is a bridging oxygen ligand and oxo was favoured over hydroxo [7,9,10].

Many di-iron complexes, reported as models for non-haem iron proteins, have been magnetically characterized. Ligand-mediated coupling between high-spin FeIII ions results in a negative isotropic exchange coupling (J) that is significantly larger for μ-oxo than for μ-hydroxo (reviewed by Solomon et al. [15]). All of the J-values discussed in the present study are referenced to the convention. Isotropic J-values of −154 cm−1 and −216 cm−1 identified oxo-bridged diferrric sites in oxy-haemrythrin [16] and RNR (ribonucleotide reductase) from Escherichia coli [17]. In contrast, FeIII–FeIII coupling in MMO (methane mono-oxygenase) from Methylosinus trichosporium of −J=15 cm−1 indicated μ-hydroxo [18]. As the bridge is responsible for mediating the coupling between haem b3 and FeB, an accurate determination of J should aid identification of this ligand in NorBC. Although the haem b3–FeB interaction has been described as antiferromagnetic [7,9], the nature of the coupling is unknown, beyond that it precludes observation of the EPR signals at the X-band. In principle, the coupling can be determined by analysing magnetic field- and temperature-dependent variations in the MCD (magnetic circular dichroism) intensity of the 595 nm band [19]. Contributions from other forms, and from the other haems, can be minimized or eliminated by working at this wavelength. Residual contributions from haems b and c can be removed entirely by careful choice of measurement conditions that exploit the fact that S=1/2 centres have a relatively simple magnetic field and temperature dependence. In the present paper we describe how we have used this approach to determine, for the first time, the magnitude of the spin coupling between FeIII haem b3 and FeBIII in the 595-form of NorBC from Paracoccus denitrificans.

EXPERIMENTAL

NorBC was purified as described from the P. denitrificans strain 93.11 [20]. Spectroscopic samples were prepared as buffered solutions in 2H2O [21]. All solutions contained 0.05% dodecyl-maltoside and 20 mM BTP {Bis-Tris propane; 1,3-bis[tris(hydroxymethyl)methylamino]propane} at pH*=7.4. pH* is the apparent pH of the 2H2O solutions measured using a standard glass pH electrode. For low-temperature spectra, an equivalent volume of glycerol was added as a glassing agent. MCD spectra were recorded on JASCO circular dichrograph models J810 and J730 used in conjunction with an Oxford Instruments SM4 split-coil superconducting solenoid capable of generating fields of up to 5 T. Intensities of spectra presented are referred to concentrations on the basis of an extinction coefficient of 300 mM−1·cm−1 for the Soret absorption band near 410 nm [12].

RESULTS

Electronic absorption

Figure 2 (solid trace) is the absorption spectrum of P. denitrificans NorBC showing the distinctive 595 nm band. The 300–580 nm region contains overlapping bands from all three haems. For high-spin FeIII haems, two CT transitions (CT1 and CT2) occur at wavelengths longer than 580 nm [21]. Both could potentially yield information on the magnetic properties of haem b3 and therefore on the nature of the haem b3–FeB interaction. CT1, at 800–1300 nm, is weak and rarely detected by absorption, but CT2 is readily observed at 580–660 nm. Preparations of oxidized P. denitrificans NorBC exhibit an electronic transition at 595 nm [6,8,22], but we note that this band shows minor variations in intensity. Preparations with low 595 nm absorption intensity instead give rise to weaker bands at longer wavelengths [23] and it needs to be ensured that the presence of these bands does not compromise analysis of the 595 nm transition. The broken trace in Figure 2 shows the spectrum of a preparation containing anomalously low 595 nm intensity, but weak intensity at ~600–660 nm. The difference spectrum (Figure 2, inset) shows the absorption to be significantly red-shifted from the 595 nm transition and centred at ~622 nm. This shift in wavelength along with differences in magnetic properties (discussed below) ensures that low levels of these other forms will not adversely affect analyses of the MCD at 595 nm.

UV–visible region electronic absorption spectra of NorBC in the 595 and the 622 forms

Figure 2
UV–visible region electronic absorption spectra of NorBC in the 595 and the 622 forms

Spectra of the 595 form (continuous line) and the 622 form (broken line). These spectra are plotted using ∊=300 mM−1·cm−1 for the Soret band at ~410 nm [12].

Figure 2
UV–visible region electronic absorption spectra of NorBC in the 595 and the 622 forms

Spectra of the 595 form (continuous line) and the 622 form (broken line). These spectra are plotted using ∊=300 mM−1·cm−1 for the Soret band at ~410 nm [12].

VFVT (variable-field variable-temperature) MCD

For a sample of predominantly 595 form NorBC {~90% as judged by RT-MCD (room temperature MCD) [23]}, visible-region MCD spectra were recorded at 18 temperatures between 1.7 and 220 K, using a field of 5 T, to give the dataset Δ∊VT (Figure 3A). At these temperatures, MCD bands of low-spin FeIII haems b and c are 1–2 orders of magnitude more intense than those of high-spin haem b3 and dominate the spectra [21]. The nIR (near IR) region, Figure 3(B), contains two low-spin haem CT bands at wavelengths characteristic of the His/His (1550 nm) and His/Met (1900 nm) ligation of haem b and haem c respectively. These bands match those reported in the RT-MCD [8]: there is no evidence for a haem b3 spin-state change at low temperature as observed for the Pseudomonas stutzeri enzyme [12]. The only exception to the dominance of low-spin bands occurs near 595 nm, where the bandshape changes with temperature, suggesting the presence of transitions from the 595 form active site (expanded in Figure 3C). The absence of low-temperature MCD bands at 610–630 nm (Figure 3C) suggests that, in the 622 form, haem b3 and FeB are strongly coupled to yield a diamagnetic active site and will not therefore compromise analysis of the 595 nm bands.

VFVT MCD spectra of P. denitrificans NorBC

Figure 3
VFVT MCD spectra of P. denitrificans NorBC

(A) VT-MCD spectra in the visible region. Spectra were recorded using a magnetic field of 5 T and temperatures of 1.73, 2.25, 2.95, 3.86, 4.22, 6.10, 7.90, 10.0, 15.0, 20.0, 29.6, 44.6, 64.5, 85.0, 110, 151, 182 and 220 K. All spectroscopic features increase in intensity with decreasing temperature. (B) VT-MCD spectra in the nIR region. Spectra were recorded using a magnetic field of 5 T and temperatures of 1.73, 4.22, 10.0 and 50.0 K. (C) Expansion of the 595 nm region of (A). (D) RD-MCD spectra in the visible region. Spectra were recorded using magnetic fields and temperature combinations (in Tesla/Kelvin) of 0.247/1.73, 0.606/4.24, 1.714/12.0, 2.857/20.0, 3.857/27.0 and 5.000/35.0. The bi-signate feature at 595 nm increases in intensity with decreasing temperatures. (E) Difference spectra of the 595 nm MCD band obtained from the spectra in (D). (F) RT-MCD spectra in the 595 nm region of NorBC (dotted line) (×1/3); hydroxide-bound FeIII-protoporphyrin IX (dashed line); and water-bound FeIII-protoporphyrin IX (continuous line). The mid-point of each bi-signate band is shown by arrows at 595, 606 and 616 nm respectively.

Figure 3
VFVT MCD spectra of P. denitrificans NorBC

(A) VT-MCD spectra in the visible region. Spectra were recorded using a magnetic field of 5 T and temperatures of 1.73, 2.25, 2.95, 3.86, 4.22, 6.10, 7.90, 10.0, 15.0, 20.0, 29.6, 44.6, 64.5, 85.0, 110, 151, 182 and 220 K. All spectroscopic features increase in intensity with decreasing temperature. (B) VT-MCD spectra in the nIR region. Spectra were recorded using a magnetic field of 5 T and temperatures of 1.73, 4.22, 10.0 and 50.0 K. (C) Expansion of the 595 nm region of (A). (D) RD-MCD spectra in the visible region. Spectra were recorded using magnetic fields and temperature combinations (in Tesla/Kelvin) of 0.247/1.73, 0.606/4.24, 1.714/12.0, 2.857/20.0, 3.857/27.0 and 5.000/35.0. The bi-signate feature at 595 nm increases in intensity with decreasing temperatures. (E) Difference spectra of the 595 nm MCD band obtained from the spectra in (D). (F) RT-MCD spectra in the 595 nm region of NorBC (dotted line) (×1/3); hydroxide-bound FeIII-protoporphyrin IX (dashed line); and water-bound FeIII-protoporphyrin IX (continuous line). The mid-point of each bi-signate band is shown by arrows at 595, 606 and 616 nm respectively.

To extract active site magnetic parameters, the challenge is one of analysing weak high-spin haem MCD against a background of intense low-spin transitions. This can be achieved using RD-MCD (ratio-data MCD) [19]. The magnetic interaction between the two high-spin FeIII ions can be described by an effective spin-Hamiltonian involving an exchange coupling tensor J:

 
formula
(2)

where []b3 and []FeB are local spin-Hamiltonians for haem b3 and FeB, each containing a Zeeman term, plus axial (D) and rhombic (E) ZFS (zero-field splitting) parameters. Figure 4 shows energy level schemes in two limiting regimes. Strong exchange coupling (|J|≫|D|) yields a diamagnetic ground state. High-spin haem ZFS parameters are substantially larger than those of non-haem iron: under weak exchange (|J|≈|D|), each of the three well-separated haem b3 doublets couples with six FeB levels to give three groups of energy levels each containing twelve states. The electronic transitions of FeIII haems are xy-polarized [24]. Consequently the MCD temperature and field dependence is wavelength invariant and the intensity can be described as a superposition of the following contributions:

 
formula
(3)

where Δ∊(λ,T,H) is total MCD intensity at wavelength λ, temperature T and field H; terms in Cls(λ) and Chs(λ) represent temperature-dependent parts due to the two low-spin haems and to high-spin haem b3 respectively; A(λ) represents the temperature-independent contributions from all three. Exchange coupling will not significantly perturb the form of haem b3 MCD bands. However, changes to ground state magnetic properties resulting from an interaction with FeB will be reflected in the temperature- and field-dependent factor . This can be calculated by diagonalization of the Hamiltonian matrix constructed from eqn (2). Expectation values of the spin operator for all resulting states can be calculated and averaged both thermally and orientationally [19], but first contributions from the terms in A(λ) and Cls(λ) must be removed to obtain Chs(λ). The temperature- and field-dependent factor (Sz(T, H))ls for either low-spin haem can be calculated from:

 
formula
(4)

where angles θ and ϕ define the orientation of an individual haem relative to the applied field, and RD-MCD exploits the fact that the dependence of the MCD intensity on T and H has this relatively simple form for an S=1/2 centre. Recording spectra at combinations of field and temperature that maintain a fixed H/T ratio results in a constant contribution to the paramagnetic term Cls(λ) from low-spin haems: variations in MCD intensity then arise only from haem with a spin state other than S=1/2.

Energy level schemes for the two S=5/2 spins of haem b3 and FeB in weak and strong coupling models

Figure 4
Energy level schemes for the two S=5/2 spins of haem b3 and FeB in weak and strong coupling models

The rhombic ZFS parameters used were

Figure 4
Energy level schemes for the two S=5/2 spins of haem b3 and FeB in weak and strong coupling models

The rhombic ZFS parameters used were

RD-MCD spectra were recorded at six combinations of H and T with a fixed ratio of 0.143 T·K−1 yielding a second dataset, Δ∊RD (Figure 3D). At most wavelengths, the scans virtually overlie because the spectra are dominated by S=1/2 low-spin haems. However, at certain wavelengths, especially in the region of the 595 nm band, spectra show RD-MCD intensity variation, demonstrating the presence of a paramagnetic chromophore with S≠1/2. A measure of A(λ), the temperature-independent contribution from all haems was obtained by extrapolating the VT-MCD (variable temperature MCD) intensity (Δ∊VT) to high temperature (1/T→0). This was subtracted from each Δ∊VT spectrum to give a corrected dataset Δ∊′VT, containing the paramagnetic intensity of the three haems. A(λ) was similarly subtracted from the Δ∊RD, after scaling appropriately for the magnetic fields used, giving Δ∊′RD. The fixed contribution from low-spin haems could then be removed by subtracting one Δ∊′RD spectrum from each of the others, resulting in the difference spectra shown in Figure 3(E). Peak-to-trough (585–593 nm) intensities of these difference spectra have been normalized and the results plotted against 1/T (Figure 5A). These data are dependent solely on the factor 〈Sz(T, H)〉hs and therefore carry active site magnetic parameters. The continuous line is the optimum simulation achieved using the spin Hamiltonian of eqn (2). For mononuclear non-haem iron proteins and model complexes, D is typically ≤2 cm−1 [2527]. Our simulations are insensitive to variations in DFeB within this range and so this was set to 1 cm−1. Similarly, changes in produced negligible effects and this was fixed at 1/3, consistent with the low-symmetry observed for other FeIII non-haem iron species. Simulations were then performed by systematically varying and J. The fit of Figure 5(A) was achieved using and an isotropic coupling constant of J=−1.7 cm−1.

RD-MCD and VT-MCD plots of the HS FeIII haem b3 CT2

Figure 5
RD-MCD and VT-MCD plots of the HS FeIII haem b3 CT2

(A) RD-MCD plot of the HS FeIII haem b3 CT2 (595 nm) band intensity extracted from the Δ∊′RT MCD dataset for NorBC. ●, extracted data points; continuous line, simulation as described in text using J=−1.7 cm−1. (B) VT-MCD plot of the HS FeIII haem b3 CT2 (595 nm) band intensity extracted from the Δ∊′VT MCD dataset for the 595 form of NorBC: ○, extracted data points; dotted line, simulated LS FeIII contributions; dashed line, simulated contribution from coupled active site using J=−1.7 cm−1; continuous line, sum of these two contributions.

Figure 5
RD-MCD and VT-MCD plots of the HS FeIII haem b3 CT2

(A) RD-MCD plot of the HS FeIII haem b3 CT2 (595 nm) band intensity extracted from the Δ∊′RT MCD dataset for NorBC. ●, extracted data points; continuous line, simulation as described in text using J=−1.7 cm−1. (B) VT-MCD plot of the HS FeIII haem b3 CT2 (595 nm) band intensity extracted from the Δ∊′VT MCD dataset for the 595 form of NorBC: ○, extracted data points; dotted line, simulated LS FeIII contributions; dashed line, simulated contribution from coupled active site using J=−1.7 cm−1; continuous line, sum of these two contributions.

For RD-MCD measurements, at the lowest temperature used (~1.7 K), the field was set sufficiently high to obtain acceptable signal intensity. With the fixed H/T ratio, the solenoid maximum current then defined 35 K as the highest temperature at which data could be collected. The VT-MCD provides a second route to parameters describing the active site, but one yielding a larger dataset, Δ∊′VT, measured over a wider temperature range. As found in a previous study of cytochrome bo3 [19], this dataset is more sensitive to Db3 than is Δ∊′RD. Figure 5(B) shows a plot of the peak-to-trough (585–593 nm) MCD intensities taken from the Δ∊′VT dataset. These data contain the temperature-dependent part of the VT-MCD of all three haems as described by the two terms Cls and Chs of eqn (3). The invariant part of the RD-MCD in Δ∊′RDwas used to fix low-spin contributions to Δ∊′VT [19]: this is shown as the dotted trace in Figure 5(B). The dashed line shows the simulated contribution from haem b3 using the Hamiltonian of eqn 2. The excellent fit of the solid line, showing the addition of these two simulated contributions, to the experimental data supports the haem b3 ZFS parameters used in the simulations of Figure 5(A).

DISCUSSION

The present study has shown that the EPR-silent 595 form active site is not diamagnetic due to strong coupling, but is paramagnetic down to 1.7 K. A measurement of the coupling has been made for the first time, resulting in an isotropic J-value of −1.7 cm−1.

In a study of 34 oxygen-bridged (O2−, HO and RO) diferric complexes, Gorun and Lippard [28] showed that J is relatively insensitive to the M-O-M angle (ϕo), but determined primarily by the length of the exchange pathway. Weihe and Güdel [29] further developed this approach for oxo-bridged complexes. They modelled the dependence of J on the average Fe–O bond length (rav) and ϕo, confirming that the latter is of minor importance:

 
formula
(5)

The experimental J-values for the 32 complexes lay within 15% of those predicted by eqn (5). This is also true for the complexes reported subsequently (Supplementary Table S1 at http://www.biochemj.org/bj/451/bj4510389add.htm). Furthermore, if the Weihe–Güdel [29] formula is applied to structurally and magnetically characterized oxo-bridged di-haem or haem/non-haem iron complexes, it predicts J as successfully as it does for non-haem iron complexes (Supplementary Table S1). Thus it can reasonably be assumed that −J for oxo-bridged haem/non-haem iron structures should also fall in the range 110–270 cm−1 [29,30]. Our experimentally determined value of −J=1.7 cm−1 is two orders of magnitude smaller and rules out an oxo-bridge in the 595 form active site.

In the Pseudomonas NorBC structure, haem b3 is bound by a proximal histidine and bridged to FeB by a single oxygen. The Fe–O bond lengths are 1.9 and 2.0 Å (1 Å=0.1 nm) [11], significantly longer than the 1.73–1.82 Å found for oxo-bridges, but consistent with hydroxide bridging. However, absorption spectra showed that reduction of the low-spin haems occurs rapidly upon exposure to X-rays [11]. The structure is therefore consistent with our previous study showing that, when haems b and c are reduced, the distal haem b3 ligand becomes hydroxide and the proximal histidine residue rebinds [8], but it provides no evidence for the presence of hydroxide bridging in the 595 form.

Hay et al. [31] suggested that decreasing the charge on the bridging oxygen lengthens the Fe–O bonds and diminishes coupling. Thus −J should be smaller for RO bridges than oxo-bridges: this is borne out by observation. Haase and co-workers analysed 34 phenoxo-, alkoxo- and hydroxo-bridged non-haem diferric complexes [30], for which −J lies in the range of 10–60 cm−1. Again, the magnitude of J was found to depend upon the coupling pathway and the influence of ϕo is negligible:

 
formula
(6)

For hydroxo-bridging, −J falls in the narrower range of 10–35 cm−1, again significantly larger than 1.7 cm−1. There are few characterized hydroxo-bridged haem complexes and so generalizations from non-haem iron should be applied with care. The reported −J values of >50 cm−1 [32] and 9 cm−1 [33] suggest that coupling may be weaker for some haem complexes, but still larger than −J=1.7 cm−1, making μ-hydroxide also very unlikely in the 595 form. A further decrease in coupling is anticipated on changing from hydroxide to water [31]. Considered in isolation, a −J of 1.7 cm−1 makes water a plausible bridge. However, the position of the CT2 band for five-co-ordinate water-ligated FeIII–protoporphyrin IX is significantly red-shifted from 595 nm (Figure 3F) as it is for analogous haem structures in proteins (e.g. [34]).

For the 595 form, magnetic coupling and the CT2 wavelength rule out bridging by simple oxygen ligands, H2O, HOor O2−. The confines of the active site and limited availability of potential ligands severely restricts candidates for the species that mediates coupling. We suggest that this role is fulfilled by the glutamate residue observed as an FeB ligand in the structure of partially reduced Pseudomonas NorBC [11] (Figure 6). Carboxylate bridges, observed in reduced non-haem iron proteins in a μ-1,3 conformation, mediate couplings of −J=1–5 cm−1 [3538]. The results of the present study rule out oxo-bridging in the 595 form, but some rationale is required for resonance Raman studies that originated this model. Both studies reported two prominent bands at 760–880 cm−1 and suggested these are due to νas(Fe-O-Fe) modes of different conformations of the oxo-bridge [9,10]. In H218O, one mode, at 811–813 cm−1, shifted to lower energy by ~30 cm−1, consistent with exchange of O2−, but it was not clear why the second mode (833 cm−1) was largely unaffected. If a μ-oxo ligand is ruled out, the H218O-sensitive mode could be due to the hydroxide-bridged 622 form, but the origin of the second mode remains unclear. A ligand that would not exchange, but could give rise to bands at these energies, is peroxide. This ligand, bound in various conformations to mononuclear or dinuclear FeIII, gives rise to an asymmetric O–O stretch, νas, in the range of 790–932 cm−1 [3942]. Although peroxide ion exists as a bridge in several dinuclear non-haem iron enzyme intermediates, it mediates strong anti-ferromagnetic coupling (e.g. [4347]), and if present in 595 form NorBC is more likely to be a ligand to FeB. Thus, in the 595-form, weak coupling could be provided by bridging glutamate with peroxide as an η2 ligand to FeB (Figure 6).

Proposed structure for the diferric active site of the 595 form of NorBC incorporating a carboxylate bridge

In summary, we have measured an unexpectedly small value of −J=1.7 cm−1 for the magnetic exchange between haem b3 and FeB in 595 form NorBC. This overturns the accepted oxo-bridged model, but is consistent with carboxylate-mediated coupling. A bridging conformation contrasts with the structure of partially reduced Pseudomonas NorBC in which the glutamate is bound only to FeB [11], suggesting a redox-linked conformational flexibility. The mechanisms proposed for NOR involve the reoxidation, by 2NO, of a diferrous active site, but are generally divided into two categories: trans, whereby a nitric oxide molecule binds to each of haem b3 and FeB; and cis, in which both substrate molecules bind to FeB [7,9,10,48]. Each category could include a number of mechanisms differing in the order of entry into the protein of the electrons, protons and substrate molecules. Dissociation from haem b3 of the glutamate residue following reduction of haems b and c constitutes a mechanism whereby NO binding to haem is blocked until electrons are present in the protein and so may have a bearing on the order of events. Furthermore this glutamate has been implicated in a proton-transfer pathway from the periplasmic surface to the active site [14], a role it would not fulfil while in a bridging conformation: electron uptake could therefore gate proton uptake. Cyanide ion (isoelectronic with NO+) binds more strongly, by three orders of magnitude, to partially reduced NorBC than to the 595 form [49], implying that the active site is inaccessible to substrate prior to electron uptake.

Abbreviations

     
  • CT

    charge-transfer

  •  
  • HCO

    haem-copper oxidase

  •  
  • MCD

    magnetic circular dichroism

  •  
  • nIR

    near IR

  •  
  • NOR

    nitric oxide reductase

  •  
  • RD-MCD

    ratio-data MCD

  •  
  • RT-MCD

    room temperature MCD

  •  
  • VFVT

    variable-field variable-temperature

  •  
  • VT-MCD

    variable temperature MCD

  •  
  • ZFS

    zero-field splitting

AUTHOR CONTRIBUTION

Jessica Van Wonderen prepared samples, recorded spectroscopic data and contributed to interpretation of data. Vasily Oganesyan performed simulations and analysis of magnetization data. Myles Cheesman and Andrew Thomson planned the experiments and interpreted the data. Myles Cheesman wrote the paper. Nicholas Watmough and David Richardson provided enzyme samples. Nicholas Watmough, Myles Cheesman, David Richardson and Andrew Thomson secured funding for the project. All authors discussed the results and commented on the paper.

M.R.C. thanks Professor Walter Zumft for helpful suggestions during the preparation of this paper.

FUNDING

This work was supported by the Biotechnology and Biological Sciences Research Council grants [grant numbers BBC0077191, B18695 and BBE0132521] and studentship [grant number 02A1B08117] and a Wellcome Trust award from the Joint Infra-structure Fund for Equipment.

References

References
1
Subbarao
G. V.
Ito
O.
Sahrawat
K. L.
Berry
W. L.
Nakahara
K.
Ishikawa
T.
Watanabe
T.
Suenaga
K.
Rondon
M.
Rao
I. M.
Scope and strategies for regulation of nitrification in agricultural systems–challenges and opportunities
Crit. Rev. Plant Sci.
2006
, vol. 
25
 (pg. 
303
-
335
)
2
Ravishankara
A. R.
Daniel
J. S.
Portmann
R. W.
Nitrous oxide (N2O): the dominant ozone-depleting substance emitted in the 21st century
Science
2009
, vol. 
326
 (pg. 
123
-
125
)
3
Zumft
W. G.
Cell biology and molecular basis of denitrification
Microbiol. Mol. Biol. Rev.
1997
, vol. 
61
 (pg. 
533
-
616
)
4
Kakishima
K.
Shiratsuchi
A.
Taoka
A.
Nakanishi
Y.
Fukumori
Y.
Participation of nitric oxide reductase in survival of Pseudomonas aeruginosa in LPS-activated macrophages
Biochem. Biophys. Res. Commun.
2007
, vol. 
355
 (pg. 
587
-
591
)
5
Saraste
M.
Castresana
J.
Cytochrome oxidase evolved by tinkering with denitrification enzymes
FEBS Lett.
1994
, vol. 
341
 (pg. 
1
-
4
)
6
Girsch
P.
de Vries
S.
Purification and initial kinetic and spectroscopic characterization of NO reductase from Paracoccus denitrificans
Biochim. Biophys. Acta
1997
, vol. 
1318
 (pg. 
202
-
216
)
7
Moenne-Loccoz
P.
de Vries
S.
Structural characterization of the catalytic high-spin heme b of nitric oxide reductase: a resonance Raman study
J. Am. Chem. Soc.
1998
, vol. 
120
 (pg. 
5147
-
5152
)
8
Grönberg
K. L. C.
Roldan
M. D.
Prior
L.
Butland
G.
Cheesman
M. R.
Richardson
D. J.
Spiro
S.
Thomson
A. J.
Watmough
N. J.
A low-redox potential heme in the dinuclear center of bacterial nitric oxide reductase: implications for the evolution of energy-conserving heme-copper oxidases
Biochemistry
1999
, vol. 
38
 (pg. 
13780
-
13786
)
9
Moënne-Loccoz
P.
Richter
O. M. H.
Huang
H. W.
Wasser
I. M.
Ghiladi
R. A.
Karlin
K. D.
de Vries
S.
Nitric oxide reductase from Paracoccus denitrificans contains an oxo-bridged heme/non-heme diiron center
J. Am. Chem. Soc.
2000
, vol. 
122
 (pg. 
9344
-
9345
)
10
Pinakoulaki
E.
Gemeinhardt
S.
Saraste
M.
Varotsis
C.
Nitric-oxide reductase–structure and properties of the catalytic site from resonance Raman scattering
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
23407
-
23413
)
11
Hino
T.
Matsumoto
Y.
Nagano
S.
Sugimoto
H.
Fukumori
Y.
Murata
T.
Iwata
S.
Shiro
S.
Structural basis of biological N2O generation by bacterial nitric oxide reductase
Science
2010
, vol. 
330
 (pg. 
1666
-
1670
)
12
Cheesman
M. R.
Zumft
W. G.
Thomson
A. J.
The MCD and EPR of the heme centers of nitric oxide reductase from Pseudomonas stutzeri: evidence that the enzyme is structurally related to the heme-copper oxidases
Biochemistry
1998
, vol. 
37
 (pg. 
3994
-
4000
)
13
Suharti, Strampraad
M. J. F.
Schröder
I.
de Vries
S.
A novel copper A containing menaquinol NO reductase from Bacillus azotoformans
Biochemistry
2001
, vol. 
40
 (pg. 
2632
-
2639
)
14
Butland
G.
Spiro
S.
Watmough
N. J.
Richardson
D. J.
Two conserved glutamates in the bacterial nitric oxide reductase are essential for activity but not assembly of the enzyme
J. Bacteriol.
2001
, vol. 
183
 (pg. 
189
-
199
)
15
Solomon
E. I.
Brunold
T. C.
Davis
M. I.
Kemsley
J. M.
Lee
S.-K.
Lehnert
N.
Neese
F.
Skulan
A. J.
Yang
Y.-S.
Zhou
J.
Geometric and electronic structure/function correlations in non-heme iron enzymes
Chem. Rev.
2000
, vol. 
100
 (pg. 
235
-
349
)
16
Dawson
J. W.
Gray
H. B.
Hoenig
H. E.
Rossman
G. R.
Schredder
J. M.
Wang
R.-H.
A magnetic susceptibility study of hemerythrin using an ultrasensitive magnetometer
Biohemistry
1972
, vol. 
11
 (pg. 
461
-
465
)
17
Petersson
L.
Gräslund
A.
Ehrenberg
A.
Sjoberg
B.-M.
Reichard
P.
The iron center in ribonucleotide reductase from Escherichia coli
J. Biol. Chem.
1980
, vol. 
255
 (pg. 
6706
-
6712
)
18
Fox
B. G.
Hendrich
M. P.
Surerus
K. K.
Andersson
K. K.
Froland
W. A.
Lipscomb
J. D.
Münck
E.
Mössbauer, EPR, and ENDOR studies of the hydroxylase and reductase components of methane monooxygenase from Methylosinus trichosporium OB3b
J. Am. Chem. Soc.
1993
, vol. 
115
 (pg. 
3688
-
3701
)
19
Cheesman
M. R.
Oganesyan
V. S.
Watmough
N. J.
Butler
C. S.
Thomson
A. J.
The nature of the exchange coupling between high-spin Fe(III) heme o3 and CuB(II) in E. coli quinol oxidase, cytochrome bo3: MCD and EPR studies
J. Am. Chem. Soc.
2004
, vol. 
126
 (pg. 
4157
-
4166
)
20
Field
S. J.
Thorndycroft
F. H.
Matorin
A. D.
Richardson
D. J.
Watmough
N. J.
The respiratory nitric oxide reductase (NorBC) from Paracoccus denitrificans
Methods Enzymol.
2008
, vol. 
437
 (pg. 
79
-
101
)
21
Cheesman
M. R.
Greenwood
C.
Thomson
A. J.
Magnetic circular dichroism of hemoproteins
Adv. Inorg. Chem.
1991
, vol. 
36
 (pg. 
201
-
255
)
22
Hendriks
J. H. M.
Prior
L.
Baker
A. R.
Thomson
A. J.
Saraste
M.
Watmough
N. J.
Reaction of carbon monoxide with the reduced active site of bacterial nitric oxide reductase
Biochemistry
2001
, vol. 
40
 (pg. 
13361
-
13369
)
23
Van Wonderen
J. H.
Mageto-optical Spectroscopic Studies of Bacterial Nitric Oxide Reductase
Ph.D. Thesis
2007
Norwich, U.K.
University of East Anglia
24
Smith
D. W.
Williams
R. J. P.
The spectra of ferric haems and haemoproteins
Struct. Bonding
1970
, vol. 
7
 (pg. 
1
-
43
)
25
Citadini
A. P. S.
Pinto
A. P. A.
Araújo
A. P. U.
Nascimento
O. R.
Costa-Filho
A. J.
EPR studies of chlorocatechol 1,2-dioxygenase: evidences of iron reduction during catalysis and of the binding of amphipatic molecules
Biophys. J.
2005
, vol. 
88
 (pg. 
3502
-
3508
)
26
Neese
F.
Solomon
E. I.
Detailed spectroscopic and theoretical studies on [Fe(EDTA)(O2)]3−: electronic structure of the side-on ferric-peroxide bond and its relevance to reactivity
J. Am. Chem. Soc.
1998
, vol. 
120
 (pg. 
12829
-
12848
)
27
Kerfeld
C. A.
Yoshida
S.
Tran
K. T.
Yeates
T. O.
Cascio
D.
Bottin
H.
Berthomieu
C.
Sugiura
M.
Boussac
A.
The 1.6 Å resolution structure of Fe-superoxide dismutase from the thermophilic cyanobacterium Thermosynechococcus elongatus
J. Biol. Inorg. Chem.
2003
, vol. 
8
 (pg. 
707
-
714
)
28
Gorun
S. M.
Lippard
S. J.
Magnetostructural correlations in magnetically coupled (μ-oxo)diiron(III) complexes
Inorg. Chem.
1991
, vol. 
30
 (pg. 
1625
-
1630
)
29
Weihe
H.
Güdel
H. U.
Angular and distance dependence of the magnetic properties of oxo-bridged iron(III) dimers
J. Am. Chem. Soc.
1997
, vol. 
119
 (pg. 
6539
-
6543
)
30
Werner
R.
Ostrovsky
S.
Griesar
K.
Haase
W.
Magnetostructural correlations in exchange coupled phenoxo-, alkoxo-, and hydroxo-bridged dinuclear iron(III) compounds
Inorg. Chim. Acta
2001
, vol. 
326
 (pg. 
78
-
88
)
31
Hay
P. J.
Thibeault
J. C.
Hoffmann
R.
Orbital interactions in metal dimer complexes
J. Am. Chem. Soc.
1975
, vol. 
97
 (pg. 
4884
-
4899
)
32
Scheidt
W. R.
Cheng
B.
Safo
M. K.
Cukiernik
F.
Marchon
J.-C.
Debrunner
P. G.
New class of bridged diiron(III) complexes with a single hydroxo bridge. The preparation and structure of (μ-hydroxo) bis((octaethylporphinato)iron(III)) perchlorate
J. Am. Chem. Soc.
1992
, vol. 
114
 (pg. 
4420
-
4421
)
33
Ghosh
S. K.
Rath
S. P.
A remarkably bent diiron(III)-μ-hydroxo bis-porphyrin: unusual stabilization of two spin states of iron in a single molecular framework
J. Am. Chem. Soc.
2010
, vol. 
132
 (pg. 
17983
-
17985
)
34
Pond
A. E.
Roach
M. P.
Sono
M.
Huff Rux
A.
Franzen
S.
Hu
R.
Thomas
M. R.
Wilks
A.
Dou
Y.
Ikeda-Saito
M.
, et al. 
Assignment of the heme axial ligand(s) for the ferric myoglobin (H93G) and heme oxygenase (H25A) cavity mutants as oxygen donors using magnetic circular dichroism
Biochemistry
1999
, vol. 
38
 (pg. 
7601
-
7608
)
35
Pulver
S. C.
Tong
W. H.
Bollinger
J. M. J.
Stubbe
J.
Solomon
E. I.
Circular dichroism and magnetic circular dichroism studies of the fully reduced binuclear non-heme iron active site in the Escherichia coli R2 subunit of ribonucleoside diphosphate reductase
J. Am. Chem. Soc.
1995
, vol. 
117
 (pg. 
12664
-
12678
)
36
Wei
P. P.
Skulan
A. J.
Mitić
N.
Yang
Y.-S.
Saleh
L.
Bollinger
J. M. J.
Solomon
E. I.
Electronic and spectroscopic studies of the non-heme reduced binuclear iron sites of two ribonucleotide reductase variants: comparison to reduced methane monooxygenase and contributions to O2 reactivity
J. Am. Chem. Soc.
2004
, vol. 
126
 (pg. 
3777
-
3788
)
37
Tomter
A. B.
Bell
C. B.
Røhr
Å.K.
Andersson
K. K.
Solomon
E. I.
Circular dichroism and magnetic circular dichroism studies of the biferrous site of the class Ib ribonucleotide reductase from Bacillus cereus: comparison to the class Ia enzymes
Biochemistry
2008
, vol. 
47
 (pg. 
11300
-
11309
)
38
Bell
C. B. I.
Calhoun
J. R.
Bobyr
E.
Wei
P. P.
Hedman
B.
Hodgson
K. O.
DeGrado
W. F.
Solomon
E. I.
Spectroscopic definition of the biferrous and biferric sites in de novo designed four-helix bundle DFsc peptides: implications for O2 reactivity of binuclear non-heme iron enzymes
Biochemistry
2009
, vol. 
48
 (pg. 
59
-
73
)
39
Vaska
L.
Dioxygen-metal complexes: toward a unified view
Acc. Chem. Res.
1976
, vol. 
9
 (pg. 
175
-
183
)
40
Tshuva
E. Y.
Lippard
S. J.
Synthetic models for non-heme carboxylatebridged diiron metalloproteins: strategies and tactics
Chem. Rev.
2004
, vol. 
104
 (pg. 
987
-
1012
)
41
Wallar
B. J.
Lipscomb
J. D.
Dioxygen activation by enzymes containing binuclear non-heme iron clusters
Chem. Rev.
1996
, vol. 
96
 (pg. 
2625
-
2657
)
42
Que
L.
Jr
Dong
Y.
Modeling the oxygen activation chemistry of methane monooxygenase and ribonucleotide reductase
Acc. Chem. Res.
1996
, vol. 
29
 (pg. 
190
-
196
)
43
Friedle
S.
Reisner
E.
Lippard
S. J.
Current challenges of modeling diiron enzyme active sites for dioxygen activation by biomimetic synthetic complexes
Chem. Soc. Rev.
2010
, vol. 
39
 (pg. 
2768
-
2779
)
44
Moënne-Loccoz
P.
Baldwin
J.
Ley
B. A.
Loehr
T. M.
Bollinger
J. M. J.
O2 activation by non-heme diiron proteins: identification of a symmetric μ-1,2-peroxide in a mutant of ribonucleotide reductase
Biochemistry
1998
, vol. 
37
 (pg. 
14659
-
14663
)
45
Broadwater
J. A.
Ai
J.
Loehr
T. M.
Sanders-Loehr
J.
Fox
B. G.
Peroxodiferric intermediate of stearoyl-acyl carrier protein Δ9 desaturase: oxidase reactivity during single turnover and implications for the mechanism of desaturation
Biochemistry
1998
, vol. 
37
 (pg. 
14664
-
14671
)
46
Song
W. J.
Behan
R. K.
Naik
S. G.
Huynh
B. H.
Lippard
S. J.
Characterization of a peroxodiiron(III) intermediate in the T201S variant of toluene/o-xylene monooxygenase hydroxylase from Pseudomonas sp
OX1. J. Am. Chem. Soc.
2009
, vol. 
131
 (pg. 
6074
-
6075
)
47
Liu
K. E.
Valentine
J. S.
Wang
D.
Huynh
B. H.
Edmondson
D. E.
Salifoglou
A.
Lippard
S. J.
Kinetic and spectroscopic characterization of intermediates and component interactions in reactions of methane monooxygenase from Methylococcus capsulatus (Bath)
J. Am. Chem. Soc.
1995
, vol. 
117
 (pg. 
10174
-
10185
)
48
Moënne-Loccoz
P.
Spectroscopic characterization of heme iron-nitrosyl species and their role in NO reductase mechanisms in diiron proteins
Nat. Prod. Rep.
2007
, vol. 
24
 (pg. 
610
-
620
)
49
Grönberg
K. L. C.
Watmough
N. J.
Thomson
A. J.
Richardson
D. J.
Field
S. J.
Redox-dependent open and closed forms of the active site of the bacterial respiratory nitric-oxide reductase revealed by cyanide binding studies
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
17120
-
17125
)

Author notes

1

This paper is dedicated to the memory of Colin Greenwood.

Supplementary data