Biochemical Journal

Research article

Observation of fast release of NO from ferrous d1 haem allows formulation of a unified reaction mechanism for cytochrome cd1 nitrite reductases

Serena Rinaldo , Katharine A. Sam , Nicoletta Castiglione , Valentina Stelitano , Alessandro Arcovito , Maurizio Brunori , James W. A. Allen , Stuart J. Ferguson , Francesca Cutruzzolà

Abstract

Cytochrome cd1 nitrite reductase is a haem-containing enzyme responsible for the reduction of nitrite into NO, a key step in the anaerobic respiratory process of denitrification. The active site of cytochrome cd1 contains the unique d1 haem cofactor, from which NO must be released. In general, reduced haems bind NO tightly relative to oxidized haems. In the present paper, we present experimental evidence that the reduced d1 haem of cytochrome cd1 from Paracoccus pantotrophus releases NO rapidly (k=65–200 s−1); this result suggests that NO release is the rate-limiting step of the catalytic cycle (turnover number=72 s−1). We also demonstrate, using a complex of the d1 haem and apomyoglobin, that the rapid dissociation of NO is largely controlled by the d1 haem cofactor itself. We present a reaction mechanism proposed to be applicable to all cytochromes cd1 and conclude that the d1 haem has evolved to have low affinity for NO, as compared with other ferrous haems.

  • haemoprotein
  • nitric oxide
  • nitrite
  • nitrite reductase
  • Pseudomonas aeruginosa
  • Paracoccus pantotrophus

INTRODUCTION

Denitrification is an anaerobic respiratory process, found widely in both autotrophic and heterotrophic micro-organisms [1], in which oxidized nitrogen compounds, such as nitrate and nitrite, are used as electron acceptors for energy production. Denitrification has been implicated in the virulence of several bacterial species, including Brucella [2], Pseudomonas [3] and Neisseria [4]. In Pseudomonas aeruginosa, denitrification is also a source of NO, a crucial signalling molecule during infection and growth of bacteria in anaerobic biofilms [5]. The biological role of NO under physiological and pathological situations is well-known [6]; a critical aspect of the biology of NO is the interaction of this molecule with the haem group. The general rule is that NO binds much more strongly to the ferrous than to the ferric haem, a consideration that imposes mechanistic constraints on many proteins that synthesize or bind NO [79].

In a wide range of denitrifying bacteria the reduction of nitrite (NO2) to NO is carried out by periplasmic cytochrome cd1 nitrite reductase (cytochrome cd1) [1], a homodimeric, haem-containing protein with one c haem and one d1 haem per monomer [10] (Supplementary Figure S1 at http://www.BiochemJ.org/bj/435/bj4350217add.htm). The d1 haem (3,8-dioxo-17-acrylate-porphyrindione) is an unusual macrocycle with partial saturation and a set of oxo, methyl and acrylate substituents, which makes it unique among tetrapyrroles (Supplementary Figure S2 at http://www.BiochemJ.org/bj/435/bj4350217add.htm). The best-studied examples of cytochrome cd1 nitrite reductases are those from Paracoccus pantotrophus and P. aeruginosa [10]. The c haem moiety of cytochrome cd1 accepts electrons from soluble electron carriers, such as c-type cytochromes (cyt c550, cyt c551 and cyt c554) or copper proteins, such as azurin or pseudoazurin [10]. The electrons are transferred from the c haem of cytochrome cd1 to the d1 haem (of the same monomer), where the substrate binds and is converted into NO. Nitrite co-ordination to P. pantotrophus cytochrome cd1 is via the N-atom as shown by crystallography [11]. The reduction of nitrite by cytochrome cd1 involves binding of the substrate to the reduced d1 haem iron (d12+), followed by one-electron reduction to yield NO, which is then released [1]. The last step in the catalytic cycle of cytochrome cd1, i.e. NO dissociation, has been the subject of much debate [1214]. Previously it was shown that NO is released rapidly from the ferrous d1 haem (koff=71 s−1) of P. aeruginosa cytochrome cd1 [15,16]. This surprising result suggested that the affinity for NO of the ferrous d1 haem in the enzyme from P. aeruginosa was 3–4 orders of magnitude smaller (Ka~107 M−1) than for many other haemoproteins such as myoglobin and haemoglobin (Ka~1011–1012 M−1) [7], and cytochrome aa3 oxidase (Ka~1010 M−1) [17].

In order to clarify whether the capability to release NO rapidly from ferrous d1 haem iron is a general property of cytochrome cd1 nitrite reductases, we have investigated the reactivity of ferrous P. pantotrophus cytochrome cd1 with NO. There are a number of significant differences between the enzymes from P. pantotrophus and P. aeruginosa. In steady-state turnover of nitrite, P. pantotrophus cytochrome cd1 displays a considerably higher kcat (72 s−1 at pH 7.0 and 25 °C) [18] than the value for P. aeruginosa cytochrome cd1 (6 s−1 at pH 7.0 and 20 °C) [15]. Also, the internal electron transfer between the c haem and the d1 haem occurs several orders of magnitude faster in P. pantotrophus cytochrome cd1 (~1000 s−1) [14,19] than in P. aeruginosa cytochrome cd1 (3 s−1) [20]. These differences raised the question as to whether NO release from ferrous d1 haem of the P. pantotrophus enzyme would be sufficiently rapid to be catalytically competent.

In the present paper we report that NO is released rapidly from P. pantotrophus cytochrome cd1; the value of the NO dissociation rate suggests that this process represents the rate-limiting step of the catalytic cycle. The finding that the reaction mechanism for cytochrome cd1 involves fast NO dissociation from ferrous d1 haem supports the idea that the chemical properties of this unique cofactor are crucial for activity. To corroborate this hypothesis we have also investigated the kinetics of NO dissociation from ferrous d1 haem complexed with sperm whale apomyoglobin, in order to probe NO reactivity in a protein environment different from cytochrome cd1. We found that ferrous d1 haem bound to apomyoglobin releases NO rapidly, confirming the leading role of this cofactor in controlling reactivity of cytochrome cd1 with NO.

EXPERIMENTAL

Preparation of P. pantotrophus cytochrome cd1

P. pantotrophus was grown in anaerobic conditions at 37 °C. Cytochrome cd1 was purified from the periplasm of the cells according to the method of Moir et al. [21] as modified by Koppenhöfer et al. [22]. The purity of the enzyme was determined by the RZ (Reinheitzahl) value (A406(Ox)/A280) and all the cytochrome cd1 used in the present study had an RZ>1.25. The concentration of the enzyme was determined at 406 nm for the oxidized enzyme and 418 nm for the reduced enzyme, using the respective molar absorption coefficients of 142.5 mM−1·cm−1 and 161.5 mM−1·cm−1 [19,22]. These molar absorption coefficients refer to the concentration of the enzyme monomer; throughout the present study, the enzyme concentration will be given as monomer concentration and thus catalytic constants will refer to turnover per monomer. Fully reduced cytochrome cd1 was prepared by reduction with sodium dithionite in an anaerobic glove box; the excess reductant was removed by passing the enzyme down a desalting column packed with P6-DG resin (Bio-Rad) and equilibrated with 50 mM potassium phosphate buffer (pH 7.0), in the presence of 5 mM sodium ascorbate and 13 μg/ml ascorbate oxidase (Sigma). The reduced NO-bound and cyanide-bound derivatives were obtained after the addition of either a stoichiometric amount of NOor excess KCN to the reduced protein. Stock solutions of NO were prepared by equilibrating NO with buffer (50 mM potassium phosphate pH 7.0) at 25 °C, assuming the aqueous concentration of NO at 1 atm (101.325 kPa) to be 1.9 mM. All of the spectra were recorded in a JASCO V650 spectrophotometer.

A summary of the relevant spectra of P. pantotrophus cytochrome cd1 derivatives, together with a scheme of the c haem and the d1 haem ligands in each derivative, is shown in Supplementary Figure S3 (at http://www.BiochemJ.org/bj/435/bj4350217add.htm). It was reported previously that, upon reduction of the c haem, the iron co-ordination changes from histidine–histidine to histidine–methionine [11]. The latter co-ordination (histidine–methionine) does not change during turnover; reversion to the histidine–histidine state occurs very slowly (minutes) and only in the absence of ligands or substrates [23]. Therefore the change in co-ordination of the c haem is not expected to occur during the time range of the kinetics reported in the present paper.

Stopped-flow measurements

All of the stopped-flow experiments described in the present paper were carried out anaerobically in 50 mM potassium phosphate buffer (pH 7.0) in the presence of 5 mM sodium ascorbate and 13 μg/ml ascorbate oxidase (Sigma), added to scavenge possible oxygen contamination. The concentration of the samples in the different experiments is given a.m. (after mixing): a 1:1 dilution was always used in the symmetric mixing apparatus. All of the experiments were performed with an Applied Photophysics stopped-flow apparatus (DX.17MV, Applied Photophysics). To follow the rate of NO dissociation, a monochromatic light source was used in the single wavelength acquisition mode; the use of the diode array acquisition mode was limited because photodissociation of NO is known to affect the rate. All of the kinetic analysis was carried out with the IgorPro program (Wavemetrics).

The spectral analysis of the events occurring at the d1 haem was carried out following the absorbance changes at wavelengths above 600 nm, in order to avoid overlap with the c haem, which does not contribute to this region of the spectrum. In the experimental conditions used in the present study, there is no evidence that ligands such as nitrite and cyanide bind to the c haem.

In order to determine the dissociation rate of NO from the reduced d1 haem, the reduced NO-bound cytochrome cd1 (3 μM) was rapidly mixed in the stopped-flow apparatus with excess potassium cyanide (0.08, 0.12, 0.2, 1, 5, 10, 15 and 100 mM a.m.), and the formation of the reduced-cyanide-bound derivative was followed at different wavelengths. Under these conditions, cyanide displaces NO by binding to the reduced d1 haem.

To probe if NO dissociation from the fully reduced enzyme can be facilitated by the substrate, the reduced NO-bound cytochrome cd1 (3.5 μM) was rapidly mixed in the stopped-flow apparatus with sodium nitrite (3 mM), and the reaction was followed in the diode array acquisition mode. This approach has been chosen in order to gain spectral information on the species involved, although we are aware that the observed rate can be slightly faster than under monochromatic light.

Under these conditions the following reaction occurs and nitrite binding is rate limited by NO dissociation: Embedded Image

As a control, the reduced ligand-free (3.5 μM) enzyme was mixed with the same amounts of nitrite as above in the presence of excess ascorbate. The similarity of the latter reaction in terms of species involved and time course to previously published results [24] indicates that excess ascorbate does not alter the reaction mechanism.

Laser photolysis measurements

A 4.5 μM protein solution in 50 mM potassium phosphate buffer (pH 7.0) at room temperature (25 °C) was reduced anaerobically as described above and then transferred to a fully filled gas-tight cuvette (Hellma) with four transparent windows; the path length along the direction of the probe light was 1 cm and the path length in the orthogonal direction was 4 mm. Different volumes of a stock NO solution were added anaerobically to produce the reduced NO-bound derivative in the concentration range 25–100 μM. The apparatus used for the photolysis experiments is home-built and has been described elsewhere [25]; every kinetic trace was the result of an average of 256 time courses at a single wavelength, each obtained after a single laser shot. The time course was followed at 460 nm, a wavelength where the spectrum of the reduced cytochrome cd1 is dominated by the contribution of d1 haem. Analysis of the results was performed by globally fitting all of the kinetic traces at different final NO concentrations.

Mbd1 (haem d1–apomyoglobin) experiments

Mbd1 was prepared after reconstitution of sperm whale apomyoglobin (prepared according to [26]) using the d1 haem extracted from wild-type P. aeruginosa cytochrome cd1, as reported previously [27]. The Mbd1 preparation and kinetic experiments were carried out in 40 mM sodium phosphate buffer (pH 6.9).

The reduced Mbd1 was prepared anaerobically by adding 5 mM sodium ascorbate to the oxidized protein in degassed buffer (40 mM sodium phosphate pH 6.9) at 20 °C, in the presence of 13 μg/ml ascorbate oxidase as an oxygen scavenger. The reduced NO-bound derivative was obtained after the addition of 100 μM NO solution (prepared as described above). All of the spectra were recorded in a JASCO V550 spectrophotometer.

To determine the NO-dissociation-rate constant, the reduced NO-bound Mbd1 was mixed anaerobically in a gas-tight cuvette containing 5 mM sodium dithionite and ~500 μM CO; the process was followed by recording the spectrum (400–700 nm). The formation of the CO-bound complex [27] was complete within 2 min. To measure the kinetics of NO dissociation, the NO-bound Mbd1 was mixed in the stopped-flow apparatus with a buffer solution containing 1 mM CO and 5 mM sodium dithionite (before mixing; three independent experiments). The kinetic process was followed at 451 nm; the time course was fitted with a single exponential equation.

RESULTS

The rate constant for the binding of NO to reduced P. pantotrophus cytochrome cd1

The association rate constant for the binding of NO to reduced P. pantotrophus cytochrome cd1 was determined by laser flash photolysis at pH 7.0 under anaerobic conditions at 25 °C (Figure 1). Time courses were followed at 460 nm in the presence of different NO concentrations (25, 50, 75 and 100 μM). The small absorbance change observed in the experiment is consistent with the low photolysis yield of the NO species, as previously observed with P. aeruginosa cytochrome cd1 [15]. The global fit of the kinetic traces at different NO concentrations yields a value of kon(NO)=3.5±0.3×108 M−1·s−1 (Table 1).

Figure 1 Time course of NO recombination with fully reduced P. pantotrophus cytochrome cd1

The reaction was initiated by photodissociation of NO with a laser pulse at 532 nm; the recombination was monitored at 460 nm at four different NO concentrations, i.e. 100, 75, 50 and 25 μM. Only traces at 100 μM and 25 μM (grey thin lines) are shown for clarity, together with the results obtained by globally fitting all the traces (black broken and continuous lines respectively). As described in the Experimental section, every kinetic trace is the result of an average of 256 time courses, each obtained after a single laser shot. Experimental conditions: cytochrome cd1 4.5 μM, 25 °C and 50 mM potassium phosphate (pH 7.0) containing 5 mM ascorbate and 13 μg/ml ascorbate oxidase.

View this table:
Table 1 Kinetic parameters for nitrite reduction and for NO and cyanide binding to and dissociation from reduced cytochrome cd1 nitrite reductases

For comparison, the catalytic parameters for proteins able to catalyse the reduction of nitrite to NO [52] and containing a b-type haem are also shown.*Cytochrome cd1 nitrite reductase binds the selected ligands to the ferrous state of the d1 haem. Anions such as nitrite or cyanide bind to ferric d1 haem with very low affinity (KD nitrite ~1 M) [53]. NO reacts slowly with the ferric enzyme by binding to the d1 haem and then reducing it to produce the d12+-NO derivative; this autoreductive phenomenon is pH dependent [54]. The kinetic parameters reported in the Table are at pH 7.0 and 20–25 °C; under these experimental conditions, none of the ligands used react with the c haem. However, at more acidic pH, nitrosylation of the c haem (by NO or nitrite after reduction via d1 haem) was reported to occur [23].†Nitrite binding occurs very rapidly and cannot be directly measured; therefore the values in the Table refer to estimates reported in the cited references. ‡Two kinetic constants are reported in all experiments where biphasic kinetics were observed (see present work and corresponding references).

The rate constant for the dissociation of NO from reduced P. pantotrophus cytochrome cd1

To measure NO dissociation, we took advantage of the unusually high affinity for cyanide of reduced cytochrome cd1, which can be exploited to displace NO from the ferrous d1 haem [15]. It was shown previously that ferrous P. pantotrophus cytochrome cd1 binds cyanide relatively tightly (Kd=0.7±0.2×10−6 M at pH 7.0) (Table 1) [28], although the corresponding association and dissociation rate constants were not determined. We determined these rate constants (Supplementary Figures S4 and S5 at http://www.BiochemJ.org/bj/435/bj4350217add.htm) since they are necessary to calculate the NO dissociation rate constant in the cyanide–NO displacement experiment and found values of: kon(CN)=5.9±0.1×105 M−1·s−1 and koff(CN)=0.4±0.02 s−1 (Table 1). These results yield a Kd of 0.68 μM, which is in excellent agreement with the equilibrium value reported previously [28].

The NO-dissociation time course at 20 °C and pH 7.0 was time-resolved by mixing reduced NO-bound P. pantotrophus cytochrome cd1 with excess cyanide (0.08–100 mM) in the stopped-flow instrument and following the kinetics at 627 nm (Figure 2), a wavelength characteristic of the ferrous d1 haem–CN complex (Figure S3) and where the c haem does not absorb. The time course of formation of the cyanide derivative (Figure 2A) was biphasic under all conditions. The biphasicity cannot be ascribed to the cyanide binding process, which is always monophasic (Figure S4).

Figure 2 Dissociation of NO from fully reduced NO-bound P. pantotrophus cytochrome cd1 obtained by displacement with excess cyanide

(A) The dissociation time course was followed after mixing fully reduced NO-bound cytochrome cd1 (6 μM) with excess potassium cyanide (0.08, 0.12, 0.2, 1, 5, 10, 15 and 100 mM) in the stopped-flow apparatus. The kinetic data were collected in the single wavelength acquisition mode in order to minimize NO photodissociation. The kinetics were followed at 627 nm (absorption maximum of the cyanide bound d1 haem; there is no absorbance from the c haem at this wavelength) and the time course fitted with a double exponential equation (broken line). (B) Plot of the observed NO displacement rate constant from reduced cytochrome cd1 as a function of cyanide concentration; each kinetic phase (see above) was plotted independently (k1=●; k2=○). The results are fitted using the replacement model (thin line) [26]; the equation contains the cyanide kon and koff and the NO kon values measured independently in this work. The values of three independent measurements at each cyanide concentration are reported. Experimental conditions: 20 °C and 50 mM potassium phosphate (pH 7.0) containing 5 mM ascorbate and 13 μg/ml ascorbate oxidase.

The values of the rate constants for both phases, k1 and k2, have been plotted as a function of ligand concentration (Figure 2B). At low CN, the two phases have equal amplitudes, suggesting that the two monomers of cytochrome cd1 possibly react at different rates; at high CN, some amplitude of the first phase is lost in the dead time of the instrument because the reaction is too fast. The data have been analyzed using the replacement model described in eqn 1 [26]: Embedded Image The two rate constants for NO dissociation from reduced P. pantotrophus cytochrome cd1 are as follows: k1=201.2±11.4 s−1 and k2=64.9±2.3 s−1 for the fast and slow processes respectively (Table 1).

Displacement of NO by nitrite from reduced P. pantotrophus cytochrome cd1

Both the ferrous NO-bound and, for comparison, the ferrous enzyme, were mixed anaerobically with excess nitrite (3 mM) in the stopped-flow apparatus (at 20 °C and pH 7.0). The time courses at 551 and 660 nm, representative of the c and d1 haems respectively, are shown in Figures 3(A) and 3(B) for the reduced-NO bound enzyme and the reduced unligated enzyme. It is clear that the kinetic phase seen within the first 20 ms in the experiment with the reduced-NO-bound species is absent in the reaction with the unligated reduced enzyme. Spectral analysis shows that this fast phase involves c haem oxidation and synchronous formation of a transient at the d1 haem, occurring at approx. 100 s−1 (Figures 3A and 3B respectively). At longer times (>20 ms) the time courses of the two experiments are superimposable. Analysis of the spectra (Figure 3C) shows that the first observable species (at 1 ms) is different in the two experiments, whereas at longer times (>0.1 s) the spectra superimpose, suggesting that the same derivative is formed on the latter timescale. The final species, which is formed at ~20 s−1, is very similar to that previously observed in the reaction of reduced P. pantotrophus cytochrome cd1 with nitrite [24]. We conclude that during the first 20 ms nitrite displaces NO and reacts with the reduced protein populating an intermediate, whose formation is rate-limited by NO dissociation. The synchronous c haem oxidation indicates that this intermediate is formed as nitrite is reduced to NO; afterwards this species decays as when the initially reduced but unligated enzyme reacts with nitrite.

Figure 3 Dissociation of NO from reduced P. pantotrophus cytochrome cd1 induced by displacement with nitrite

Time course of the reaction of 3.5 μM cytochrome cd1, either reduced NO-bound (●) or reduced (△), mixed with 3 mM nitrite. The reaction was followed at wavelengths representative of the c haem (551 nm; A) and of the d1 haem (660 nm; B). In the reaction of the reduced-NO-bound enzyme, the decrease at 551 nm (●, A) indicates oxidation of the c haem, whereas the increase of absorption at 660 nm (●, B) indicates the formation of a transient species. The continuous lines represent fits of the data with a single- or double-exponential equation respectively. The arrows indicate 20 and 400 ms after mixing (see below). (C) Optical spectra collected at 1, 20 and 400 ms after mixing with nitrite the reduced NO-bound (continuous line) and the reduced cytochrome cd1 (broken line). Experimental conditions: 20 °C and 50 mM potassium phosphate (pH 7.0) containing 5 mM ascorbate and 13 μg/ml ascorbate oxidase.

The rate of dissociation of NO from reduced d1 haem complexed with apomyoglobin

The low affinity towards NO displayed by cytochrome cd1 could be due to the presence of the unique d1 haem and/or the structure of the haem-binding pocket. In order to assess the role of d1 haem in controlling the affinity for NO, we have measured the NO-dissociation-rate constant from ferrous d1 haem complexed with Mbd1.

As published previously, the d1 haem cofactor combines with apomyoglobin to form a stable complex; several ferric and ferrous derivatives have been spectroscopically characterized [27]. Steup and Muhoberac [27] suggested that the d1 haem was co-ordinated by a strong-field ligand that they assigned as a histidine residue; however, no static or kinetic characterization with NO was reported. Since the peculiarity of the reactivity with NO of the d1 haem in cytochrome cd1 resides in the fast release of the ligand, we have measured the dissociation rate of NO from the reduced Mbd1 complex. Sperm whale apomyoglobin was complexed with the d1 haem; the spectral properties of the complex were identical to those published previously [27]. Figure 4(A) shows the spectra of the ascorbate-reduced state and of the stable derivative obtained by reacting the latter species with 100 μM NO, with peaks at 445 and 644 nm. To determine the NO-dissociation-rate constant, the NO derivative was mixed anaerobically in a gas-tight cuvette with 5 mM sodium dithionite and ~500 μM CO (final concentrations) and the spectrum (400–700 nm) was recorded. The first spectrum (obtained 1.5 min a.m.) shows peaks at 452 and 642 nm (Figure 4A) suggesting that the CO-bound species had already formed. The spectrum did not change significantly during the time of the experiment (~30 min), but the shoulder at 590 nm broadened (results not shown), suggesting that two CO-bound species were formed which slowly re-equilibrated, in agreement with previous evidence [27]. Since the binding of CO, and thus dissociation of NO, could not be time-resolved in the spectrophotometer, the same reaction was followed by stopped-flow at 451 nm (Figure 4B). The reaction had a half-time of ~350 ms and the observed absorbance change corresponded to 92% of that expected (Figure 4A). The majority of the transition (Figure 4B) occurred at koff=2.0±0.4 s−1; this reaction was found to be CO-concentration independent (results not shown).

Figure 4 Dissociation of NO from ferrous d1 haem complexed with sperm whale apomyoglobin

The displacement of NO was measured in the presence of excess CO and dithionite [26] (at 20 °C and pH 6.9); under these conditions CO competes with NO for binding to the reduced d1 haem iron and the kinetics are rate-limited by NO dissociation. (A) Absorbance spectra of ascorbate-reduced Mbd1 (thin line) and of the corresponding NO-bound derivative (bold line). The latter species was mixed anaerobically in a gas-tight cuvette with 5 mM sodium dithionite as an NO scavenger and ~500 μM CO, and the spectrum (400–700 nm) was recorded (broken line). This spectrum did not change significantly over the following ~30 min. (B) Reaction of ascorbate-reduced Mbd1–NO bound species mixed with 1 mM CO and 5 mM dithionite (before mixing) in the stopped-flow apparatus. The observed time course at 451 nm (grey circles) represents the decay of the reduced NO-bound derivative; the time course was fitted with a single-exponential equation (continuous line). Experimental conditions: 20 °C and 40 mM potassium phosphate (pH 6.9) containing 5 mM ascorbate and 13 μg/ml ascorbate oxidase. The final protein concentrations (a.m.) were the same in (A) and (B).

DISCUSSION

The reaction of NO with haems plays an important role in controlling cellular physiopathological processes [79,17]. The fact that NO binds strongly to ferrous haem, albeit reversibly at extremely low free NO concentrations, imposes mechanistic constraints on many proteins that synthesize or bind NO. The enzymatic turnover of cytochrome cd1 nitrite reductase, in which nitrite is converted into NO, also demands the fast release of the product to avoid inhibition.

In the present paper, we aim at understanding whether the unique d1 haem, where nitrite reduction occurs, has a role in controlling the release of NO. Therefore we have measured the reactivity of the d1 haem with NO in the physiological context, i.e. the cytochrome cd1 enzyme from P. pantotrophus and in a non-native protein environment, provided by apomyoglobin. The results of the present study here provide novel mechanistic insights on the reaction mechanism of cytochrome cd1 and allow us to draw a common scheme that is applicable to this whole class of enzymes (Figure 5).

Figure 5 Proposed reaction mechanism for the catalytic cycle of cytochrome cd1 nitrite reductases

The catalytic cycle (A) is illustrated starting from the fully reduced enzyme (species 1), which binds the substrate nitrite yielding the Michaelis complex (species 2) [10,11,31]. Nitrite binding and reduction involves two conserved histidine residues [11], whose mutation was shown to affect the activity of P. aeruginosa cytochrome cd1 [35] (B for a scheme of the active site). Transfer of one electron from the d1 haem to nitrite produces NO bound to the oxidized haem iron (species 3) and a water molecule, which is released (oxidized haems are shown in grey). NO release from species 3 was not observed [14]. All available data suggest that NO dissociation occurs from the fully reduced NO-bound enzyme (species 5) in both P. pantotrophus and P. aeruginosa cytochromes cd1 (present paper and [15] respectively). Species 5 is formed via species 4 in two steps: transfer of an electron from the c haem to the d1 haem, followed by reduction of the c haem by an external electron donor. The physiological reductants of cytochrome cd1 are small cytochromes c (such as P. aeruginosa cytochrome c551 or P. pantotrophus cytochrome c550) or copper proteins (such as pseudoazurin in P. pantotrophus) [10], which specifically transfer the electron only to the c haem. After NO dissociation from species 5, the fully reduced enzyme is reformed and enters a new catalytic cycle. The present results does not allow us to rule out that NO dissociation may occur from the mixed valence nitrosylated protein (species 4); this would be difficult to measure due to the intrinsic redox instability of the mixed valence species [54]. The scheme described in the present paper refers to a single monomer of the enzyme; various results (see the Discussion section and [31]) imply the possibility of co-operativity between monomers during the catalytic cycle, but this is thought to affect the relative rates of catalytic steps in each monomer, not the steps themselves.

NO release is rapid in ferrous cytochrome cd1: relevance to the catalytic cycle

The major novel conclusion of the present study is that fast NO dissociation from the ferrous d1 haem is a common feature of cytochrome cd1 nitrite reductases. This conclusion is supported by the observation that NO dissociation from the reduced d1 haem of P. pantotrophus cytochrome cd1 is fast (Figure 2 and Table 1; koff=200 s−1 and 65 s−1 for the biphasic dissociation); these data agree well with previous results for the P. aeruginosa cytochrome cd1 where NO dissociation occurred at 70 s−1 [15] (Table 1). As a consequence of the greater koff from reduced d1 haem, the Kd for NO in cytochrome cd1 is orders of magnitude larger than in other haemoproteins [7,17,29] (Table 1).

The rapid NO dissociation from cytochrome cd1 nitrite reductases strongly suggests that NO release from ferrous d1 haem of cytochrome cd1 is a plausible step in the catalytic cycle. Therefore the mechanism previously proposed [15] and summarized in Figure 5 is a reasonable general catalytic cycle for all cytochromes cd1. Notably, it is also essentially the simplest reaction cycle that one can draw for these enzymes. This interpretation explains why the d13+-NO derivative of P. pantotrophus cytochrome cd1 is a very long-lived species in the absence of excess reductant [14,30]; it also agrees with previous ultra-fast (microsecond resolution) kinetic studies on the same enzyme showing that intramolecular c haem to d1 haem electron transfer triggered product release [31]. All available results indicate that P. pantotrophus cytochrome cd1 only works efficiently in the presence of substrate and electron donors [18,30,32,33], i.e. it gets ‘stuck’ with NO bound in the absence of reducing equivalents. It has long been a puzzle why an enzyme that catalyses a one-electron reduction requires two redox centres (the c and d1 haems). Electron donation to cytochrome cd1 from partner proteins occurs only via the c haem. A reaction mechanism where nitrite is first reduced to NO at the d1 haem, oxidizing that haem, but in which the d1 haem must be re-reduced (by the c haem) to allow NO release, would explain the need for two redox centres.

The results of the present study also suggest that, after the release of NO, the ferrous enzyme is competent to start a new catalytic cycle; therefore the reduced NO-bound P. pantotrophus cytochrome cd1 is not irreversibly inhibited. This is supported by the kinetic experiment in which the fully reduced-NO-bound cytochrome cd1 is mixed with nitrite: in this experiment, a complex reaction involving both electron transfer between the two redox centres (as seen by the oxidation of c haem, Figure 3A) and chemistry at the d1 haem (Figure 3B) is observed. The relevant observation is that an intermediate species (populated at 100 s−1) could be observed with spectroscopic features (around 660 nm) similar to those observed previously when the reduced P. pantotrophus cytochrome cd1 was reacted with nitrite in a freeze-quench experiment [31]. In the kinetic experiment, the formation of this transient species (Figure 3) supports the idea that, after the release of NO, the ferrous enzyme can start a new catalytic cycle. The ability of the P. pantotrophus cytochrome cd1 to release NO without undergoing significant product inhibition is also consistent with earlier steady-state experiments in which a linear rate of NO production occurred up to tens of micromolar concentrations [34].

Another novel mechanistic insight into the reaction with nitrite catalysed by cytochrome cd1 is the assignment of the rate-determining step. In P. pantotrophus cytochrome cd1 the slower phase of NO dissociation (koff=65 s−1; Figure 2) occurs with a rate that is very similar to the overall kcat (72 s−1 per d1 haem [18]), indicating that this process probably represents the rate-limiting step of the reaction. However, for P. aeruginosa cytochrome cd1 the rate-determining step is probably the intramolecular electron transfer from the c haem to the d1 haem, since the reported electron transfer rate (1–4 s−1) is similar to the turnover number (6 s−1) [5,20]. Thus in P. aeruginosa cytochrome cd1, the relatively slow formation of ferrous d1 haem–NO would be followed by a more rapid dissociation of NO.

Although more speculative, we note that analysis of NO dissociation from P. pantotrophus cytochrome cd1 may also suggest that, in this enzyme, the two monomers dissociate NO at different rates (NO dissociation is biphasic and each kinetic phase corresponds to ~50% of the total absorbance change). These results are in agreement with previous kinetic experiments with nitrite, which were interpreted in terms of anti-co-operativity of the two monomers [31]. Even though the mechanism of such anti-co-operativity is yet to be explained, intrinsic asymmetry of cytochrome cd1 is suggested and supported by evidence obtained from ligand-binding experiments [28,36], a pre-steady-state kinetic study with nitrite [31], intramolecular electron transfer [20,37] and potentiometric titrations [38,39]. Moreover, in the many X-ray crystal structures of P. pantotrophus cytochrome cd1 and its ligand-bound derivatives, the two monomers are always different [11,28,40].

Role of the d1 haem in NO release

The other novel outcome of the present study is that fast NO dissociation from ferrous haem iron is largely due to the presence of the unique d1 haem cofactor. The role of ferrous d1 haem in controlling NO dissociation was analysed in a different protein environment, i.e. the d1 haem complexed with apomyoglobin. In ferrous Mbd1, we observe a single NO-dissociation process occurring at 2 s−1 (Figure 4), a rate which is approx. 3–5 orders of magnitude greater than that measured for native ferrous myoglobin which contains b haem (Fe-protoporphyrin IX) (Table 1) [7]. Given these results and taking into account that both Mbd1 and cytochrome cd1 have a common proximal histidine residue ligand, we conclude that the d1 haem is designed to release NO relatively rapidly. It is known that NO is a good π-acceptor and, in general, an important contribution to the strength of NO binding to the ferrous haem iron can arise from electron donation from the t2g orbitals on the iron. The two electron withdrawing carbonyl groups on the d1 haem ring (see Supplementary Figure S2) might weaken such electron donation to NO.

There are previous observations indicating that the ring structure of the d1 haem confers on the Fe atom other properties that are distinct from those of protoporphyrin IX. The ferrous state of d1 haem binds anions, including the substrate nitrite and the inhibitor cyanide, unusually strongly and, strikingly, more strongly than the ferric state (Table 1) [32]. This also correlates with the presence of two electron-withdrawing carbonyl groups on the d1 haem ring. Another feature of the d1 haem ring is that in the ferric state this cofactor shows a peculiar ordering of the energy levels of the d orbitals [41]; the latter property is shared with other haems in which the porphyrin ring is partially saturated (haem d in Escherichia coli and sirohaem) [42,43]. Clearly, a range of advanced spectroscopic and theoretical studies, outside the scope of the present work, needs to be applied to fully understand how the d1 haem ring is tuned, in the ferrous state, to bind anions unusually strongly, but NO unusually weakly, in order to catalyse the reduction of nitrite. The consequence is that whereas a priori one might have expected a haem enzyme to bind nitrite to, and release NO from, the ferric state, the opposite strategy has evolved.

The higher NO-dissociation rate measured for cytochromes cd1 (up to 200 s−1) (as compared with Mbd1) may in principle reflect to some extent a feature in the d1 haem pocket that optimizes NO release. A possible mechanism may include the conformational change seen in P. pantotrophus cytochrome cd1 and thought to be implicated in catalysis [44]; this change involves Tyr197 on the proximal side of the d1 haem pocket and the hinge region (residues 132–136) connecting the c and the d1 haem domains. However, mutation of these residues (Y197F and P134F/P135F) has no significant effect on the enzymatic activity [55] and thus this conformational change is unlikely to control NO release. Therefore the (relatively small) contribution of the protein moiety to the control of NO release is yet to be identified at the molecular level.

Conclusion

The present study allows the formulation of a straightforward, unified reaction mechanism for cytochrome cd1 nitrite reductases, a class of enzymes studied for over 50 years (Figure 5). The experiments reported in the present paper strongly suggest that fast NO dissociation from ferrous haem iron is largely due to the presence of the unique d1 haem cofactor, and that enzyme inhibition by NO is unlikely to be relevant under physiological conditions. The evidence underscores that the reactivity of porphyrins with NO can be modulated very extensively by the functional groups present on the haem macrocycle, information of general significance in light of the emerging important biological functions of nitrite as a source of NO under hypoxic conditions [45].

AUTHOR CONTRIBUTION

Serena Rinaldo and Katharine A. Sam performed most of the experiments on P. pantotrophus cytochrome cd1 with the necessary assistance of Nicoletta Castiglione and Valentina Stelitano. Alessandro Arcovito performed the laser photolysis experiments; Serena Rinaldo and Valentina Stelitano performed the experiments on Mbd1. The results analysis was carried out by Serena Rinaldo with the contribution of Francesca Cutruzzolà. Francesca Cutruzzolà directed the study and together with Stuart J. Ferguson, James W.A. Allen and Maurizio Brunori conceptualized the project. Francesca Cutruzzolà wrote the manuscript, and all authors contributed to the final version(s) prior to publication. Maurizio Brunori, James W.A. Allen, Francesca Cutruzzolà and Stuart J. Ferguson wrote the grants that provided funding.

FUNDING

This work was supported by the Ministero della Università of Italy [grant numbers 20074TJ3ZB and RBRN07BMCT]; and the University of Rome La Sapienza to F.C. and M.B. K.A.S. thanks St Edmund Hall, Oxford for the award of a W.R. Miller Junior Research Fellowship. The Biotechnology and Biological Sciences Research Council supported J.W.A.A. through a David Phillips Fellowship [grant number BB/D019753/1] and S.J.F. through project grants.

Abbreviations: a.m., after mixing; Mbd1, haem d1–apomyoglobin; RZ, Reinheitzahl

References

View Abstract