Kinetics of Nitrite Reduction and Peroxynitrite Formation by Ferrous Heme in Human Cystathionine β-Synthase

Abstract

Cystathionine β-synthase (CBS) is a pyridoxal phosphate-dependent enzyme that catalyzes the condensation of homocysteine with serine or with cysteine to form cystathionine and either water or hydrogen sulfide, respectively. Human CBS possesses a noncatalytic heme cofactor with cysteine and histidine as ligands, which in its oxidized state is relatively unreactive. Ferric CBS (Fe(III)-CBS) can be reduced by strong chemical and biochemical reductants to Fe(II)-CBS, which can bind carbon monoxide (CO) or nitric oxide (NO^•), leading to inactive enzyme. Alternatively, Fe(II)-CBS can be reoxidized by O₂ to Fe(III)-CBS, forming superoxide radical anion (O₂^˙̄). In this study, we describe the kinetics of nitrite (NO₂⁻) reduction by Fe(II)-CBS to form Fe(II)NO^•-CBS. The second order rate constant for the reaction of Fe(II)-CBS with nitrite was obtained at low dithionite concentrations. Reoxidation of Fe(II)NO^•-CBS by O₂ showed complex kinetic behavior and led to peroxynitrite (ONOO⁻) formation, which was detected using the fluorescent probe, coumarin boronic acid. Thus, in addition to being a potential source of superoxide radical, CBS constitutes a previously unrecognized source of NO^• and peroxynitrite.

Introduction

Cystathionine β-synthase (CBS)³ is a key enzyme in the metabolism of sulfur amino acids in mammals. It catalyzes the pyridoxal 5′-phosphate (PLP)-dependent condensation of serine with homocysteine to give cystathionine and water in the first step of the trans-sulfuration pathway that leads to cysteine. CBS also condenses homocysteine and cysteine to form cystathionine and H₂S, a newly recognized physiological modulator with a broad range of effects in the cardiovascular, gastrointestinal, and nervous systems (1–3). An increased level of homocysteine in plasma constitutes a risk factor for cardiovascular diseases and is correlated with neural tube defects and Alzheimer's disease (4–7). Inborn errors of metabolism caused by mutations in CBS represent the most common cause of inherited hyperhomocysteinemia (8). Full-length human CBS has a subunit molecular mass of ∼63 kDa and exists as a homodimer or as higher order oligomers (9). Each monomer has a modular organization that comprises an N-terminal domain that binds heme, followed by a catalytic domain that binds PLP, and a C-terminal regulatory domain that binds the allosteric activator, S-adenosyl-l-methionine (AdoMet). CBS activity increases in the presence of AdoMet or by limited proteolysis that results in the separation of the regulatory domain and formation of a truncated catalytic core (10–13).

The b-type heme in human CBS has a low spin and six-coordinate iron in both the ferric and ferrous states (14), with Cys⁵² and His⁶⁵ serving as the axial ligands (10, 12, 15, 16). The heme environment is conserved in CBS among several eukaryotic species but does not resemble any other known heme protein (17, 18). Although a direct role for the heme in the PLP-dependent enzymatic mechanism was excluded (19), a regulatory role has been proposed on the basis of the observation that perturbations in the heme coordination environment can modulate enzyme activity through shifts in the PLP tautomeric equilibrium. Communication between the heme and PLP sites is transmitted via an α-helix that interacts at one end with the Cys⁵² heme ligand via Arg²⁶⁶ and at the other end with PLP via two conserved threonine residues (Thr²⁵⁷ and Thr²⁶⁰) (18, 20–24).

Ferric CBS is quite stable and inert to exogenous ligands (16, 25). It is sensitive, however, to the thiophilic cation mercury(II) and to the strong oxidant and nitrating agent peroxynitrite (ONOO⁻)⁴ (26, 27). Despite the fairly low heme reduction potential of −0.350 V (21), CBS reduction can potentially be accomplished in vivo by enzymatic oxidoreductases. We have previously demonstrated that the cytosolic diflavin-containing methionine synthase reductase, in the presence of NADPH as electron donor, is able to reduce the heme in CBS (28, 29). In the reduced form, Fe(II)-CBS can react with O₂, and it can also bind CO and NO^•. The reaction of Fe(II)-CBS with O₂ occurs rapidly with a second order rate constant of (1.13 ± 0.05) × 10⁵ m⁻¹ s⁻¹ (pH 7.4, 25 °C), leading to the formation of Fe(III)-CBS and superoxide radical anion (O₂^˙̄) (30). Fe(II)-CBS can bind CO, which replaces the axial cysteine ligand, forming Fe(II)CO-CBS. The observed rate constants increase hyperbolically with CO concentration, reaching limiting values of 0.012–0.017 s⁻¹ at 25 °C (31, 32) or 0.0031 s⁻¹ at 1 mm CO (28). Relevantly, Fe(II)CO-CBS displays decreased catalytic activity, with a K_i for CO of ∼5.6 μm, which might be within a physiologically relevant concentration range (31, 33). Exposure of Fe(II)CO-CBS to O₂ results in its rapid reoxidation to Fe(III)-CBS concomitant with recovery of enzyme activity (29). In contrast to CO, NO^• binding to Fe(II)-CBS results in a pentacoordinate iron-nitrosyl Fe(II)NO^•-CBS species with a Soret maximum at 394 nm, indicating the loss of both endogenous axial ligands (34). It was recently reported that NO^• binds with a K_d ≤ 0.23 μm and with kinetic rate constants k_on of ∼8 × 10³ m⁻¹ s⁻¹ and k_off of 0.003 s⁻¹ (32). In addition, Fe(II)-CBS can slowly decay to an inactive Fe(II)424-CBS form with a Soret maximum at 424 nm, in which the cysteine is replaced by an unidentified neutral ligand (35, 36). Although Fe(II)424-CBS formation is promoted under nonphysiological conditions such as high temperature (36), the biological significance of this process is not clear. We have demonstrated recently that Fe(II)-CBS has the ability to reduce nitrite (NO₂⁻) to NO^•, leading to Fe(II)NO^•-CBS formation, suggesting a possible new role for CBS in NO^• signaling (37).

In this study, we have characterized the kinetics of nitrite reduction by the ferrous form of human CBS using low dithionite concentrations. We have also investigated the kinetics of Fe(II)NO^•-CBS reoxidation by O₂ and examined the formation of peroxynitrite in the process. Overall, our results provide new insights into the kinetics of heme-mediated regulation of human CBS.

Experimental Procedures

Materials

All reagents were purchased from Sigma unless otherwise specified. Coumarin boronic acid (CBA) and peroxynitrite were from Cayman Chemical (Ann Arbor, MI). CBA was first dissolved in DMSO and then added directly to the samples to obtain the desired final concentration. The concentration of peroxynitrite was determined at 302 nm (ϵ = 1,670 m⁻¹ cm⁻¹ (38)), and the stock solution was diluted in 10 mm NaOH immediately before use. Recombinant full-length wild-type and T257I human CBS and the truncated ΔC143-CBS variant were purified as described previously (24, 39). Protein concentration was determined by the Bradford method using bovine serum albumin as standard (40) and by using the reported extinction coefficients for CBS protein and heme (30). Reduction of Fe(III)-CBS was performed under a nitrogen atmosphere by the addition of known aliquots of sodium dithionite (Na₂S₂O₄) and confirmed by the appearance of the 449-nm Soret peak in the UV-visible absorption spectrum. Dithionite stock solutions were prepared in nitrogen-degassed 0.1 m NaOH and quantified by ferricyanide reduction (ϵ₄₂₀ = 1,020 m⁻¹ cm⁻¹) (41) assuming a 2:1 (ferricyanide:dithionite) stoichiometry, in agreement with the actual reductant being the dithionite dissociation product SO₂^˙̄. Solutions of sodium nitrite were prepared under anaerobic conditions and kept under a positive pressure of argon.

Kinetics of Fe(II)-CBS Reaction with Nitrite

An anaerobic solution of Fe(III)-CBS (1–10 μm) in phosphate buffer (0.1 m, pH 7.4, 0.1 mm diethylenetriaminepentaacetic acid (DTPA)) was reduced with low dithionite concentrations (10–15 μm) in an anaerobic chamber (Vacuum Atmospheres Co., Hawthorne, CA; 95/5 N₂/H₂ and <0.3 ppm O₂). The resulting Fe(II)-CBS solution was then mixed with a series of sodium nitrite concentrations at 37 °C, and the UV-visible absorption spectra were recorded outside the anaerobic chamber. Alternatively, stoppedflow mixing was performed using a stopped flow spectrophotometer (SX.MV18; Applied Photophysics) equipped with a diode array detector, placed inside an anaerobic chamber. To avoid photochemical artifacts, the amount of incident white light was limited by adjusting the slit to 1 mm.

Kinetics of Fe(II)NO^•-CBS Reoxidation by O₂

Fe(II)NO^•-CBS solutions were prepared inside the anaerobic chamber. An anaerobic solution of Fe(III)-CBS (10 μm) in phosphate buffer (0.1 m, pH 7.4, with 0.1 mm (DTPA) was reduced with dithionite (90 μm), and heme reduction was monitored spectrophotometrically by a shift in the Soret peak from 428 to 449 nm. Sodium nitrite was added, and Fe(II)NO^•-CBS formation was confirmed by a shift of the Soret peak to 394 nm. Excess dithionite and nitrite were removed by gel filtration using Micro Bio-Spin columns (Bio-Rad). The reaction mixtures containing the anaerobic solution of Fe(II)NO^•-CBS were placed in cuvettes, sealed with a rubber septum, and purged with certified 100% O₂ (final concentration, ∼1.25 mm), air (21% O₂), or mixed with solutions of increasing O₂ concentrations obtained by mixing different volumes of anaerobic and O₂-equilibrated water. The experiments were performed at 25.0 ± 0.1 °C.

Peroxynitrite Formation by Fe(II)NO^•-CBS Exposed to O₂

The generation of peroxynitrite was determined using the boronate-based probe, CBA, and monitored by the increase in fluorescence intensity (excitation at 330 nm, emission at 450 nm) because of formation of the corresponding hydroxylated compound, 7-hydroxycoumarin (COH) (42). The procedure was calibrated using authentic peroxynitrite, and fluorescence spectroscopy was similar to that previously reported (42), with the intensity of the excitation and emission bands increasing linearly with peroxynitrite concentrations (data not shown). Typical reaction mixtures were prepared in the anaerobic chamber and contained Fe(III)-CBS (15 μm) in phosphate buffer (0.1 m, pH 7.4, with 0.1 mm DTPA) to which dithionite (100 μm) was added. After heme reduction, sodium nitrite (4 mm) was added, and Fe(II)NO^•-CBS formation was confirmed spectrophotometrically, followed by gel filtration to avoid possible interference from nitrite or dithionite. Then Fe(II)NO^•-CBS was placed in a fluorescence cuvette, and CBA was added. The cuvette was sealed with a rubber septum, and outside the anaerobic chamber, water-equilibrated with O₂ (100%) was added. Final concentrations were ∼ 7.5, 20, and 625 μm for Fe(II)NO^•-CBS, CBA, and O_2, respectively. Florescence was registered at 25 °C in a Shimadzu 5031 PC spectrofluorimeter. To minimize photochemical artifacts, the incident light was diminished by decreasing the slit width to 1.5 mm and by recording the fluorescence every 10 min. At the end of the experiment, the cuvette was placed in the spectrophotometer to record the UV-visible spectrum of the heme. Controls showed that Fe(II)NO^•-CBS and Fe(III)-CBS did not interfere with COH fluorescence intensity; neither did dithionite (below 600 μm). In some experiments, cysteine (5 mm), glutathione (20 mm), superoxide dismutase (875 units/ml, ∼ 5.5 μm), or catalase (0.1 μm) was added.

Simulations and Data Analysis

Kinetic traces were analyzed and fitted using OriginPro 8. Data from the stopped flow experiments were fitted using the Pro-Data Viewer Software (Applied Photophysics). Computer-assisted kinetic simulations were performed with Gepasi (43). All experiments were repeated at least three times, and representative data are shown.

Results

Kinetics of Fe(II)-CBS Reaction with Nitrite

When an anaerobic solution of CBS was reduced with dithionite and mixed with nitrite, the characteristic Soret maximum of Fe(II)-CBS at 449 nm was converted to the five-coordinate Fe(II)NO^•-CBS species, confirmed by the appearance of a 394 nm peak as reported previously (32, 34, 37) (Fig. 1A). This is consistent with the reaction of Fe(II)-CBS with nitrite to form Fe(III)-CBS and NO^•. In the presence of reducing equivalents, this is followed by the reaction of Fe(II)-CBS with NO^• to yield Fe(II)NO^•-CBS (reactions 1–3), in analogy to the mechanism proposed for globins (44–47).

This reaction sequence implies that the ferric heme, once formed, should be reduced back to the ferrous state to react with NO^• to form Fe(II)NO^•-CBS. Accordingly, when the concentration of dithionite was lower than CBS so that the reductant was limiting, Fe(III)-CBS was the final product, and Fe(II)NO^•-CBS formation was not detected (Fig. 1B). The reaction of ferric CBS with NO^• is very slow as reported previously (32) and not expected to be a contributing factor under our experimental conditions.

FIGURE 1.

Effect of dithionite concentration on Fe(II)-CBS reaction with nitrite. A, Fe(III)-CBS (10 μm) was reduced with dithionite (90 μm) in phosphate buffer (0.1 m, pH 7.4, with DTPA 0.1 mm) inside an anaerobic chamber. Then sodium nitrite (1 mm) was added, the cuvette was sealed, and UV-visible absorption spectra were recorded outside the anaerobic chamber immediately after mixing (black trace) and every minute (gray traces) for 35 min (dotted trace). The arrows indicate the direction of the absorbance change over time. B, Fe(III)-CBS (10 μm) was reduced with a substoichiometric dithionite concentration (8 μm) and a similar procedure as described above was then followed. Selected UV-visible absorption spectra recorded immediately after mixing (black trace); at 1, 2, 3, 4, and 5 min (gray traces); and at 12 min (dotted trace) are shown. C, an anaerobic solution of Fe(III)-CBS (5 μm) was reduced with dithionite (50 μm) and mixed with sodium nitrite (10 mm) in the presence of various dithionite concentrations in sodium phosphate buffer (0.1 m, pH 7.4, with 0.1 mm DTPA) at 37 °C. The lines represent the kinetic traces of Fe(II)-CBS decay at 449 nm obtained with different dithionite concentrations: 75 μm (solid line), 275 μm (dotted line), and 3.0 mm (dashed line). The concentrations are those after mixing. The experiments were repeated at least twice, and representative data are shown.

An effect of dithionite on the reaction of Fe(II)-CBS with nitrite was observed. The time course of the reaction changed when the concentration of dithionite increased from 75 μm to 3.0 mm (Fig. 1C). A likely explanation for this observation is interference from the reaction between NO^• and dithionite (k = ∼1.4 × 10³ m⁻¹ s⁻¹) (48). Computer-assisted simulations were performed using the reactions and rate constants listed in Table 1, and initial concentrations of the reagents, Fe(II)-CBS, dithionite, and nitrite as used in the experiments. Although the experimental and simulated results were not identical, the simulations provided insights into the reaction kinetics. The interference of dithionite in the reaction was revealed by exclusion of the direct reaction between dithionite and NO^• from the simulation (Fig. 2, A and B). It is important to note that dithionite does not effectively reduce nitrite (49). Dithionite can compete with Fe(II)-CBS for NO^• because the formation of Fe(II)NO^•-CBS is a relatively slow process (32). Accordingly, when a 10-fold higher value for the k_on for the reaction between Fe(II)-CBS and NO^• was used in the simulation, the interference due to dithionite decreased (Fig. 2C).

TABLE 1.

Reactions involved in the anaerobic reaction of dithionite-reduced CBS with nitrite

Reaction	Rate constanta	References
S2O42− ⇌ 2 SO2˙̄	2.5 s−1 (f),b 1.8 × 109 m−1 s−1 (r)	Refs. 72 and 73
Fe(III)-CBS + SO2˙̄ → Fe(II)-CBS + SO2	1.58 × 105 m−1 s−1	Refs. 28 and 74
Fe(II)-CBS + NO2− + H+ → Fe(III)-CBS + NO• + OH−	0.66 m−1 s−1	This work and Ref. 37
Fe(II)-CBS + NO• ⇌ Fe(II)NO•-CBS	8 × 103 m−1 s−1 (f), 3 × 10−3 s−1 (r)	Ref. 32
NO• + S2O42− → products	1.4 × 103 m−1 s−1	Ref. 48
Fe(II)-CBS + HSO3−/SO32− → Fe(III)-CBS + SO2˙̄ + OH−	17 m−1 s−1	Ref. 32
SO2 + H2O ⇌ HSO3−/SO32− + H+	3.4 × 106 s−1 (f); 2.0 × 108 m−1 s−1 (r)	Ref. 75

^a Rate constants are reported at pH 7.4 and 25 °C except for the homolysis of dithionite, at pH 6.5 (72, 73) and for the reaction of Fe(II)-CBS with nitrite, reported at 37 °C (this work). For simulations, the rate constants at 25 °C were multiplied by 2.

^b (f) and (r) represent forward and reverse kinetic constants, respectively.

FIGURE 2.

Simulated profile of Fe(II)-CBS reaction with nitrite. Computer-assisted simulations of the transient concentrations of Fe(II)-CBS (solid line), Fe(III)-CBS (dotted line), and Fe(II)NO^•-CBS (dashed line). Simulations were performed using as initial concentrations of Fe(II)-CBS, dithionite, and nitrite 5 μm, 15 μm, and 10 mm, respectively. A, simulation obtained using the reactions and rate constants listed in Table 1. B, simulations obtained when the direct reaction between dithionite and NO^• was omitted from the list of reactions. C, simulation obtained when the reaction between dithionite and NO^• was included, but a 10-fold higher k_on value for NO^• binding to Fe(II)-CBS was used.

To minimize the interference of dithionite, the kinetics of Fe(II)-CBS reaction with nitrite and Fe(II)NO^•-CBS formation were further characterized using relatively low dithionite concentrations (∼10–15 μm) but in excess over CBS. In the UV-visible spectra, clear isosbestic points were observed at 410 and 587 nm up to 300 s (Fig. 3A). At longer times, the isosbestics were lost, and the absorbance at 394 nm decayed slowly. The Fe(III)-CBS intermediate was not detected because it is reduced back to Fe(II)-CBS by dithionite. Using the rate constants shown in Table 1, kinetic simulations showed that during the time course of the reaction, the concentration of Fe(III)-CBS would remain low, representing <10% of total CBS, whereas the decay of Fe(II)-CBS is paralleled by an increase in Fe(II)NO^•-CBS. The reaction course during the first 300 s monitored by the decay of the signal at 449 nm characteristic of Fe(II)-CBS or that at 394 nm because of formation of Fe(II)NO^•-CBS were fitted to a single exponential plus straight line function (Fig. 3A, inset). The observed rate constants at 449 nm increased linearly with nitrite concentration, showing no evidence of saturation behavior up to 50 mm nitrite (Fig. 3B). From the slope of the plot, the second order rate constant for full-length CBS was estimated to be 0.66 ± 0.03 m⁻¹ s⁻¹ at pH 7.4 and 37 °C. This value for the second order rate constant between nitrite and Fe(II)-CBS is an approximation because of the multiple reactions required to obtain Fe(II)NO^•-CBS from Fe(II)-CBS and nitrite. In the presence of the allosteric activator AdoMet, the second order rate constant obtained from fits at 449 nm is 1.32 ± 0.06 m⁻¹ s⁻¹, ∼2-fold greater than in its absence and in agreement with a previous report (37). Interestingly, the T257I mutant of CBS exhibited a >10-fold increase in the value of the second order rate constant than wild-type CBS in the absence of AdoMet (Table 2). For truncated CBS, we also determined a higher reactivity with nitrite compared with full-length CBS, with estimates for the second order rate constant of 2.37 ± 0.07 m⁻¹ s⁻¹.

FIGURE 3.

Kinetics of Fe(II)-CBS reaction with nitrite. A, UV-visible absorption spectra obtained in a stopped flow instrument after mixing full-length CBS (reduced with dithionite) with sodium nitrite in phosphate buffer (0.1 m, pH 7.4, 0.1 mm DTPA) at 37 °C. The concentrations of CBS, dithionite, and nitrite after mixing were 1 μm, ∼10 μm, and 10 mm, respectively. Spectra were collected from 0.5 to 500 s every 0.5 s. For clarity, only selected spectra are shown. The arrows indicate the direction of the absorbance change over time. Inset, absorbance at 394 nm (open circles) and 449 nm (closed circles) were plotted against time up to 300 s. The solid lines represent the fits to single exponential plus a straight line functions. B, an anaerobic solution of Fe(III)-CBS (5 μm) in phosphate buffer (0.1 m, pH 7.4, 0.1 mm DTPA) was reduced with low dithionite concentrations (15 μm). The resulting Fe(II)-CBS solution was then mixed with increasing sodium nitrite concentrations. The absorbance at 449 nm was recorded outside the anaerobic chamber at 37 °C and fitted to a single exponential plus straight line function to obtain the observed rate constants (k_obs). The results are represented as the means ± standard deviation (n ≥ 3).

TABLE 2.

Comparison of reaction rate constants of mammalian hemeproteins with nitrite

Mammalian hemeproteins	Rate constanta	References
	m−1 s−1
CBS full length	0.66 ± 0.03	This work and Ref. 37
CBS full length + AdoMet	1.32 ± 0.06	This work
CBS full length T257I	7.70 ± 0.62b	This work
CBS truncated	2.37 ± 0.07	This work
Hemoglobin (human, T state)	0.12	Ref. 76
Hemoglobin (human, R state)	6	Ref. 76
Myoglobin (horse)	2.9	Ref. 47
Neuroglobin (human)	0.26	Ref. 47
Cytoglobin (human)	0.14	Ref. 77

^a Rate constants reported at pH 7.4 and 37 °C, except for myoglobin and cytoglobin reported at pH 7 and 25 °C.

^b Rate constant was determined at 394 nm.

Kinetics of Fe(II)NO^•-CBS Reaction with O₂

As reported previously, air exposure of Fe(II)NO^•-CBS results in reoxidation to Fe(III)-CBS concomitant with recovery of enzyme activity (32, 37). When a solution of Fe(II)NO^•-CBS was mixed with excess O₂, the nitrosylated enzyme with the characteristic absorption peak at 394 nm slowly converted to Fe(III)-CBS with a maximum at 428 nm (Fig. 4A). Clear isosbestic points were not observed in the UV-visible spectra. The increase in absorbance at 428 nm was biphasic and fitted a double exponential function (Fig. 4B, inset). The first phase had k_obs values on the order of 10⁻³ s⁻¹ that depended linearly on oxygen concentration, leading to an apparent second order rate constant of 2.0 ± 0.4 m⁻¹ s⁻¹. The second phase had a k_obs value of ∼2 × 10⁻⁴ s⁻¹ and was constant within the range of O₂ concentrations (Fig. 4B). In some experiments, an initial fast phase could be detected. This kinetic phase has been reported previously (32) but presented high variability and was not pursued further. It is possible that this fast phase was due to the presence of a small amount of reduced CBS without bound NO^• that auto-oxidized quickly upon oxygenation. Under anaerobic conditions, Fe(II)NO^•-CBS decayed spontaneously but extremely slowly to the ferric species, with a half-life of ∼11 h at room temperature (data not shown).

FIGURE 4.

Oxidation of Fe(II)NO^•-CBS by O₂. A, an anaerobic solution of Fe(III)-CBS (40 μm) was reduced with dithionite (200 μm) and mixed with sodium nitrite (4 mm) in phosphate buffer (0.1 m, pH 7.4, with 0.1 mm DTPA) inside the anaerobic chamber. The resulting solution was gel-filtered to remove excess dithionite and nitrite, and after removing the sample from the chamber, Fe(II)NO^•-CBS was mixed with an equal volume of a solution containing O₂ obtained by equilibrating buffer with 100% O₂ (final concentration, 50% O_2, ∼0.625 mm). The cuvette headspace was also bubbled with 50% O₂. UV-visible absorption spectra were registered every 4 min up to 250 min at 25 °C. B, an anaerobic solution of Fe(III)-CBS (10 μm) was reduced with dithionite (90 μm) in phosphate buffer (0.1 m, pH 7.4, with DTPA 0.1 mm) inside the anaerobic chamber. Then sodium nitrite (1 mm) was added, and after removing the sample from the anaerobic chamber, the resulting Fe(II)NO^•-CBS solution was bubbled ∼1 min with 100% O₂ (1.25 mm), air (21% O₂) or mixed with solutions of increased O₂ concentrations obtained from mixing known volumes of anaerobic and O₂-equilibrated water at 25 °C. The k_obs (s⁻¹) values, determined from the fit of the increase of Fe(III)-CBS at 428 nm to a double exponential function, were plotted against O₂ concentrations (0.263–1.25 mm) for the first (closed circles) and second (open circles) phases at each O₂ concentration. Inset, kinetic trace of the reaction monitored by the increase of absorbance at 428 nm. The line represents the best fit to double exponential function. The experiment was repeated at least three times, and representative data are shown.

Peroxynitrite Formation by Fe(II)NO^•-CBS Exposed to O₂

The formation of peroxynitrite was evaluated using the boronate probe CBA, which reacts fast with peroxynitrite (k = 1.1 × 10⁶ m⁻¹ s⁻¹) to yield COH as the major product, which can be followed by fluorescence spectroscopy (42). As shown in Fig. 5A, upon exposure of Fe(II)NO^•-CBS to O₂ in the presence of CBA, an increase in fluorescence intensity over time corresponding to COH accumulation was observed, indicating peroxynitrite formation. The kinetics of peroxynitrite formation were relatively slow and, with a global k_obs of (1.5 ± 0.4) × 10⁻³ s⁻¹, mirrored the kinetics of Fe(II)NO^•-CBS decay to Fe(III)-CBS (0.8–2.8 × 10⁻³ s⁻¹). Controls lacking CBS did not result in significant COH formation. Artifacts derived from reactions of dithionite were avoided by passing the Fe(II)NO^•-CBS solution over a gel filtration column before CBA addition and O₂ exposure. Furthermore, photochemical artifacts were minimized by reducing the slit width and by taking measurements every 10 min rather than continuous exposure of the sample to incident light. The presence of catalase did not alter the yield of COH, confirming that H₂O₂ is not involved in its formation, which is consistent with the much greater rate constant for the reaction of catalase (0.1 μm) with H₂O₂ (∼10⁷ m⁻¹ s⁻¹) versus CBA (20 μm) with H₂O₂ (1.5 m⁻¹ s⁻¹). In the presence of superoxide dismutase (5.5 μm), which reacts with O₂^˙̄ with a rate constant of 2 × 10⁹ m⁻¹ s⁻¹ (50), COH fluorescence was partially inhibited (∼ 40%). The inhibition increased further to ∼60%, in the presence of both superoxide dismutase and catalase. This is expected because catalase reacts with the H₂O₂ formed from superoxide dismutase-catalyzed disproportionation of O₂^˙̄. Taken together, the results suggest that diffusible O₂^˙̄ is implicated in the formation of most but not all peroxynitrite.

FIGURE 5.

Detection of peroxynitrite produced by reaction of Fe(II)NO^•-CBS with O₂ in the presence of CBA. An anaerobic solution of Fe(III)-CBS (15 μm) was reduced with dithionite (100 μm) and mixed with sodium nitrite (4 mm) in phosphate buffer (0.1 m, pH 7.4, with 0.1 mm DTPA). After 60 min, the solution of Fe(II)NO^•-CBS was purified using a gel filtration column and mixed with CBA and water-equilibrated with O₂ (100%). The final concentrations were ∼7.5, 20, and 625 μm for Fe(II)NO^•-CBS, CBA, and O₂, respectively. The same procedure was repeated in the presence of cysteine (5 mm), glutathione (20 mm), catalase (0.1 μm), superoxide dismutase (SOD, 5.5 μm), and both superoxide dismutase and catalase. A control lacking CBS, containing cysteine (5 mm) in buffer was also included. A, the increase in the fluorescence intensity emission corresponding to COH formation was recorded every 10 min at 25 °C (λ_ex = 330 nm, λ_em = 450 nm) to minimize the incident light. B, the same procedure as described above, but the fluorescence intensity emission was recorded continuously to 6000 s. The experiment was repeated at least three times, and representative data are shown.

In the presence of 20 mm glutathione or 5 mm cysteine, which scavenge peroxynitrite with rate constants of 6 × 10² m⁻¹ s⁻¹ and 3.5 × 10³ m⁻¹ s⁻¹, respectively, at 25 °C and pH 7.4 (51, 52), COH formation was not inhibited (Fig. 5A). Controls showed that COH fluorescence observed with cysteine alone was negligible and that cysteine and glutathione, at the mentioned concentrations, inhibited COH fluorescence from authentic peroxynitrite to ∼52 and 60%, respectively, as expected from kinetic considerations. Thus, the lack of inhibition by cysteine and glutathione in the case of Fe(II)NO^•-CBS supports the possibility that the peroxynitrite formed was nitrogen-bound to the heme in CBS and was able to perform a nucleophilic attack of the terminal peroxidatic oxygen on boron but was unable to react with thiols as an electrophile. In contrast, when fluorescence was continuously recorded instead of being measured every 10 min, an inhibitory effect (∼45%) of cysteine on COH formation was observed (Fig. 5B). This is likely to have been triggered by a photochemically induced dissociation of NO^• from Fe(II)NO^•-CBS, leading to peroxynitrite formation in the presence of O₂^˙̄.

Discussion

In the present study, we have characterized the kinetics of nitrite reduction by Fe(II)-CBS that leads to formation of Fe(II)NO^•-CBS (pathways a and b in Scheme 1). We found that the dithionite concentration can affect the process of nitrite reduction by CBS. However, dithionite has been used at millimolar concentrations with other hemeproteins, apparently without this complication. We attribute this difference to the fact that other ferrous proteins such as hemoglobin, myoglobin, and neuroglobin react with NO^• very fast and form very tight complexes, with k_on values ∼10⁷–10⁸ m⁻¹ s⁻¹ and k_off values of ∼10⁻⁴ s⁻¹ (53, 54). In contrast, Fe(II)-CBS reacts several orders of magnitude more slowly with NO^•, with a k_on of 8 × 10³ m⁻¹ s⁻¹ and k_off of 3 × 10⁻³ s⁻¹ (32). Thus, hemoglobin, myoglobin, and neuroglobin are able to outcompete high dithionite concentrations and efficiently scavenge the NO^• formed, whereas CBS cannot, necessitating the use of lower dithionite concentration in our experiments.

SCHEME 1.

Interaction of human CBS with nitrite.

The second order rate constant for the reaction of nitrite and Fe(II)-CBS obtained with full-length CBS was estimated to be 0.66 ± 0.03 m⁻¹ s⁻¹ at pH 7.4 and 37 °C. This value, which represents an approximation of the rate constant, is higher than those of several hemeproteins that function as nitrite reductases including hemoglobin (T state), neuroglobin, and cytoglobin (Table 2). In addition, Fe(II)-CBS binds NO^• more loosely than other hemeproteins (K_d = ∼0.23 μm (32)). This looser binding means that the NO^• formed from nitrite reduction would be available to react with potential targets instead of remaining bound to CBS. Together with the relatively high rate constant for nitrite reduction, the low affinity for NO^• suggests that CBS could be a potential source of NO^• in vivo. The increased rate constants for the full-length CBS in the presence of the allosteric activator AdoMet and for the truncated variant lacking the AdoMet binding domain suggest that the regulatory domain affects the heme environment. In this regard, it has recently been reported that AdoMet binding enhances association of CO and, to a lesser extent, NO^•, to Fe(II)-CBS (55). Remarkably, the mutant T257I CBS exhibited enhanced nitrite reduction kinetics. Mutations at this conserved threonine residue located in the PLP binding pocket drastically affect enzyme activity (24). A pathogenic T257M mutation has been described in a homocystinuric patient (56). In addition, these mutations at Thr²⁵⁷ perturb the heme environment, as observed by the enhanced propensity to form the Fe(II)424-CBS species (15, 24) and, as shown herein, for T257I to enhance reduction of nitrite and generation of NO^•.

From a mechanistic point of view, reduction of nitrite by six-coordinate Fe(II)-CBS could involve an outer sphere mechanism. Consistent with this possibility is the observation that dissociation of the cysteine ligand to heme does not appear to limit Fe(II)-NO^•-CBS formation because plots of k_obs versus nitrite concentration are linear up to 50 mm nitrite. Furthermore, the k_obs values for Fe(II)-NO^•-CBS formation (0.006–0.032 s⁻¹; Fig. 3B) are similar to or higher than the values for dissociation of the cysteine ligand, which was estimated from the maximum rate constant for CO binding to be ∼0.0166 s⁻¹ (31), 0.012 s⁻¹ (32), and 0.0031 s⁻¹ (28). The outer sphere proposal is further consistent with the very negative reduction potential of CBS (−0.350 V, (21)) that would favor electron transfer according to Marcus theory. As precedent, the reaction of Fe(II)-CBS with O₂, which occurs with a k_obs of 20–80 s⁻¹ and depends linearly on O₂ concentration (30). Alternatively, the reaction with nitrite would implicate a pentacoordinate (pentacoordinate ferrous CBS) species. This has been proposed in the case of human neuroglobin, another hexacoordinate hemeprotein for which thorough kinetic analyses have been reported (47, 57). The first step of the reaction is the dissociation of an axial His ligand, followed by nitrite (or nitrous acid) binding to the ferrous iron, either through the nitrogen or the oxygen atoms. Proton addition and electron transfer occur, yielding ferric heme, NO^•, and OH⁻. Neuroglobin is a bis-His ligated protein with a reduction potential of −0.118 V; the dissociation of the axial His is estimated to occur at 0.4 s⁻¹ (58), and the k_obs for the reaction with nitrite is ∼2 orders of magnitude lower (47) than this value. Thus, in the case of neuroglobin, it is likely that nitrite reduction is preceded by conversion to the pentacoordinate state. However, in the case of CBS, an outer sphere process seems to be the more probable pathway for explaining the reactivity of this unusual and low potential hemeprotein.

In contrast to the rapid monophasic O₂-mediated reoxidation of Fe(II)-CBS to Fe(III)-CBS (30), O₂-dependent reoxidation of Fe(II)NO^•-CBS is slow and showed complex kinetic behavior with a biphasic time course and no clear isosbestic points. The results obtained for Fe(II)NO^•-CBS can be interpreted in terms of two parallel mechanisms, one that involves the slow dissociation of NO^• as a rate-limiting step (0.003 s⁻¹ (32)) followed by a fast reaction with O₂ (1.11 × 10⁵ m⁻¹ s⁻¹ (30)), and another one with linear dependence on O₂ concentration in which O₂ first reacts with Fe(II)NO^•-CBS, albeit slowly. Biphasic processes without isosbestic points have also been reported for the reactions of nitrosyl derivatives of myoglobin (59), hemoglobin (60), and neuroglobin (61) with O₂. In the case of myoglobin, both phases were slow (10⁻⁴ s⁻¹), and one of them showed hyperbolic dependence on O₂ concentration, whereas the other was independent. In the case of hemoglobin, the rate of ferric heme formation appeared limited by the dissociation of NO^•, which occurred on a similar (10⁻⁴ s⁻¹) time scale. In the case of neuroglobin, the first phase depended linearly on O₂ concentration with a second order rate constant of 16 m⁻¹ s⁻¹, and the second phase was independent of O₂ concentration with k_obs of 5 × 10⁻⁴ s⁻¹. In contrast, for the heme thiolate protein, nitric oxide synthase, the reaction of the Fe(II)NO^• species with O₂ occurs through a rapid direct reaction, but in this case the hexacoordinate Fe(II)NO^• complex is very different from the pentacoordinate one observed in CBS (62, 63). In this context, it is important to note that, unlike other hemeproteins (e.g. hemoglobin, myoglobin, bacterial flavohemoglobins (64)), typical nitric oxide dioxygenation is unlikely to occur in the case of CBS. Such a process would require the reaction of NO^• with an oxyferrous Fe(II)O₂ heme to form nitrate and ferric heme. Fe(II)-CBS is unable to form a stable oxyferrous species (30), so typical nitric oxide dioxygenase activity can be excluded.

Our results with CBA demonstrate that peroxynitrite is formed during reoxidation of Fe(II)NO^•-CBS. Although CBA can also react with H₂O₂ (1.5 m⁻¹ s⁻¹) (42) and with amino acid hydroperoxides (7–23 m⁻¹ s⁻¹) (65), the rate constants are several orders of magnitude smaller than with peroxynitrite (k = 1.1 × 10⁶ m⁻¹ s⁻¹). Peroxynitrite formation is likely to be mediated by two parallel mechanisms that mirror the kinetics of O₂-dependent reoxidation of Fe(II)NO^•-CBS. The first is mediated by dissociation of the precursor radical NO^• followed by the fast reaction of Fe(II)-CBS with O₂ to form O₂^˙̄ (pathway c in Scheme 1). NO^• and O₂^˙̄ would then react with a diffusion-limited rate constant of ∼10¹⁰ m⁻¹ s⁻¹. The second is dependent on the direct reaction of O₂ with Fe(II)NO^•-CBS (pathway d in Scheme 1). The fact that superoxide dismutase (in the presence of catalase to decompose the H₂O₂ formed) did not completely inhibit COH formation suggests that diffusible O₂^˙̄ cannot account for all the peroxynitrite formed and that some peroxynitrite is generated directly at the heme site by reaction of O₂ with heme-bound NO^•, generating a nitrogen-bound heme-peroxynitrite species. In addition, the observation that glutathione and cysteine, which are efficient peroxynitrite scavengers, did not inhibit formation of COH fluorescence is also consistent with the in situ formation of peroxynitrite at the heme site.

Fe(III)-CBS can be reduced to Fe(II)-CBS by a physiological reducing system (28, 29). Thus, in addition to being a source of O₂^˙̄ by a fast reaction with O₂ (30), under low O₂ tension conditions, Fe(II)-CBS could react with nitrite leading to the transient formation of Fe(II)NO^•-CBS. This process might constitute not only a mechanism for decreasing the yield of cystathionine, cysteine, and CBS-derived H₂S but also might represents a potential source of NO^•. Moreover, O₂-mediated Fe(II)NO^•-CBS reoxidation could provide a previously unrecognized source of peroxynitrite, a strong biological oxidizing and nitrating agent that can react with CO₂, thiols, selenium compounds, and metal centers (66–68). In addition, CBS is susceptible to peroxynitrite. We have previously reported that peroxynitrite reacts with Fe(III)-CBS with a second order rate constant of (2.4–5.0) × 10⁴ m⁻¹ s⁻¹ (pH 7.4 and 37 °C), leading to nitration of protein residues, loss of heme thiolate coordination, bleaching, and inactivation of the enzyme (26), with possible relevance to hyperhomocysteinemia and pathogenesis. Increased oxidative stress caused by CBS deficiency was reported to correlate with the development of liver disease (69). In addition, a decrease in CBS activity with elevation in homocysteine concentrations was observed to be associated with an increase in oxidative stress in rat kidney ischemia-reperfusion injury, and it was hypothesized that an increase in NO^• levels during reperfusion might play a role in regulating CBS activity (70, 71). In summary, this study adds important new insights to our understanding of heme reactivity in CBS and adds to the evidence that CBS represents a point of interplay between nitrite, peroxynitrite, NO^•, O₂^˙̄, CO, and H₂S.

Author Contributions

S. C. designed, performed, and analyzed the experiments and wrote the manuscript. E. C. performed experiments and analyzed the data. C. G. and P. K. Y. helped conceive the experiments and contributed to analyze the data. D. P. B. provided technical assistance, analyzed the results, and edited the manuscript. B. A. and R. B. helped conceive the experiments, analyzed the data, and co-wrote the manuscript. All authors approved the final version of the manuscript.