Mechanistic Studies on the Reaction of Nitrocobalamin with Glutathione: Kinetic evidence for formation of an aquacobalamin intermediate

Abstract

The essential but also toxic gaseous signaling molecule nitric oxide is scavenged by the reduced vitamin B₁₂ complex cob(II)alamin. The resulting complex, nitroxylcobalamin (NO⁻-Cbl(III)), is rapidly oxidized to nitrocobalamin (NO₂Cbl) in the presence of oxygen; however it is unlikely that nitrocobalamin is itself stable in biological systems. Kinetic studies on the reaction between NO₂Cbl and the important intracellular antioxidant, glutathione (GSH), are reported. In this study, a reaction pathway is proposed in which the β-axial ligand of NO₂Cbl is first substituted by water to give aquacobalamin (H₂OCbl⁺), which then reacts further with GSH to form glutathionylcobalamin (GSCbl). Independent measurements of the four associated rate constants k₁, k₋₁, k₂, and k₋₂ support the proposed mechanism. These findings provide insight into the fundamental mechanism of ligand substitution reactions of cob(III)alamins with inorganic ligands at the β-axial site.

Introduction

Two essential enzymes in mammals, L-methylmalonyl-CoA mutase, and methionine synthase, and numerous bacterial enzymes require the vitamin B₁₂ derivatives (= cobalamins, Cbls) adenosylcobalamin (AdoCbl) and methylcobalamin (MeCbl) as cofactors, Figure 1.^[1] MeCbl-dependent methionine synthase transfers a methyl group from methyltetrahydrofolate to homocysteine to generate tetrahydrofolate and methionine, whereas AdoCbl-dependent L-methylmalonyl-CoA mutase catalyzes the isomerization of L-methylmalonyl-CoA to succinyl-CoA.^[1] Cobalamins may also have additional roles in biological systems, including regulating the immune response and protecting against intracellular oxidative stress.^[2]

Figure 1.

Structure of cobalamins showing the two axial sites (upper = β, lower = α) with respect to the corrin ring.

The signaling molecule nitric oxide (^•NO) plays a key role in the immune response, vasodilation, and neurotransmission.^[3] However, high levels of NO can be deleterious and can result in sepsis and septic shock,^[4] leading to organ failure and even death. Importantly, administering cobalamins suppresses ^•NO-induced relaxation of smooth muscle,^{[5] •}NO-induced vasodilation^[6] and ^•NO-mediated inhibition of cell proliferation.^[7] Cobalamins also reverse ^•NO-induced neural tube defects.^[8] Both mammalian B₁₂-dependent enzymes are inhibited by ^•NO.^[9] With the exception of glutathionylcobalamin (X = glutathione, Figure 1), ^•NO does not directly react with cob(III)alamins,^[10] whereas the rate of the reaction between cob(II)alamin and nitric oxide to form nitrosylcobalamin (NOCbl) is almost diffusion controlled and essentially irreversible.^[11] Given that all cob(III)alamins are readily reduced by intracellular reductases,^[12] it is likely that cob(II)alamin reacts with ^•NO in biological systems, to form NOCbl.

In discussions of the biological relevance of NOCbl formation, we have observed that authors neglect to mention that NOCbl itself is not a stable entity.^{[5a, 6–8, 9c–e]} However, in the presence of even minute amounts of air, orange NOCbl rapidly oxidizes to form red nitrocobalamin, NO₂Cbl.^[13] Indeed, we have used this property of NOCbl in our laboratory numerous times to check the condition of valves and taps used in strictly anaerobic experimental setups. The intracellular fate of NO₂Cbl is unclear. Possibilities include NO₂Cbl being reduced by intracellular reductases, and/or NO₂Cbl reacting with the strong nucleophile and reductant glutathione (GSH), which is present in mM concentrations in cells.^[14] In this work we report kinetic studies on the reaction of NO₂Cbl with glutathione. Interestingly, our kinetic data show that GSH does not directly react with NO₂Cbl, but instead reacts with aquacobalamin, which is always present in equilibrium, albeit typically in small amounts, with cobalamins incorporating inorganic ligands at the β-axial cobalamin site (Figure 1).

Results and Discussion

Kinetic data were collected for the reaction of NO₂Cbl with glutathione (GSH). Experiments were initiated by adding a small aliquot of concentrated aqueous NO₂Cbl solution (final concentration 5.0 × 10⁻⁵ M) to a buffered GSH solution (3.00 mL) at a specific pH condition (I = 1.0 M (NaCF₃SO₃)). Control experiments established that rate constants were identical within experimental error in the absence and presence of air; hence all experiments were carried out under aerobic conditions. Importantly, dissolving NO₂Cbl in water rather than in buffer minimized the decomposition of NO₂Cbl to aquacobalamin (H₂OCbl⁺) prior to the addition of an aliquot of this solution into a buffered GSH solution.

Figure 2(a) shows UV-vis spectral changes observed upon the addition of NO₂Cbl to a buffered GSH solution (5.00 × 10⁻² M) at pH 4.00. NO₂Cbl (λ_max 354, 413 and 532) is converted to GSCbl (λ_max 333, 372, 428 and 534^[15]) with isosbestic points at 336, 367, 452 and 543, indicating that a single reaction occurs. The corresponding plot at 354 nm versus time is given in Figure 2(b). The data fit well to the first-order rate equation, giving an observed rate constant, k_obs= (1.15 ± 0.07) × 10⁻² s⁻¹.

Figure 2.

(a) UV-vis spectra for the reaction of GSH (5.00 × 10⁻² M) with NO₂Cbl (5.0 × 10⁻⁵ M) at pH 4.00 (25.0 °C, 0.020 M NaOAc, I = 1.0 M, NaCF₃SO₃). Selected spectra for the reaction are shown every 1.00 min. (b) Plot of absorbance at 354 nm versus time for the experiment shown in 2(a). Data were fitted to a first-order rate equation, giving k_obs = (1.15 ± 0.07) × 10⁻² s⁻¹.

Kinetic data were collected at pH 4.00 and 7.00, in order to determine whether the rate of the reaction is pH dependent. Plots of k_obs versus total GSH concentration are shown in Figures 3(a) and 3(b). These plots indicate that the observed rate constant reach a limiting value at high GSH concentrations and that there is no pH dependence in this pH region. There are two plausible mechanisms by which NO₂Cbl reacts with GSH to give a plot exhibiting curvature to reach a limiting value of k_obs at high GSH concentrations (saturation kinetics). The first involves rapid equilibration to form a NO₂Cbl•GSH association complex prior to rate-determining substitution of the β-axial NO₂⁻ ligand of NO₂Cbl by GSH to give GSCbl. The other alternative is a two step process, in which H₂OCbl⁺, which is in equilibrium with NO₂Cbl, reacts with GSH, Scheme 1. Given that all Cbls with β-axial inorganic ligands exist in equilibrium with H₂OCbl⁺ and that Cbls undergo β-axial ligand exchange via a dissociative interchange mechanism,^[16] the latter possibility is more likely to occur. Kinetic studies on the reaction between H₂OCbl⁺/HOCbl and GSH have been reported,^[17] and show that H₂OCbl⁺ reacts rapidly with GSH to form GSCbl under the conditions of our study.

Figure 3.

Plots of k_obs vs [GSH] at pH 4.00 (a) and 7.00 (b) for the reaction between NO₂Cbl (5.00 × 10⁻⁵ M) and glutathione (25.0 °C, 0.020 M NaOAc (a) or 0.020 M KH₂PO₄ (b), I = 1.0 M (NaCF₃SO₃)). Data in (a) were fitted to eq (2) in the text fixing k₋₂ = 7.4 × 10⁻⁴ s⁻¹, giving k₁ = (1.75 ± 0.02) × 10⁻² s⁻¹ and K = 94.5 ± 3.7 M⁻¹ at pH 4.00. Data in (b) were fitted to eq (2) fixing k₋₂ = 0 s⁻¹, giving k₁ = (1.73 ± 0.05) × 10⁻² s⁻¹ and K = 102.1 ± 9.7 M⁻¹ at pH 7.00.

Scheme 1.

Proposed reaction pathway for the reaction of NO₂Cbl with GSH. Note that in aqueous solution H₂OCbl⁺ exists in equilibrium with HOCbl (pK_a(H₂OCbl⁺) = 7.8 ^[17]); however at the pH conditions of our kinetic experiments HOCbl formation is unimportant, since the values of the rate constants k₋₁ and k₂ are the same at pH 4.00 and 7.00.

The rate expression corresponding to the reaction pathway shown in Scheme 1 is ^[18]

$${\mathrm{k}}_{\text{obs}}=({\mathrm{k}}_1{\mathrm{k}}_2[\text{GSH}]/({\mathrm{k}}_{-1}[{{\text{NO}}_2}^{-}]))+{\mathrm{k}}_{-2})/(1+{\mathrm{k}}_2[\text{GSH}]/({\mathrm{k}}_{-1}[{{\text{NO}}_2}^{-}]))$$ (1)

$$=({\mathrm{k}}_1\mathrm{K}[\text{GSH}])+{\mathrm{k}}_{-2})/(1+\mathrm{K}[\text{GSH}])$$ (2)

where K = k₂/(k₋₁[NO₂⁻]). The rate constant, k₋₂, for decomposition of GSCbl to H₂OCbl⁺ and GSH was independently determined at pH 4.00 and found to be (7.4 ± 0.5) × 10⁻⁴ s⁻¹, Figure S1, Supporting Information. Previous studies have shown that the observed equilibrium constant for formation of GSCbl, K_obs(GSCbl), = k₂/k₋₂, increases by approximately one order of magnitude for each unit change in pH.^[17] Since k₂ is pH independent,^[17] k₋₂ is therefore negligible at pH 7.00. Data in Figure 3(a) were fitted to eq (2) fixing k₋₂ = 7.4 × 10⁻⁴ s⁻¹, giving k₁= (1.75 ± 0.02) × 10⁻² s⁻¹ and K = 94.5 ± 3.7 M⁻¹. Data in (b) were fitted to the same equation fixing k₋₂ = 0 s⁻¹, giving k₁ = (1.73 ± 0.05) × 10⁻² s⁻¹ and K = 102.1 ± 9.7 M⁻¹. K is therefore also pH independent in the pH 4–7 region, within experimental error.

Importantly, if our model is correct, then K = k₂/(k₋₁[NO₂⁻]). Note, however, that the free nitrite concentration is not strictly constant during the reaction, but will vary from 0 to a maximum value of 5.0 × 10⁻⁵ M as the reaction proceeds, as NO₂Cbl (5.0 × 10⁻⁵ M) reacts with GSH to give GSCbl plus NO₂⁻. Hence K is not strictly constant during the reaction, as reflected in the error associated with K. In order to provide support for our model, new data was therefore collected at pH 7.00 under the same conditions as the experiments summarized in Figure 3(b), except that 5.00 × 10⁻⁴ M NaNO₂ was added to each solution, so the nitrite concentration is constant (pseudo-first order conditions) during the reaction. These data are shown in Figure 4. Fitting this data to eq (2) fixing k₋₂ = 0 s⁻¹ gave k₁ = (1.60 ± 0.05) × 10⁻² s⁻¹ and K = 16.2 ± 0.2 M⁻¹. The data now fit considerably better to eq (2) as expected (the error in K is now one order of magnitude smaller), since the nitrite concentration remains constant during the reaction.

Figure 4.

Plot of k_obs vs [GSH] for the reaction between NO₂Cbl (5.0 × 10⁻⁵ M) and GSH in the presence of 5.00 × 10⁻⁴ M NaNO₂ at pH 7.00 (25.0°C, 0.020 M KH₂PO₄, 5.00×10⁻⁴ M NaNO₂, I = 1.0 M, NaCF₃SO₃). Data were fitted to eq (2) in the text fixing k₋₂ = 0, giving k₁ = (1.60 ±0.05) × 10⁻²s⁻¹ and K = 16.2 ± 0.2 M⁻¹.

The rate constant k₂ for the reaction of H₂OCbl⁺ with GSH was also independently determined at pH 4.00 and found to be 12.00 ± 0.25 M⁻¹ s⁻¹ (25.0 °C, 0.020 M NaCH₃COO, I = 1.00 M (NaCF₃SO₃)), Figure S2, Supporting Information. This value is in good agreement with a value reported by us under slightly different conditions (k₂ = 18.5 M⁻¹ s⁻¹ at pH 4.50, 25.0 °C, 0.10 M NaOAc, I = 0.50 M (KNO₃)^[17]), and is pH independent in the pH 4–7 range.^[17] The rate constant k₋₁ for the reaction between H₂OCbl⁺ and NO₂⁻ was independently determined to be (1.25 ± 0.02) × 10³ and (1.20 ± 0.02) × 10³ M⁻¹ s⁻¹ at pH 4.00 and 7.00, respectively, Figures S3 and S4, Supporting Information. Hence k₋₁ is also independent of the pH (pH 4–7), as expected, as the ionization of the reactants is essentially unchanged (pK_a(HNO₂) ~ 3.2; pK_a(H₂OCbl⁺) = 7.8 ^[17]) in this pH region. The value of k₋₁ agrees well with a value reported by others (k₋₁ = 99.8 × 10² M⁻¹ s⁻¹ at 25 °C, I = 2.2 M (NaNO₃)).^[19] Using k₂ = 12.00 M⁻¹ s⁻¹, k₋₁ = 1.23 × 10³ M⁻¹ s⁻¹ and NO₂⁻ = 5.00 × 10⁻⁴ M gives K ~20 M⁻¹, which is in very good agreement with the experimental value of K obtained from the best fit of the data (16.2 ± 0.2 M⁻¹), providing strong support for the reaction pathway proposed in Scheme 1.

Finally, the rate constant k₁ for decomposition of NO₂Cbl to H₂OCbl⁺ was also independently determined by obtaining kinetic data for the decomposition of NO₂Cbl to H₂OCbl⁺ upon dissolving NO₂Cbl in buffer. Figure S5 in the Supporting Information shows the absorbance change (ΔAbs = 0.013 at 350 nm) that occurs at pH 4.00. Only a small fraction of NO₂Cbl is converted to H₂OCbl⁺; however the absorbance change is sufficient to allow calculation of k₁. The results at different pH conditions are summarized in Table S1, and give a mean value for k₁ of (1.48 ± 0.22) × 10⁻² M⁻¹ s⁻¹ (pH 3.5 – 6.0; k₁ is pH independent). The observed rate constant for NO₂Cbl partially decomposing to give H₂OCbl⁺ is actually k₁ + k₋₁[NO₂⁻] for the pseudo-first-order reversible process. A separate experiment showed that an absorbance difference of 0.232 is observed upon completely converting H₂OCbl⁺ to NO₂Cbl upon the addition of a slight excess of NO₂⁻ (Cbl_T = 5.0 × 10⁻⁵ M). Hence an absorbance change of 0.013 corresponds to formation of 5.6% H₂OCbl⁺ (2.8 × 10⁻⁶ M H₂OCbl⁺ and 2.8 × 10⁻⁶ M NO₂⁻) upon dissolving NO₂Cbl in H₂OCbl⁺; that is, the maximum NO₂⁻ is 2.8 × 10⁻⁶ M. Using k₁ = 1.48 × 10⁻² M⁻¹ s⁻¹, k₋₁ = (1.25 ± 0.02) × 10³ and [NO₂⁻] = 2.8 × 10⁻⁶ M shows that k₁ is ~ 5 times larger than k₋₁[NO₂⁻], validating the assumption that k_obs ~ k₁ upon dissolving NO₂Cbl in H₂OCbl⁺. Using our values of k₁ and k₋₁, the equilibrium constant for formation of NO₂Cbl, K(NO₂Cbl), = k₋₁/k₁, is 8.5 × 10⁴ M⁻¹, which is in reasonable agreement with a value reported by others under different ionic strength conditions (K(NO₂Cbl) = 2.2 × 10⁵ M⁻¹, 25 °C, I = 2.2 M ^[20]).

Conclusions

Kinetic studies on the reaction between NO₂Cbl and GSH show that the rate of the reaction is pH independent in the pH 4–7 region. By independently determining values of k₁, k₋₁, k₂ and k₋₂, we have shown that the data fits a model involving an H₂OCbl⁺ intermediate, which then rapidly reacts with GSH to form GSCbl, Scheme 1. To our knowledge this is the first time that the reaction pathway of β-axial inorganic ligand exchange for cob(III)alamins via an H₂OCbl⁺ intermediate has been unequivocally demonstrated. This may have important consequences for free and potentially even protein-bound cob(III)alamins incorporating inorganic ligands (X-ray structures of cobalamins bound to B₁₂ transport proteins show that the β-axial site can be readily accessed by solvent and small molecules^[21]) that is, the amount of each of these species may reflect the concentrations and binding constants to aquacobalamin of the various inorganic ligands present. As such, GSCbl would be expected to be the major intracellular non-alkylcob(III)alamin, given that intracellular GSH concentrations are mM,^[14] and only CN⁻ binds stronger than GSH to H₂OCbl⁺(K_CNCbl ~ 10¹⁴ M^{−1 [22]}). Finally, at 0.5 mM GSH the rate constant for the reaction between NO₂Cbl and GSH is ~ 8 × 10⁻³ s⁻¹ at pH 7.0 (25 °C), corresponding to a half life of ~ 1.4 min. Hence formation of GSCbl is one possible reaction pathway by which NO₂Cbl decomposes in biological systems.

Experimental Section

General

Hydroxycobalamin hydrochloride (HOCbl•HCl, 98% stated purity by manufacturer) was purchased from Fluka. Glutathione (98%), acetic acid (sodium salt, ≥99%), sodium nitrite (97%) and CF₃SO₃H (99%) were obtained from Acros Organics. TES buffer (98%) was purchased from MP Biomedicals Inc. Potassium dihydrogen phosphate was purchased from Sigma. NaCF₃SO₃ was prepared by neutralizing a concentrated, aqueous solution of CF₃SO₃H with NaOH, reducing it to dryness by rotary evaporation, and drying it overnight in a vacuum oven at 70.0 °C. Nitrocobalamin was synthesized and characterized according to a published procedure.^[15] The purity was ≥ 95%, as determined by ¹H NMR spectroscopy.

UV-visible spectra and kinetic data for slower reactions were recorded on a Cary 5000 spectrophotometer equipped with a thermostated (25.0 ± 0.1°C) cell changer operating with WinUV Bio software (version 3.00). Reactant solutions were thermostated for 15 min prior to measurements. Kinetic data for rapid reactions were obtained at 25.0 ± 0.1°C using an Applied Photophysics SX20 stopped-flow spectrophotometer equipped with a photodiode array detector in addition to a single wavelength detector. Data were collected with Pro-Data SX (version 2.1.4) and Pro-Data Viewer (version 4.1.10) software, and a 1.0 cm pathlength cell was utilized. All data were analyzed using Microcal Origin version 8.0.

pH measurements were carried out using an Orion model 710A pH meter equipped with a Wilmad 6030–02 pH electrode. The electrode was filled with a 3 M KCl/saturated AgCl solution (pH 7.0) and standardized with standard BDH buffer solutions at pH 6.98, 4.01 and 2.02. Solution pH was adjusted using 50% v/v aqueous CF₃SO₃Hand NaOH (~ 5 M).

¹H NMR spectra was recorded on a Bruker Avance 400 MHz spectrometer equipped with 5 mm probe. Solutions for NMR measurements were prepared in D₂O. ¹H NMR spectra were internally referenced to TSP (0 ppm).

Kinetic measurements

The rates of the reaction between NO₂Cbl and glutathione (GSH) were determined under pseudo-first-order conditions with excess GSH. Stock solutions of GSH (0.500 M) in the presence or absence of sodium nitrite (5.00 × 10⁻⁴ M) were prepared in the appropriate buffer (0.020 M) at pH 4.00 and 7.00 and diluted as appropriate. A small aliquot of concentrated NO₂Cbl (final concentration 5.0 × 10⁻⁵ M) in water was added to initiate the reaction and the absorbance at 354 nm was recorded as a function of time.

Kinetic data for the reaction between H₂OCbl⁺ (5.0 × 10⁻⁵ M) with varying concentrations of NO₂⁻ were obtained at pH 4.00 and 7.00. Stock solutions of NaNO₂ (0.010 M) were prepared in the appropriate buffer (0.020 M) and diluted as needed. Data were collected at 354 nm. Kinetic data for the reaction of H₂OCbl⁺ (5.0 × 10⁻⁵ M) with varying concentrations of glutathione at pH 4.00 were collected at 354 nm in acetate buffer (0.020 M).

The rate of decomposition of NO₂Cbl to H₂OCbl⁺ was determined in the pH 3.5–6.0 range by adding solid NO₂Cbl directly to the appropriate buffer solution (0.200 M; the solution was thermostated at 25.0 °C for 10 min prior to the addition of NaNO₂). The solution was quickly filtered through a micropore filter (0.45 μm) and data collection initiated at 354 nm.

The rate of decomposition of GSCbl to H₂OCbl⁺ was determined at pH 4.00. An aliquot of GSCbl (4.0 × 10⁻⁵ M) was added to pH 4.00 acetate buffer (0.020 M; the buffer solution was thermostated at 25.0 °C for 10 min prior to the addition of GSCbl) and data were collected at 354 nm.

The total ionic strength was maintained at 1.0 M using NaCF₃SO₃ for all solutions.