Introduction

Nitroxylcobalamin (CblNO, formally CoIII–NO−) [1] is one of the most interesting forms of Vitamin B12 that was shown to be stable in biological systems [2, 3]. Under physiological conditions, CblNO can be produced in a very efficient reaction between the major intracellular form of Vitamin B12r, viz. cob(II)alamin [4], and nitric oxide (·NO) for which k = 7.4 × 108 M−1 s−1 and KNO ≈ 1 × 108 M−1 at 25 °C [5,6,7]. It has been postulated that cobalamins show the potential to eliminate excess ·NO from organisms [8, 9] since the CoIII–NO− complex can be protonated at neutral pH to form CoIII–NOH, which in turn can undergo aquation to release HNO. The latter species is known to dimerize and decompose to water and gaseous N2O in aqueous solution [10].

CblNO is extremely air sensitive [6, 11, 12] and in the presence of oxygen it rapidly oxidizes to nitrocobalamin (CblNO2) [6, 11,12,13]. However, the oxidation of CblNO is not just a simple reaction [14]. According to Brasch et al., the mechanism of CblNO oxidation under certain conditions is rather complex, with multiple products that can be formed [15]. Oxidation of CblNO is a reversible process in which CblNO2 can be reduced by strong reducing agents to reform CblNO. Brasch et al. studied the reaction between CblNO2 and one of the strongest intracellular antioxidants glutathione (GSH), as one of the possibilities to form CblNO [16]. An important conclusion from this work is that for a pH between 4 and 7, reduction of CblNO2 to CblNO was not observed! Furthermore, the main product of this reaction is glutathionylcobalamin (CblGS). The reported kinetic data suggest that the observed reaction is a two-step process which involves aquacobalamin (CblOH2) as an intermediate that rapidly reacts with GSH to form CblGS (Scheme 1) [16].

Scheme 1
scheme 1

Schematic presentation of the reaction between CblNO2 and GSH

Full size image

Our recent study [17] showed that at neutral pH (pH 7.2, 0.1 M Tris buffer, 25 °C), CblNO2 can react with another biological reducing agent, ascorbic acid (HAsc−), to form reduced cobalamin (Cbl(II)) as product, but not CblNO as was hoped. Kinetic data showed that both CblNO2 and CblOH2 present in an equilibrium mixture react with HAsc−, but CblOH2 reacts ca. two orders of magnitude faster than CblNO2 (Scheme 2), i.e. k2 ≫ k3. Thus, on addition of ascorbic acid to the equilibrium mixture of CblOH2 and CblNO2, it reacts rapidly with CblOH2 which in turn is reformed through the aquation of CblNO2 (k1 = 1.4 × 10−2 s−1 at 25 °C) and so represents the major reactive species in solution. Furthermore, in the presence of a large excess of nitrite the reduction of CblOH2 can be drastically inhibited due to the formation of CblNO2 [17]. Recent work in our laboratories indicated that this system shows a more interesting behavior under slightly different experimental conditions.

Scheme 2
scheme 2

Suggested mechanism for the reaction between CblNO2 and HAsc− at pH ~ 7

Full size image

Experimental section

Materials

Hydroxocobalamin hydrochloride (HOCbl·HCl, ≥ 98%) was purchased from Sigma-Aldrich, sodium nitrite was purchased from LPPH and ascorbic acid was obtained from Polfa Kraków. Acetic acid (CH3COOH,  \(\ge\) 99.5–99.9%) and sodium hydroxide (NaOH, \(\ge\) 98.8%) were obtained from a range of suppliers (Sigma-Aldrich, Merck, Fisher Scientific or POCH). All chemicals used throughout this study were of analytical grade or better.

General methods

All solutions were prepared in de-ionized water using a water purification system. Strictly anaerobic solutions were prepared using appropriate air-free techniques and handling the solutions in appropriate glassware. Oxygen-free argon or nitrogen was used to deoxygenate the reactant solutions. UV–Vis spectral measurements were carried out in screw-cap cuvettes equipped with a silicone septum. pH measurements were carried out at room temperature using a HI 221 (Hanna Instruments) pH-meter equipped with an AmpHel glass electrode filled with a 3 M KCl solution.

UV–Vis spectroscopy

UV–Vis spectra and kinetic data were recorded on Perkin Elmer Lambda 25 spectrophotometer equipped with a thermostated (25.0 ± 0.1 °C) cell holder (Perkin Elmer PTP-6 Peltier System). All data were analyzed using Origin Lab software.

Results and discussion

After completion of our studies at pH ~ 7, we decided to check the influence of ascorbic acid on the reaction with CblNO2 at a more acidic pH ≤ 5, where ascorbic acid is mainly present in the mono-protonated form HAsc− and the much weaker reducing agent H2Asc, depending on the selected pH (pKa1 = 4.1 and pKa2 = 11.3 [18]). Under these conditions, mixtures of H2Asc and HAsc− have significantly weaker reducing properties than at pH ~ 7, where ascorbic acid is mainly present as HAsc− with traces of the stronger reducing agent Asc2− [18,19,20,21,22,23]. UV–Vis spectra recorded during the reaction between CblNO2 and HAsc− at pH < 5 (0.1 M acetate buffer, 25 °C, Ar atmosphere) indicated that CblNO2 (λmax = 354, 413 and 532 nm) is not converted to Cbl(II) as before [17], but appears to be converted directly to nitoxylcobalamin, CblNO (λmax = 316, 344 and 475 nm) [6] with isosbestic points at 335, 373, 490 nm (Fig. 1a). Under conditions of pH 4.3 (0.1 M acetate buffer, 25 °C, Ar atmosphere), [CblNO2] = 8.6 × 10−5 M (obtained by mixing CblOH2 and NO2−, [NO2−]/[CblOH2] = 5) and [HAsc−] = 8.6 × 10−4 M, CblNO was formed within 100 min from the start of the reaction. During the next 300 min practically no spectral changes were observed, from which we conclude that CblNO is stable in solution under the selected conditions (Fig. 1b). However, during the next 400 min the absorbance at 476 nm decreased to the initial value and the UV–Vis spectra clearly showed that CblNO2 was reformed fully (isosbestic points at 335, 373, 490 nm) (Fig. 1c, d).

Fig. 1
figure 1

a Spectral changes observed for the reaction between CblNO2 (8.6 × 10−5 M, obtained by mixing CblOH2 and NO2−, [NO2−]/[CblOH2] = 5) and HAsc− (8.6 × 10−4 M) at pH 4.3 (0.1 M acetate buffer, 25 °C, Ar atmosphere) during the first 100 min, b between 100 and 400 min, c between 400 and 1200 min from the start of the reaction. Spectra were recorded every 1 min. d Plot of absorbance at 476 nm vs. time

Full size image

Further studies showed that CblNO was formed faster and remained for longer times in the reaction mixture when higher concentrations of HAsc− were used (compare Fig. 2a, b and d), whereas increasing the nitrite concentration slowed down the conversion of CblNO2 to CblNO (compare Fig. 2b and c). In a more detailed series of experiments, the nitrite concentration was varied over a wider concentration range while keeping the ascorbate concentration constant. The results reported in Figure S1 (Supporting Information) show that the initial phase of the reaction indeed slows down significantly on increasing the nitrite concentration. The kinetic data could be fitted best with a zero-order process for the initial changes in absorbance with time, and the slopes of such plots as a function of CblOH2 concentration show typical saturation kinetics (see Figure S1e). From the value of K = 1 × 105 M−1 (see Scheme 2) reported at 25 °C [17], it can be estimated that the concentration of CblOH2 in solution decreases from 1.95 × 10−6 to 4.97 × 10−7 M on increasing the nitrite concentration from 4.3 × 10−4 to 1.72 × 10−3 M for the experiments reported in Figure S1. This clearly demonstrates the effect of the nitrite concentration on the rate of formation of CblNO. In the presence of a large excess of nitrite, no reaction was observed at all, suggesting that nitrite completely suppresses the formation of CblNO.

Fig. 2
figure 2

Plot of absorbance at 476 nm vs. time for the reaction between CblNO2 (8.6 × 10−5 M, obtained by mixing CblOH2 with NO2−) and HAsc− at pH 4.3 (0.1 M acetate buffer, 25 °C, Ar atmosphere). Concentration ratio for [CblOH2]:[NO2−]:[HAsc−]: a 1:5:10, b 1:5:20, c 1:10:20 and d 1:10:30

Full size image

Additional studies were performed to check whether after almost 20 h from the start of the reaction, it is possible to reduce the final CblNO2 product by the addition of an extra amount of ascorbic acid. Surprisingly, addition of an extra portion of HAsc− resulted once again in the formation of CblNO as shown in Fig. 3b by the increase in absorbance at 476 nm from 1100 to 1400 min.

Fig. 3
figure 3

a Plot of absorbance at 476 nm vs. time for the reaction between CblNO2 and HAsc−. b Plot of absorbance at 476 nm vs. time for the reaction between CblNO2 and HAsc− with the addition of extra HAsc− after 1100 min from the start of the reaction

Full size image

A blank experiment was performed in the absence of cobalamin in which solutions of nitrite and HAsc− were mixed under exactly the same conditions as we used in the experiments with cobalamin. The absorbance maximum at 262 nm which comes from HAsc− decreases with time to reach the pre-reaction value after 20 h from the start of the reaction (Figure S2, Supporting Information). We ascribe these findings to slow side reactions between ascorbic acid and nitrite, and/or slow diffusion of oxygen into the sealed cuvettes over longer periods of time (see further discussion).

In addition, we could show that by changing the sequence of mixing the reagents this resulted in totally different spectral changes. On mixing first CblOH2 with HAsc− (before addition of nitrite), CblOH2 was fully reduced to Cbl(II) as shown in Fig. 4a. Subsequent addition of nitrite led to the rapid formation of CblNO2 and not CblNO, i.e. Cbl(II) is oxidized to CblOH2 which reacts rapidly with nitrite to form CblNO2 followed by its reduction to CblNO (Fig. 4b). The CblNO complex remains stable for some time (Fig. 4c) before it slowly converts back to CblNO2 due to the depletion of ascorbate (Fig. 4d).

Fig. 4
figure 4

Spectral changes observed for the reaction between CblOH2 (8.6 × 10−5 M) and HAsc− (8.6 × 10−4 M) at pH 4.3 (0.1 M acetate buffer, 25 °C, Ar atmosphere: a), and for the reaction of Cbl(II) obtained in a with NO2− (4.3 × 10−4 M; b, c and d) between 0 and 60 min (b), between 60 and 600 min (c) and between 600 and 1200 min from the start of the reaction (d). In a and b spectra are recorded every 1 min, whereas in c and d  every 30 min

Full size image

We also used FeII(EDTA) as a very efficient trap for the intermediate formation of ·NO [6, 24]. On mixing typical concentrations of nitrite and ascorbate under Ar atmosphere and allowing them to react for 2 h, the addition of FeII(edta) immediately resulted in the formation of Fe(edta)NO as shown in Figure S3 (Supporting Information). This is clear evidence for the intermediate formation of ·NO during the reduction of nitrite by ascorbate under the selected conditions of this study.

The results reported above show that it is indeed possible to observe the formation of CblNO under milder reducing conditions with ascorbic acid at pH < 5. In terms of the biological relevance of these findings, we repeated a series of measurements where the pH was systematically decreased from 7.2 to 5.0, to see where the changeover from Cbl(II) to CblNO as reaction product occurs. In these experiments a typical concentration ratio of [CblOH2]:[NO2−]:[HAsc−] = 1:5:10 was selected as done in Fig. 2. On decreasing the pH the reaction product changed from only Cbl(II) (pH 7.2, Figure S4, Supporting Information) to a mixture of Cbl(II) and CblNO (pH 5.5, Figure S5, Supporting Information), to only CblNO (pH 5.0, Figure S6, Supporting Information). It follows that as we go to milder reducing conditions by lowering the pH, only Cbl(NO) is formed at pH ≤ 5.0, which must be related to the pH dependence of the redox potential for the two-electron ascorbic acid/dehydroascorbate transformation. According to the Pourbaix diagram for this transformation [18], the redox potential of H2Asc at pH 0 is + 0.4 V, of H2Asc/HAsc− at pH 4.1 (pKa1) is + 0.16 V, for HAsc− at pH 7.0 (8.0) is + 0.07 (+ 0.04 V), and for HAsc−/Asc2− at pH 11.3 (pKa2) is − 0.15 V. These data clearly show the large change in redox potential to a significantly stronger reducing agent on increasing the pH of the solution.

The challenge now will be to find a biologically relevant reducing agent that under mild reaction conditions will reduce CblNO2 to CblNO at pH 7.4.

Mechanistic interpretation

The results of this study have clearly demonstrated that it is possible to obtain stable solutions of CblNO in the presence of a reducing agent starting from CblNO2 under well-selected reaction conditions. The remaining question is how can we account for the different reaction steps observed?

Our initial idea was that coordinated nitrite can be reduced by ascorbic acid to form the nitroxyl complex CblNO, since this is what the observed spectral changes tell us, thus a direct reaction from CblNO2 to CblNO. However, in these experiments performed at a pH < 5 we noticed that the conversion of CblNO2 to CblNO slowed down on increasing the nitrite concentration. In the presence of a large excess of nitrite, no reaction was observed at all. Thus, nitrite suppresses the formation of CblNO, from which we can conclude that ascorbic acid cannot reduce coordinated nitrite. This means that the remaining low concentration of CblOH2 in solution is reduced by the added HAsc− as found in our earlier report [17] and presented in reactions (1) and (2), where DHAsc represents dehydroascorbate.

$${\text{CblOH}}_{2} + {\text{ HAsc}}^{ - } \to {\text{Cbl}}\left( {\text{II}} \right) \, + {\text{ Asc}}^{ \cdot - }$$
(1)
$${\text{CblOH}}_{2} + {\text{ Asc}}^{ \cdot - } \to {\text{Cbl}}\left( {\text{II}} \right) \, + {\text{ DHAsc}}$$
(2)

In a subsequent reaction, Cbl(II) can be oxidized by nitrite to form CblOH2 and ·NO.

$${\text{Cbl}}\left( {\text{II}} \right) \, + {\text{ NO}}_{ 2}^{ - } \to {\text{CblOH}}_{2} + {}^{ \cdot }{\text{NO}}$$
(3)
$${\text{Cbl}}\left( {\text{II}} \right) \, + {}^{ \cdot }{\text{NO }} \to {\text{CblNO}}$$
(4)

The CblOH2 formed in reaction (3) is immediately reduced by HAsc− in reactions (1) and (2) and the nitroxyl product is formed in the very fast radical coupling of Cbl(II) and ·NO (k = 7.4 x 108 M−1 s−1 [6]), reaction (4). Thus, reactions (1) to (4) account for the observation that CblNO was formed faster and remained for longer times in the reaction mixture when higher concentrations of HAsc− were used, whereas increasing nitrite concentration slowed down the conversion of CblNO2 to CblNO. At the point where ascorbic acid is depleted, CblNO can react with excess ·NO to yield CblOH2 and N2O, similar to that found for the reaction between free HNO and ·NO to form NO2− and N2O [25]. Subsequently, CblOH2 is converted to CblNO2, the final reaction product in the presence of excess nitrite.

There are some side reactions (5) and (6) that can contribute to the formation of ·NO on a longer timescale and add to the redox cycling that causes the overall decomposition of ascorbic acid to DHAsc and the reformation of CblNO2. These reactions have been known for almost 60 years [26].

$${\text{NO}}_{2}^{ - } + {\text{ HAsc}}^{ - \cdot } \to {}^{ \cdot }{\text{NO }} + {\text{ Asc}}^{ \cdot - }$$
(5)
$${\text{NO}}_{2}^{ - } + {\text{ Asc}}^{ \cdot - } \to {}^{ \cdot }{\text{NO }} + {\text{ DHAsc}}$$
(6)

In this case, the produced ·NO will react rapidly with Cbl(II) to form CblNO in reaction (4).

In addition, traces of oxygen can oxidize ·NO to ·NO2 and reform nitrite according to reactions (7)–(9).

$$2{}^{ \cdot }{\text{NO }} + {\text{ O}}_{2} \to 2{}^{ \cdot }{\text{NO}}_{2}$$
(7)
$$^{ \cdot } {\text{NO }} + {}^{ \cdot }{\text{NO}}_{2} \to {\text{N}}_{2} {\text{O}}_{3}$$
(8)
$${\text{N}}_{2} {\text{O}}_{3} + {\text{ H}}_{2} {\text{O}} \to 2{\text{HONO}} \rightleftarrows 2{\text{H}}^{ + } + \, 2{\text{NO}}_{2}^{ - }$$
(9)

Reactions (5)–(9) can account for the reformation of nitrite and CblNO2, as well as the overall depletion of ascorbic acid from the solution over longer reaction times.

Traces of oxygen can also oxidize Cbl(II) to form CblOH2, which in turn will be reduced by HAsc− back to Cbl(II) [27,28,29], by which more ascorbic acid will be used. But at the same time, oxygen will be reduced to form superoxide, peroxide and OH·, which are stronger oxidants and can account for more redox cycling to occur during which ascorbic acid will be depleted, reactions (10)–(12). Finally, CblNO can also react with dioxygen to form CblNO2 and CblOH2 in a complex process involving OH· and ·NO2 intermediates [15].

$${\text{Cbl}}\left( {\text{II}} \right) \, + {\text{ O}}_{2} \to {\text{CblOH}}_{2} + {\text{ O}}_{2}^{ - }$$
(10)
$${\text{Cbl}}\left( {\text{II}} \right) \, + {\text{ O}}_{2}^{ - } \to {\text{CblOH}}_{2} + {\text{ O}}_{2}^{2 - }$$
(11)
$${\text{O}}_{2}^{2 - } + \, {\text{ H}}^{ + } \rightleftarrows {\text{HO}}_{2}^{ - } + {\text{ H}}^{ + } \rightleftarrows{\text{H}}_{2} {\text{O}}_{2} \rightarrow 2{\text{OH}}^{ \cdot }$$
(12)

The overall reaction sequence for the redox cycling of CblNO2, CblNO and CblOH2 in the presence of ascorbate is presented in Scheme 3. The main redox cycling components are marked in red and blue, whereas the oxidation of ascorbate in green. Complications caused by traces of oxygen that could not be avoided over very long reaction times are not included in the scheme for clarity reasons, but are given in reactions (7)–(12).

Scheme 3
scheme 3

Overall reaction sequence for the redox cycling of CblNO2, CblNO and CblOH2 in the presence of ascorbate

Full size image

The suggested reaction sequence in Scheme 3 is based on a one-electron reduction process by which nitrite is reduced to ·NO by ascorbate. From the recent literature [30, 31] it is known that free ·NO can be reduced by ascorbic acid to form HNO, which in turn can react with CblOH2 to form CblNO. Brasch and coworkers [32] demonstrated that using Angeli’s salt (HN2O3−) as source of HNO, the formation of CblNO can occur at pH > 10.8, where the rate-determining step is the release of HNO by Angeli’s salt, such that no mechanistic details about the mechanism of the reaction between CblOH2/CblOH and HNO/NO− could be revealed. At present, it is questionable whether further reduction of ·NO to HNO and a direct reaction of CblOH2 with HNO to form CblNO can account for the results presented in this study.

All in all, our goal to find suitable reaction conditions to produce CblNO from CblNO2 in the presence of a reducing agent over a long period of time was successful and has added to the overall understanding of the complex reaction system.

Conclusions

The reaction of CblNO2, one of the naturally occurring forms of cobalamin, with ascorbate has been studied by UV–Vis spectroscopy. The present study provides mechanistic information on this reaction at pH < 5. Under this condition, the only product of the reaction is CblNO. However, for the reduction of CblNO2 by ascorbate, no direct evidence for the reduction of coordinated nitrite could be found. On the contrary, we showed that excess of nitrite suppressed the formation of CblNO, from which we can conclude that ascorbic acid/ascorbate cannot reduce coordinated nitrite. We suggest that the studied system is not just a simple reaction, but a rather complex chemical process. During the reaction of ascorbic acid with nitrocobalamin, the first products formed are the reduced form of Vitamin B12 (Cbl(II)) and nitric oxide (·NO) that subsequently react rapidly to form CblNO. Our results show that the studied reactions are extremely oxygen sensitive due to the reverse oxidation of both Cbl(II) and CblNO to CblOH2 and CblNO2, respectively.