|Home | About | Journals | Submit | Contact Us | Français|
Kinetic evidence is reported for the role of the peroxymonocarbonate, HOOCO2−, as an oxidant for reduced Cu,Zn-superoxide dismutase-Cu(I) (SOD1) during the peroxidase activity of the enzyme. The formation of this reactive oxygen species results from the equilibrium between hydrogen peroxide and bicarbonate. Recently, peroxymonocarbonate has been proposed to be a key substrate for reduced SOD1 and has been shown to oxidize SOD1-Cu(I) to SOD1-Cu(II) much faster than H2O2. We have reinvestigated the kinetics of the reaction between SOD1-Cu(I) and HOOCO2− by using conventional stopped-flow spectrophotometry and obtained a second-order rate constant of k = 1600 ± 100 M−1s−1 for SOD1-Cu(I) oxidation by HOOCO2−. Our results demonstrate that peroxymonocarbonate oxidizes SOD1-Cu(I) to SOD1-Cu(II) and is in turn reduced to the carbonate anion radical. It is proposed that the dissociation of His61 from the active site Cu(I) in SOD-Cu(I) contributes to this chemistry by facilitating the binding of larger anions, such as peroxymonocarbonate.
It is well known that Cu,Zn-superoxide dismutase (SOD1) has additional catalytic activities other than its dismutase activity, such as peroxidase activity [1–4]. The SOD1 peroxidase cycle involves several steps. Reduction of SOD1-Cu(II) in the active site to SOD1-Cu(I) by H2O2 is followed by oxidation of SOD1-Cu(I) by a second molecule of H2O2, thus generating a hypothetical hypervalent or hydroxyl radical-like bound oxidant at the active site. This oxidant has been proposed to involve various high energy intermediates such as SOD-Cu(II)(•OH), SOD-Cu(I)=O, or SOD1-Cu(III) [3,5]. This oxidant, produced at the active site and derived from the reaction of SOD-Cu(I) with H2O2, has been demonstrated to oxidize histidines at the enzyme’s active site [6,7], ultimately leading to loss of the catalytic copper ion and SOD1 inactivation , and in the oxidation of nitrite [1,9].
It has also been found that the addition of CO2 causes marked increases in peroxidase activity [3–5], which led to the proposal that bicarbonate is oxidized to the diffusible CO3•− [1,10]. The CO3•− is believed to mediate the oxidations promoted by the SOD peroxidase activity [1,11–13]. Alternatively, it has been proposed that CO2 reacts directly with the hypervalent/bound hydroxyl radical oxidant produced at the SOD active site, producing the carbon dioxide cation radical CO2•+, which then hydrolyzes to produce CO3•−  according to Eqns. (3) and (4) or to produce CO3•− directly (5):
The primary difficulty with the proposed mechanism according to Eqns. (3) and (4) is that the ionization potential (IP) of CO2 has been determined to be 14.4 V [14,15]; in general, only very high-energy radiation is capable of producing CO2•+.
Alternative mechanisms for CO3•− formation from SOD/H2O2 in bicarbonate buffer have been proposed by some investigators [16–21] and criticized by others . We have hypothesized that peroxymonocarbonate (HCO4−), which is a small anionic species, can diffuse through the anionic channel to the enzyme active site and be reduced to the CO3•− radical by the SOD1-Cu(I) (Scheme 1) . Peroxymonocarbonate ion (HOOCO2−) is a well characterized covalent adduct of CO2 and hydroperoxide anion with a corresponding equilibrium constant of Keq = 0.32 M−1 . Very recent kinetic simulations were shown to be consistent with a mechanism in which CO2 forms from bicarbonate and subsequently reacts with hydrogen peroxide via perhydration  (Eqn. 6):
Deprotonation of the percarbonic acid (HOOCO2H) forms peroxymonocarbonate (Eqn. 7):
Equivalently, CO2 can react directly with the conjugate base HOO− via base-catalyzed perhydration (Eqn. 8):
Richardson and co-workers published an estimation of the rate constants for the forward reactions (6) and (7) at 25°C (0.002 M−1s−1 and 280 M−1s−1, respectively) and proposed that the contributions of the HOO− and H2O2 pathways are equivalent at pH 8, but above this value HOO− is dominant in the formation of HOOCO2− .
It has been shown that peroxymonocarbonate is able to oxidize organic sulfides [23,25] and sulfur-containing biomolecules [26,27], and in biological systems the formation of HOOCO2− is also accelerated by the presence of carbonic anhydrase [13,24]. On the basis of recent ESR and 13C-NMR studies , it was proposed that once SOD1-Cu(II) is reduced to SOD1-Cu(I) via the reaction shown in Eqn. (1) [3,5], the enzyme can be oxidized back to SOD1-Cu(II) by peroxymonocarbonate more efficiently than by H2O2 itself or HOO− (Scheme 1):
Previous estimated rates for Eqn. (9) were derived from ESR measurements and disagreed with the data published by Medinas et al. . In the present investigation, we have reexamined the rates and mechanisms of the bicarbonate-catalyzed hydrogen peroxide oxidation of SOD1-Cu(I) using a conventional stopped-flow technique. We provide kinetic evidence that peroxymonocarbonate (HOOCO2−) resulting from the equilibrium between carbon dioxide (CO2) and peroxide anion (HOO−) reacts with reduced SOD1-Cu(I).
Bovine kidney superoxide dismutase (SOD1) was purchased from Calzyme Laboratories, Inc. (San Luis Obispo, CA). Sodium phosphate, sodium bicarbonate, hydrogen peroxide (obtained as a 30% solution), ascorbic acid, and diethylenetriaminepentaacetic acid (DTPA) were from Sigma (St Louis, MO). The hydrogen peroxide concentration was determined from its absorbance at 240 nm (ε= 39.4 M−1cm−1).
The kinetics of the reactions of SOD1 with hydrogen peroxide, bicarbonate (HCO3−), and peroxymonocarbonate (HOOCO2−) were measured using a double-mixing stopped-flow apparatus model SF-61DX2 (HiTech Scientific, Bradford-on-Avon, UK) equipped with a rapid scanning diode array spectrophotometer and Kinet-Asyst software for data acquisition and analysis. Phosphate buffer, pH 7.4, was used with or without bicarbonate and treated with Chelex and 500 μM DTPA to protect the enzyme against H2O2-induced, copper-catalyzed SOD1 non-active site fragmentation . The concentration of SOD1 was determined from the band at 680 nm (ε= 300 M−1cm−1), which results from the d-d transitions of the Cu atom [29–32]. For the reactions, SOD1 was loaded into one syringe, and H2O2 or H2O2/HCO3− was loaded into a second syringe. The drive syringes and the reaction cell were equilibrated at 25 °C, and equal volumes of reactants were mixed to initiate the reaction. Spectra were acquired from 350 to 700 nm as a function of time to follow the reduction or oxidation of SOD1. The change in absorbance was monitored at 680 nm, and the data were fit to first-order-reaction kinetics. For reduction of SOD1-Cu(II) by H2O2, pseudo-first-order conditions were achieved by keeping the hydrogen peroxide concentration to at least a fivefold excess over the enzyme.
To obtain the second-order rate constant of SOD1 oxidation by peroxymonocarbonate (HOOCO2−), 400 μM SOD1-Cu(I) was prepared by incubating 400 μM SOD1-Cu(II) with a fivefold excess of ascorbic acid (2 mM), and the complete reduction was monitored at 680 nm with a Cary 100 spectrophotometer (Varian Inc., Palo Alto, CA) using a 500 μl quartz cuvette . After the complete reduction, the ascorbate was removed by using 7 ml/9K MWCO concentrators (Pierce, Rockford, IL), and the reduced enzyme was resuspended in 100 mM phosphate buffer, pH 7.4. Concentrations of HOOCO2− were at least eightfold greater than SOD1-Cu(I) and were calculated from the equation [HOOCO2−] = Keq × [HCO3−] × [H2O2], where Keq = 0.32 M−1 is the equilibrium constant of Eqn. (8) . The initial concentration of HCO3− was kept constant at 500 mM, but in order to preform different amounts of peroxymonocarbonate, the initial [H2O2] was varied, being set to 20 mM, 30 mM, 40 mM, or 50 mM.
When a 10-fold excess of hydrogen peroxide (2 mM final) was added to SOD1-Cu(II) (200 μM final), the absorbance maximum at 680 nm quickly disappeared as the enzyme was reduced to Cu(I). Our stopped-flow data (k = 45 ± 5 M−1s−1) for the reduction of SOD1-Cu(II) by H2O2 via Eqn. (1) (data not shown) confirmed previous results obtained by both ESR  and optical spectrometry [20,34]. Next, we examined mild reduction of the enzyme by using ascorbate . This cellular reductant was able to reduce the enzyme, but more slowly than H2O2, and the complete reduction was achieved within 40 min.
To demonstrate the effect of bicarbonate on the reduction of SOD1-Cu(II), 20 mM and 500 mM HCO3− were equilibrated for 30 min with 4 mM H2O2 at pH 7.4 before mixing with the protein (1:1 volume ratio). Upon mixing, these solutions yielded approximately 12.8 μM and 320 μM peroxymonocarbonate, respectively. Figure 1 shows the results of the stopped-flow experiments in which 400 μM SOD1-Cu(II) (initial concentration) was mixed in a 1:1 volume ratio with HCO3− (20 mM or 500 mM)/H2O2 (4 mM) at pH 7.4. As demonstrated in time trace b, 10 mM HCO3− (final concentration) had no effect on the initial rate of reduction of Cu(II), but it significantly decreased the yield of SOD1-Cu(I). The presence of bicarbonate decreases the net reduction of SOD1-Cu(II) by approximately 50%. This result is in agreement with previously published data , which showed that 10 mM HCO3− had no effect on the reduction rate of the enzyme by 2 mM H2O2. The authors also observed the decreased formation of SOD1-Cu(I) (Fig. 3 in reference ), but did not discuss it.
At the higher HCO3− concentration, the reduction of SOD1-Cu(II) was significantly diminished upon mixing with the enzyme (Fig. 1, trace c). Apparently, because of the presence of higher HOOCO2− (320 μM), SOD1-Cu(I) cannot accumulate in its reduced form due to the fast oxidation of SOD1-Cu(I) back to SOD1-Cu(II) .
To demonstrate that SOD1-Cu(I) is oxidized by peroxymonocarbonate, we reduced the resting SOD1 enzyme (400 μM) with 1 mM ascorbate. After the removal of the reductant (see Materials and Methods), we used stopped-flow experiments to calculate the rate of reoxidation of SOD1-Cu(I) back to SOD1-Cu(II). To obtain the equilibrium concentrations of peroxymonocarbonate, we separately mixed 500 mM and 200 mM HCO3− with 20 mM, 30 mM, 40 mM, or 50 mM H2O2 30 min prior to the experiments. Based on Eqn. (9) and the previously published value for Keq of the equilibrium represented by Eqn. (8), we estimated the steady-state initial concentration of HOOCO2− for the first set of data with 500 mM HCO3− to be 3.2 mM, 4.8 mM, 6.4 mM, and 8 mM. For the second set of data when using 200 mM HCO3− and the same concentrations of H2O2, the equilibrium concentrations of peroxymonocarbonate were calculated as 1.28 mM, 1.92 mM, 2.56 mM, and 3.2 mM. Figures 2A and 2B show the time courses of the reaction of SOD1-Cu(I) with the two sets of peroxymonocarbonate concentrations. Under these pseudo-first-order conditions, the initial rates of the reaction of SOD1-Cu(I) with peroxymonocarbonate were measured for each particular concentration of HOOCO2− by using a linear fit of the first 500 msec of the time course. The initial rates were divided by the initial concentration of the reduced enzyme to provide kobs rates. The second-order rate constant for SOD1-Cu(II) formation was determined from the kobs rates as a function of HOOCO2− concentration (Fig. 2C) and was estimated to be 1600 ± 100 M−1s−1, which is ~10-fold more rapid than the rate previously determined based on our ESR experiments  but similar to the rate reported by Medinas et al. (2.0 × 103 M−1s−1) . As shown in Fig. 2C, there is a linear dependence between both sets of data. Since we varied the concentrations of hydrogen peroxide to obtain the equilibrium concentrations of HOOCO2− and the pseudo-first-order conditions were kept in respect to H2O2, we plotted the determined kobs values as a function of H2O2. As shown in Fig. 2D, each set of data could be fitted linearly with H2O2 concentration, but different rate constants were obtained depending on the two different bicarbonate concentrations, which demonstrates that in the presence of bicarbonate, no reaction of SOD1-Cu(I) with hydrogen peroxide occurs on this time scale.
To confirm that SOD1-Cu(I) was oxidized by peroxymonocarbonate, we used stopped-flow experiments without any preequilibration. First, as a control, we mixed reduced SOD1 (200 μM final) with 250 mM bicarbonate, and the typical time trace showed no oxidation of Cu(I), as expected (Fig. 3, trace a). Next, to characterize the ability of H2O2 to oxidize SOD1-Cu(I), we mixed the protein with the lowest and highest concentrations of H2O2 (10 mM and 25 mM) that we used for the experiments with preequilibration described above. The very slow absorbance increase at 680 nm (traces b and c) indicated a negligible oxidation of Cu(I) by H2O2, which is in agreement with our previously published ESR data showing that H2O2 is not an efficient mediator of SOD1-Cu(I) oxidation [Fig. 2, trace c in reference ]. To further demonstrate that SOD1-Cu(I) is oxidized by HOOCO2−, we prepared 400 μM SOD1-Cu(I) in 100 mM phosphate buffer containing 500 mM bicarbonate at pH 7.4 in one of the stopped-flow reactors and filled the other reactor with 50 mM H2O2. Upon mixing (1:1 volume ratio), no detectable SOD1-Cu(II) was formed. During the time course of 100 s, the known inactivation of SOD by H2O2 with a rate of 3.1 M−1s−1 would result in the loss of >90% of the SOD activity . In the initial absence of peroxymonocarbonate, the SOD1 would be rapidly destroyed by H2O2, accounting for the lack of significant SOD1-Cu(II) formation in Fig. 3, trace d.
Recently, we proposed that the SOD1 peroxidase cycle proceeds via two steps . First, hydrogen peroxide reduces SOD1-Cu(II) and second, the reduced SOD1-Cu(I) is oxidized back to SOD1-Cu(II) by peroxymonocarbonate anion (Scheme 1). It has been previously proposed that peroxymonocarbonate anion, like other small anions, can diffuse through the anion channel of SOD to the catalytically active site . Structural studies indicate that small anions such as azide and thiocyanate are able to bind directly to the active-site copper ion [36,37], while larger anions such as acetate, borate, formate, perchlorate, phosphate, and sulfite interact with the anion-binding channel but do not appear to bind directly to the copper [38–40]. Nevertheless, larger ions such as dithionite (in equilibrium with SO2•−) and sulfite are able to reduce the active site SOD-Cu(II) to SOD-Cu(I) [37,41].
However, recent theoretical analyses of SOD activity indicate that the bond connecting the bridging His61 (His63 in the human enzyme) with the copper is ruptured upon formation of the SOD-Cu(I) species . In this type of structure, the anion binding site becomes able to accommodate substantially larger ligands. For example, in the analogous two-Zn2+ structure in which the His63 imidazole does not bind directly with the active-site Zn2+ (PDB code: 1P1V ), a sulfate ion is bonded directly with the active-site Zn2+ ion. Thus, it seems quite possible for the peroxymonocarbonate ion to interact directly with the SOD-Cu(I) ion in a structure lacking the Cu(I)-His61 bond. This direct interaction between the Cu(I) ion and the peroxymonocarbonate may be important for the observed redox chemistry; however, it is not essential for electron transfer.
It has also been suggested that in the absence of bicarbonate buffer, SOD1-Cu(I) can be oxidized to Cu(I) peroxide complex, which probably decays to a copper-bound hydroxyl radical complex, SOD1-Cu(II) (•OH), ultimately causing enzyme inactivation. Originally, it was assumed that in the presence of bicarbonate buffer, the hydroxyl radical-like SOD intermediate would react with HCO3−, leading to CO3•− production [1,34,43]. More recently it was proposed that CO2 may be the substrate for the enhanced peroxidase activity of SOD1 on the basis of the carbonic anhydrase-induced inhibition of the stimulation of NADPH oxidation by SOD1 in the presence of H2O2 and CO2-saturated water . However, the oxidation of SOD1-Cu(I) by peroxymonocarbonate would form the CO3•− (E = 1.78 V)  directly, without the need for the formation of the much higher energy hydroxyl radical (E = 2.31 V)  or its equivalent SOD intermediate, SOD1-Cu(II)(•OH). The substantially lower oxidation potential of the CO3•− radical relative to the hydroxyl radical-like species is likely reflected in a much more favored reaction .
13C NMR experiments [19,20] have confirmed the existence of a peroxymonocarbonate (HOOCO2−) intermediate as the product of the reaction of CO2 with hydrogen peroxide in agreement with Richardson’s studies [23–26]. The stopped-flow experiments performed by Medinas et al. also supported a direct reaction between peroxymonocarbonate and human SOD1 . Medinas et al. favor a related mechanism in which SOD1 reacts with HOOCO2− produced as a result of carbon dioxide addition to SOD1-Cu(I)(OOH−) rather than the reaction of SOD1-Cu(I) with freely diffusing HOOCO2− which, in these experiments, is preformed outside the active site. However, our results are fully consistent with the conclusion that HOOCO2− increases human SOD1 turnover with production of CO3•− .
Further support for the conclusion that SOD1-Cu(I) is oxidized by HOOCO2− produced in the bulk solvent is derived from our nonequilibrated stopped-flow studies (Fig. 3). These studies show that very slow, if any, oxidation of native SOD1-Cu(I) by H2O2 or unequilibrated mixtures of H2O2 with HCO3− occurs. However, the results with preequilibrated mixtures of HCO3− and H2O2 (Fig. 2) are consistent with peroxymonocarbonate being the reactive oxygen species which oxidizes SOD1-Cu(I) with a rate constant of 1600 ± 100 M−1s−1 at 25 °C, pH 7.4.
From previously published ESR data we calculated a rate constant of 150 M−1s−1 for SOD1-Cu(I) oxidation by preequilibrated H2O2/HCO3− mixtures . In contrast, published kinetic simulations by Medinas et al.  estimated a more than 10-fold higher second-order rate constant for human SOD1-Cu(I) oxidation by peroxymonocarbonate. This discrepancy led us to reexamine the kinetic experiments, but using a stopped-flow technique to minimize any inaccuracy. By reexamining the kinetic experiments with the stopped-flow technique, we estimated the second-order rate to be 1600 M−1s−1. This rate constant is close to the rate estimated for human SOD1  and is 10-fold higher than the rate previously published , which was apparently underestimated due to inaccuracies with the relatively slow manual freezing for ESR studies. Regardless of the previously published value or the differences in the proposed mechanism, the rate constants determined here and by Medinas et al.  indicate that the oxidation of SOD1-Cu(I) by HOOCO2− is more than 100-fold faster than the oxidation of SOD1-Cu(I) by H2O2 :
At the physiological concentration of bicarbonate (25 mM) and a concentration of H2O2 of 10 mM, the equilibrium concentration of peroxymonocarbonate is 80 μM. After substitution of the initial concentrations of the reactants and estimated rate constants, the rates can be compared according to:
which gives υ9 ≈ υ2. The fact that the reaction rate is the same for both Eqns. (2) and (9) raises the question of whether they are, in fact, two interpretations of the reaction we presented in Eqn. (9). The rate of formation (υ2) of the SOD1-bound hydroxyl radical was determined indirectly by the oxidation of NADPH by CO3•− , whereas our data from direct measurement of SOD1-Cu(II) formation demonstrated that peroxymonocarbonate is the oxidant for SOD1-Cu(I) via Eqn. (9). Thus, our results indicate that Eqn. (9) rather than Eqns. (3) and (4) or (5) is responsible for the formation of CO3•−. As a result, the reaction of H2O2 and SOD1-Cu(I) with a reported rate constant 13 M−1s−1 [22,34] had, in fact, not been measured at all and may be near zero, and, in any case, is much slower than the inactivation of SOD1 by H2O2 as we have found (Fig. 4). Fridovich and Liochev  have argued that CO2 pulls the equilibrium of Eqn. (3) or (5) to the right by reacting with the SOD1-bound oxidant. Since we have determined the initial rates of our reaction (9), subsequent reactions such as Eqn. (3) can have no effect.
Regardless of the merit of our mechanism (Scheme 1), the oxidation of CO2 by the SOD oxidant, Eqn. (3), formed at the active site, Eqn. (2), is thermodynamically very unfavorable [14,15]. Even assuming that the hypervalent/bound hydroxyl radical SOD oxidant is as strong as the free hydroxyl radical (E = 2.32 V) , the oxidation of CO2 will be essentially zero as the following calculations show (vide infra). As discussed in the Introduction, the ionization potential for CO2•+ was determined by Mackay as 14.4 V [14,15]. Previous studies have correlated the ionization and oxidation potentials of many compounds [46,47]. For example, Miller et al.  used the equation E = 0.89 × (IP) − 6.04 with a correlation coefficient of 0.95. Based on this equation, we estimated that the oxidation potential for Eqn. (10) is 6.78 V:
Given ΔG0reaction = − nFΔE0 = − 1 × 23.06 × [(− 6.78 V) − (− 2.32 V)] = 102.8 kcal = 430 kJ, where n is the number of electrons transferred per mole, F is the Faraday constant = 23.06 kcal/V, and ΔE0reaction = E010 − E011 , we calculated the Gibbs energy for Eqn. (3), showing that the reaction is thermodymanically very unfavorable (ΔG 0). Knowing the Gibbs energy for Eqn. (3) allows one to calculate the equilibrium constant as Keq = e−ΔG/RT, which we determined to be Keq = 2.71 × 10−76. Assuming that Eqn. (3) is reversible [3,22] and that the reverse reaction can not exceed the diffusion limit 1 × 1010 M−1s−1 requires that the forward reaction can not be faster than 2.71 × 10−66 M−1s−1, i.e., effectively zero.
In summary, using direct optical measurements of SOD1-Cu(II) formation, we show that SOD1-Cu(I) is oxidized by HOOCO2− generated in the equilibration of HCO3−/CO2 and H2O2. We propose that bicarbonate is unique among the small anions known to interact with the copper of SOD1 in that it alone is known to interact with hydrogen peroxide to form an inorganic peroxide (peroxymonocarbonate) that will be rapidly reduced by SOD1-Cu(I) to form the oxidizing carbonate anion radical.
We would like to acknowledge Mrs. Mary J. Mason and Dr. Ann Motten for their help in the preparation of the manuscript. The authors also thank Dr. Richard S. Magliozzo from Brooklyn College for the use of his stopped-flow spectrophotometer. This work has been supported by the Intramural Research Program of the NIH, NIEHS.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.