Reexamination of the mechanism of hydroxyl radical adducts formed from the reaction between familial amyotrophic lateral sclerosis-associated Cu,Zn superoxide dismutase mutants and H₂O₂

Abstract

Amyotrophic lateral sclerosis (ALS) involves the progressive degeneration of motor neurons in the spinal cord and motor cortex. Mutations to Cu,Zn superoxide dismutase (SOD) linked with familial ALS are reported to increase hydroxyl radical adduct formation from hydrogen peroxide as measured by spin trapping with 5,5′-dimethyl-1-pyrrolline N-oxide (DMPO). In the present study, we have used oxygen-17-enriched water and H₂O₂ to reinvestigate the mechanism of DMPO/^⋅OH formation from the SOD and SOD mutants. The relative ratios of DMPO/^⋅17OH and DMPO/^⋅16OH formed in the Fenton reaction were 90% and 10%, respectively, reflecting the ratios of H₂¹⁷O₂ to H₂¹⁶O₂. The reaction of the WT SOD with H₂¹⁷O₂ in bicarbonate/CO₂ buffer yielded 63% DMPO/^⋅17OH and 37% DMPO/^⋅16OH. Similar results were obtained from the reaction between familial ALS SOD mutants and H₂¹⁷O₂: DMPO/^⋅17OH (64%); DMPO/^⋅16OH (36%) from A4V and DMPO/^⋅17OH (62%); and DMPO/^⋅16OH (38%) from G93A. These results were confirmed further by using 5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide spin trap, a phosphorylated analog of DMPO. Contrary to earlier reports, the present results indicate that a significant fraction of DMPO/^⋅OH formed during the reaction of SOD and familial ALS SOD mutants with H₂O₂ is derived from the incorporation of oxygen from water due to oxidation of DMPO to DMPO/^⋅OH presumably via DMPO radical cation. No differences were detected between WT and mutant SODs, neither in the concentration of DMPO/^⋅OH or DEPMPO/^⋅OH formed nor in the relative incorporation of oxygen from H₂O₂ or water.

Amyotrophic lateral sclerosis (ALS), also known as Lou Gehrig’s disease or motor neuron disease, is a condition in which there is degeneration of the motor neurons of the spinal cord, brain stem, and cerebral cortex (1–8). Approximately 10% of ALS cases are familial with the remainder being sporadic. The genetic defect in 20% of familial ALS (FALS) cases now has been linked to Sod 1, the gene that encodes the cytosolic Cu,Zn superoxide dismutase (SOD) enzyme (2, 9, 10). To date, the only humans known to exhibit a genetic defect in the coding regions of SOD are those with the FALS disorder.

The mechanisms by which FALS-linked Sod 1 mutants cause selective degeneration of motor neurons remain unclear. Cu, ZnSOD is an antioxidant enzyme that primarily catalyzes the dismutation of O_2⨪ to O₂ and H₂O₂ (11, 12). The remarkably high reaction rate of Cu,ZnSOD is thought to be caused by the electrostatic guidance of the charged species (O_2⨪) into the active site by conserved charged amino acid residues (13). It has been hypothesized that mutations at various sites in Sod 1 may push open the loops that form the superoxide-binding pocket and expose the active site to reaction with other oxidants (14). The gain-of-function hypotheses in FALS pathogenesis are based on the premise that the mutant SOD has an increased ability to react with H₂O₂ to generate higher oxidants (^⋅OH or peroxidase activity) or an increased ability to react with peroxynitrite to form nitrated tyrosines (14–20). Evidence for increased peroxidase activity or ^⋅OH formation from FALS-associated SOD mutants (A4V and G93A) was first obtained through the electron spin resonance (ESR) spin-trapping technique (16–23). The spin-trapping technique involves trapping of a reactive radical (i.e., superoxide anion or hydroxyl radical) by a nitrone spin trap to yield a more persistent nitroxide spin adduct that can be detected by ESR. By using this technique, Stadtman, Yim, and coworkers (18–20) found that FALS mutants generated increased formation of ^⋅OH upon reaction with H₂O₂. The intensity of the ESR signal due to DMPO/^⋅OH was much greater with FALS-associated SOD mutants than with the wild-type (WT) Cu,ZnSOD (16–20). Recently, Fridovich criticized these spin-trapping interpretations and suggested that the DMPO/^⋅OH adduct was formed from the peroxidase activity of SOD and not from trapping of free hydroxyl radical (24, 25). Bredesen and Valentine and co-investigators also suggested that the peroxidase activity of SOD was responsible for oxidation of DMPO to DMPO/^⋅OH and that FALS-associated SOD mutants exhibit considerably greater peroxidase activity (16, 17). However, no evidence for increased hydroxyl radical formation was obtained by using salicylate hydroxylation and lipid peroxidation assays in transgenic mice that develop progressive motor neuron disease from expressing human FALS-linked Sod 1 mutations (26). Marklund et al. also did not find any evidence for increased reactivity of the FALS mutant SOD D90A with hydrogen peroxide (27). Because of these conflicting reports, we decided to reexamine the previous spin-trapping investigations (16, 20). In the present study, we used two spin traps: 5,5′-dimethyl-1-pyrroline N-oxide (DMPO) and its phosphorylated analog 5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide (DEPMPO) (28–30). To determine the source of the oxygen atom in the hydroxyl radical adducts, we used oxygen-17-enriched hydrogen peroxide ([¹⁷O]-H₂O₂) and water ([¹⁷O]-H₂O). Spin-trapping experiments were performed in a loop-gap resonator (31, 32), which enabled us to use exceedingly small volumes of expensive spin traps and ¹⁷O-labeled compounds. The mutants studied were the following: A4V (Ala⁴ → Val) and G93A (Gly⁹³ → Ala) (33). A4V causes the most aggressive form of FALS (34), and G93A has been used previously in several transgenic mouse studies (34).

EXPERIMENTAL PROCEDURES

Materials.

[¹⁷O]-H₂O (45%) and [¹⁷O]-H₂O₂ (89%) were obtained from ICON Isotopes (Summit, NJ). DMPO was obtained from Sigma and double distilled to remove the paramagnetic impurities. DEPMPO was synthesized and purified as described (30). The bicarbonate/CO₂ buffer solutions were prepared by treating with Chelex 100 resin (Bio-Rad) to remove trace polyvalent metal ions. The pH of the Chelex-treated buffer was readjusted to pH 7.4 by bubbling with 95% N₂/5% CO₂ mixture.

Preparation of SOD Mutants.

Both WT and mutant SODs were expressed in a bacterial system and purified by ammonium sulfate fractionation and anion exchange chromatography as described (35). The metal content of the purified SODs was determined by the colorimetric 4-pyridylazaresorcinol assay (36). The SODs were suspended in 10 mM sodium acetate (pH 5.0). Sufficient cupric citrate and zinc sulfate were added to bring the metal content to 110% of the available copper and zinc binding sites on SOD, and the SODs were incubated overnight at 4°C. The SODs were repurified over a high resolution HQ10 anion exchange resin (PerSeptive Biosystems) by using 20 mM Tris⋅HCl with a 0–200 mM NaCl gradient over 10 column volumes. The second protein peak generally contained the highest activity with 95–103% of copper and zinc. The protein was concentrated in water and frozen until use. Protein concentrations were determined by the bicinchoninic acid method (Pierce). A stock of the WT human SOD was used as the protein standard, whose activities were determined by the cytochrome c method (11) and were ≈5,200 units/mg protein. SOD activity in the metal-replete mutant proteins was equivalent to that of the WT protein as reported (36) and was directly proportional to the copper content in partially metal containing fractions. Zinc-deficient SOD was prepared by dialyzing 1 mg/ml Cu,Zn SOD exhaustively against Chelex 100-treated 100 mM potassium phosphate (pH 3.5) until the zinc content was reduced to <10%, as described (36, 37).

ESR Spin-Trapping.

ESR spectra were recorded at room temperature on a Varian E-109 spectrometer operating at 9.5 GHz and with a 100-KHz field modulation equipped with a TE₁₀₂ cavity or a loop-gap resonator. Reactions were initiated by the addition of H₂O₂ to the incubation mixtures containing SOD (100 μg/ml), DMPO (100 mM), and diethylenetriaminepentaacetic acid (DTPA) (100 μM). A typical reaction mixture for ESR analysis consisted of 1.5 μl of WT-SOD or SOD mutants (1 mg/ml), 1 μl of DMPO (1 M) or DEPMPO (0.5 M), 1 μl of [¹⁶O]-H₂O₂ or [¹⁷O]-H₂O₂ (15 mM), and 11.5 μl of bicarbonate/CO₂ buffer (25 mM, pH 7.4) containing DTPA (100 μM) in a total volume of 15 μl. For experiments with [¹⁷O]-H₂O, the reaction mixture consisted of 1.5 μl of WT-SOD or SOD mutants (1 mg/ml), 1 μl of DMPO (1 M) or DEPMPO (0.5 M), 7.5 μl of [¹⁷O]-H₂O (45%), 1 μl of [¹⁶O]-H₂O₂, and 4 μl of bicarbonate/CO₂ buffer (100 mM, pH 7.4) containing DTPA (0.4 mM) in a total volume of 15 μl. The sample was transferred to a capillary tube (0.64 mm i.d. × 0.84 mm o.d. and 100 mm long) that was sealed with miniseal (Baxter Scientific Products, McGaw Park, IL) and placed into the loop-gap resonator. Computer-based simulations of ESR spectra were performed by using software written by Duling (38). The correlation coefficient (r) for the spectral simulation was 0.994 ± 0.001. Spin adduct concentrations were obtained by double integration using 3-carbamoyl-2,2,5,5-tetramethyl-3-pyrroline-1-yloxy as a standard. Spectrometer conditions used in spin trapping studies were: microwave power, 2 mW; time constant, 0.128 s; modulation amplitude, 0.5 G; and scan time, 2 min.

RESULTS

Formation of DMPO- and DEPMPO-Hydroxyl Radical Adducts.

Addition of H₂O₂ (0.5–5 mM) to a solution containing WT-Cu,ZnSOD (100 μg/ml) or the FALS SOD mutant (100 μg/ml) in a bicarbonate buffer (pH 7.4) containing the metal ion chelator DTPA produced a four-line ESR spectrum (a^N = a^H = 15 G) with an intensity ratio of 1:2:2:1 corresponding to the DMPO-hydroxyl adduct (DMPO/^⋅OH) (Fig. 1A). The spectral intensity of DMPO/^⋅OH increased to a steady-state concentration (≈50 μM) within 10 min. The spectral intensity of DMPO/^⋅OH obtained from the reaction between H₂O₂ and the FALS mutants G93A or A4V was nearly the same as that obtained with WT-SOD (Fig. 1A). The concentration of DMPO/^⋅OH produced was 93 ± 5% for A4V and 88 ± 6% for G93A relative to the WT-SOD and was found to be independent of DTPA concentration (0.1–1 mM). The yields of DMPO/^⋅OH from zinc-deficient WT-SOD or FALS mutant protein in the presence of H₂O₂ were found to be identical (data not shown). The rate of Cu²⁺ release during the reaction among WT-SOD, A4V, and G93A (100 μg/ml) with H₂O₂ (1 mM) in PBS at 37°C also was not significantly different (1.5 nM/s for WT and A4V and 1.1 nM/s for G93A) (data not shown).

Figure 1.

Formation of DMPO/^⋅OH and DEPMPO/^⋅OH adducts during the reaction between H₂O₂ and WT-SOD or SOD mutants. (A) H₂O₂ (5 mM), DMPO (50 mM), DTPA (100 μ M), and WT-SOD (100 μg/ml) or A4V or G93A (100 μg/ml) were incubated in bicarbonate/CO₂ buffer (25 mM, pH 7.4). (B) H₂O₂ (5 mM), DEPMPO (50 mM), DTPA (100 μM), and WT-SOD (100 μg/ml) or A4V or G93A (100 μg/ml) were incubated in bicarbonate/CO₂ buffer (25 mM, pH 7.4). The spectra were recorded immediately after the reaction was initiated with the enzyme. Spectra are representative scans of three independent experiments.

Next, we verified whether DEPMPO, a structural analog of DMPO, could trap the hydroxyl radical-like oxidant formed during the reaction between H₂O₂ and SOD or SOD mutants. Addition of WT-SOD or SOD mutant to a solution containing H₂O₂ (5 mM) and DEPMPO (25 mM) in a bicarbonate buffer (25 mM, pH 7.4) produced an eight-line spectrum with an intensity ratio of 1:2:2:1:1:2:2:1 (Fig. 1B). Based on the literature data (28, 29), this spectrum (a^P = 47.3 G, a^H = 13.2 G, and a^N = 14.0 G) was assigned to the DEPMPO-hydroxyl adduct (DEPMPO/^⋅OH). Formation of this spectrum was absolutely dependent on all of the components. The spectral intensity of DEPMPO/^⋅OH obtained from the reaction between H₂O₂ and the FALS mutants G93A or A4V was nearly the same as that obtained with WT-SOD (Fig. 1B). Contrary to previous reports (16–20), the present spin-trapping data obtained using two different spin traps demonstrate that FALS SOD mutants (A4V and G93A) do not generate a significant increase in hydroxyl radical adduct formation compared with the WT-SOD.

Incorporation of ¹⁷O Atom from ¹⁷O-H₂O₂ into DMPO/^⋅OH Adduct.

To investigate further the mechanism of formation of DMPO/^⋅OH, spin-trapping experiments were carried out in a solution containing 89% ¹⁷O-labeled hydrogen peroxide (39). The rationale for using the isotopically enriched H₂O₂ is as follows: The nuclear spin quantum number (I) for the ¹⁷O atom is 5/2 in contrast to the ¹⁶O, which has I = 0, and, thus, additional couplings from the ¹⁷O atom will be observed if there is any incorporation of oxygen atom into the DMPO/^⋅OH adduct from hydrogen peroxide.

By using [¹⁷O]-H₂O₂, it was shown that nearly all of the DMPO/^⋅OH formed in the Fenton system (Fe²⁺ + H₂O₂) came from hydrogen peroxide as reported (39). The ESR spectrum of DMPO/^⋅OH obtained from incubations containing [¹⁷O]-H₂O₂, DMPO, and Fe(II)-EDTA in a bicarbonate/CO₂ buffer consists of 90% DMPO/^⋅17OH and 10% DMPO/^⋅16OH (Fig. 2) matching the isotopic distribution of the hydrogen peroxide. In contrast to the Fenton system, the WT-SOD and the mutants (G93A and A4V) yielded considerably lower amounts of oxygen-17-incorporated hydroxyl adducts (Table 1). We estimate that nearly 35% of DMPO/^⋅OH arises from the incorporation of oxygen from a source other than H₂O₂, presumably water, and that 65% of DMPO/^⋅OH is derived from incorporation of oxygen from H₂O₂ in a reaction catalyzed by SOD or SOD mutants.

Figure 2.

The effect of ¹⁷O-labeled H₂O₂ on DMPO/^⋅OH formation. Incubation mixtures contained H₂O₂ (5 mM), DMPO (50 mM), and DTPA (100 μM) in a bicarbonate/CO₂ buffer. The reaction was initiated with either Fe(II)-EDTA (100 μM) or 100 μg of SOD or SOD mutants (G93A and A4V) as shown. Line positions from ¹⁷O and ¹⁶O couplings are shown by • and ○, respectively. The dotted lines show computer simulations of ESR spectra using the parameters a^N = 15.0 G and a^H = 15.0 G for DMPO/^⋅16OH and a^N = 15.0 G, a^H = 15.0 G, and a^17O = 4.6 G for DMPO/^⋅17OH.

Table 1.

The relative concentrations of ¹⁷O and ¹⁶O spin adducts of DMPO/^⋅OH formed during the reaction of SOD with 89% [¹⁷O]-H₂O₂

Conditions	% of DMPO/⋅OH spin adducts from H2O2	% of DMPO/⋅OH spin adducts from H2O
Fe(II)-EDTA	100	0
WT-SOD	63  ±  0.5	37  ±  0.5
A4V	64  ±  1.4	36  ±  1.4
G93A	62  ±  0.8	38  ±  0.8

Data were normalized with respect to the ¹⁷O content in H₂¹⁷O₂ by using the formula: %DMPO/^⋅OH from water = 100(%H₂¹⁷O₂ − %DMPO/^•17OH)/%H₂¹⁷O₂. 

Incorporation of ¹⁷O Atom from [¹⁷O]-H₂O into DMPO/^⋅OH Adduct.

To determine the source of the remaining fraction of DMPO/^⋅OH, we investigated the effect of H₂¹⁷O. The Fenton reaction (Fe²⁺ + H₂¹⁶O₂) was carried out in H₂¹⁷O in the presence of DMPO (Fig. 3A). The ESR spectrum of DMPO/^⋅OH recorded at a higher gain was identical to that obtained in H₂¹⁶O (Fig. 3A, Lower). This result clearly established that there is no exchange between the ¹⁷O in water and the ¹⁶O present in DMPO/^⋅16OH (DMPO/^⋅16OH + H₂¹⁷O –X→ DMPO/^⋅17OH + H₂¹⁶O) (39). In contrast to the Fenton reaction, incubation mixtures containing WT-SOD or SOD mutants, H₂O₂, and DMPO in bicarbonate/CO₂ buffer prepared in 45% H₂¹⁷O produced additional hyperfine couplings from DMPO/^⋅17OH (marked • in Fig. 3B, Upper). Computer simulation of the total signal (dotted line in Fig. 3B) from both the DMPO/^⋅16OH and DMPO/^⋅17OH adducts showed that ≈35% of DMPO-hydroxyl adduct was formed from the addition of water to DMPO in a reaction catalyzed by SOD and H₂O₂ (39). Similar ESR spectra were obtained from incubations containing FALS SOD mutants, H₂O₂, and DMPO in bicarbonate/CO₂ buffer prepared in 45% H₂¹⁷O (Fig. 3C).

Figure 3.

The effect of ¹⁷O-labeled H₂O on DMPO/^⋅OH formation. (A, Upper) Incubation mixture contained Fe(II)-EDTA (100 μM), H₂O₂ (5 mM), and DMPO (50 mM) in a bicarbonate/CO₂ buffer made with H₂¹⁷O; (Lower) obtained at gain × 10. (B, Upper) Incubation mixtures contained H₂O₂ (5 mM), DMPO (50 mM), WT-SOD (100 μg/ml), and DTPA (100 μM) in a bicarbonate/H₂O₂ buffer made with H₂¹⁷O (pH 7.4); (Lower) same as Upper but obtained at gain × 10. The dotted lines show a computer simulation of ESR spectra using the parameters a^N = 15.0 G and a^H = 15.0 G for DMPO/^⋅16OH and a^N = 15.0 G, a^H = 15.0 G, and a^17O = 4.6 G for DMPO/^⋅17OH. (C) Conditions were the same as in B (Lower) using SOD mutants.

Incorporation of ¹⁷O Atom from [¹⁷O]-H₂O₂ into DEPMPO/^⋅OH Adduct.

The oxidation potentials of DMPO and DEPMPO are ≈1.87 V (vs. NHE) and 2.24 V (vs. NHE), respectively. We surmised that the 370-mV difference in oxidation potential could help differentiate between the cation radical and hydroxyl radical-mediated formation of hydroxyl adducts. However, similar results were also obtained with DEPMPO trap. The ESR spectrum of DEPMPO/^⋅OH obtained from incubations containing [¹⁷O]-H₂O₂, DEPMPO, and Fe(II)-EDTA in a bicarbonate/CO₂ buffer consists of 90% DEPMPO/^⋅17OH and 10% DEPMPO/^⋅16OH matching the distribution of the hydrogen peroxide. (Fig. 4). In contrast to the Fenton system, the WT-SOD and the mutants (G93A and A4V) yielded considerably lower amounts of oxygen-17 incorporated in DEPMPO/^⋅OH (Fig. 4 and Table 2). As with DMPO, we estimate that ≈35–40% of DEPMPO/^⋅OH is derived from incorporation of oxygen from water, and the rest (60–65%) of DEPMPO/^⋅OH arises from incorporation of oxygen from H₂O₂ in a reaction catalyzed by SOD or SOD mutants.

Figure 4.

The effect of ¹⁷O-labeled H₂O₂ on DEPMPO/^⋅OH formation. Incubation mixtures contained H₂O₂ (5 mM), DEPMPO (20 mM), and DTPA (100 μM) in a bicarbonate/CO₂ buffer. The reaction was initiated with either Fe(II)-EDTA (100 μM) or 100 μg of SOD or SOD mutants (G93A and A4V) as shown. Line positions from ¹⁷O and ¹⁶O couplings are shown by • and ○, respectively. The dotted lines show a computer simulation of ESR spectra using the parameters a^N = 14.0 G, a^H = 13.2 G, a^P = 47.0 G for DEPMPO/^⋅16OH and a^N = 14.0 G, a^H = 13.1 G, a^P = 47.0 G, and a^17O = 4.1 G for DEPMPO/^⋅17OH).

Table 2.

The relative concentrations of ¹⁷O and ¹⁶O spin adducts of DEPMPO/^⋅OH formed during the reaction of SOD with 89% [¹⁷O]-H₂O₂

Conditions	% of DEPMPO/⋅OH spin adducts from H2O2	% of DEPMPO/⋅OH spin adducts from H2O
Fe(II)-EDTA	100	0
WT-SOD	60  ±  2.82	40  ±  2.82
A4V	66  ±  0.47	34  ±  0.47
G93A	63  ±  0.47	37  ±  0.82

Data were normalized with respect to the ¹⁷O content in H₂¹⁷O₂ by using the formula: %DEMPMPO/^⋅OH from water = 100(%H₂¹⁷O₂ − %DEPMPO/^⋅17OH)/%H₂¹⁷O₂. 

DISCUSSION

SOD-Dependent Radical Reactions.

Table 3 shows the antioxidant and prooxidant reactions of SOD. Cu,ZnSOD is an antioxidant enzyme that primarily catalyzes the dismutation of O_2⨪ to O₂ and H₂O₂ (11, 12). In addition to this dismutase activity, SOD also exhibits prooxidant activities, which include the peroxidase activity, hydroxyl radical generating activity and nitration of tyrosine (15–20, 40, 41). A novel superoxide-dependent peroxidase activity recently has been reported for the FALS mutant H48Q in which histidine 48 has been replaced by glutamine (42).

Table 3.

Free radical reactions of

Dismutase activity
1	SOD−Cu2+ + O2⨪ → SOD−Cu+ + O2
2	SOD−Cu+ + O2⨪ → SOD−Cu2+ + H2O2
3
Peroxidase activity
4	SOD−Cu2+ + H2O2 ⇌ SOD−Cu+ + O2⨪ + 2H+
5	SOD−Cu+ + H2O2 → SOD−Cu2+−⋅OH + OH−
6	SOD−Cu2+−⋅OH + His-61 → His⋅ + SOD−Cu2+ + H2O
7	SOD−Cu2+−⋅OH + HCO2− → SOD−Cu2+ + CO2⨪ + H2O
8	SOD−Cu2+−⋅OH → inactive SOD (Enz−Cu2+ + 2-oxohistidine)
Hydroxyl radical generating activity
9	Enz−Cu2+ + H2O2 → Enz−Cu+ + O2⨪ + 2H+
10	Enz−Cu+ + H2O2 → Enz−Cu2+ + ⋅OH + OH−
Nitration of tyrosine
11	SOD−Cu2+ + ONOO− → SOD−CuO....NO2+
12	SOD−CuO....NO2+ + Tyr → SOD−Cu2+ + OH− + NO2-Tyr

The peroxidase activity refers to the “bound” hydroxyl radical to the copper atom at the active site (i.e., SOD−Cu²⁺−^⋅OH) and not to the conventional compound I-type oxidant derived from heme proteins. For example, the horseradish peroxidase/H₂O₂ system, which forms compound I, does not oxidize DMPO to DMPO/^⋅OH.

Mechanism of Formation of DMPO/^⋅OH.

The mechanism of formation of hydroxyl radical adduct was attributed to the trapping of enzyme-bound oxidant SOD − Cu²⁺ − ^⋅OH by DMPO (SOD − Cu²⁺ − ^⋅OH + DMPO → SOD − Cu²⁺ + DMPO/^⋅OH) (15). However, the present data show only a partial transfer of oxygen-17 from H₂¹⁷O₂. We interpret the present spin-trapping data by a mechanism involving both oxidation and hydroxylation of DMPO by Cu²⁺ − ^⋅OH (43–45).

The incorporation of ¹⁷O from water into DMPO/^⋅OH can occur by at least two known mechanisms. In the first, a Cu²⁺-catalyzed nucleophilic addition of H₂¹⁷O to DMPO results in the formation of the corresponding hydroxylamine of DMPO/^⋅17OH with subsequent air oxidation forming DMPO/^⋅17OH (45). This mechanism is not favored in the SOD/H₂O₂ system because DTPA did not prevent the incorporation of ¹⁷O from water into DMPO/^⋅OH. In the second, a strong oxidant oxidizes DMPO to its cation radical DMPO^⋅+, which reacts with H₂¹⁷O to form DMPO/^⋅17OH directly (44). The significant difference in these two mechanisms is the need for a very strong oxidant to form DMPO^⋅+, which has an oxidation potential greater than 1.8 V. The oxidation of the corresponding hydroxylamine of radical adducts occurs even in air and is rapid with ferricyanide (46). Direct incorporation of ¹⁷O-atom into DMPO/^⋅OH in the presence of H₂¹⁷O or H₂¹⁷O₂ may be explained by the following reaction mechanism:

The significance of this reaction, however, remains unclear, because DMPO (≅200 mM) did not prevent the inactivation of SOD by H₂O₂ (data not shown).

In the presence of formate, the spectrum due to the DMPO–carbon dioxide anion radical adduct (DMPO/^⋅CO₂⁻) or DEPMPO/^⋅CO₂⁻ was obtained (29, 30). This result is attributed to trapping of the carbon dioxide radical anion (^⋅CO₂⁻) produced from oxidation of formate by copper-bound hydroxyl radical (reaction 7 in Table 3) (47–49). The ESR spectra of DMPO/^⋅CO₂⁻ (a^N = 15.7 G, a^H = 18.8 G) and DEPMPO/^⋅CO₂⁻ (a^N = 14.5 G, a^H = 17.3 G and a^P = 51.6 G) were observed in incubations containing WT-SOD (100 μg/ml), H₂O₂ (5 mM), and formate (200 mM) in a bicarbonate/CO₂ buffer (25 mM, pH 7.4) (data not shown).

Oxidative Mechanisms in FALS Pathogenesis.

Many mechanisms have been suggested to explain how mutations to SOD cause the selective degeneration of motor neurons. SOD has been implicated in increasing oxidative injury by mechanisms involving the increased generation of hydroxyl radical, increased peroxidase activity, or increased nitration of tyrosine by peroxynitrite. Based on the present data, we can rule out the formation of free hydroxyl radical from loosely bound copper (reactions 9 and 10 in Table 3). The present data also do not favor the previous hypothesis that FALS mutants exhibit an increased peroxidase activity relative to WT-SOD. Furthermore, the mechanism of generating DMPO/^⋅OH adduct in the SOD/H₂O₂ system is more complex than previously reported.

Previously, the gain in function of FALS mutants was attributed to increased toxicity caused by released copper from the active site of FALS mutants. The ALS mutant SODs have reduced affinity for zinc (36). Furthermore, neurofilament subunits bind zinc with sufficient affinity that they can remove zinc from SOD (36). The increased loss of zinc from SOD will diminish the scavenging of superoxide and increase the catalysis of tyrosine nitration by peroxynitrite (36). The accumulation of nitrotyrosine has been demonstrated in both familial and sporadic ALS patients as well as in ALS–SOD transgenic mice (50). These results suggest that the loss of zinc from SOD may enhance peroxynitrite formation and motor neuron death in ALS.

In summary, the ¹⁷O results demonstrate that approximately two-thirds of the DMPO/^⋅OH arises from reaction with an oxidant from SOD, with the remainder being derived from the oxidation of the spin trap with subsequent addition of water. No differences were detected between WT and mutant SODs, neither in the concentration of DMPO/^⋅OH or DEPMPO/^⋅OH formed nor in the relative incorporation of oxygen from H₂O₂ or water.