Oxidation of 3,4-Dihydroxyphenylacetaldehyde, a Toxic Dopaminergic Metabolite, to a Semiquinone Radical and an ortho-Quinone*

The oxidation and toxicity of dopamine is believed to contribute to the selective neurodegeneration associated with Parkinson disease. The formation of reactive radicals and quinones greatly contributes to dopaminergic toxicity through a variety of mechanisms. The physiological metabolism of dopamine to 3,4-dihydroxyphenylacetaldehyde (DOPAL) via monoamine oxidase significantly increases its toxicity. To more adequately explain this enhanced toxicity, we hypothesized that DOPAL is capable of forming radical and quinone species upon oxidation. Here, two unique oxidation products of DOPAL are identified. Several different oxidation methods gave rise to a transient DOPAL semiquinone radical, which was characterized by electron paramagnetic resonance spectroscopy. NMR identified the second oxidation product of DOPAL as the ortho-quinone. Also, carbonyl hydration of DOPAL in aqueous media was evident via NMR. Interestingly, the DOPAL quinone exists exclusively in the hydrated form. Furthermore, the enzymatic and chemical oxidation of DOPAL greatly enhance protein cross-linking, whereas auto-oxidation results in the production of superoxide. Also, DOPAL was shown to be susceptible to oxidation by cyclooxygenase-2 (COX-2). The involvement of this physiologically relevant enzyme in both oxidative stress and Parkinson disease underscores the potential importance of DOPAL in the pathogenesis of this condition.

exact causes of this selective degeneration are unknown (2), the role of affected cells as centers of dopamine (DA) synthesis, storage, and metabolism suggests that DA may be an endogenous neurotoxin (3,4) that contributes to the pathogenesis of PD. DA may act as a source of cellular oxidative stress and is known to undergo oxidation (Scheme 1) to cytotoxic radicals and quinones (5)(6)(7)(8). Such oxidations can occur spontaneously or via metal-or enzyme-catalyzed mechanisms (9). One-electron oxidation of DA produces a radical capable of interfering with DA storage and causing oxidative protein and DNA modifications (5,6,10). Similarly, two-electron oxidation of DA to an ortho-quinone results in reactivity with cellular nucleophiles such as thiols and proteins (8,11). Both species are capable of redox cycling, which could deplete cellular oxidative defenses. Also, in the presence of transition metals and/or O 2 , such oxidations could result in the production of ROS capable of inducing lipid peroxidation and damage to other cellular macromolecules (12). Another potential mechanism of toxicity for DA is its physiological metabolism to 3,4-dihydroxyphenylacetaldehyde (DOPAL) (13).
DOPAL is a very reactive aldehyde that is 100 -1000-fold more toxic than DA both in vivo and in vitro (14,15). Physiological levels of DOPAL in the substantia nigra are ϳ2 M; levels as low as 6 M can exert significant toxicity (16). Reactivity with proteins, presumably via Schiff base formation, is an important mechanism of toxicity for DOPAL (17,18). Protein modification can result in inhibition of enzyme activity and loss of function for cellular proteins. Other toxicity mechanisms such as mitochondrial dysfunction (19), hydroxyl radical production (20), and interference with DA storage (21) have also been implicated in vitro. However, these mechanisms do not fully account for the high levels of toxicity associated with DOPAL. Therefore, we and others have hypothesized that DOPAL may be capable of undergoing oxidation to a quinone (16,17,20,22).
Oxidation of DOPAL to an ortho-quinone is predicted to have deleterious effects on cells and could result in additional mechanisms of toxicity (16,17,20,22). Work from our laboratory suggests that protein reactivity associated with DOPAL is partially mediated via catechol oxidation (18). Similarly, oxidation of DOPAL to a radical could help explain its high toxicity. However, direct investigations into the oxidation of DOPAL to either a radical or a quinone have not been reported.
In this study, we are the first to report the novel and systematic oxidation of DOPAL to a semiquinone radical and an orthoquinone. Following UV-visible spectrophotometric identification of two unique oxidation products of DOPAL, spin stabilization electron paramagnetic resonance spectroscopy (EPR) was used to characterize a DOPAL semiquinone radical, whereas two-dimensional NMR experiments were used to identify the structure of the DOPAL quinone. DOPAL was also shown to undergo carbonyl hydration in aqueous media; the quinone seems to exist exclusively in this form. Furthermore, oxidation of DOPAL resulted in enhanced reactivity with protein and ROS production. Also, DOPAL is shown here for the first time to be a substrate for cyclooxygenase-2 (COX-2), providing a potential mechanism for in vivo formation of the oxidized products identified here.
Water and buffers used in experiments were treated with Chelex 100 to remove opportune transition metals (27). NaH 2 PO 4 and Tricine were from Research Products International (Mount Prospect, IL), and Cambridge Isotope Laboratories (Andover, MA) provided deuterated solvents and trimethylsilylpropanoate-d 4 . Human recombinant COX-2 (EC 1.14.99.1; 10,636.25 units/mg) and arachidonic acid were from Cayman Chemicals (Ann Arbor, MI), and DMSO, K 3 Fe(CN) 6 , and perchloric acid were from Thermo Fisher Scientific. All other chemicals were from Sigma-Aldrich.
UV-visible Characterization of Oxidized DOPAL-A Hewlett-Packard 8453 diode array spectrometer (standard settings) was used to obtain spectra and max of DOPAL (0.5 mM) oxidized by tyrosinase (45 units) in 25 mM NaH 2 PO 4 , pH 7.4, over 60 min. Samples (1.5 ml) in quartz cuvettes were stirred with a small stir bar to ensure proper mixing for all readings, and temperature (25°C) was maintained with a water jacket. Buffer with heat-denatured (80°C, 5 min) tyrosinase served as a blank, and time-dependent absorbance maxima ( max ) were analyzed versus time to monitor product formation and disappearance. Results are representative of at least three trials.
EPR Spectroscopy-A Bruker EMX spectrometer with an Aqua-X sample holder and high sensitivity cavity equipped with a microwave frequency counter was used for EPR. Sample volumes were 1.0 ml, and spectra were recorded at room temperature. Typical EPR parameters were as follows: 3510.75 G center field; 24 G sweep width; 9.854 GHz microwave frequency; 20 milliwatt power; 2.52 ϫ 10 4 receiver gain; 100 kHz modulation frequency; 0.20 G modulation amplitude; 40.96 ms conversion time; 81.92 ms time constant; 41.93 s sweep time; and up to 10 X-scans for each 1024-point spectrum. The resultant spectra are representative of at least three trials. Relative radical concentrations were determined by comparing double integrations (using WinEPR) with a stable radical (2,2,5,5-tetramethyl-1-pyrrolidinyloxy-3-carboxylic acid) of known concentration (data not shown).
High pH (40 mM NaH 2 PO 4 pH 10.5) was used to induce autooxidation (30) of the catechol substrates (DOPAL, DOPAC, and DA, 400 M each) under non-spin-stabilizing conditions (i.e. without Mg 2ϩ ). The spectra were simulated with WinSim and WinSim 2002 software (31) to determine the correlation coefficients and splitting constants. DOPAL (400 M in 0.5 M MgCl 2 , 25 mM Tricine buffer, pH 7.4) was also incubated with 2.0 eq of H 2 O 2 to test the contribution of H 2 O 2 to DOPAL oxidation. The g values for DA, DOPAL, and DOPAC were determined using 1,4-benzosemiquinone as a standard (g ϭ 2.00469 (in aqueous solution)) (32).
For initial oxidation experiments, DOPAL in D 2 O (ϳ20 mM) was carefully oxidized with 1.00 eq of NaIO 4 (stock solution dissolved in D 2 O), and transferred into an NMR tube. 1 H spectra were then collected. The reaction was repeated with 10 mM NaH 2 PO 4 (pD 7.5 in D 2 O), and with a 4-fold excess of acetoned 6 . For structure determination experiments, the reaction mixture was diluted 1:4 with acetone-d 6 (prechilled on dry ice) 60 s after the addition of NaIO 4 , and then transferred into an NMR tube and immediately analyzed at Ϫ20°C. The oxidation product was characterized via a variety of one-and two-dimensional homonuclear and heteronuclear experiments ( 1 H, 1 H-1 H correlated spectroscopy (COSY), 1 H-1 H nuclear Overhauser effect spectroscopy (NOESY), 1 H-13 C heteronuclear multiple quan- . TD, NS, and DS refer to time domain data points, number of scans, and dummy scans, respectively, and f 1 and f 2 refer to sweep width. NMR data were processed with TOPSPIN 1.3 software. One-dimensional 1 H data were processed with zerofilling to 64,000 data points and 0.2 Hz exponential line broadening, whereas two-dimensional NMR data were processed with zero-filling to 4096 points in acquisition and 1024 points in second dimension.
The formation of a DOPAL hydrate was evident from the spectra collected in D 2 O prior to oxidation. For determination of the K eq of hydration, stock DOPAL solutions (8,20, and 24 mM) were prepared from three different synthetic batches in both D 2 O and 10 mM NaH 2 PO 4 buffer (pD 7.5 in D 2 O). 1 H spectra were collected, and integration ratios (hydrate:aldehyde) for all protons were calculated for each sample on three different days.
SDS-PAGE Protein Cross-linking-Oxidized or non-oxidized DOPAL (50 M) was reacted with glyceraldehyde 3-phosphate dehydrogenase (GAPDH, 0.3 mg/ml) in 50 mM NaH 2 PO 4 , pH 7.4, at 37°C for 4 h and then analyzed via SDS-PAGE (4% stacking, 10% resolving gels, and 9 g of protein loaded/well) and densitometry (NIH ImageJ software) as described previously (18). DOPAL oxidized by various methods (1.0 eq of NaIO 4 , 36 units/ml tyrosinase, and 2.0 eq of K 3 Fe(CN) 6 ) was added to the protein at specific time points (10 or 60 s following oxidation). The resultant protein reactivity and cross-linking were indicated by a decrease in intensity of the control protein band with corresponding formation of higher molecular weight species. The results are representative of two separate trials.
DOPAL Auto-oxidation and ROS Production-Superoxide formation during DOPAL auto-oxidation was investigated using the method of Eyer (30,33). UV-visible spectrophotometry (kinetic settings) was used to measure the change in absorbance at 400 nm during auto-oxidation of DOPAL (0.5 mM) in 50 mM NaH 2 PO 4 , pH 7.8. The reaction was repeated in the presence of SOD (8 M, 1200 units/ml; or 40 M, 6000 units/ml). The samples (1.5 ml) were stirred at 37°C and blanked against buffer with or without SOD. The readings were taken every 3 min for 12 h, and the reported results are representative of at least three runs.
Oxidation of DOPAL by COX-2-The initial linear slope and k cat of DOPAL (50 M) oxidized by human recombinant COX-2 (100 units/ml) were determined by measuring the formation of oxidized products and the disappearance of DOPAL over time. The assay conditions of Wangpradit et al. (34) were utilized for all reactions (100 mM KH 2 PO 4 , pH 7.4, 100 M potassium arachidonate, 1 M hematin, 1 ml volume). 100-l aliquots were removed at various time points, treated with perchloric acid (6 l), centrifuged (10,000 ϫ g, 5 min), and analyzed quantitatively with an Agilent 1200 series capillary system HPLC with photodiode array detector. The concentration of DOPAL (non-oxidized) was determined by the calibration curve method described previously (18,24). To estimate the concentrations of oxidized DOPAL in each aliquot, a calibration curve (not shown) was constructed using 50 M DOPAL oxidized by various concentrations of NaIO 4 . This allowed the correlation of the total HPLC peak area with the concentration of oxidized DOPAL.
The amounts of DOPAL oxidized by COX-2 (aliquots at 0-, 1.5-, and 5-min time points) were plotted to determine the initial linear slope; k cat was then calculated by accounting for the amount of enzyme used in each assay and the molecular mass of COX-2 (72 kDa). The assay was repeated three times for DOPAL and once with DA (65 M) for comparison. Also, a trial, with the enzyme omitted but all other components present, was utilized to determine base-line (non-enzymatic) DOPAL oxidation.

RESULTS
UV-visible Characterization-Rearrangement of DOPAL upon oxidation was evident colorimetrically. The treatment of DOPAL with tyrosinase catalyzed the stepwise formation of two unique chromophores, an early and transient pink species followed by a relatively more stable yellow species. UV-visible composite scans (Fig. 1A) allowed the correlation of the early transient species with a max of 520 nm and the later species with a max of 400 nm. The relative intensity of these species over time is demonstrated graphically in Fig. 1B. Oxidation of DOPAL with NaIO 4 followed a similar course, although it show the formation of two unique species. The early oxidation product is transient and centered at 520 nm, whereas the later species is relatively more stable and centered at 400 nm. Only the first several minutes are displayed for clarity. B, relative absorbance of the two oxidation products over time.
occurred too quickly to capture changes spectroscopically (not shown). Because of the ability of tyrosinase and NaIO 4 to oxidize DA to a quinone (7), we expected DOPAL to undergo a similar oxidation. However, the max of the initial species was higher than expected for a quinone (35). Its transient nature and the susceptibility of quinones to undergo comproportionation (30) in the presence of their reduced counterparts (i.e. hydroquinones and catechols) suggested that the pink species was a radical (Scheme 2).
EPR Spectroscopy-The use of redox-inactive divalent metal cations, known to stabilize semiquinone radicals (28,29), allowed the spectra of a DOPAL semiquinone radical to be collected for the first time. Both two-electron (tyrosinase and NaIO 4 ) and one-electron oxidizing agents (HRP/H 2 O 2 and K 3 Fe(CN) 6 ) produced the same DOPAL semiquinone radical as seen by identical EPR spectra (Fig. 2). Regardless of the oxidation method, the radical levels were below the detection limit at pH 7.4 in the absence of Mg 2ϩ .
Auto-oxidation (via high pH) of DOPAL also resulted in the formation of a DOPAL semiquinone radical (Fig. 3B, column i). As expected, there are small differences in hyperfine splittings observed in this system, as no divalent cation is present. DA and DOPAC also readily auto-oxidized at pH 10.5 to semiquinone radicals and displayed additional hyperfine splittings (Fig. 3, A and C, column i). Computer simulations reproduced well the experimental spectra (correlation coefficients generally Ͼ0.99) (Fig. 3, column iii). DA, DOPAL, and DOPAC oxidized by tyrosinase under spin stabilization conditions (Fig. 3, column ii) are also shown for comparison. The spin-stabilized radicals were also simulated successfully (spectra not shown). Because DOPAL, DOPAC, and DA each has five similarly located hydrogens that can provide hyperfine splittings, they produce similar EPR spectra. The hyperfine splitting constants, reported for the first time for DOPAL, are shown in Table 1. The coupling constants determined for DA and DOPAC are in good agreement with those reported in the literature (6,36). The g values for DA, DOPAL, and DOPAC are reported in Table 1. Experimental DOPAL radical concentrations were similar to DA for all tested conditions, whereas DOPAC radical levels were severalfold higher (data not shown). Also, incubation of DOPAL (400 M) with 2.0 eq of H 2 O 2 did not result in the production of a DOPAL radical.
Careful oxidation of DOPAL by NaIO 4 resulted in the formation of a unique species (Fig. 4B) Fig. 4C). As these peaks diminished, a new single peak at 8.44 ppm increased concurrently, which may have been due to the formation of a polymer; this interpretation is also consistent with the formation of a broad hump under the aromatic resonances during the time course of the reaction. Use of either unbuffered or buffered D 2 O gave rise to the same oxidized product; however, unbuffered D 2 O minimized the formation of the 8.44-ppm polymer-SCHEME 2. Comproportionation of quinone and catechol to radicals. like peak, which enhanced the intensities of the other resonances. The special conditions developed (Ϫ20°C, 1:4 dilution with cold acetone) enhanced the lifetime of the species significantly, allowing completion of COSY, NOESY, and other heteronuclear two-dimensional NMR experiments. Importantly, these conditions did not change the course of the reaction or the product formed. Peak assignments and bond correlations apparent from NMR experiments are indicated schematically in Fig. 4D. 13 )). This structure is best described as the ortho-quinone of DOPAL with a hydrated aldehyde. These experiments confirmed for the first time the formation of a DOPAL quinone.
Protein Reactivity-The oxidation of DOPAL enhanced its ability to induce protein cross-linking of the model protein, GAPDH (Fig. 5). The different oxidation conditions represent different active components (NaIO 4 , 10 s: radical-quinone mixture; NaIO 4 , 60 s: quinone only; tyrosinase, 10 s: radical-catechol mixture; tyrosinase, 60 s: radical-quinone mixture; K 3 Fe(CN) 6 : radical only). Non-oxidized DOPAL caused ϳ17% loss of protein as compared with the control (Table 3). Oxidation with NaIO 4 at the early and late time points resulted in further losses of 11 and 6%, respectively. Tyrosinase oxidation increased protein losses by 30% when tyrosinase was added at 10 s and 22% when added at 60 s. Unlike NaIO 4 and tyrosinase, control reactions of K 3 Fe(CN) 6 with GAPDH (without DOPAL) did cause some loss of protein (ϳ12%). Therefore, DOPAL oxidized by K 3 Fe(CN) 6 was compared with K 3 Fe(CN) 6 -treated protein instead of untreated protein (Fig. 5,  lane 1), indicating a 29% greater protein loss. Concurrent with the disappearance of the control band protein is the formation of higher molecular weight aggregates indicative of protein cross-linking (note arrows in Fig. 5). Overall, the oxidation of DOPAL enhanced cross-linking by up to 30%.
DOPAL Auto-oxidation and ROS Production-DOPAL undergoes relatively slow auto-oxidation to quinone (Fig. 6A). In the presence of SOD (8 M), the rate of auto-oxidation is

TABLE 1 Experimental hyperfine coupling constants and g-values for simulated dopaminergic radicals
Computer-generated simulations (WinSim) of experimental EPR spectra (see Fig. 3) were used to determine a-values. Benzosemiquinone (32) 6B). Equilibrium for this reaction generally favors reactants (30). However, the presence of SOD can shift the equilibrium toward the products by rapid dismutation of superoxide, effectively "pulling" the reaction. COX-2 Oxidation of DOPAL-The turnover numbers (k cat ) for the oxidation of DOPAL and DA by COX-2 were calculated by determining the initial linear slopes (Fig. 7 and Table 4). The results indicate that a small but significant portion of DOPAL is oxidized non-enzymatically by assay components. The reported rates were corrected to account for this base-line oxidation. DOPAL had a k cat of 6.96 Ϯ 0.40 and 5.54 Ϯ 0.89 min Ϫ1 as calculated by substrate loss and product formation, respectively; the difference in the values between the two methods arises due to the reactive and unstable nature of radicals and quinones. In comparison, DA had a k cat of 4.18 min Ϫ1 , which is in good agreement with the literature (38). Furthermore, oxidized DOPAL demonstrated sensitivity to the acidic conditions utilized in the work-up procedure for this assay.    Table 3.

DISCUSSION
Cytosolic DA can undergo catechol oxidation (2,39) to form radicals and quinones that exert cytotoxicity through a variety of mechanisms (6,12). Alternatively, DA may undergo monoamine oxidase-catalyzed oxidative deamination to produce DOPAL, which can be further metabolized by aldehyde dehydrogenase to DOPAC (16,40). Previously, we showed that lipid peroxidation, which is closely associated with oxidative stress, can inhibit aldehyde dehydrogenase causing elevated intracellular DOPAL levels (17,24). DOPAL is highly toxic and is hypothesized to be a causative factor in PD (15,22). Furthermore, catechol oxidation may be important for the toxicity of DOPAL.
Tyrosinase and NaIO 4 are known to oxidize DA to a quinone. The oxidation of DOPAL by these methods produced two unique species. EPR confirmed that the initial product was a novel DOPAL semiquinone radical (Scheme 2) produced experimentally at concentrations similar to the DA radical (data not shown). Importantly, the DA semiquinone is a known toxic agent. It interferes with DA storage, oxidatively damages DNA and proteins, produces ROS, and undergoes redox cycling (5,6).
The DOPAC radical was present in much higher concentrations than either DA or DOPAL, under identical experimental conditions. However, DOPAC has limited toxicity. This is likely because of its further metabolism or selective removal via a transporter (41). Therefore, the DOPAC radical formation reported here may have limited biological consequences.
NMR was used to identify the second oxidation product of DOPAL as the ortho-quinone, allowing us to uncover the carbonyl hydration of DOPAL. Although hydration (gem-diol formation) in aqueous media is known for some aldehydes (42), it has been previously unreported for DOPAL. Importantly, the prevalence of the hydrate over the aldehyde (K eq ϭ 2.50) calls into question the role of the aldehyde versus the gem-diol in the behavior of DOPAL, at both normal and toxic levels. 1 H NMR resonances of the later DOPAL oxidation product were consistent with an ortho-quinone structure, which was verified with NOESY, COSY, HMQC, and HMBC experiments. Interestingly, upon oxidation of DOPAL to the quinone, aldehyde resonances were no longer present or were below the detection limit of the NMR. This disappearance is likely because of a shift in the equilibrium for the hydration reaction. Generally, K eq shifts away from the aldehyde are catalyzed by strongly electronegative substituents. For example, the K eq for acetaldehyde hydration is 1.06, but for trifluoroacetaldehyde it is 2.9 ϫ 10 4 (42). The K eq of 2.50 for DOPAL hydration indicates that the catechol is a strong electron-withdrawing group. However, the apparent loss of the aldehyde resonance following Tyrosinase c Radical-Quinone 22 7 K 3 Fe(CN) 6 Radical 29 a Non-oxidized DOPAL ("DOPAL alone") caused a 17% loss of protein. Results are the amount beyond 17%. b 10 s pre-oxidation (mimics early oxidation product). c 60 s pre-oxidation (mimics later oxidation product). FIGURE 6. Enhancement of the rate of DOPAL auto-oxidation by SOD indicates that superoxide is being formed. A, auto-oxidation of DOPAL at pH 7.8 proceeds slowly but is greatly accelerated in the presence of increasing concentrations of SOD. B, although superoxide is formed during spontaneous (i.e. air) oxidation of catechols and hydroquinones, the reverse reaction is much faster. Sufficient removal of superoxide by SOD allows the reaction to proceed much more quickly. The results, based on the concentration of substrate oxidized, were calculated using both product formation and substrate loss. DA (65 M) was also shown for comparison using substrate disappearance to calculate the initial linear slope. The assay system without the added enzyme functioned as a control and allowed the results to be corrected for base-line DOPAL oxidation. oxidation indicates that the quinone is much more electronegative than the catechol. The K eq of hydration for the DOPAL quinone is likely several orders of magnitude higher than the catechol. These results indicate that the DOPAL quinone is highly electrophilic and may be susceptible to nucleophilic attack and/or reduction. Because of the similar electrophilic nature and cellular localization, the DA and DOPAL quinones likely have comparable biological effects. Toxicity mechanisms for the DA quinone include nucleophilic reactivity with proteins and thiols, depletion of cellular reducing equivalents, and ROS production (12).
Protein reactivity is an important mechanism of toxicity for both DOPAL and oxidized DA (43). We have shown here that DOPAL oxidation enhances its ability to induce cross-linking of GAPDH. Because of its role as a redox-sensitive enzyme involved in glucose metabolism and multiple external thiol residues (12) and its cellular proximity to DA metabolism, GAPDH serves as an appropriate target for modification by DOPAL. Also, DA-induced aggregation of GAPDH results in increased oxidative stress and cytotoxicity to dopaminergic cell models (44,45). Although both the DOPAL radical and quinone increased protein cross-linking, this effect was more pronounced for the radical (29%) than for the quinone (6%). This may indicate that GAPDH is more susceptible to radical-induced oxidative damage than quinone-induced nucleophilic reactivity. Although quinones are susceptible to Michael-type ring addition reactions with nucleophiles such as thiols, this mechanism did not seem to contribute greatly to overall protein reactivity. Also, reductive gel conditions indicate that radical-induced cross-linking is due to oxidative damage above and beyond simple disulfide rearrangements. The protein aggregates formed here may involve stable sulfhydryl cross-links (46) or other types of oxidative damage such as peroxidized proteins or protein carbonyls. A high increase (22%) in protein reactivity was evident with the radical-quinone mixture, indicating that redox cycling and ROS production may contribute to the observed results. Furthermore, GAPDH is known to be sensitive to ROS modification, and GAPDH aggregation is involved in dopaminergic models of oxidative stress-induced cell death (45). This makes the cross-linking observed here potentially relevant as a specific mechanism of toxicity for DOPAL.
Another toxicity mechanism for DOPAL may be ROS production. Our results indicate that superoxide is produced during DOPAL auto-oxidation. Superoxide can be formed via sequential one-electron reductions of O 2 (Fig. 6B), although the reverse reaction is usually favored (47). The varying lag periods and the biphasic nature of the reaction (Fig. 6A) are consistent with the need to generate enough DOPAL quinone to allow sufficient semiquinone to be formed to bring about the "rapid" oxidation of DOPAL (33,47). That SOD reduces the lag time is consistent with the formation of superoxide in this system via the semiquinone radical.
Also, DOPAL reportedly will react with H 2 O 2 , producing a hydroxyl radical (20). The proposed mechanism involves oneelectron oxidation of DOPAL and concurrent reduction of H 2 O 2 to the hydroxyl radical. However, EPR experiments of DOPAL incubated with H 2 O 2 did not result in the formation of a DOPAL radical. Therefore, H 2 O 2 production by SOD was not a significant contributor to DOPAL auto-oxidation in these experiments.
Enhancement of DOPAL auto-oxidation is highly relevant given that SOD is up-regulated in the substantia nigra of PD patients (48,49). Superoxide dismutation may be beneficial for cells but also result in the deleterious consequence of increased formation of DOPAL radicals and quinones.
COX-2 is a bifunctional, inducible, xenobiotic-metabolizing enzyme (50) capable of oxidizing catechols and catecholamines as cosubstrates during prostaglandin biosynthesis (39). DA is a known COX-2 substrate and can be oxidized to a reactive semiquinone radical or quinone capable of modifying DNA and proteins. Elevated activity and expression of COX-2 occurs in the substantia nigra in PD and may contribute to neurodegeneration via oxidation of biogenic amines to endogenous neurotoxins. Reported sites of COX-2 up-regulation vary conflictingly between dopaminergic neurons, non-astrocytic glial cells, or both (51)(52)(53). The COX-2 catalyzed oxidation of DA may also be involved in accumulation of ␣-synuclein, the primary component of Lewy bodies (39). Furthermore, certain COX-2 inhibitors prevent or attenuate PD in human epidemiological studies and neurotoxin-induced in vitro and animal models. This suggests a pivotal role for COX-2 in the pathogenesis of PD (54 -57). Here, we have shown for the first time that DOPAL is also a substrate for COX-2. This development suggests a relevant mechanism of formation for the DOPAL radical and quinone in cells affected by PD. Although Schiff base formation is assumed to be essential for the reactivity of DOPAL with proteins and the resulting cellular dysfunction, catechol oxidation is likely also important for the mediation of DOPAL toxicity. Furthermore, some of the neurodegeneration previously associated with DA oxidation may involve the oxidation of other species such as DOPAL. These findings have implications for the role of DOPAL in PD pathogenesis and therapeutic development.
We have shown here for the first time the oxidation of DOPAL to a semiquinone radical and an ortho-quinone. Such oxidations result in ROS formation and enhanced cross-linking of proteins. The reactivity of oxidized DOPAL with GAPDH serves as an example of a potentially deleterious biological outcome. Also, DOPAL was demonstrated to be a COX-2 substrate, indicating a method for the possible biological formation of the oxidized products identified here. The findings of this work highlight the potential importance of DOPAL oxidation by COX-2 (or other methods) in PD.