1090
Chem. Res. Toxicol. 1999, 12, 1090-1097
New Reaction Pathways of Dopamine under Oxidative Stress Conditions: Nonenzymatic Iron-Assisted Conversion to Norepinephrine and the Neurotoxins 6-Hydroxydopamine and 6,7-Dihydroxytetrahydroisoquinoline Alessandra Napolitano, Alessandro Pezzella, and Giuseppe Prota* Department of Organic and Biological Chemistry, University of Naples “Federico II”, Via Mezzocannone 16, I-80134 Naples, Italy Received May 12, 1999
Aerial oxidation of dopamine at concentrations as low as 50 µM in the presence of ferrous ions in phosphate buffer (pH 7.4) led in the early stages (6-8 h) to the formation of the quinone of the neurotoxin 6-hydroxydopamine, 2, followed (24 h) by a complex product pattern comprising main components norepinephrine (5), 3,4-dihydroxybenzaldehyde (4), and the neurotoxic alkaloid 6,7-dihydroxy-1,2,3,4-tetrahydroisoquinoline (3). Product formation required the assistance of metal ions such as Mn(II), Zn(II), and iron, in either the ferrous or ferric form. Product yields were shown to vary linearly with iron and dopamine concentration in the early phases of the reaction (2 h). Biologically relevant antioxidants, like glutathione and ascorbate, and metal chelators, e.g., 2,2′-bipyridyl, inhibited dopamine conversion to products 2-5, but not substrate consumption, while hydroxyl radical scavengers such as DMSO and mannitol did not alter the course of the reaction. On the contrary, mannitol increased product yields, an effect seen for other monosaccharides. Catalase exhibited a significant inhibitory effect particularly on the formation of 3 and 4. By using 18O2, evidence was obtained for incorporation of the label into the carbonyl oxygen of 4, but not into the hydroxyl group of 5. On the basis of these and other results, a complete mechanistic picture of the oxidation is drawn involving conversion of dopamine to the corresponding o-quinone and its quinonemethide tautomer with concomitant reduction of O2 to H2O2. Nucleophilic attack by H2O to the quinonemethide gives rise to 5, while H2O2 addition leads to benzaldehyde 4 via a β-aminohydroperoxide intermediate. This latter reaction path also gives formaldehyde which yields the isoquinoline 3 by Pictet-Spengler condensation with dopamine. The quinone 2 results from H2O2 attack at the 6-position of dopamine o-quinone in agreement with previous studies. These results provide an insight into new routes of nonenzymatic conversion of dopamine to its metabolite norepinephrine and neurotoxic species which may become operative under conditions relevant to neurodegeneration.
Introduction A marked alteration of cellular redox equilibria is commonly recognized as a critical condition associated with Parkinson’s disease and related neurodegenerative disorders (1). Although the root cause of the selective loss or dysfunction of nigrostriatal dopamine neurons in Parkinson’s disease has not yet been definitely assessed, a persisting condition of oxidative stress has been suggested as a primary factor contributing to the progression of neurodegeneration (2). This is the result of an intricate cascade of intra- and extraneuronal events involving, inter alia, an alteration of GSH and ascorbate levels (3, 4), a marked increase in lipid peroxidation and malondialdehyde levels (5), mitochondrial complex I deficiency (6), and an accelerated cytosolic dopamine deamination by MAO1 (monoamino oxidase) in the surviving neurons * To whom all correspondence should be addressed: Department of Organic and Biological Chemistry, University of Naples “Federico II”, Via Mezzocannone 16, I-80134 Naples, Italy. Phone: +39-817041249. Fax: +39-81-5521217. E-mail:
[email protected].
giving rise to abnormally high fluxes of hydrogen peroxide (7, 8). Selective iron accumulation and decreased ferritin levels have been documented at autopsy in parkinsonian substantia nigra compared to controls (9), and have been ascribed to the peculiar ability of neuromelanin in pigmented nigrostriatal neurons to act as a sink of metal ions probably at the catechol sites (10). Coupled with the overproduction of hydrogen peroxide, this abnormal iron load has provided the basis of theories which advocate generation of highly cytotoxic hydroxyl radical (OH•) and superoxide anion (O2•-) by Fenton and Haber-Weiss-type reactions, capable of triggering lipid peroxidation leading to membrane damage and eventually to neuronal death (7, 11). Following the discovery by Tranzen and Thoenen (12) that an oxidized dopamine derivative, 6-hydroxydopamine, caused selective destruction of peripheral catechola1 Abbreviations: MAO, monoamine oxidase; DOPA, 3-(3,4-dihydroxyphenyl)alanine; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; EI/MS, electron impact mass spectrometry; HRMS, high-resolution mass spectrometry.
10.1021/tx990079p CCC: $18.00 © 1999 American Chemical Society Published on Web 10/22/1999
Neurotoxin Formation by Iron-Assisted Dopamine Oxidation
minergic nerve endings, the hypothesis has been formulated that a chronic diversion of dopamine metabolism toward aberrant oxidative routes may be induced under oxidative stress conditions leading to neurotoxins capable of targeting critical cell structures (13, 14). Dopaminederived alkaloids, and particularly 1-methyl-6,7-dihydroxy-1,2,3,4-tetrahydroisoquinoline (salsolinol, 1), were found in urine and brain of parkinsonian patients treated with L-3-(3,4-dihydroxyphenyl)alanine (DOPA) and shown to induce behavioral changes similar to those observed in Parkinson’s disease when administred in rat brain (15, 16). Like that of the exogenous toxin 1-methyl-4-phenyl1,2,3,6-tetrahydropyridine (MPTP), the enzymatic or nonenzymatic oxidation to catechol isoquinolines proved to be a prerequisite for the selective and potent neurotoxicity to exist (17). However, certain critical differences between idiopathic Parkinson’s disease and MPTP, or chemically induced parkinsonism, have been recognized in other studies (4), and the issue has remained controversial (13). These findings stimulated research work aimed at investigating likely routes of endogenous generation of neurotoxins as a potential contributory factor in the etiopathogenesis of Parkinson’s disease. Model in vitro studies demonstrating formation of 6-hydroxydopamine by oxidation of dopamine mediated by various agents have been reported (18, 19). In a reexamination of this issue, we showed the formation of substantial amounts of the quinone of 6-hydroxydopamine, 2, from dopamine under conditions of relevance to neurodegenerative processes, i.e., by oxidation in the presence of hydrogen peroxide (20), and by reaction with polyunsaturated fatty acid hydroperoxides and ferrous ions (21). Recently, we obtained evidence that the quinone 2 is also formed in the early stages of the iron-assisted aerial oxidation of dopamine (22). Most notably, the oxidation reaction leads in the further steps to a complex reaction pattern comprising the major component neurotoxic tetrahydroisoquinoline alkaloid 3 (norsalsolinol) arising by Pictet-Spengler condensation of the catecholamine with formaldehyde generated by oxidative fission of the dopamine side chain.
As an extension of these studies, we have carried out a detailed analysis of the products arising from aerial dopamine oxidation in the presence of ferrous ions, and report now the formation, in addition to the tetrahydroisoquinoline 3, of 3,4-dihydroxybenzaldehyde and, notably, of the dopamine metabolite norepinephrine by a nonenzymatic route. The effects of a variety of parameters and additives on the kinetics and course of the reaction have also been surveyed. An overall picture of the mechanisms of formation of the reaction products emerging from 18O labeling experiments is presented.
Experimental Procedures Electrospray mass spectra were recorded with a quadrupole mass spectrometer. For electron impact (EI/MS) and highresolution (HR/MS) mass spectra, samples were ionized with a
Chem. Res. Toxicol., Vol. 12, No. 11, 1999 1091 70 eV beam, with the source at 230 °C. Main fragmentation peaks are reported with their relative intensity (percent values are in brackets). 1H NMR spectra were recorded at 400 MHz. Analytical and preparative HPLC were performed with an instrument equipped with a variable-wavelength UV detector and an electrochemical detector. The UV detector was set at 280 nm, while the electrochemical detector was set with electrode 1 at -0.10 V and electrode 2 at 0.70 V versus the internal reference electrode. A 4.6 mm × 250 mm octadecylsilane-coated column (5 µm particle size, rate of 1 mL/min) and a 22 mm × 250 mm column (10 µm, rate of 18 mL/min) were used for analytical and preparative runs, respectively. Elution conditions were as follows: 0.05 M phosphate buffer (pH 3.0) containing 10 mM sodium 1-octanesulfonate/acetonitrile (93:7) (eluant A), 5% formic acid/acetonitrile (85:15) (eluant B), and 1% AcOH/acetonitrile (70:30) (eluant C). Materials. Dopamine hydrochloride, 3,4-dihydroxybenzaldehyde, 2,4,5-trihydroxyphenethylamine hydrobromide (6-hydroxydopamine), (()-1-(3,4-dihydroxyphenyl)-2-aminoethanol (norepinephrine), D-glucose, D-fructose, D-ribose, ascorbic acid, reduced glutathione, ferrous sulfate heptahydrate, ferric chloride, EDTA disodium salt, 2,2′-bipyridyl, nonstabilized hydrogen peroxide (35% solution in water), and catalase from bovine liver (H2O2:H2O2 oxidoreductase, EC 1.11.1.6, 9740 units/mg) were used as they were obtained from commercial sources. 18O2 (50 at. % 18O) was from Cambridge Isotope Laboratory (Andover, MA). 2-Hydroxy-5-(2-aminoethyl)-1,4-benzoquinone (2) was prepared by sodium periodate oxidation of 6-hydroxydopamine as reported previously (23). 6,7-Dihydroxy-1,2,3,4-tetrahydroisoquinoline (3) was obtained by reaction of dopamine with formaldehyde according to a general procedure for the synthesis of tetrahydroisoquinolines (24). General Procedure for the Oxidation of Dopamine. A solution of dopamine hydrochloride (50-200 µM) in 0.1 M phosphate buffer (pH 7.4) was treated with appropriate volumes of a freshly prepared stock solution of ferrous sulfate in water up to the desired concentration (10-100 µM) while being stirred vigorously. The mixture was incubated while being stirred at 37 °C in a thermostated water bath. Aliquots of the reaction mixture were periodically withdrawn, acidified to pH 1.0 with 3 M HCl, and analyzed by HPLC. When required, solutions of ferric salts were used in place of ferrous salts. In a series of experiments, aliquots of a 0.03% hydrogen peroxide solution were added to the oxidation mixture over the first 8 h of the reaction at 1 h time intervals up to 3 molar equiv with respect to the substrate, and product analysis was carried out at 24 h. Additives, e.g., enzymes, metal chelators, and antioxidants, were added to the desired concentration, as indicated in the table or the figure legends. Identification and quantification of dopamine and reaction products were carried out by comparing retention times and integrated peak areas with external calibration curves for authentic samples. All experiments were carried out at least in triplicate. Statistical parameters were determined by linear least-squares fitting. Isolation or Identification of Products 2-5. Reaction products 2 and 3 were identified by comparison of the UV spectra and retention times under elutographic conditions A and B with those of authentic samples prepared as indicated in Materials. For isolation of 4 and 5, dopamine (1.9 g, 10 mmol) and ferrous sulfate (1.4 g, 5 mmol) in 0.1 M phosphate buffer (pH 7.4) (1 L) were allowed to stand while being stirred vigorously in air at 37 °C. After 72 h, the mixture was filtered to remove pigmented materials and, after acidification to pH 3 with 6 M HCl, extracted with ethyl acetate (3 × 250 mL). The organic layers were dried over sodium sulfate, and the residue was fractionated by preparative HPLC, eluant B, to give 4. 1H NMR and EI/MS spectra of the isolated fraction compared well with those of a commercial sample. The aqueous phase was fractionated by ion exchange chromatography (DOWEX 50WX4 100-200 mesh) with an HCl gradient of 0.1 to 2 M. Fractions that eluted with 1-2 M HCl were reduced to small volume and purified by preparative HPLC using eluant B as the mobile
1092
Chem. Res. Toxicol., Vol. 12, No. 11, 1999
Napolitano et al.
phase to obtain a fraction corresponding to 5. 1H NMR and electrospray mass spectra of the compound thus purified compared well with those of a commercial sample. Oxygen Labeling Experiments. The oxidation reaction was carried out as described above using 20 mg of the starting material. The solution was incubated with 18O2 after four cycles of evacuation and N2 reflushing. The N2 gas that was used was passed through an alkaline pyrogallol solution. Aliquots of the mixture were withdrawn at 6 and 24 h, acidified, and worked up as described above to obtain 4. The aqueous phase of the mixture at 24 h was taken to dryness and the residue treated with acetic anhydride (2 mL) and pyridine (100 µL) overnight. After removal of the volatile components, the mixture was fractionated by preparative HPLC, eluant C, to yield 5. Compounds 4 and 5 were subjected to EI/MS analysis. 4: m/z (relative intensity) 140 (M+, 17), 139 (6), 138 (M+, 15), 137 (26), 112 (12), 111 (3), 110 (6), 109 (12), 84 (62), 83 (50), 82 (23), 81 (10), 58 (42), 57 (99), 56 (94), 55 (100); HREI/MS exact mass calcd for C7H6O3 138.0317, found 138.0321; HREI/MS exact mass calcd for C7H616O218O 140.0359, found 140.0357. 5: m/z (relative intensity) 337 (M+, 1), 295 (33), 278 (100), 236 (68), 223 (80), 194 (96), 181 (74), 139 (98); HREI/MS exact mass calcd for C16H19NO7 337.1161, found 337.1164.
Results Previous studies (21, 22, 25) have shown that oxidation of dopamine in air in phosphate buffer at pH 7.4 is relatively slow, the extent of consumption of the catecholamine not exceeding 20% over a period of up to 24 h, but is markedly accelerated in the presence of ferrous ions (22, 25). Under the latter conditions, the course of the reaction is diverted, leading to a well defined product pattern in place of the pigmented ill-defined materials which are ultimately obtained by enzymatic or chemical oxidation (26-28). A quite complex course of the ironpromoted oxidation showing a rise and decay of different products over a prolonged reaction time of up to 72 h was evidenced in a preliminary report (22) using the catecholamine at a concentration of 1 mM. On the basis of these observations, a systematic analysis of the product patterns was carried out in the investigation described here with dopamine at concentrations as low as 50 µM which are closer to those occurring in the neuronal environment (29). Panels a and b in Figure 1 show the HPLC profiles of the aerobic oxidation of 100 µM dopamine in the presence of 45 µM ferrous ions in phosphate buffer at pH 7.4, as monitored by electrochemical detection, at reaction times of 2 and 24 h, respectively. In the initial phases of the reaction (panel a), in addition to the starting material, the mixture showed a major component identified as the quinone of the neurotoxin 6-hydroxydopamine, 2, by comparison of its spectrophotometric and chromatographic properties with those of an authentic sample (23), while at later stages (panel b), a much more complex product pattern prevailed. The reaction component corresponding to the elutographic peak labeled B was identified as 3,4-dihydroxybenzaldehyde (4) by performing the oxidation on a preparative scale. At a reaction time of 72 h, the ethyl acetate extractable fraction was purified by preparative HPLC and the major component characterized as 4 by electron impact mass and proton NMR spectral analysis. Compounds A and C were identified as norepinephrine (5) and the tetrahydroisoquinoline 3, respectively, by co-injection with authentic samples. The structure of compound A as 5 was confirmed by electrospray mass spectrometric and proton NMR analy-
Figure 1. HPLC elution profile of the products formed by aerial oxidation of dopamine (100 µM) in the presence of ferrous ions (45 µM) in 0.1 M phosphate buffer (pH 7.4) at reaction times of 2 (a) and 24 h (b). Analysis was carried out with electrochemical detection, using eluant A as the mobile phase; all other conditions were as described in Experimental Procedures.
Figure 2. Time course of dopamine consumption (dashed line, left axis) and product formation (right axis, solid lines) by aerial oxidation of dopamine in the presence of ferrous ions. Reaction conditions were as described in the legend of Figure 1: dopamine (0), 5 (9), 4 (2), 3 (b), and 2 ([).
sis of the product isolated by ion exchange chromatography combined with preparative HPLC.
Figure 2 illustrates the time course of dopamine oxidation and product formation for the reaction carried out under the conditions described above. A substrate
Neurotoxin Formation by Iron-Assisted Dopamine Oxidation
Figure 3. Plot of dopamine (dashed line, left axis) and product (right axis, solid lines) concentration vs Fe(II) concentration: dopamine (0), 5 (9), 4 (2), 3 (b), and 2 ([) at reaction times of 2 (a) and 24 h (b).
consumption level of around 80% is observed in the first 12 h with associated formation of the quinone 2 which peaks at 2 h and decreases below detection limits after ca. 10 h; on the other hand, the concentration of 5 as well of the tetrahydroisoquinoline 3 increases without reaching a plateau over the time period that was investigated. Variation of ferrous ion concentration in the range of 10-100 µM affected dopamine oxidation at a reaction time of 2 h (Figure 3a), with a significant increase in the level of substrate consumption from 60 to 40% at low iron to dopamine molar ratios (0.1-0.25), whereas at higher metal concentrations, the level of dopamine consumption did not increase proportionally. The dependence of product yields on iron concentration varied significantly and was more pronounced in the case of 5. By contrast, no significant effects of Fe(II) concentrations on either substrate consumption or product formation were observed at longer reaction times, e.g., 24 h (Figure 3b), with the exception of 5, whose level apparently declines at high metal to substrate ratios. No substantial changes in the reaction kinetics and course were observed using Fe(III) in place of Fe(II) ions. Under an oxygen-depleted atmosphere, in the presence of either Fe(III) or Fe(II) ions, the reaction did not take place. Plots in Figure 4 report the formation yields of the oxidation products as a function of dopamine concentration in the range of 50-200 µM at different reaction times. Both at 2 h (upper plot) and at longer reaction times (e.g., 24 h, lower plot), the extent of formation of products varies proportionally with dopamine concentration, while the extent of dopamine consumption decreases with increasing dopamine concentration. Periodic addition of hydrogen peroxide up to 3 molar equiv with respect to dopamine in the first hours (6-8 h) of the reaction produced a decrease in the product yields with the associated formation of chromatographically ill-defined materials. In the presence of catalase, the course of dopamine oxidation is modified with respect to product distribution, the effect being more apparent
Chem. Res. Toxicol., Vol. 12, No. 11, 1999 1093
Figure 4. Plot of dopamine (dashed line, left axis) and product (right axis, solid lines) concentration vs dopamine concentration: dopamine (0), 5 (9), 4 (2), 3 (b), and 2 ([) at reaction times of 2 (a) and 24 h (b).
Figure 5. Effect of catalase on the course of dopamine oxidation. Reaction conditions were like those described in the legend of Figure 1. Dopamine and product concentration: no additive (white bars) and catalase added (stippled bars) at a reaction time of 10 h. Catalase was added in three aliquots over the reaction time up to a final concentration of 210 units/mL.
at a reaction time of 10 h. Plots of Figure 5 show that catalase induces a decrease of the formation yields of 4 and 3, while those of 5 are scarcely modified. Table 1 shows the effect of various additives on the extent of dopamine oxidation and product formation determined at reaction times of 2 and 24 h. Metal chelators, i.e., 2,2′-bipyridyl and EDTA, did not affect dopamine consumption to a significant extent, but inhibited product formation, the effect being significantly greater for 2,2′-bipyridyl. Ascorbate and glutathione, which reach high levels in brain tissues (30), showed a pronounced inhibition on accumulation of the reaction products, particularly at the 2 h reaction time, but only slightly affected dopamine consumption. The effect was generally greater for glutathione and was more appreciable in the formation of 2, 3, and 5. Some metal ions which are present in trace amounts in tissues and are relevant to several pathologies, particularly neurodegenerative disorders (31), were also examined. Only Mn(II) exhibited a catalyzing action on the reaction comparable
1094
Chem. Res. Toxicol., Vol. 12, No. 11, 1999
Napolitano et al.
Table 1. Effect of Additives on Dopamine Oxidation and Product Yieldsa yield (%)b 2
d
hc
24 hc
additives
dopamine
2
3
4
5
dopamine
3
4
5
none DMSO (200 mM) mannitol (200 mM) EDTA (40 mM) 2,2′-bipyridyl (120 µM) ascorbate (10 mM) glutathione (10 mM) Cu (40 µM)d Zn (40 µM)d Mn (40 µM)d
44 47 52 53 47 50 56 37 88 63
3.3 3.0 2.0 0.89 0.6 1.6 1.2 1.0 1.0
0.4 0.5 2.0 0.36 0.14 0.34 0.17 0.3 0.98
3.4 3.0 3.5 1.4 0.83 1.9 2.6 2.8 2.9
5.3 4.0 3.8 2.0 1.7 1.5 1.7 2.6 4.9
20 16 17 20 18 23 20 32 20
7.0 5.0 19 5.0 3.0 5.6 5.9 5 3.3
16 15 16 14 11 12.3 15 12 17
15 13 14 10 5.1 8.8 11 10 14
a Reaction conditions like those described in the legend of Figure 1. b Average of three determinations; SD e 10%. c Reaction time. No iron added.
Scheme 1. Fragmentation Pattern of Compound 4 As Obtained by Fe(II)-Assisted Dopamine Oxidation in the Presence of 50% 18O2
Figure 6. Effect of different monosaccharides on the course of dopamine oxidation. Reaction conditions were like those described in the legend of Figure 1. Dopamine and product concentration: no additive (white bars), 1 mM glucose added (black bars), fructose added (stippled bars), and ribose added (cross-hatched bars) at a reaction time of 24 h.
to Fe(II) ions, while Zn(II) was less active; Cu(II) induced a far higher level of consumption of the substrate at 2 h, but did not give rise to detectable amounts of products 3-5, leading exclusively to insoluble pigmented material. Particularly worthy of note was the effect of the radical scanvengers DMSO and mannitol. While 200 mM DMSO proved to be virtually inactive during the course of the reaction, mannitol at a concentration of 200 mM produced a 4-fold increase in the yields of 3 with respect to the control at both reaction times, although it did not vary markedly the level of dopamine consumption. On the basis of this finding, investigation was extended to other monosaccharides, including glucose, fructose, and ribose. As shown in Figure 6, the yields of 3 were significantly enhanced and the level of dopamine consumption reduced in the presence of all the monosaccharides examined at a concentration of 1 mM, while the yields of the other products were scarcely changed. Further experiments were aimed at ascertaining the mechanism of formation of the oxidation products under the reaction conditions that were investigated. In our previous study (22), we investigated the origin of C-1 of tetrahydroisoquinoline by use of R- and β-13C-labeled dopamine. To probe the origin of the carbonyl oxygen of 4 and the hydroxyl group of 5, the oxidation reaction was carried out in an atmosphere of 50% 18O2 as described in Experimental Procedures. At reaction times of 6 and 24 h, aliquots of the reaction mixture were extracted, and after purification, the aldehyde was subjected to EI/MS analysis. A peak at m/z 140 at a 1:1 ratio with respect to the molecular ion peak of 4 at m/z 138 was very apparent in the spectra from both aliquots, providing evidence for incorporation of 18O in the compound. Further support
was gained by the fragmentation pattern summarized in Scheme 1. The base peak at m/z 109 arises from the loss of the formyl group carrying the label. A contribution to this peak is also provided by the loss of CO at the catechol site with H migration from the molecular ion peak, as is apparent from the presence of a prominent m/z 111 peak due to the 18O-labeled compound. The m/z 81/83 couple is generated by such a fragmentation route by further CO loss. An alternative fragmentation involving sequential CO loss at the catechol sites gives rise to peaks at m/z 110/112 and 82/84 from the 16O/18O compound. Intense peaks were also present at m/z 55/57 arising from the molecular ion peak at m/z 140/138 by loss of a C4H3O2 unit. In view of the relatively low amounts of oxidation mixtures that are available in the labeling experiments, analysis of norepinephrine was carried out at a reaction time of 24 h after acetylation treatment which made isolation of the compound easier. The EI/MS spectrum exhibited a weak molecular ion peak at m/z 337 accompanied by fragmentation patterns due to loss of an acetoxyl group (m/z 278, M - OCOCH3) and two sequential losses of ketene (42 amu) from the acetyl groups, with peaks at m/z 236 and 194, or four sequential losses of ketene with peaks at m/z 295, 223, 181, and 139. None of the latter series of fragments, which should include the oxygen of the original hydroxyl group, exhibited a comparable peak greater by 2 amu which ruled out the incorporation of the 18O from oxygen in 5.
Discussion The role played by species arising from aberrant oxidation pathways of dopamine has been considered
Neurotoxin Formation by Iron-Assisted Dopamine Oxidation
with increasing interest with a view of accounting for the selective loss of dopaminergic nigrostriatal neurons observed in Parkinson’s disease and other neurodegenerative disorders. However, the nature of these species and the molecular mechanisms underlying their formation under the oxidative stress conditions associated with the onset of neurodegeneration remain an open issue. Most attention (25, 32) has focused on the quinone of dopamine and further oxidation products such as dopaminochrome whose toxicity has been ascribed to the electrophilic character making these species prone to attack by nucleophilic sites of cellular constituents or enzymes, notably, sulfhydryl groups on free cysteine, reduced glutathione, and cysteinyl residues in proteins (33, 34). A novel hint is offered by the results of the study presented here, providing evidence for an unusual oxidative reactivity of dopamine at the ethylamine side chain which may result in the formation of toxic metabolites. These reaction routes become operative under conditions which can be straightforwardly related to those documented in neurodegenerative processes, i.e., the presence of oxygen and accumulation of iron in the reaction milieu (7, 10, 30). The ability of iron ions to divert the oxidation reactivity of dopamine has been described in our previous study (21). The effect is apparently due to the formation of a chelate which involves the catechol moiety of the molecule. This is very consistent with the effect of metal chelators which inhibited product formation, but did not affect dopamine consumption resulting from aerial oxidation. Under the reaction conditions that were adopted, a rapid and quantitative oxidation of Fe(II) to Fe(III) ions, which are actually involved in chelate formation, takes place which accounts for the close similarity of the reaction course observed with either species. In addition to iron ions, other metal ions proved to be able to direct the reaction course toward formation of products 2-5, notably, Mn(II) which is commonly regarded as an enviromental neurotoxin because of the Parkinson-like syndrome reported in manganese miners (35). By contrast, the enhanced dopamine conversion to melanin-like pigments observed with copper ions may be taken as evidence that the oxidation-promoting effect of these ions supersedes their chelating capacity. It should be cautioned, however, that though relatively low, dopamine concentrations used in the investigation presented here exceed those occurring in extraneuronal environments, and hence, the relevance of the observed effects of metal ions on the dopamine oxidation pathways in vivo is difficult to assess on the basis of the available evidence. Another peculiar aspect of the reaction course which may be worthy of comment is the involvement of hydrogen peroxide. Evidence for a direct role of hydrogen peroxide in product formation derives primarily from the experiments using catalase which showed a marked inhibitory effect on product formation. A most likely mode of generation of hydrogen peroxide involves reduction of molecular oxygen by the catechol species. The failure to speed the reaction or enhance product yields by slow external addition of hydrogen peroxide may be interpreted in terms of the intervention of degradative processes involving nucleophilic addition to quinonoid intermediates and subsequent muconic-type ring fission (36). It seems therefore that only a much slower, yet continuous, production of hydrogen peroxide, which is
Chem. Res. Toxicol., Vol. 12, No. 11, 1999 1095 Scheme 2. Proposed Mechanism of Formation of Compounds 2-5 in the Aerial Oxidation of Dopamine in the Presence of Metal Ions Highlighting the Role of Dopamine o-Quinone and Quinonemethide as Reaction Intermediates
difficult to mimic by external addition, warrants formation of products 2-5. On the basis of these data and the 18O incorporation experiments, a mechanism accounting for the formation of reaction products 2-5 can be envisaged involving as an initial step oxidation of dopamine to the corresponding o-quinone with concomitant generation of hydrogen peroxide from molecular oxygen as depicted in Scheme 2. The quinone is at equilibrium with the tautomer quinonemethide to which further stabilization is possibly offered by chelation by metal ions. Formation of 5 would then result from addition of water as indicated by the lack of label incorporation from 18O2. An alternate reaction path of the quinonemethide involves nucleophilic attack by hydrogen peroxide which gives a reactive β-aminohydroperoxide which subsequently decomposes via a carbonyl-forming fragmentation (37) to formaldimine and 4, this latter incorporating the original 18O from molecular oxygen. Pictet-Spengler condensation of dopamine with CH2dNH or the resulting formaldehyde leads to the tetrahydroisoquinoline 3 which, as shown by our previous 13C labeling experiments (22), incorporates at the 1-position the R-carbon of another dopamine molecule. In this picture, 6-hydroxydopamine quinone 2 would result from hydrogen peroxide addition to the electron-deficient 6-position of the o-quinone of dopamine, by a mechanism analogous to that shown in our previous study of oxidation of dopamine and model catechol compounds in the presence of hydrogen peroxide (20). In this vein, it is interesting to note that several candidate mechanisms have been proposed for endogenous formation of 6-hydroxydopamine or its quinone, including the Fenton- or Udenfriend-mediated ring hydroxylation of dopamine (19) and the reaction of the dopamine-Fe(II) chelate with lipid hydroperoxides (21). Although basically different, these reaction pathways are not mutually exclusive and may all contribute to production of significant amounts of a quinone like 2. The reported toxicity
1096
Chem. Res. Toxicol., Vol. 12, No. 11, 1999
(13, 38) of this compound may be well identified in terms of the lower tendency with respect to dopamine o-quinone to self-deactivate by intramolecular cyclization (39) which makes the attack by nucleophilic sites of cellular constituents a most likely reaction pathway (21). The rapid decay of the quinone 2 during the course of aerial dopamine oxidation is to be ascribed primarily to the tendency to undergo oxidative fission under the reaction conditions (unpublished results) and to a lesser extent to its intramolecular cyclization leading to dopaminochrome and hence to melanin pigments (20). In view of the mechanistic scheme depicted above, the assistance offered by metal ions to the formation of norepinephrine may be ascribed to the stabilization of dopamine quinonemethide by iron chelation, which allows the attack of a nucleophile as weak as water to become significant. It should be emphasized in this vein that this mechanism, which is also relevant to the chemistry of insect cuticle sclerotization (40), is not unprecedented (41, 42), although in most studies hydration or addition of nucleophiles other than water at the benzylic position of dopamine quinonemethide was observed on N-blocked derivatives. Under the conditions used in the study presented here, however, formation of 5 by oxidation of dopamine did not produce detectable yields unless it was carried out in the presence of iron ions. On the other hand, the decrease in norepinephrine yields at high metal to dopamine ratios at long reaction times is likely due to the reactivity of the norepinephrine-iron complex as evidenced in separate experiments (unpublished results). Additional support for the mechanistic route to 3 presented in Scheme 2 is the marked dependence of the formation yields on dopamine concentration, which is more appreciable at long reaction times as expected considering the kinetics of degradative fission of the ethylamine side chain. The inhibitory effect on dopamine oxidation exerted by antioxidants such as GSH and ascorbate is in line with the postulated intermediacy of quinonoid species. It should be considered, however, that the kinetics of the oxidation reaction leading to 3-5 is slow with respect to the life span of these antioxidants in the reaction medium, accounting for the significant decrease in the magnitude of the inhibitory effect observed at 24 h. In accord with the mechanism presented in Scheme 2 which rules out the intervention of a Fenton-type chemistry in the formation of norepinephrine, as previously suggested (19), is also the failure of hydroxyl radical scavengers such as DMSO and mannitol to produce a detectable inhibitory effect. By contrast, the ease of undergoing degradation in the presence of Fe(II)/H2O2 with generation of aldehyde products, most notably formaldehyde (43), may readily explain the promoting action of mannitol as well as that of the other monosaccharides that were investigated. Apart from the possible involvement in neuronal degeneration, it is tempting to speculate here that the observed oxidative degradation of the dopamine side chain leading eventually to norsalsolinol (3) represents a previously unrecognized route to isoquinoline alkaloids. The occurrence of these compounds in human brain is widely documented (17), especially under conditions of continuous dopamine refuelling as in parkinsonian patients on L-DOPA medication (44). Metabolic routes to 3 have been demonstrated in vivo, involving condensation of dopamine either with formaldehyde derived from
Napolitano et al.
sources such as N5,N10-methylene tetrahydrofolate (45) or with glyoxalate (46) with subsequent decarboxylation. Although it still awaits in vivo demonstration, the reaction to 3 presented in this paper may contribute as well to the metabolic pool of this alkaloid.
Conclusions Investigation of the oxidation chemistry of dopamine under physiologically relevant conditions has traditionally focused on the reactivity of the catechol functionality, particularly in relation to the biosynthesis of neuromelanin (47) which is regarded as a specific vulnerability factor for pigmented nigrostriatal neurons, which are preferentially lost in aging and Parkinson’s disease. The study presented here discloses novel modifications of the dopamine side chain under oxidative stress conditions which offer new clues to the interpretation of dopaminederived toxicity. When the fact that iron and hydrogen peroxide concentrations exceed basal levels at advanced stages of substantia nigra degeneration in Parkinson’s disease is considered, it is conceivable that these reaction paths become operative only after the primary toxic processes are underway, perhaps contributing to the exacerbation of the damaging events by release of highly toxic molecules. The reactivity described may also be regarded as a new mechanism of toxicity related to iron accumulation which adds to that revolving around the ability of neuromelanin-iron complexes to promote membrane lipid peroxidation (48). While the neurotoxicity related to generation of 6-hydroxydopaminequinone and tetrahydroisoquinoline alkaloids seems fairly well established, the potential cytotoxicity of the benzaldehyde 4 should not be overlooked also in view of the reported ability of this compound to inhibit DNA synthesis (49). The implication of norepinephrine generation from dopamine by a route alternative to that mediated by dopamine β-hydroxylase is not clear at present. It is possible however that the final outcome of the activation of this route under oxidative stress conditions is an alteration of dopamine networks by receptorial and autoreceptorial interference.
Acknowledgment. This work was supported by grants from Ministero dell’Universita` e della Ricerca Scientifica e Tecnologica (Rome, Italy) and Consiglio Nazionale delle Ricerche (CNR, Rome, Italy). Mass spectra were performed by the Centro di Spettrometria di Massa del CNR e dell’Universita` di Napoli. The assistance of the staff is gratefully acknowledged. We thank Miss Silvana Corsani for technical assistance.
References (1) Olanow, C. W. (1990) Oxidation reaction in Parkinson’s disease. Neurology 40, 32-37. (2) Fhan, S., and Cohen, G. (1992) The oxidant stress hypothesis in Parkinson’s disease: evidence supporting it. Ann. Neurol. 32, 804-812. (3) Perry, T. L., Godin, D. A., and Hansen, S. (1982) Parkinson’s disease: a disorder due to nigral glutathione deficiency? Neurosci. Lett. 33, 305-310. (4) Riederer, P., Sofic, E., and Rausch, W. D. (1989) Transition metals, ferritin, glutathione and ascorbic acid in parkinsonian brain. J. Neurochem. 52, 512-520. (5) Dexter, D. T., Carter, C. J., Wells, F. R., Javoy-Agid, F., Agid, Y., Lees, A., Jenner, P., and Marsden, C. D. L. (1990) Basal lipid peroxidation in substantia nigra is increased in Parkinson’s disease. J. Neurochem. 52, 381-389.
Neurotoxin Formation by Iron-Assisted Dopamine Oxidation (6) Schapira, A. H. V., Cooper, J. M., Dexter, D., Clark, J. B., Jenner, P., and Marsden, C. D. (1990) Mitocondrial complex I deficiency in Parkinson’s disease. J. Neurochem. 54, 823-827. (7) Spina, M. B., and Cohen, G. (1989) Dopamine turnover and glutathione oxidation: implication for Parkinson’s disease. Proc. Natl. Acad. Sci. U.S.A. 86, 1398-1400. (8) Cohen, G. (1986) Monoamine oxidase, hydrogen peroxide and Parkinson’s disease. In Advances in Neurology (Yahr, M. D., and Bergmann, K. J., Eds.) Vol. 45, pp 119-125, Raven Press, New York. (9) Sofic, E., Paulus, W., Jellinger, K. A., Riederer, P., and Youdim, M. B. H. (1991) Selective increase of iron in substantia nigra zona compacta of parkinsonian brain. J. Neurochem. 56, 978-982. (10) Youdim, M. B. H., Ben-Shachar, D., and Riederer, P. (1989) Is Parkinson’s disease a progressive siderosis of substantia nigra resulting in iron and melanin induced neurodegeneration? Acta Neurol. Scand. 126, 47-54. (11) Adam, J. D., Jr., and Odunze, I. N. (1991) Oxygen free radicals and Parkinson’s disease. Free Radical Biol. Med. 10, 161-169. (12) Tranzer, J. P., and Thoenen, H. (1968) An electron microscopic study of selective acute degeneration of sympathetic nerve terminals after administration of 6-hydroxydopamine. Experientia 24, 155-156. (13) Langston, J. W., Irwin, I., and Ricaurte, G. A. (1987) Neurotoxins, parkinsonism and Parkinson’s disease. Pharmacol. Ther. 32, 1949. (14) Seiden, L. S., and Vosmer, G. (1984) Formation of 6-hydroxydopamine in caudate nucleus of the rat brain after a single large dose of methylamphetamine. Pharmacol., Biochem. Behav. 21, 29-31. (15) Yoshida, M., Ogawa, M., Suzuki, K., and Nagatsu, T. (1993) Parkinsonism produced by tetrahydroisoquinoline (TIQ) or the analogues. Adv. Neurol. 60, 207-211. (16) Naoi, M., Dostert, P., Yoshida, M., and Nagatsu, T. (1993) N-Methylated tetrahydroisoquinolines as dopaminergic neurotoxins. Adv. Neurol. 60, 212-216. (17) Naoi, M., Maruyama, W., and Dostert, P. (1995) Dopaminederived 6,7-dihydroxy-1,2,3,4-tetrahydroisoquinolines: oxidation and neurotoxicity. Prog. Brain Res. 106, 227-239. (18) Garcia-Moreno, M., Rodriguez-Lopez, J. N., Martinez-Ortiz, F., Tudela, J., Varon, R., and Garcia-Canovas, F. (1991) Effect of pH on the oxidation pathway of dopamine catalysed by tyrosinase. Arch. Biochem. Biophys. 288, 427-434. (19) Slivka, A., and Cohen, G. (1985) Hydroxyl radical attack on dopamine. J. Biol. Chem. 29, 15466-15472. (20) Napolitano, A., Crescenzi, O., Pezzella, A., and Prota, G. (1995) Generation of the neurotoxin 6-hydroxydopamine by peroxidase/ H2O2. J. Med. Chem. 38, 917-922. (21) Pezzella, A., d’Ischia, M., Napolitano, A., Misuraca, G., and Prota, G. (1997) Iron-mediated generation of the neurotoxin 6-hydroxydopamine quinone by reaction of fatty acid hydroperoxides with dopamine: a possible contributory mechanism for neuronal degeneration in Parkinson’s disease. J. Med. Chem. 42, 20112017. (22) Napolitano, A., Pezzella, A., and Prota, G. (1999) 6,7-Dihydroxy1,2,3,4-tetrahydroisoquinoline formation by iron mediated dopamine oxidation: a novel route to endogenous neurotoxins under oxidative stress conditions. Tetrahedron Lett. 40, 2833-2836. (23) Wehrli, P. A., Pigott, F., Fischer, U., and Kaiser, A. (1972) Oxidationsprodukte von 6-Hydroxy-dopamin. (Oxidation products of 6-Hydroxydopamine). Helv. Chim. Acta 55, 3057-3061. (24) Whaley, W. M., and Govindachari, T. R. (1951) The PictetSpengler synthesis of TQ and related compounds. In Organic Reactions (Adams, R., Ed.) Vol. 6, pp 151-190, John Wiley & Sons, New York. (25) Tse, D. C. S., McCreery, R. L., and Adams, R. N. (1976) Potential oxidative pathways of brain catecholamines. J. Med. Chem. 19, 37-40. (26) Zhang, F., and Dryhurst, G. (1994) Effects of L-cysteine on the oxidation chemistry of dopamine: new reaction pathways of potential relevance to idiopathic Parkinson’s disease. J. Med. Chem. 37, 1084-1098. (27) Crescenzi, O., Kro¨sche, C., Hoffbauer, W., Jansen, M., Napolitano, A., Prota, G., and Peter, M. G. (1994) Synthesis of dopamines labelled with 13C in the R- or β-side chain positions, and their application for structural studies on melanins by solid-state NMR spectroscopy. Justus Liebigs Ann. Chem. 1994, 563-567. (28) Swan, G. A. (1974) Structure, chemistry and biosynthesis of the melanins. In Fortschritte der Chemie Organischer Naturstoffe (Hertz, W., Grisenbach, H., and Kirby, G. W., Eds.) Vol. 31, pp 521-582 Springer-Verlag, Wien, Germany.
Chem. Res. Toxicol., Vol. 12, No. 11, 1999 1097 (29) Dorris, R. L., and Shore, P. A. (1971) Amine uptake and storage mechanisms in the corpus striatum of rat and rabbit. J. Pharmacol. Exp. Ther. 179, 15-19. (30) Linert, W., Herlinger, E., Jameson, R. F., Kienzl, E., Jellinger, K., and Youdim, M. B. H. (1996) Dopamine, 6-hydroxydopamine, iron, and dioxygen-their mutual interaction and possible implication in the development of Parkinson’s disease. Biochim. Biophys. Acta 1316, 160-168. (31) Kienzl, E., Puchinger, L., Jellinger, K., Linert, W., Stachelberger, H., and Jameson, R. F. (1995) The role of transition metals in the pathogenesis of Parkinson’s disease. J. Neurol. Sci. 134, 6978. (32) Smythies, J., and Galzigna, L. (1998) The oxidative metabolism of catecholamines in the brain: a Review. Biochim. Biophys. Acta 1380, 159-162. (33) Rabinovic, A. D., and Hastings, T. G. (1998) Role of endogenous glutathione in the oxidation of dopamine. J. Neurochem. 71, 2071-2078. (34) Spencer, J. P., Jenner, P., Daniel, S. E., Lees, A. J., Marsden, D. C., and Halliwell, B. (1998) Conjugates of catecholamines with cysteine and GSH in Parkinson’s disease: possible mechanisms of formation involving reactive oxygen species. J. Neurochem. 71, 2112-2122. (35) Barbeau, A. (1984) Manganese and extrapyramidal disorders. Neurotoxicology 5, 13-36. (36) Sawaky, Y., and Foote, C. F. (1983) Mechanism of C-C cleavage of cyclic diketones with alkaline hydrogen peroxide. The acyclic mechanism and its application to the basic autoxidation of pyrogallol. J. Am. Chem. Soc. 105, 5035-5040. (37) Becker, K. B., and Grob, C. A. (1977) The formation of unsaturated groups by heterolytic fragmentation In The Chemistry of DoubleBonded Functional Groups (Patai, S., Ed.) Supplement A, pp 653723, John Wiley & Sons, New York. (38) Graham, D. G., Tiffany, S. M., Bell, W. R., Jr., and Gutknecht, W. F. (1978) Autoxidation versus covalent binding of quinones as the mechanism of toxicity of dopamine, 6-hydroxydopamine and related compounds toward C1300 neuroblastoma cells in vitro. Mol. Pharmacol. 14, 644-653. (39) Blank, C. L., McCreery, R. L., Wightman, R. M., Chey, W., and Adams, R. N. (1976) Intracyclization rates of 6-hydroxydopamine and 6-aminodopamine analogs under physiological conditions. J. Med. Chem. 19, 178-180. (40) Peter, M. G. (1989) Chemical modification of biopolymers by quinone and quinone methides. Angew. Chem., Int. Ed. 28, 555570. (41) Thompson, D. C., Thompson, J. A., Sugumaran, M., and Moldeus, P. (1993) Biological and toxicological consequences of quinone methide formation. Chem.-Biol. Interact. 86, 129-162. (42) Senoh, S., Creveling, C. R., Udenfriend, S., and Witkop B. (1959) Chemical, enzymatic and metabolic studies on the mechanism of oxidation of dopamine. J. Am. Chem. Soc. 81, 6236-6240. (43) Mlakar, A., Batna, A., Dudda, A., and Spiteller, G. (1996) Iron(II) ions induced oxidation of ascorbic acid and glucose. Free Radical Res. 25, 525-539. (44) Niwa, T., Takeda, N., Yoshizumi, H., Tatematsu, A., Yoshida, M., Dostert, P., Naoi, M., and Nagatsu, T. (1993) Presence of tetrahydroisoquinoline-related compounds, possible MPTP-like neurotoxins, in parkinsonian brains. Adv. Neurol. 60, 234-237. (45) Melchior, C., and Collins, M. A. (1982) The route and significance of endogenous synthesis of alkaloids in animals. Crit. Rev. Toxicol. 9, 313-356. (46) Tasaki, Y., Makino, Y., Ohta, S., and Hirobe, M. (1993) Biosynthesis of 1-methyl-1,2,3,4-tetrahydroisoquinoline (1MeTIQ), a possible antiparkinsonism agent. Adv. Neurol. 60, 231-233. (47) d’Ischia, M., and Prota, G. (1997) Biosynthesis, structure and function of neuromelanin and its relation to Parkinson’s disease: a critical update. Pigment Cell Res. 10, 370-376. (48) Dexter, D. T., Carter, C. J., Wells, F. R., Javoy-Agid, F., Agid, Y., Lees, A., Jenner, P., and Marsden, C. D. (1989) Basal lipid peroxidation in substantia nigra is increased in Parkinson’s disease. J. Neurochem. 52, 382-389. (49) Nishimura, M., Umeda, M., Kusanagi, A., Urakabe, R., Yukimatsu, K., and Otawara, S. (1981) Cytomorphological study of cultured mammalian cells treated with benzaldehyde, catechol and their derivatives with special reference to inhibition of DNA synthesis. Gann 72, 272-279.
TX990079P
