Oxidation Chemistry of Norepinephrine - American Chemical Society

Sep 25, 2007 - following transient cerebral ischemia, and the onset and progression of idiopathic vitiligo. An oxidative pathway of 1 is also believed...
1 downloads 0 Views 177KB Size
Download PDF
Chem. Res. Toxicol. 2007, 20, 1549–1555

1549

Oxidation Chemistry of Norepinephrine: Partitioning of the O-Quinone between Competing Cyclization and Chain Breakdown Pathways and Their Roles in Melanin Formation Paola Manini, Lucia Panzella, Alessandra Napolitano,* and Marco d’Ischia Department of Organic Chemistry and Biochemistry, UniVersity of Naples Federico II, Via Cinthia 4, I-80126 Naples, Italy ReceiVed July 11, 2007

Aberrant oxidation of norepinephrine (1) via the transient o-quinone has been implicated as a critical pathogenetic mechanism underlying the degeneration of noradrenergic cell bodies in the locus coeruleus in Parkinson’s disease, the degeneration of noradrenergic nerve terminals in Alzheimer’s disease and following transient cerebral ischemia, and the onset and progression of idiopathic vitiligo. An oxidative pathway of 1 is also believed to account for the slow deposition of neuromelanin in pigmented neurons of the locus coeruleus. Remarkably, after extensive investigations spanning over several decades, there is still a lack of knowledge of the oxidation chemistry of 1 beyond the classic cyclization route leading to aminochrome and lutin intermediates. We report herein that oxidation of 1 in the 50–500 µM concentration range with H2O2-dependent oxidizing agents, such as the Fenton reagent (Fe2+-EDTA/ H2O2) and the horseradish peroxidase (HRP)/H2O2 system, leads not only to the known cyclization products, such as noradrenochrome and 5,6-dihydroxyindole (3), but also to a significant proportion of chain breakdown products, including 3,4-dihydroxybenzaldehyde, 3,4-dihydroxybenzoic acid, 3,4dihydroxymandelic acid, and 3,4-dihydroxyphenylglyoxylic acid, which has never been described among the oxidation products or metabolites of 1. Analysis of the brown melanin-like pigment obtained by oxidation of 1 with HRP/H2O2 gave pyrrole-2,3-dicarboxylic acid and pyrrole-2,3,5-tricarboxylic acid, diagnostic markers of 3-derived units in eumelanins. Comparison with reference pigments prepared by similar oxidation of dopamine and 3 indicated that in the case of 1 oxidative polymerization of indole units through the 2-position contributes only to a minor extent to melanin formation. Overall, the results of this study provide a complete characterization of the oxidative chain fission pathways of 1, highlight 3,4-dihydroxyphenylglyoxylic acid as a novel possible metabolic product of this catecholamine, and yield an insight into norepinephrine-melanin, a putative component of locus coeruleus neuromelanin. Introduction Norepinephrine (noradrenaline, 1) is a major catecholamine neurotransmitter in brain areas concerned with physical and mental arousal and elevated mood. Under oxidative stress conditions, diversion of the normal metabolism of 1 towards aberrant pathways has been implicated as a potential etiopathogenetic factor and/or contributory mechanism in several disease states (1, 2). In Parkinson’s disease (PD)1 , neuropathological studies and biochemical findings have documented the occurrence of a degeneration of noradrenergic cells in the locus ceruleus (LC) concomitant with the established loss of dopaminergic neurons of the substantia nigra (SN) (3). Noradrenergic neurons of the LC contain a neuromelanin pigment formed by the oxidative polymerization of 1 and analogous to the black dopamine-based neuromelanin in the SN (4). Noradrenergic neurons that project from the LC to the cortex and hippocampus also degenerate in Alzheimer’s disease (AD) and following a transient ischemic insult (5), and it has been * To whom correspondence should be addressed. Phone: +39-081674133. Fax: +39-081674393. E-mail: [email protected]. 1 Abbreviations: PD, Parkinson’s disease; LC, locus coeruleus; SN, substantia nigra; AD, Alzheimer’s disease; EDTA, ethylenediaminetetraacetic acid; HRP, horseradish peroxidase; TMS, tetramethylsilane; COMT, catechol-O-methyl transferase; MAO, monoamino oxidase.

suggested that the vulnerability of noradrenergic nerve terminals might be related to aberrant oxidation of 1 to an o-quinone that reacts with endogenous L-cysteine and/or reduced glutathione to give neurotoxic adducts (6, 7). Increased urinary and plasma levels of 1 (up to 1 nM), epinephrine, and their metabolites and elevated production of hydrogen peroxide are found in patients with vitiligo, an acquired idiopathic hypomelanosis characterized by the appearance of depigmented areas on the skin (8–11). Although aberrant oxidative changes of 1 are now widely accepted as a possible underlying cause of toxicity (12), knowledge of the oxidative pathways of 1 remains poor and limited to the early stages of the process. These involve the formation of an unstable o-quinone, via semiquinone free radicals (13), and its cyclization to give leuconoradrenochrome and then noradrenochrome (2) (14). The subsequent steps of the pathway are believed to involve isomerization of 2 to a fluorescent indoxyl derivative, known as noradrenolutin (3,5,6trihydroxyindole) (15, 16). Apart from 2, no other oxidation product of 1 preceding the deposition of the final dark pigment has so far been identified. Recently, we found that oxidation of 1 (5 × 10-3 M) in aqueous phosphate buffer, pH 7.4, with Fe2+-EDTA/H2O2 (Fenton reagent) or with other chemical oxidants leads to a main reaction product, which was isolated as the acetyl derivative and was identified as the unusual 4-[bis-(1 H-5,6-dihydroxyin-

10.1021/tx700254q CCC: $37.00  2007 American Chemical Society Published on Web 09/25/2007

1550 Chem. Res. Toxicol., Vol. 20, No. 10, 2007

dol-2-yl)methyl]-1,2-dihydroxybenzene (17). Other components of the oxidation mixture, isolated as the acetyl derivatives, included noradrenolutin and 5,6-dihydroxyindole (3), originating from cyclization/dehydration of the o-quinone of 1. Inspection of the aqueous phase revealed the presence of 3,4-dihydroxymandelic acid (4) and 3,4-dihydroxybenzaldehyde (5), derived from the oxidative breakdown of the 2-amino-1-hydroxyethyl chain via a quinonemethide intermediate.

As an extension of that study, we report herein an in vitro study of the reactivity of 1 with the Fenton reagent (Fe2+-EDTA/ H2O2) and horseradish peroxidase (HRP)/H2O2, two oxidizing systems commonly used to mimic a physiological inflammatory response. The study was carried out with the catecholamine at micromolar concentrations, that is, close to those of physiological relevance but still compatible with the scale of the chemical investigation. The structures of new main products have been elucidated along with their mechanism of formation, and the relevance of competing oxidative pathways to the mechanism of formation of the neuromelanin-like pigment has been addressed for the first time.

Experimental Procedures Materials and Methods. (() Norepinephrine hydrochloride, 3,4dihydroxybenzoic acid, and horseradish peroxidase (HRP) (donor: H2O2 oxidoreductase, EC 1.11.1.7) type II were purchased from Sigma. 3,4-Dihydroxybenzaldehyde and ethylenediaminetetraacetic acid (EDTA) was from Fluka. DL-3,4-Dihydroxymandelic acid, (() epinephrine, sodium dithionite, and potassium ferricyanide were obtained from Aldrich. Hydrogen peroxide (water solution, 30%), Fe(NH4)2(SO4)2 × 6H2O, and sodium periodate were purchased from Carlo Erba. Adrenochrome and noradrenochrome (18), 5,6dihydroxyindole (19), pyrrole-2,3-dicarboxylic acid, and pyrrole2,3,5-tricarboxylic acid (20) were prepared according to reported procedures. Organic solvents were HPLC quality; 0.1 M phosphate buffer, pH 7.4, was treated with Chelex-100 resin before use to remove transition metal contaminants. UV spectra were obtained using a Beckman DU 640 spectrophotometer. NMR spectra were recorded with a Bruker DRX-400 MHz instrument. 1H and 13C NMR spectra were recorded at 400.1 and 100.6 MHz, respectively, in CD3OD or in D2O/DCl using TMS and t-butylalcohol, respectively, as the internal standards. HPLC analysis were carried out on a Shimadzu apparatus (SCL-10AV VP) equipped with a UV detector (SPD-10AV VP). The following systems were used: (I) Sphereclone ODS (5 µm, 4.6 × 250 mm) column, flow rate of 1 mL/min, 0.2 M formic acid/acetonitrile 95:5 (v/v) as the eluant, and the UV detector set at 280 nm; (II) Sphereclone ODS (5 µm, 4.6 × 250 mm) column, flow rate of 1 mL/min, 0.2 M H3PO4 with 10 mM sodium octansulphonate (pH 3)/acetonitrile 99:1 (v/v) as the eluant, and the UV detector set at 490 nm; (III) Synergi Hydro-RP 80A (4 µm, 4.6 × 250 mm) column (Phenomenex, Torrance, CA, USA), flow rate of 0.7 mL/

Manini et al. min, 1% formic acid (pH 2.8)/methanol 97:3 (v/v) as the eluant, and the UV detector set at 254 and 280 nm; (IV) Econosil (10 µm, 22 × 250 mm) column, flow rate of 15 mL/min, 0.1 M formic acid/acetonitrile 90:10 (v/v) as the eluant, and the UV detector set at 280 nm. Prior to HPLC analysis, all the mixtures were filtered through a 0.45 µm nylon membrane (Alltech Associates, Deerfield, IL, USA). Oxidation of Norepinephrine: General Procedure. A solution of 1 (50–500 µM) in 0.1 M phosphate buffer, pH 7.4, at 37 °C was treated with Fe(NH4)2(SO4)2 × 6H2O (one molar equivalent), EDTA (one molar equivalent), and H2O2 (one molar equivalent). Aliquots of the reaction mixture (300 µL) were periodically withdrawn, acidified with 3 M HCl (100 µL), and subjected to HPLC analysis for the detection of 1, 4–7 (system I) and of 2 (system II). Because of the high instability of this latter compound, its concentration was determined using adrenochrome as the chromatographic standard, in the assumption that 2 had a comparable molar extinction coefficient. Similar experiments were performed using HRP (2 U/mL) and H2O2 (one molar equivalent), or potassium ferricianyde (one molar equivalent). When required, a 1:10 molar ratio of the Fe(NH4)2(SO4)2 × 6H2O–EDTA complex with respect to substrate and hydrogen peroxide was used. Oxidation of 3,4-Dihydroxymandelic Acid (4) or 3,4Dihydroxybenzoic Acid (6) with the Fenton Reagent. A solution (500 µM) of 4 or 6 in 0.1 M phosphate buffer, pH 7.4, at 37 °C was treated with Fe(NH4)2(SO4)2 × 6H2O (500 µM), EDTA (500 µM), and H2O2 (500 µM). Aliquots of the reaction mixture (300 µL) were periodically withdrawn, acidified with 3 M HCl (100 µL), and subjected to HPLC analysis (system I). In the case of 6, the reaction mixture was analysed at 60 h (about 70% consumption of the substrate) after acidification to pH 3. Isolation of 3,4-Dihydroxymandelic Acid (4), 3,4-Dihydroxybenzaldehyde (5), 3,4-Dihydroxybenzoic Acid (6), and 3,4-Dihydroxyphenylglyoxylic Acid (7). A solution of 1 (720 mg, 3.5 mmol) in 0.1 M phosphate buffer, pH 7.4 (720 mL), at 37 °C was treated under vigorous stirring with Fe(NH4)2(SO4)2 × 6H2O (1.39 g, 3.5 mmol), EDTA (1.30 g, 3.5 mmol), and H2O2 (360 µL, 3.5 mmol). After 5 min, the reaction mixture was acidified with 3 M HCl (pH 3), was extracted with ethyl acetate (3 × 300 mL), and the aqueous and the organic layers were separated. The organic layers were dried over anhydrous sodium sulphate and evaporated under reduced pressure, and the residue was subjected to preparative HPLC (system IV) to afford pure 4 (26 mg, 4% yield, tR ) 4 min, system I), 5 (10 mg, 2% yield, tR ) 17.5 min, system I), and 6 (11 mg, 2% yield, tR ) 12 min, system I). The aqueous phase was lyophilized and fractionated by preparative HPLC (system IV) to afford pure 7 (70 mg, 11% yield, tR ) 6.3 min, system I). 4: 1H NMR (400 MHz, CD3OD) (21), δ (ppm): 4.97 (1H, s), 6.75–6.76 (2H, m), 6.89 (1H, d J ) 1.6 Hz). 5: 1H NMR (400 MHz, CD3OD), δ (ppm): 6.91 (1H, d J ) 8.4 Hz, H-5), 7.30 (1H, d J ) 1.6 Hz, H-2), 7.31 (1H, dd J ) 8.4, 1.6 Hz, H-6), 9.68 (1H, s, CHO). 6: 1H NMR (400 MHz, CD3OD) (22), δ (ppm): 6.80 (1H, d J ) 8.4 Hz, H-5), 7.42 (1H, dd J ) 8.4, 1.6 Hz, H-6), 7.43 (1H, d J ) 1.6 Hz, H-2). 7: 1H NMR (400 MHz, D2O/DCl), δ (ppm): 6.99 (1H, d J ) 7.6 Hz, H-5), 7.46 (1H, d J ) 1.6 Hz, H-2), 7.51 (1H, dd J ) 7.6, 1.6 Hz, H-6); 13C NMR (100 MHz, D2O/DCl), δ (ppm): 115.9 (CH), 116.4 (CH), 123.8 (CH), 126.5 (C), 144.8 (C), 151.8 (C), 169.8 (COOH), 192.5 (CO). Oxidation of 1, 3 and Dopamine with the HRP/H2O2 System: Isolation of Melanin-Like Pigments. A solution of the appropriate compound (500 µM) in 0.1 M phosphate buffer, pH 7.4, was treated with HRP (2 U/mL) and H2O2 (one molar equivalent). After 1 h, the mixture was acidified to pH 3 and the pigment collected by centrifugation at 1258 × g at 4 °C, washed repeatedly (4–5 times) with water, and lyophilized. Oxidative Degradation of Melanin Pigments. The pigments (5 mg) obtained by the oxidation of 1, 3, and dopamine were subjected to two different degradation reactions: (A) with H2O2 in 1 M K2CO3 according to previously described procedures (20, 23)

H2O2-Promoted Oxidation of Norepinephrine

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1551

Figure 2. Effect of the initial concentration of 1 on product yields; 4, shaded bars; 6, black bars; 5, grey bars; 7, white bars. The values were determined at 5 min reaction time.

readily identified as 2 by comparison of its behaviour with that of an authentic sample prepared by a standard procedure (18). To obtain sufficient amounts of products A–D for structural characterization, the reaction was carried out on a preparative scale, and the mixture was extracted with ethyl acetate. Preparative HPLC (system IV) fractionation of the organic phase led to isolation in pure form of products A, C, and D, whereas product B was obtained from the aqueous phase. Products A–D were then subjected to complete spectral analysis. Products A, C, and D were readily identified as 4, 5, and 3,4-dihydroxybenzoic acid (6) by spectral analysis (21, 22) and structural assignment was secured by comparison with authentic samples. The 1H NMR spectrum of compound B displayed signals for an aromatic ABX spin system at δ 6.99, 7.46, and 7.51, whereas the 13C NMR spectrum exhibited, besides the carbon signals of the catechol ring, two resonances at δ 169.8 and 192.5 diagnostic for carboxyl and keto groups, respectively. On this basis, the compound was formulated as 3,4-dihydroxyphenylglyoxylic acid (7).

Figure 1. HPLC-UV elutograms of the mixture obtained by oxidation of 1 with the Fenton reagent at 5 min reaction time. (a) System I; (b) system II. For details see Experimental Procedures.

with the following modifications. Prior to the addition to the K2CO3 solution, the pigment suspended in water was homogenized by a glass/glass potter. The reaction was allowed to stand under stirring for 18 h. After the addition of 5% NaHSO3, the mixture was taken to pH 4 with 85% H3PO4, filtered through nylon membranes, and analyzed by HPLC (system III). (B) The other was with acidic permanganate as described (24). After ethyl acetate extraction of the oxidation mixture (3 × 10 mL), the combined organic layers were dried over sodium sulphate and taken to dryness. The residue was dissolved in water (400 µL) and analyzed by HPLC (system III).

Results Oxidation of 1 (50–500 µM) with the Fenton reagent was carried out in 0.1 M phosphate buffer, pH 7.4, by reacting the catecholamine with Fe2+-EDTA and H2O2 at equimolar levels with respect to the substrate. The reaction course was monitored by reverse phase HPLC with detection wavelengths set at 280 nm for catechol compounds (system I) and at 490 nm for aminochrome species (system II). The elutogram in Figure 1a reveals after 5 min some residual 1 (tR ) 4.6 min) and four main products, A (tR ) 4.0 min), B (tR ) 6.3 min), C (tR ) 11.5 min), and D (tR ) 17.5 min). Figure 1b shows the elutogram of the same mixture recorded at 490 nm, in which only a single product was apparent (tR ) 6.3 min). This was

Traces of 3, a primary building block of cutaneous eumelanins and a major oxidative cyclization product of dopamine, were also found in the Fenton-promoted reaction by HPLC analysis (not shown). Figure 2 shows changes in product formation with the initial concentration of the substrate and equimolar levels of the Fenton reagent. With 50 µM 1, the oxidation proceeded at a slow rate and gave relatively high yields of 4, with little amounts of 5 and 7, but no detectable 6. With increasing concentration of 1, yields of 4 dropped while those of 7 increased steadily. Yields of 5 and 6 attained maximum values with 100 µM 1 and then decreased probably because of subsequent oxidation. In control experiments, it was found that Fenton-induced oxidation of 6 proceeds smoothly to give soluble yellowish products with no dark melanin deposited. Given the marked dependence of product distribution and fate on the concentration of 1, in further experiments, the time course of the reaction was monitored at two different concentrations of 1, i.e., 50 µM (top plot) and 100 µM (bottom plot) with equimolar amounts of the Fenton reagent (Figure 3). At 50 µM concentration, the substrate was almost completely consumed (>90%) within the first 5 min, and the rapid formation of 4 was observed, attaining maximum concentration at 5 min (44 µM) and then gradually decreasing with concomitant increase of 2 and 7. At 100 µM 1, however, the rate of

1552 Chem. Res. Toxicol., Vol. 20, No. 10, 2007

Figure 3. Time course of the oxidation of 1 at 50 µM (top panel) and 100 µM (bottom panel) with the Fenton reagent in 0.1 M phosphate buffer at pH 7.4 at 37 °C. The concentration of 4 (9) is reported on the left axis. The concentrations of 2 (*), 5 (2), 6 (∆), and 7 (0) are on the right axis.

substrate consumption was lower (50% after 5 min) and so was the rate of formation of compound 4, attaining a concentration of about 30 µM at 5 min and then gradually decreasing. When the reaction was performed on 100 µM 1 but with 10 µM Fenton reagent, the decay of 1 was much slower, though a similar product distribution was observed after 24 h by HPLC. In a separate experiment, 4 was oxidized with the Fenton reagent and was found to afford 5 as main product (80%), in line with early observations (25, 26). However, no detectable 7 was formed, thus ruling out a possible origin of the latter product by oxidation of 4. To determine whether the observed product pattern is specific to the Fenton reagent, oxidation of 1 was investigated using an enzymatic system, HRP/H2O2, and a chemical oxidant, potassium ferricyanide (K3Fe(CN)6). HPLC analysis showed in both cases the formation of 4, 5 (in trace amounts), and 6; 3 was also present as an additional minor component. Figure 4 shows the yields of 4 and 5 from 500 µM 1 with the Fenton reagent, the HRP/H2O2 system and K3Fe(CN)6 as determined after 5 min of reaction time. In all cases, under comparable conditions, 4 was formed in higher yields compared to that of 5, which presumably derives from 4. Increased formation and accumulation of 4 with HRP/ H2O2 or K3Fe(CN)6 therefore suggests that the latter oxidants are less efficient than the Fenton reagent in promoting the conversion of 4 and other late oxidation products to 5. Judicious analysis of these data, however, must take into due account the different nature and mechanisms of action of the oxidizing systems tested. With these results available, a final series of experiments was directed to compare the chemical properties of the melanin pigment produced by the oxidation of 1, dopamine, and 3 under the specific conditions of this study, and to discuss the results in light of previous data (27). The aims of the experiments were (i) to determine the proportion of cyclized indole units and (ii) to assess the degree of polymerization of the indole units, that is, to determine whether they participate in oligomer/polymer structures or whether are simply incorporated into the pigment. Melanins were prepared by oxidation of 1 (500 µM) with HRP/H2O2, a most convenient oxidizing system for melanin

Manini et al.

Figure 4. Yields of formation of 4 (A) and 5 (B) by oxidation of 1 with the Fenton reagent (black bars), the peroxidase/H2O2 system (open bars), and K3Fe(CN)6 (gray bars). [1] ) 500 µM, reaction time ) 5 min.

formation in vitro, and the final pigment was collected after about 1 h. Data in Table 1 show that melanins from 1 and dopamine are obtained in relatively low yields (ca. 20%) compared to those from 3. Considering that consumption of 1 is about 70% at 1 h, it can be inferred that some 50% of the oxidation mixture consists of nonmelanizing species, such as chain-cleaved products. The same holds also for dopamine, whereas 3 is completely consumed within 1 h, leading to melanin pigments almost quantitatively. Pigment analyses were carried out by oxidative degradation followed by quantitative determination of two main fragmentation products serving as markers of indole units in melanins, i.e., pyrrole-2,3-dicarboxylic acid (8) and pyrrole-2,3,5-tricarboxylic acid (9). Compound 8 is a marker of indole units unsubstituted at C-2, chiefly terminal units, whereas 9 is an index of indole units linked through the 2-position, according to the prevalent mode of coupling of 3 (Scheme 1). Since the relative proportion of terminal units unsubstituted at C-2 gradually decreases with increasing oligomer length, the 8/9 ratio would decrease with increasing polymer length and can be taken as a rough index of the degree of polymerization of 3 in the pigment. Chemical degradation of melanins was performed according to two established procedures involving oxidative degradation with alkaline H2O2 (20) and with acidic KMnO4 (24). The former gives a better estimate of 8, which is partly degraded in acids, whereas the latter provides a more reliable determination of 9 since in acids artefactual indoleforming cyclization of catecholamine units during oxidative degradation is limited. However, the use of both sets of data is recommended for more confident conclusions about melanin indole composition, especially in those cases in which the yields of both 8 and 9 are low, or whenever the extent of artifactual formation of these markers has to be determined. In any case, similar qualitative trends are observed with both procedures, which should be preferably used on a comparative basis. Yields of 8 and 9 from either alkaline or acidic degradation of melanins are reported in Table 1 in comparison with those from the starting materials. The results are in satisfactory agreement with previous data (27) and indicate that (a) yields of 8 are higher under alkaline rather than acidic conditions; (b) 1-melanin gives the lowest overall yields of 8 + 9, melanin from 3 gives the highest, whereas dopamine–melanin stands in

H2O2-Promoted Oxidation of Norepinephrine

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1553

Table 1. Yields of 8 and 9 by Degradation of Melanins from 1, 3 and Dopaminea yield (µg/mg melanin)b degradation procedure acidic KMnO4 sample

8

9

8

20 94 20

0.15 ( 0.02 0.12 ( 0.03 0.16 ( 0.02 0.06 ( 0.01 0.08 ( 0.01 nd

0.24 ( 0.02 2.2 ( 0.2 1.1 ( 0.1 nd 0.34 ( 0.03 nd

8.8 ( 0.5 15.0 ( 1.0 11.0 ( 0.6 3.9 ( 0.2 50d 0.90 ( 0.04

1-melanin 3-melanin dopamine-melanin 1 3 dopamine a

alkaline H2O2

melanin yield (%)

9 c

nd 15.0 ( 1.0 7.0 ( 0.3 nd 0.1d nd

Data from precursors are included for comparison. b Average of three separate experiments ( SD. c nd, nondetectable. d Taken from ref 20.

Scheme 1. Origin of 8 and 9 from Oxidative Cleavage of 3-Units of the Melanic Polymer

Scheme 2. Mechanism Proposed for the Formation of the Reaction Products by Oxidation of 1 with the Fenton Reagent

between; (c) little 8 was obtained from 1 even with rigorous exclusion of oxygen, indicating a certain degree of oxidative cyclization induced by the alkaline degradation procedure. The value of the 8/9 ratio as an index of indole polymerization is well illustrated by comparing data referring to 3 and melanin from 3.

Discussion This is the first study delineating competing oxidation pathways of 1 of possible pathophysiological relevance. Much of the complexity of this chemistry can be traced to the mechanistic choices available to the initially formed o-quinone and other key intermediates at later stages, which are summarized in Scheme 2. In this scheme, the o-quinone partitions between the indole route to leuconoradrenochrome (I) and the alternative isomerization pathway to quinonemethide II. Formation of quinonemethides from catecholamine quinones is usually much slower than cyclization and only occurs when the latter process is unfavourable. It is known that the o-quinone of 1 undergoes cyclization at a relatively slow rate compared, for example, to epinephrine quinone or DOPA quinone (28), and is therefore susceptible to tautomerization, a process favored by the hydroxyl group enhancing the acidity of the β-proton. As previously reported (17), I can undergo dehydration to give 3, an event that competes with oxidative conversion to 2. However, the quinone-methide II partitions further between two isomerization paths giving rise to noradrenalone (III) or the imine IV. Hydrolysis of IV would give 3,4-dihydroxymandelic aldehyde (VI), which is oxidized to the corresponding acid 4. Oxidation of 4 would give the unstable o-quinone, which undergoes spontaneous decarboxylation to give 5 and then 6. Formation of 5 from 4 has already been described (26, 27) and would represent an important nonenzymatic route of oxidative chain breakdown of 1. The oxidative deamination of III would give the aldehyde V, which conceivably produces 7 by an oxidative process. To the best of our knowledge, 7 has never been reported

as an oxidation product of 1, and its generation by a nonenzymatic route is worthy of note. No straightforward explanation can be offered to the increased formation of 7 with increasing concentrations of 1 since the observed yields reflect a balance of formation and degradation pathways difficult to assess. However, the preferential accumulation of 4 with HRP/H2O2 and K3Fe(CN)6 can be explained, considering that these latter systems are less effective in oxidizing 4. The chemistry emerging from this study may be of relevance to the metabolic fate of 1 under oxidative stress conditions. In mammals, 1 is rapidly degraded to various metabolites, including normetanephrine (via the enzyme catechol-O-methyl transferase, COMT), 4, 3-methoxy-4-hydroxymandelic acid, and 3-methoxy4-hydroxyphenylglycol (via monoamino oxidase, MAO). The identification in biological samples of the oxidative chain breakdown products of 1 described in the present study might provide new clues as to the fate of 1 in vivo and the relative involvement of MAO-dependent versus MAO-independent metabolic pathways. Another important outcome of this study is the first detailed characterization of the melanin pigment produced by oxidation of 1. This pigment is virtually unknown from a structural viewpoint, though it is commonly regarded as a possible model of the neuromelanin present in human LC, which has so far been much less investigated compared to its SN counterpart. One view is that neuromelanin from LC is structurally and functionally different from that in SN and that it is unspecifically

1554 Chem. Res. Toxicol., Vol. 20, No. 10, 2007 Scheme 3. Competing Oxidative Pathways of 1 in Melanin Formation

formed by autoxidation of 1, although a genetic program underlying the biogenesis of neuromelanin has been proposed (29). Data in Table 1 allow one to draw the following conclusions: (a) 1-melanin has a lower content of cyclized indole units related to 3 compared to that of dopamine–melanin and (b) a large proportion of the indole units is unsubstituted at C-2, due perhaps to a low degree of polymerization or to the incorporation of leuconoradrenochrome in the pigment prior to its conversion to 3. Examination of Scheme 2 shows that the formation of 3 requires simple dehydration of leuconoradrenochrome (I), a route that competes with further oxidation to 2. This entails a significant, hitherto overlooked difference between the melaninforming pathways from 1 and dopamine, whereas in the case of dopamine, the leucoaminochrome must be oxidized to the aminochrome, which is then isomerized to 3. In the case of 1, a different channel is made available to the leucoaminochrome by the presence of the β-hydroxyl group, leading to the formation of 3. As an interesting observation, it has been reported that 1-melanin differs significantly from dopamine– melanin with regard to free radical properties (30). Whereas dopamine–melanin displays a single EPR line (∆Bpp, 0.50 mT) and appears to contain o-semiquinone free radicals with the characteristic g-value of 2.004, 1-melanin exhibits a superposition of two lines (∆Bpp, 0.45 and 0.81 mT). o-Semiquinone free radicals have been proposed to be responsible for the narrower component, while nitrogen free radicals with a g-factor of 2.003 were suggested to account for the broader component. It is tempting to speculate that this latter reflects the presence in the pigment polymer of uncyclized and/or leuconoradrenochrome-derived units differing from the indole units related to 3 supposed to be present in dopamine melanin. The relative importance of the competing oxidative pathways of 1 in melanin formation is illustrated in Scheme 3, which is clearly an oversimplification of the actual situation but highlights the main conclusions emerging from chemical analysis. Central to this scheme is leuconoradrenochrome, which is envisioned to partition between the dehydration route to 3 and the oxidative route to 2. Because of their relatively oxidizable nature, leuconoradrenochrome, 2, and 3 may all take part in the polymerization process leading to melanin, though only leuconoradrenochrome and 3 may give rise to 8/9-forming units.

Manini et al.

By contrast, tautomerization routes leading to oxidative chain breakdown would provide only a minor contribution, if any, to melanin formation since most of the resulting products are water soluble, and some, such as 5 and 6, are less oxidizable than 3 and would therefore be less proclive to take part in pigment build up. This has been demonstrated by showing that the oxidation of 6 does not result in any dark melanin-like pigment. The data reported in this study revealed unexpected changes in the kinetic course and product distribution with substrate concentration: at higher levels of 1, i.e., under conditions favoring dimerization and coupling, the o-quinone is mainly directed toward pigment-forming routes by cyclization to 2, whereas at lower concentrations, the quinone is shifted to chainbreaking pathways by way of 4. This is clearly apparent from the formation ratio of 2/4, which on passing from 100 µM to 50 µM 1 decreased from 0.50 to 0.08 at 90 min of reaction time. To fulfil the primary aim of the work, that is, to elucidate the main oxidation pathways of 1 on exposure to H2O2-based oxidizing systems, relatively high concentrations of 1, i.e., 100 µM, were required to allow for product isolation. Average plasma levels of 1 are in the nanomolar range (31), but higher concentrations were measured in some arterial compartments (32) and in pathological states such as vitiligo in the active phase (10) whereby operation of the same mechanisms can confidently be predicted at the catecholamine and H2O2 levels found in pathophysiological settings. The competing oxidative pathways described in this study may have a bearing on the mechanisms of toxicity of 1 in settings of oxidative stress. Putting aside the potential cytotoxicity of 1-quinone itself, it may be relevant to note that benzaldehyde 5 generated by the oxidation of 4 along the chainbreaking channel of 1-quinone has been reported to elicit a toxic response apparently via the inhibition of DNA synthesis (33–35). The present data may also have a bearing on the established significance of neuromelanin as a metal chelator in connection with the possible toxic actions of Fenton-type chemistry in pigmented neuron degeneration and loss in Parkinson disease (36–38). Although potentially protective because of its metal sequestering properties, neuromelanin may have also the potential for exacerbating oxidative stress, for example, by generating H2O2 when it is intact or by releasing redox-active metal ions (especially iron) if it loses its integrity (39). Evidence for iron accumulation by chelation to melanin pigments in LC has been obtained by EPR studies (40); however, comparative analysis suggests lower iron mobilization and toxicity in LC neurons than in dopaminergic pigmented SN neurons (41). It is tempting to suggest that this difference reflects the larger proportion of uncyclized catecholamine units or carboxylated catecholic moieties in 1-melanin relative to dopamine–melanin, which may account for the more efficient metal-binding capacity. The reported higher affinity of 1 for iron (II) compared to that for dopamine is also noted in this connection (42). In conclusion, the results of this model study may be used to construct an improved detailed description of the oxidation pathways of 1 and their relevance to melanin formation. Emerging chemistry highlights two main branching points at the o-quinone and leuconoradrenochrome level, and factors operating on these two points may dictate the final outcome of the oxidation of 1, including the nature of the melanin polymer. Verification of the metabolic transformations of 1 predicted by this chemistry may disclose new mechanisms of catecholamine toxicity and neuromelanin formation in oxidative stress-induced disease states.

H2O2-Promoted Oxidation of Norepinephrine

Acknowledgment. This work was carried out in the frame of the research programs IFO n. 121 (Italian Ministry of Health), IFO n. 125 (Italian Ministry of Health), and the PRIN 2006 project “Oxidative chemistry of 5,6-dihydroxyindoles and their oligomers: mechanistic and computational studies of eumelanin build up” financed by Italian Ministry of University and Research (MIUR). We thank the ‘Centro Interdipartimentale di Metodologie ChimicoFisiche′ (CIMCF, University of Naples Federico II) for NMR facilities, and Mrs. Silvana Corsani for technical assistance.

Chem. Res. Toxicol., Vol. 20, No. 10, 2007 1555

(19) (20) (21) (22)

References (1) Kalinin, S., Gavrilyuk, V., Polak, P. E., Heneka, M. T., and Feinstein, D. L. (2007) Noradrenaline deficiency in brain increases beta-amyloid plaque burden in an animal model of Alzheimer’s disease. Neurobiol. Aging 28, 1206–1214. (2) Srinivasan, J., and Schmidt, W. J. (2003) Potentiation of parkinsonian symptoms by depletion of locus coeruleus noradrenaline in 6-hydroxydopamine-induced partial degeneration of substantia nigra in rats. Eur. J. Neurosci. 17, 2586–2592. (3) Gesi, M., Soldani, P., Giorgi, F. S., Santinami, A., Bonaccorsi, I., and Fornai, F. (2000) The role of the locus coeruleus in the development of Parkinson’s disease. Neurosci. BiobehaV. ReV. 24, 655–668. (4) Zucca, F. A., Bellei, C., Giannelli, S., Terreni, M. R., Gallorini, M., Rizzio, E., Pezzoli, G., Albertini, A., and Zecca, L. (2006) Neuromelanin and iron in human locus coeruleus and substantia nigra during aging: consequences for neuronal vulnerability. J. Neural Transm. 113, 757–767. (5) Szot, P., White, S. S., Greenup, J. L., Leverenz, J. B., Peskind, E. R., and Raskind, M. A. (2006) Compensatory changes in the noradrenergic nervous system in the locus ceruleus and hippocampus of postmortem subjects with Alzheimer’s disease and dementia with Lewy bodies. J. Neurosci. 26, 467–478. (6) Shen, X-M., and Dryhurst, G. (1996) Oxidation chemistry of (–)norepinephrine in the presence of L-cysteine. J. Med. Chem. 39, 2018–2029. (7) Xin, W., Shen, X-M., Li, H., and Dryhurst, G. (2000) Oxidative metabolites of 5-S-cysteinylnorepinephrine are irreversible inhibitors of mitochondrial complex I and R-ketoglutarate dehydrogenase and pyruvate dehydrogenase complexes: possible implications for neurodegenerative brain disorders. Chem. Res. Toxicol. 13, 749–760. (8) Dell’Anna, M. L., and Picardo, M. (2006) A review and a new hypothesis for non-immunological pathogenetic mechanisms in vitiligo. Pigm. Cell Res. 19, 406–411. (9) Schallreuter, K. U., Moore, J., Wood, J. M., Beazley, W. D., Gaze, D. C., Tobin, D. J., Marshall, H. S., Panske, A., Panzig, E., and Hibberts, N. A. (1999) In vivo and in vitro evidence for hydrogen peroxide (H2O2) accumulation in the epidermis of patients with vitiligo and its successful removal by a UVB-activated pseudocatalase. J. InVest. Dermatol. 4, 91–96. (10) Cucchi, M. L., Frattini, P., Santagostino, G., and Orecchia, G. (2000) Higher plasma catecholamine and metabolite levels in the early phase of non-segmental vitiligo. Pigm. Cell Res. 13, 28–32. (11) Cucchi, M. L., Frattini, P., Santagostino, G., Preda, S., and Orecchia, G. (2003) Catecholamines increase in the urine of non-segmental vitiligo especially during its active phase. Pigm. Cell Res. 16, 111– 116. (12) Solano, F., Hearing, V. J., and Garcia-Borron, J. C. (2000) Neurotoxicity due to o-quinones: neuromelanin formation and possible mechanisms for o-quinone detoxification. Neurotox. Res. 1, 153–169. (13) Terland, O., Almas, B., Flatmark, T., Andersson, K. K., and Sorlie, M. (2006) One-electron oxidation of catecholamines generates free radicals with an in vitro toxicity correlating with their lifetime. Free Radical Biol. Med. 41, 1266–1271. (14) Jimenez, M., Garcia-Canovas, F., Garcia-Carmona, F., Lozano, J.A., and Iborra, J.L. (1984) Kinetic study and intermediates identification of noradrenaline oxidation by tyrosinase. Biochem. Pharmacol. 33, 3689–3697. (15) Bouchilloux, S., and Kodja, A. (1958) The characterization and isolation by chromatography and paper electrophoresis of substances observed in melanogenesis in vitro. Compt. Rend. 247, 2484–2486. (16) Giulivi, C., and Cadenas, E. (1998) Oxidation of adrenaline by ferrylmyoglobin. Free Radical Biol. Med. 25, 175–183. (17) Manini, P., Pezzella, A., Panzella, L., Napolitano, A., and d’Ischia, M. (2005) New insight into the oxidative chemistry of noradrenaline: competitive o-quinone cyclisation and chain fission routes leading to an unusual 4-[bis-(1H-5,6-dihydroxyindol-2-yl)methyl]-1,2-dihydroxybenzene derivative. Tetrahedron 61, 4075–4080. (18) Remião, F., Milhazes, N., Borges, F., Carvalho, F., Bastos, M., LemosAmado, F., Domingues, P., and Ferrer-Correia, A. (2003) Synthesis

(23)

(24) (25)

(26) (27) (28) (29)

(30) (31) (32) (33) (34)

(35)

(36) (37)

(38) (39)

(40)

(41) (42)

and analysis of aminochromes by HPLC-photodiode array. Adrenochrome evaluation in rat blood. Biomed. Chromatogr. 17, 6–13. Benigni, J. D., and Minnis, R. L. (1965) The syntheses of 5,6dihydroxyindole and some of its derivatives. J. Heterocycl. Chem. 2, 387–392. Napolitano, A., Pezzella, A., Vincensi, M. R., and Prota, G. (1995) Oxidative degradation of melanins to pyrrole acids: a model study. Tetrahedron 51, 5913–5920. Sugumaran, M., Dalit, H., and Semensi, V. (1992) Mechanistic studies on tyrosinase-catalysed oxidative decarboxylation of 3,4-dihydroxymandelic acid. Biochem. J. 281, 353–357. An, L. J., Guan, S., Shi, G. F., Bao, Y. M., Duan, Y. L., and Jiang, B. (2006) Protocatechuic acid from Alpinia oxyphylla against MPP+induced neurotoxicity in PC12 cells. Food Chem. Toxicol. 44, 436– 443. Napolitano, A., Pezzella, A., d’Ischia, M., and Prota, G. (1996) New pyrrole acids by oxidative degradation of eumelanins with hydrogen peroxide. Further hints to the mechanism of pigment breakdown. Tetrahedron 52, 8775–8780. Ito, S., and Fujita, K. (1985) Microanalysis of eumelanin and pheomelanin in hair and melanomas by chemical degradation and liquid chromatography. Anal. Biochem. 144, 527–536. Martinez Ortiz, F., Tutela Serrano, J., Rodriguez Lopez, J. N., Varon Castellanos, R., Lozano Teruel, J. A., and Garcia-Canovas, F. (1988) Oxidation of 3,4-dihydroxymandelic acid catalyzed by tyrosinase. Biochim. Biophys. Acta 957, 158–163. Sugumaran, M. (1986) Tyrosinase catalyzes an unusual oxidative decarboxylation of 3,4-dihydroxymandelate. Biochemistry 25, 4489– 4492. Ito, S., and Wakamatsu, K. (1998) Chemical degradation of melanins: application to identification of dopamine-melanin. Pigm. Cell Res. 11, 120–126. Chavdarian, C. G., Karashima, D., and Castagnoli, N., Jr. (1978) Oxidative and cardiovascular studies on natural and synthetic catecholamines. J. Med. Chem. 21, 548–554. Tribl, F., Marcus, K., Meyer, H. E., Bringmann, G., Gerlach, M., and Riederer, P. (2006) Subcellular proteomics reveals neuromelanin granules to be a lysosome-related organelle. J. Neural Transm. 113, 741–749. Pilawa, B., Chodurek, E., and Wilczok, T. (2003) Types of paramagnetic centers in Cu2+ complexes with model neuromelanins. Appl. Magn. Res. 24, 417–422. Hasking, G. J., and Esler, M. D. (1990) Plasma norepinephrine and treatment of heart failure. Am. J. Cardiol. 66, 386–387. Esler, M., and Kaye, D. (2000) Measurement of sympathetic nervous system activity in heart failure: the role of norepinephrine kinetics. Heart Failure ReV. 5, 17–25. Tseng, T. H., Tsheng, Y. M., and Lee, Y. J. (2001) Cytotoxicity effects of di- and tri-hydroxybenzaldehydes as a chemopreventive potential agent on tumor cells. Toxicology 161, 179–187. Nishimura, M., Umeda, M., Kusanagi, A., Urakabe, R., Yukimatsu, K., and Otawara, S. (1981) Antiproliferative activity of protocatechualdehyde on Chinese hamster cells grown in culture. Gann. 72, 868– 879. 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. Youdim, M. B. H., Double, K. L., Gerlach, M., and Riederer, P. (2006) Neuromelanin. J. Neural Transm. 113, 63. Shamoto-Nagai, M., Maruyama, W., Yi, H., Akao, Y., Tribl, F., Gerlach, M., Osawa, T., Riederer, P., and Naoi, M. (2006) Neuromelanin induces oxidative stress in mitochondria through release of iron: mechanism behind the inhibition of 26S proteasome. J. Neural Transm. 113, 633–644. Goetz, M. E., Double, K., Gerlach, M., Youdim, M. B. H., and Riederer, P. (2004) The relevance of iron in the pathogenesis of Parkinson’s disease. Ann. N.Y. Acad. Sci. 1012, 193–208. Enochs, W. S., Sarna, T., Zecca, L., Riley, P. A., and Swartz, H. M. (1994) The roles of neuromelanin, binding of metal ions, and oxidative cytotoxicity in the pathogenesis of Parkinson’s disease: a hypothesis. J. Neural Transm. 7, 83–100. Zecca, L., Shima, T., Stroppolo, A., Goj, C., Battiston, G. A., Gerbasi, R., Sarna, T., and Swartz, H. M. (1996) Interaction of neuromelanin and iron in substantia nigra and other areas of human brain. Neurosci. 73, 407–415. Fasano, M., Bergamasco, B., and Lopiano, L. (2006) Modifications of the iron-neuromelanin system in Parkinson’s disease. J. Neurochem. 96, 909–916. Rajan, K. S., and Davis, J. M. (1976) Interactions of iron with biogenic amines and ATP. J. Inorg. Nucl. Chem. 38, 897–905.

TX700254Q