Synthesis, Redox Properties, in Vivo Formation, and Neurobehavioral

pigmented dopaminergic cell bodies in the substantia nigra (SN) in which cytoplasmic dopamine ... These include a large decrease of nigral glutathione...
0 downloads 0 Views 568KB Size
Download PDF
Chem. Res. Toxicol. 1996, 9, 1117-1126

1117

Synthesis, Redox Properties, in Vivo Formation, and Neurobehavioral Effects of N-Acetylcysteinyl Conjugates of Dopamine: Possible Metabolites of Relevance to Parkinson’s Disease Xue-Ming Shen, Bing Xia, Monika Z. Wrona, and Glenn Dryhurst* Department of Chemistry and Biochemistry, University of Oklahoma, Norman, Oklahoma 73019 Received March 25, 1996X

A very early event in the pathogenesis of idiopathic Parkinson’s disease (PD) has been proposed to be an elevated translocation of L-cysteine (CySH) and/or glutathione (GSH) into pigmented dopaminergic cell bodies in the substantia nigra (SN) in which cytoplasmic dopamine (DA) is normally autoxidized to DA-o-quinone as the first step in a reaction leading to black neuromelanin polymer. Such an elevated influx of CySH and GSH would be expected to initially result in formation of 5-S-cysteinyldopamine (5-S-CyS-DA) and 5-S-glutathionyldopamine (5S-Glu-DA), respectively, and might account for the massive irreversible loss of GSH and progressive depigmentation of SN cells that occurs in the Parkinsonian brain. However, 5-SGlu-DA has not been detected in the Parkinsonian brain. Furthermore, although the 5-SCyS-DA/DA and 5-S-CyS-DA/homovanillic acid concentration ratios increase significantly in the SN and cerebrospinal fluid, respectively, of PD patients, the absolute concentrations of 5-S-CyS-DA are extremely low and similar to those measured in age-matched control patients. One explanation for these observations is that 5-S-CyS-DA might be intraneuronally oxidized to more complex cysteinyldopamines and a number of dihydrobenzothiazines (DHBTs) and benzothiazines (BTs). Similarly, 5-S-Glu-DA might be intraneuronally oxidized to more complex glutathionyldopamines. In this investigation, however, it is demonstrated that 5-S-Glu-DA is rapidly metabolized in rat brain to 5-S-CyS-DA and 5-S-(N-acetylcysteinyl)dopamine (5) in reactions mediated by γ-glutamyl transpeptidase (γ-GT) and cysteine conjugate N-acetyltransferase. Similarly, 5-S-CyS-DA is metabolized to 5 in rat brain although more slowly than 5-S-Glu-DA. These reactions occur most rapidly in the midbrain, a region that contains the SN. Furthermore, 5, 2-S-(N-acetylcysteinyl)dopamine (6) and 2,5-di-S-(N-acetylcysteinyl)dopamine (9) are toxic when administered into mouse brain having LD50 values of 14, 25, and 42 µg, respectively, and evoke a profound hyperactivity syndrome. These results suggest that the failure to detect 5-S-Glu-DA and the presence of only very low levels of 5-S-CyS-DA in Parkinsonian SN tissue and CSF might be related to both their intraneuronal oxidation and extraneuronal metabolism to N-acetylcysteinyl conjugates of DA. Furthermore, the toxic properties and neurobehavioral responses evoked by 5, 6, and 9 raise the possibility that these N-acetylcysteinyl conjugates of DA, in addition to certain cysteinyldopamines, DHBTs and BTs, might include endotoxins that contribute to SN cell death and other neuronal damage that occurs in PD. Methods are described for the synthesis of several N-acetylcysteinyl conjugates of DA, and their redox behaviors have been studied using cyclic voltammetry.

Introduction The biochemical basis of Parkinson’s disease (PD)1 is a severe deficiency of dopamine (DA) in the striatum resulting from a rather selective and progressive degeneration of nigrostriatal dopaminergic neurons (1). The degeneration of these neurons results from pathological processes that occur in their neuromelanin-pigmented cell bodies in the substantia nigra (SN) pars compacta (2) which include oxidative stress (3) and inhibition of mitochondrial respiration at the complex I stage (4). The chemical, biochemical, and genetic mechanisms that underlie these pathological processes are unknown. Nevertheless, several very characteristic chemical and biochemical changes occur in the Parkinsonian SN. These include a large decrease of nigral glutathione (GSH) that is not accompanied by a corresponding * Address correspondence to this author. Tel: (405) 325-4811; Fax: (405) 325-6111; E-mail: [email protected]. X Abstract published in Advance ACS Abstracts, September 1, 1996.

S0893-228x(96)00052-5 CCC: $12.00

increase of glutathione disulfide (GSSG) (5, 6), a significant increase in the 5-S-cysteinyldopamine (5-S-CyS-DA)/ 1 Abbreviations: Parkinson’s disease [PD]; dopamine [DA]; substantia nigra [SN]; glutathione [GSH]; glutathione disulfide [GSSG]; 5-S-cysteinyldopamine [5-S-CyS-DA]; γ-glutamyl transpeptidase [γ-GT]; L-cysteine [CySH]; 5-S-glutathionyldopamine [5-S-Glu-DA]; dihydrobenzothiazine [DHBT]; 5-S-glutathionyl-R-methyldopamine [5-SGlu-RMe-DA]; 3,4-(methylenedioxy)amphetamine [MDA]; 3,4-(methylenedioxy)methamphetamine [MDMA]; 5-S-cysteinyl-R-methyldopamine [5-S-CyS-RMe-DA]; 5-S-(N-acetylcysteinyl)-R-methyldopamine [5-SNAc-CyS-RMe-DA]; intracerebroventricular [icv]; N-acetyl-L-cysteine [NAc-CySH]; 2,5-di-S-glutathionyldopamine [2,5-di-S-Glu-DA]; 2,5,6tri-S-glutathionyldopamine [2,5,6-tri-S-Glu-DA]; high performance liquid chromatography with electrochemical detection [HPLC-EC]; pyrolytic graphite electrode [PGE]; saturated calomel reference electrode [SCE]; fast atom bombardment mass spectrometry [FAB-MS]; trifluoroacetic acid [TFA]; acetonitrile [MeCN]; 5-S-(N-acetylcysteinyl)dopamine [5]; 2-S-(N-acetylcysteinyl)dopamine [6]; 2,5-di-S-(N-acetylcysteinyl)dopamine [9]; 5,6-di-S-(N-acetylcysteinyl)dopamine [10]; 2,5,6tri-S-(N-acetylcysteinyl)dopamine [13]; S,S′-[2-[[2-(acetylamino)ethenyl]thio]-3-(2-aminoethyl)-5,6-dihydroxy-1,4-phenylene]bis[N-acetyl-L-cysteine] [16]; sweep rate [ν]; 5,6-dihydroxyindoline [2]; dopaminochrome [3]; 5,6-dihydroxyindole [5,6-DHT]; peak potential [Ep]; peak current [ip]; formal potential [E°′]; benzothiazine [BT]; cerebrospinal fluid [CSF].

© 1996 American Chemical Society

1118 Chem. Res. Toxicol., Vol. 9, No. 7, 1996

dopamine (DA) concentration ratio (7), upregulation of γ-glutamyl transpeptidase (γ-GT) (8), and progressive depigmentation of dopaminergic SN cells (7, 9-12). The

neuromelanin that pigments SN neurons is formed as a consequence of autoxidation of cytoplasmic DA to DA-oquinone that subsequently cyclizes and oxidatively polymerizes (13). In vitro studies, however, have demonstrated that the oxidation of DA to insoluble black melanin polymer is diverted or completely blocked by both GSH (14) and L-cysteine (CySH) (15, 16). This is so because GSH and CySH scavenge DA-o-quinone to give a number of soluble glutathionyl or cysteinyl conjugates of DA, respectively (15, 16). The 5-S-CyS-DA detected in the Parkinsonian SN has been proposed to be formed either by nucleophilic addition of CySH to DAo-quinone or, in view of the much larger concentrations of GSH in the brain, by addition of GSH to give 5-Sglutathionyldopamine (5-S-Glu-DA) that is subsequently hydrolyzed by peptidase enzymes (7). However, that the latter reaction occurs in vivo has not been established experimentally. Histochemical evidence indicates that neuronal cell bodies throughout the brain normally contain little GSH or CySH. Rather, GSH and CySH are primarily located in glial cells, nerve terminals, and axons (17). Based on these and related lines of evidence, a new hypothesis has recently been advanced that might contribute to an understanding of the pathogenesis of PD (15, 18). This hypothesis proposes that chronic exposure of individuals genetically equipped with certain defective hepatic detoxication enzymes (19, 20) to environmental toxicants may permit some of these substances to enter the brain and evoke elevated biosynthesis of CySH and GSH (21) in glial cells (22, 23) and upregulation of γ-GT (24). In the SN the latter two responses might potentiate elevated translocation of CySH and/or GSH (25, 26) into pigmented dopaminergic neurons with resultant diversion of the neuromelanin pathway, irreversible loss of GSH, and an increased 5-S-CyS-DA/DA ratio, all of which occur in the Parkinsonian brain. Furthermore, the CySH/GSH-mediated diversion of the neuromelanin pathway would account for the significantly decreased pigmentation of surviving pigmented cells in the SN of PD patients (7, 9-12). However, 5-S-CyS-DA and 5-S-GluDA, the initial major in vitro products of the reaction between DA-o-quinone and CySH or GSH, respectively, are more easily oxidized than DA in reactions that lead to more complex cysteinyl or glutathionyl conjugates of the neurotransmitter (14, 15, 18). Furthermore, cysteinyl conjugates of DA are also readily further oxidized to a number of dihydrobenzothiazine (DHBT) compounds (15, 18). At least one cysteinyl conjugate of DA and several DHBTs are toxic when administered into the brains of mice and rats (15, 18). Accordingly, it has been speculated that such putative metabolites might include the endotoxins that contribute to the degeneration of dopaminergic SN neurons in PD (15, 18).

Shen et al.

Recently, it has been reported that 5-S-glutathionylR-methyldopamine (5-S-Glu-RMe-DA), a metabolite of the serotonergic neurotoxins 3,4-(methylenedioxy)amphetamine (MDA) and 3,4-(methylenedioxy)methamphetamine (MDMA) (27, 28), is rapidly transformed into 5-Scysteinyl-R-methyldopamine (5-S-Cys-RMe-DA) in rat brain in a reaction mediated by γ-GT (29). The latter cysteinyl conjugate is then transformed into 5-S-(Nacetylcysteinyl)-R-methyldopamine (5-S-NAc-CyS-RMeDA) by cysteine conjugate N-acetyltransferase. However, 5-S-NAc-CyS-RMe-DA is quite slowly cleared from the brain (29). A very large intracerebroventricular (icv) injection of 5-S-Glu-RMe-DA (720 nmol; 340 µg) to the rat evoked behavioral alterations that are similar to those caused by MDA and MDMA, and larger doses were lethal. Much lower icv doses of 5-S-NAc-CyS-RMe-DA (7 nmol; 2.3 µg) evoked similar neurobehavioral responses (29). These observations suggest that 5-S-Glu-RMe-DA and particularly 5-S-NAc-CyS-RMe-DA might be metabolites of MDA and MDMA that contribute to the neurotoxicity of these drugs, although this remains to be experimentally confirmed. Furthermore, such observations also suggest that γ-GT might also mediate the hydrolysis of glutathionyl conjugates of DA, formed in the Parkinsonian SN, to the corresponding cysteinyl conjugates, and that the latter conjugates might not only be further oxidized to toxic DHBTs but also transformed into N-acetylcysteinyl conjugates. The N-acetylcysteinyl conjugates of DA, in addition to DHBTs, therefore, might together represent a family of endotoxins that contribute to SN cell death and other secondary damage, to serotonergic pathways, for example, which occurs in PD. In this article, therefore, methods are described for the synthesis of several N-acetylcysteinyl conjugates of DA. The redox properties of these conjugates have also been elucidated. Evidence is also presented to demonstrate that 5-S-Glu-DA is rapidly metabolized in rat brain to 5-S-CyS-DA and 5-S-(N-acetylcysteinyl)dopamine. Additionally, several N-acetylcysteinyl conjugates of DA are toxic when administered into mouse brain and evoke a profound neurobehavioral response.

Materials and Methods Caution. The following chemicals were toxic (lethal) when administered into the brains of laboratory mice and hence should be regarded as potentially hazardous in humans: 5-S-(Nacetylcysteinyl)dopamine (5), 2-S-(N-acetylcysteinyl)dopamine (6), and 2,5-di-S-(N-acetylcysteinyl)dopamine (9). Accordingly, these compounds should always be handled with protective gloves and in a hood. Chemicals. Dopamine hydrochloride (DA‚HCl), L-cysteine (CySH), and glutathione (GSH) were obtained from Sigma (St. Louis, MO). N-Acetyl-L-cysteine (NAc-CySH) was obtained from Aldrich (Milwaukee, WI). All other chemicals were of the highest commercially available grade. 5-S-Cysteinyldopamine (15, 18), 5-S-glutathionyldopamine (5-S-Glu-DA), 2,5-di-S-glutathionyldopamine (2,5-di-S-Glu-DA), and 2,5,6-tri-S-glutathionyldopamine (2,5,6-tri-S-Glu-DA) (14) were synthesized as described previously. Electrochemistry. Voltammograms were obtained at a pyrolytic graphite electrode (PGE; Pfizer Minerals, Pigments and Metals Division, Easton, PA) having an approximate surface area of 4 mm2. A conventional three-electrode cell was employed for voltammetry and contained a platinum wire counter electrode and a saturated calomel reference electrode (SCE). Voltammograms were recorded using a BAS-100A electrochemical analyzer (Bioanalytical Systems, West Lafeyette, IN) and were all fully corrected for ohmic losses. The PGE was always

N-Acetylcysteinyl Conjugates of Dopamine resurfaced prior to recording each voltammogram (15). Controlled potential electrolysis employed a Princeton Applied Research Corp. (Princeton, NJ) Model 173 potentiostat. A threecompartment cell was used with the working, counter, and reference electrode compartments separated by a Nafion membrane (type 117, DuPont Co., Wilmington, DE). The working electrode compartment had a capacity of 30 mL. The working electrode consisted of several plates of pyrolytic graphite having a total surface area of ca. 180 cm2. The counter electrode was platinum gauze and the reference electrode an SCE. The solution in the working electrode compartment was vigorously bubbled with N2 and stirred with a Teflon-coated magnetic stirring bar. All potentials are referenced to the SCE at ambient temperature (22 ( 2 °C). Spectroscopy. 1H NMR spectra were recorded on a Varian (Palo Alto, CA) XL-300 spectrometer. Fast atom bombardment mass spectrometry (FAB-MS) employed a VG Instruments (Manchester, U.K.) ZAB-E spectrometer. UV-visible spectra were recorded on a Hewlett-Packard (Palo Alto, CA) 8452A diode array spectrophotometer. High Performance Liquid Chromatography (HPLC). HPLC method I employed a Gilson (Middleton, WI) gradient system equipped with dual Model 302 pumps (10 mL pump heads), a Rheodyne (Cotati, CA) Model 7125 equipped with a 10.0 mL sample loop, and a Waters (Milford, MA) Model 440 UV detector set at 254 nm. Two mobile phase solvents were employed. Solvent A was prepared by adding concentrated trifluoroacetic acid (TFA) to deionized water until the pH was 2.15. Solvent B was prepared by adding TFA to a mixture of 2 L of HPLC grade acetonitrile (MeCN) and 2 L of deionized water until the measured pH was 2.15. HPLC method I employed a reversed phase column (Bakerbond C18, 10 µm, 250- × 21.2mm, P.J. Cobert Associates, St. Louis, MO) and the following mobile phase gradient: 0-80 min, linear gradient from 100% solvent A to 25% solvent B; 80-84 min, linear gradient to 100% solvent B; 84-96 min, 100% solvent B. The flow rate was constant at 7.0 mL min-1. HPLC with electrochemical detection (HPLC-EC) employed a Bio-Rad (Hercules, CA) Model 1300 pump, a Rheodyne 7125 injector with a 5.0 µL sample loop, and a BAS LC-4C electrochemical detector equipped with a glassy carbon detector electrode set at 800 mV vs an Ag/AgCl reference electrode. The detector was set at a full scale sensitivity of 10 nA. A reversed phase column (BAS, Phase-II ODS, 3 µm, 100- × 3.2-mm) was employed. The mobile phase consisted of 0.085% diethylamine (v/v), 0.05 M ethylenediaminetetraacetic acid, 0.47 mM soidium octyl sulfate, and 0.1 M citric acid in 7.5% MeCN/deionized water at pH 2.37. The flow rate was constant at 0.6 mL min-1. Synthesis of 2-S-(N-Acetylcysteinyl)dopamine (6) and 5-S-(N-Acetylcysteinyl)dopamine (5). DA‚HCl (5.7 mg; 1.0 mM) was dissolved in 30 mL of 0.1 M hydrochloric acid (HCl) solution and electrolyzed at 1.0 V for 30 min. The reaction solution turned from colorless to bright yellow characteristic of DA-o-quinone (30). HPLC (method I) analysis revealed that DA was converted (>90%) to DA-o-quinone. Addition of NAc-CySH (24.5 mg; 5.0 mM) to DA-o-quinone caused the solution to rapidly change from bright yellow to colorless. The entire reaction solution was pumped directly onto the preparative reversed phase column, and products were separated using HPLC method I. 2- (6), 5- (5), and 2,5-di-S-(N-acetylcysteinyl)DA (9) eluted at retention times (tR) of 38, 54, and 59 min, respectively, and were formed in the ratio 5:60:1. The same procedure was repeated several times, and the eluents containing 5 and 6 were combined individually and freeze-dried. The resulting solids were dissolved in 2-10 mL of deionized water adjusted to pH 2.15 with TFA and were purified using HPLC method I. The purified eluents were collected and freeze-dried. Oxidation Reaction Procedure, Isolation, and Purification of Products. DA‚HCl (5.7 mg; 1.0 mM) and NAc-CySH (24.5 mg; 5.0 mM) were dissolved in 30 mL of phosphate buffer (pH 7.4; ionic strength, µ ) 0.2) and electro-oxidized for, typically, 30 min at 50 mV. Upon termination of the reaction, the entire solution was pumped onto the reversed phase column,

Chem. Res. Toxicol., Vol. 9, No. 7, 1996 1119 and products and reactants were separated using HPLC method I. The solution eluted under the chromatographic peak corresponding to each product was collected and immediately frozen at -80 °C (dry ice bath). Following several repetitive experiments, the combined solutions containing each product were purified using HPLC method I. The resulting eluents were freeze-dried to give solid, dry products. Spectroscopic evidence in support of the proposed structures of each product is presented below. Assignments of proton resonances observed in the 1H NMR spectra of products were based on comparisons with the spectra of DA and NAc-CySH and were confirmed in all cases by two-dimensional correlated spectroscopy (COSY) experiments. For simplicity, the atom numbering system shown in the structures presented in Scheme 1 was employed. 5-S-(N-Acetylcysteinyl)dopamine (5). Compound 5 was isolated as a white solid. In the HPLC (method I) mobile phase (pH 2.15) 5 exhibited UV bands at λmax ) 292 and 256 nm. In pH 7.4 phosphate buffer, the UV spectrum showed λmax, nm (log max, M-1 cm-1), at 294 (3.43) and 254 (3.61). FAB-MS (3nitrobenzyl alcohol matrix) gave m/z ) 315.1013 (MH+, 100, C13H19N2O5S; calcd m/z ) 315.1015). Elemental analysis: C 41.15, H 4.77, N 5.76, S 7.18, F 14.40, corresponding to 5‚TFA‚H2O (C15H21N2O8S1F3); theory: C 40.36, H 4.71, N 6.27, S 7.17, F 12.87. 1H NMR (D2O): δ 6.82 (d, J2,6 ) 2.4 Hz, 1H, C(2)-H), 6.73 (d, J2,6 ) 2.4 Hz, 1H, C(6)-H), 4.34 (dd, J ) 8.4, 4.2 Hz, 1H, C(b)-H), 3.34 (dd, J ) 14.4, 4.2 Hz, 1H, C(a)-H), 3.15 (t, J ) 7.2 Hz, 2H, C(β)-H2), 3.13 (dd, J ) 14.4, 8.4 Hz, 1H, C(a)-H), 2.77 (t, J ) 7.2 Hz, 2H, C(R)-H2), 1.86 (s, 3H, CH3). 2-S-(N-Acetylcysteinyl)dopamine (6). Compound 6 was isolated as a whilte solid. In the HPLC (method I) mobile phase (pH 2.15) 6 exhibited UV bands at λmax ) 294 and 258 nm. In pH 7.4 phosphate buffer, the UV spectrum showed λmax, nm (log max, M-1, cm-1), at 294 (3.47) and 258 (3.48), calculated at 1.5TFA salt. FAB-MS (3-nitrobenzyl alcohol matrix) gave m/z ) 315.1014 (MH+, 100, C13H19N2O5S; calcd m/z ) 315.1015). 1H NMR (D O): δ 6.87 (d, J 2 5,6 ) 8.4 Hz, 1H, C(5)-H), 6.75 (d, J5,6 ) 8.4 Hz, 1H, C(6)-H), 4.33 (dd, J ) 8.4, 3.9 Hz, 1H, C(b)H), 3.35 (dd, J ) 14.4, 3.9 Hz, 1H, C(a)-H), 3.13 (dd, J ) 14.4, 8.4 Hz, 1H, C(a)-H), 3.11-3.05 (m, 4H, C(R)-H2, C(β)-H2), 1.84 (s, 3H, CH3). 2,5-Di-S-(N-acetylcysteinyl)dopamine (9). Compound 9 was isolated as a white solid. In the HPLC mobile phase (pH 2.15) 9 exhibited bands at λmax ) 302 (sh) and 276 nm. In pH 7.4 phosphate buffer, the UV spectrum showed λmax, nm (log max, M-1 cm-1), at 304 (sh, 3.51) and 276 (3.99). FAB-MS (3nitrobenzyl alcohol matrix) gave m/z ) 476.1177 (MH+, 100, C18H26N3O8S2; calcd m/z ) 476.1161). Elemental analysis: C 38.15, H 4.47, N 5.82, S 8.74, corresponding to 9‚2TFA (C22H27N3O12S2F6); theory: C 37.55, H 3.84, N 5.97, S 9.10. 1H NMR (D2O): δ 6.92 (s, 1H, C(6)-H), 4.39 (dd, J ) 8.4, 4.2 Hz, 1H, C(b)-H), 4.36 (dd, J ) 8.4, 4.2 Hz, 1H, C(b′)-H), 3.38 (dd, J ) 14.4, 4.2 Hz, 1H, C(a)-H), 3.32 (dd, J ) 14.4, 4.2 Hz, 1H, C(a′)H), 3.17 (dd, J ) 14.4, 8.4 Hz, 1H, C(a)-H), 3.13 (dd, J ) 14.4, 8.4 Hz, 1H, C(a′)-H), 3.12-3.02 (m, 4H, C(R)-H2, C(β)-H2), 1.87 (s, 3H, CH3), 1.84 (s, 3H, CH3). 5,6-Di-S-(N-acetylcysteinyl)dopamine (10). Compound 10, a very minor product, was isolated as a white solid that in the HPLC mobile phase (pH 2.15) exhibited a UV band at λmax ) 308 nm. FAB-MS (3-nitrobenzyl alcohol matrix) gave m/z ) 476.1170 (MH+, 92, C18N26N3O8S2; calcd m/z ) 476.1161). 1H NMR (D2O): δ 6.89 (s, 1H, C(2)-H), 4.34 (dd, J ) 8.4, 3.9 Hz, 1H, C(b)-H), 4.24 (dd, J ) 8.4, 4.2 Hz, 1H, C(b′)-H), 3.46 (dd, J ) 14.1, 4.2 Hz, 1H, C(a′)-H), 3.39 (dd, J ) 14.1, 3.9 Hz, 1H, C(a)-H), 3.27 (dd, J ) 14.1, 8.4 Hz, 1H, C(a′)-H), 3.15 (dd, J ) 14.1, 8.4 Hz, 1H, C(a)-H), 3.14-3.11 (m, 4H, C(R)-H2, C(β)-H2), 1.93 (s, 3H, CH3), 1.85 (s, 3H, CH3). Long-range 2D COSY experiments revealed coupling between the resonances at δ 6.89 (C(2)-H) and 3.12 (C(R)-H2), in agreement with that expected for the structure of 10. 2,5,6-Tri-S-(N-acetylcysteinyl)dopamine (13). Compound 13 was isolated as a white solid that in the HPLC mobile phase (pH 2.15) exhibited UV bands at λmax ) 312 (sh), 288 (sh), and

1120 Chem. Res. Toxicol., Vol. 9, No. 7, 1996

Shen et al. Scheme 1

224 nm. In pH 7.4 phosphate buffer, the UV spectrum showed λmax, nm (log max, M-1 cm-1), at 322 (3.59), 292 (sh, 3.72), 262 (sh, 4.12), 240 (sh, 4.28), and 224 (4.32). FAB-MS (3-nitrobenzyl alcohol matrix) gave m/z ) 637.1292 (MH+, 45, C23H33N4O11S3; calcd m/z ) 637.1308). Elemental analysis: C 38.13, H 4.28, N 6.66, S 11.37, F 10.22, corresponding to 13‚1.5TFA (C26H33.5N4O14S3F4.5); theory: C 38.66, H 4.15, N 6.94, S 11.90, F 10.59. 1H NMR (D O): δ 4.38 (dd, J ) 8.4, 4.5 Hz, 1H, C(b)-H), 4.37 2 (dd, J ) 8.4, 4.2 Hz, 1H, C(b′)-H), 4.24 (dd, J ) 8.4, 4.5 Hz, 1H, C(b′′)-H), 3.57-3.48 (m, 2H, C(β)-H2), 3.44 (dd, J ) 14.4, 4.5 Hz, 1H, C(a′′)-H), 3.34 (dd, J ) 14.1, 4.2 Hz, 1H, C(a′)-H), 3.33 (dd, J ) 14.1, 4.5 Hz, 1H, C(a)-H), 3.26 (dd, J ) 14.4, 8.4 Hz, 1H, C(a′′)-H), 3.17 (dd, J ) 14.1, 8.4 Hz, 1H, C(a′)-H), 3.14 (dd, J ) 14.1, 8.4 Hz, 1H, C(a)-H), 3.03-2.96 (m, 2H, C(R)-H2), 1.91 (s, 3H, CH3), 1.89 (s, 3H, CH3), 1.87 (s, 3H, CH3). S,S′-[2-[[2-(Acetylamino)ethenyl]thio]-3-(2-aminoethyl)5,6-dihydroxy-1,4-phenylene]bis[N-acetyl-L-cysteine] (16). Compound 16, a very minor product, was isolated as a white solid that in the HPLC mobile phase (pH 2.15) exhibited UV bands at λmax ) 316 (sh), 270, 244 (sh), and 222 nm. FAB-MS (3-nitrobenzyl alcohol matrix) gave m/z ) 591.1281 (MH+, 16, C22H31N4O9S3; calcd m/z ) 591.1253). 1H NMR (D2O): δ 6.78 (d, J ) 7.8 Hz, 1H, C(b′′)-H), 5.32 (d, J ) 7.8 Hz, 1H, C(a′′)-H), 4.33 (dd, J ) 8.4, 4.2 Hz, 1H, C(b)-H), 4.32 (dd, J ) 8.4, 4.2 Hz, 1H, C(b′)-H), 3.57-3.50 (m, 2H, C(β)-H2), 3.45 (dd, J ) 14.4, 4.2 Hz, 1H, C(a′)-H), 3.41 (dd, J ) 14.4, 4.2 Hz, 1H, C(a)-H), 3.24 (dd, J ) 14.4, 8.4 Hz, 1H, C(a′)-H), 3.20 (dd, J ) 14.4, 8.4 Hz, 1H, C(a)-H), 3.09-3.03 (m, 2H, C(R)-H2), 2.12 (s, 3H, CH3), 1.86 (s, 3H, CH3), 1.83 (s, 3H, CH3). Animals. Drugs were screened for behavioral response and LD50 determinations carried out using outbred adult male mice of the Hsd: ICR albino strain (Harlan Sprague-Dawley, Madison, WI) weighing 30 ( 3 g. Experimental animals were treated with test drugs dissolved in 5 µL of isotonic saline (0.9% NaCl in deionized water). Mice were first anesthesized with ether for 45-55 s. Injections were performed freehand using a 10

µL microsyringe, the point of puncture being 3 mm anterior to the interaural lines, 1 mm left lateral of the midline, and 3 mm perpendicular to the scalp according to a procedure described in detail previously (31). Control animals were treated with 5 µL of vehicle alone under otherwise identical conditions. Intracerebroventricular (icv) Administration of 5-SGlu-DA and 5-S-CyS-DA to Rats. Sprague-Dawley rats weighing 360-380 g were anesthesized with a mixture of ketamine (50 mg/kg) and xylasine (4 mg/kg) administered ip and were then placed in a stereotaxic apparatus with the inciser bar positioned 3.3 mm below the interaural line. A small burr hole was made with a drill on the leveled skull using the coordinates: AP-0.8 mm and L +1.3 mm relative to bregma. Drugs were injected by means of a 10 µL Hamilton microsyringe equipped with a 31 gauge needle, the tip of which was lowered 3.8 mm ventral to bregma. A 5.0 µL injection was made over a period of 3 min, the needle being maintained in place for an additional 1 min prior to withdrawal. The doses of 5-S-GluDA and 5-S-CyS-DA dissolved in 5 µL of artificial cerebrospinal fluid (CSF; 147 mM NaCl, 4 mM KCl, 1.2 mM CaCl2, and 1.2 mM MgSO4 dissolved in deionized water) were 20 µg (44 nmol) and 11.8 µg (44 nmol) (expressed as free base), respectively. Rats were sacrificed by guillotine decapitation 15 min after drug administration was completed, brains removed, and dissected into striatum, cortex, hippocampus, medulla and pons, and midbrain on an ice-chilled glass plate. Each brain part was immediately wrapped in aluminum foil and stored at -80 °C until assayed. Immediately prior to HPLC-EC analysis, brain samples were homogenized in a medium consisting of 0.1 M HClO4 containing 134 mM Na2EDTA, 263 mM Na2S2O3, and 1.0 mM Nω-methyl-5-hydroxytryptamine (N-Met; internal standard) in deionized water (1 mg brain sample in 5 µL of homogenization solution) using a Polytron PT-1200 tissue homogenizer (Kinematica AG, Switzerland). The resulting homogenate was centrifuged (14000g, 60 min, 4 °C), and aliquots of the supernatant were then immediately injected into the

N-Acetylcysteinyl Conjugates of Dopamine

Chem. Res. Toxicol., Vol. 9, No. 7, 1996 1121

Figure 1. Cyclic voltammograms at the PGE of (A) 1.0 mM DA, (B) 1.0 mM DA and 0.5 mM NAc-CySH, (C) 1.0 mM DA and 1.0 mM NAc-CySH, (D) 1.0 mM DA and 2.0 mM NAc-CySH, (E) 1.0 mM DA and 5.0 mM NAc-CySH, and (F) 5.0 mM NAc-CySH in phosphate buffer (pH 7.4; µ ) 1.0). Sweep rate: 100 mV s-1. HPLC-EC system for analysis. Concentrations of 5-S-Glu-DA, 5-S-CyS-DA, and 5 were determined from standard calibration curves (ratio of the peak height of the compound of interest to that of N-Met) prepared with authentic samples of these compounds. All animal procedures were approved by the Institutional Animal Care and Use Committee at the University of Oklahoma.

Results Addition of a large molar excess of NAc-CySH to DAo-quinone at low pH provides a convenient route to synthesis of the 5-S- (5) and 2-S-(N-acetylcysteinyl) (6) conjugates of DA (see Materials and Methods). However, oxidation of DA in the presence of GSH (14) and CySH (15, 18) at pH 7.4 yields a number of additional more highly substituted glutathionyl and cysteinyl conjugates of the neurotransmitter. Accordingly, similar oxidations of DA were carried out in the presence of NAc-CySH. As in earlier studies (14, 15, 18), electrochemical methods were employed to oxidize DA to its proximate oxidation product, DA-o-quinone. Voltammetric Studies. A representative cyclic voltammogram of DA (1.0 mM) in pH 7.4 phosphate buffer at a sweep rate (ν) of 100 mV s-1 is presented in Figure 1A. Previous investigations have established that peak Ia, at a peak potential (Ep) of 130 mV, corresponds to the oxidation (2e, 2H+) of DA to DA-o-quinone (32, 33). Following scan reversal, peak Ic (Ep ) 101 mV) corresponds to the reversible reduction of DA-o-quinone to DA. Deprotonation of DA-o-quinone to 1 followed by intramolecular cyclization gives 5,6-dihydroxyindoline (2) that is immediately oxidized by DA-o-quinone to give dopaminochrome 3 (Figure 1A) (32). The latter compound is

responsible for peak IIc at which it is reduced (2e, 2H+) to 2 that, on the subsequent anodic sweep, is reversibly oxidized at peak IIa to give 3. The reactions associated with each peak observed in cyclic voltammograms of DA are presented in Figure 1A. Under more prolonged oxidation conditions dopaminochrome (3) rearranges to 5,6-dihydroxyindole (5,6-DHI) that is further oxidized to p-quinone imine 4 which subsequently polymerizes to black insoluble melanin. Addition of NAc-CySH (0.55.0 mM) to solutions of DA (1.0 mM) at pH 7.4 results in several changes to the cyclic voltammetric behavior of the neurotransmitter (Figure 1B-E). Thus, in the presence of 0.5 mM NAc-CySH (Figure 1B), oxidation peak IIIa appears and forms a shoulder on the rising portion of peak Ia of DA. At higher concentrations of NAc-CySH peak IIIa (Ep ) 95 mV) becomes the dominant peak on the initial anodic sweep and peak Ia is virtually eliminated. Peak Ic, corresponding to the reduction of DA-oquinone in cyclic voltammograms of DA, has an Ep ) 101 mV (Figure 1A). In the presence of 0.5, 1.0, and 2.0 mM NAc-CySH, cyclic voltammograms of DA (1.0 mM) exhibit reduction peak IIIc at Ep values of 88, 82, and 78 mV, respectively. This indicates that peak IIIc does not correspond to reduction of DA-o-quinone. With increasing NAc-CySH concentrations the peak current (ip) for reduction peak IIIc decreases; at NAc-CySH concentrations g5.0 mM peak IIIc disappears (Figure 1E). In the presence of NAc-CySH the peak IIc/peak IIa couple (E°′ ) -295 mV) is also eliminated and a new reversible couple, peak IVc/peak IVa (E°′ ) -311 mV), appears. Cyclic voltammograms of DA in the presence of NAcCySH also exhibit an apparently irreversible oxidation peak Va (Ep ≈ 570 mV). At very high NAc-CySH

1122 Chem. Res. Toxicol., Vol. 9, No. 7, 1996

Shen et al.

Figure 3. Relative yields of products formed upon controlled potential electro-oxidation of 1.0 mM DA in the presence of 5.0 mM NAc-CySH at 50 mV in phosphate buffer (pH 7.4; µ ) 0.2) as a function of reaction time. Peak heights using HPLC method I were employed to estimate relative product yields.

Figure 2. HPLC chromatograms (method I) of the product solutions formed following controlled potential electro-oxidation of 1.0 mM DA in the presence of (A) 1.0 mM, (B) 2.0 mM, and (C) 5.0 mM NAc-CySH at 50 mV for 10 min in phosphate buffer (pH 7.4; µ ) 0.2). Injection volume: 2.0 mL.

concentrations (g5.0 mM) oxidation peak VIa appears (Ep ≈ 400 mV). At such concentrations, NAc-CySH exhibits a broad, irreversible oxidation peak at the same Ep as peak VIa (Figure 1F). Controlled Potential Electro-oxidation. The appearance of peak IIIa at less positive potentials than peak Ia in cyclic voltammograms of DA in the presence of NAcCySH (Figure 1B-E) indicates that the latter sulfhydryl compound, similar to CySH (15, 18), facilitates the oxidation of the neurotransmitter. Accordingly, controlled potential electro-oxidations of DA in the presence of NAc-CySH were performed at an applied potential of 50 mV in order to minimize secondary reactions. Under these conditions, DA alone is oxidized to only a very minor extent (18). Figure 2 presents a series of HPLC chromatograms of the product solutions formed following controlled potential electro-oxidation of DA (1.0 mM) in the presence of increasing concentrations of NAc-CySH (1-5 mM) for 10 min at pH 7.4. These chromatograms reveal that at a low (1:1) NAc-CySH/DA concentration ratio 9 is the major product with lower yields of 5 > 13 > 6 > 10 (structures are shown in Figure 2). However, with increasing NAc-CySH: DA ratio yields of 13 increase dramatically and a new minor product, 16, appears. The relative changes in product yields (based upon the chromatographic peak heights using HPLC method I) throughout the course of a more prolonged oxidation of DA (1.0 mM) in the presence of NAc-CySH (5.0 mM) are shown in Figure 3. Thus, with increasing reaction time the yields of the 5-S-(N-acetylcysteinyl) conjugate 5 systematically decrease, and after 2 h, this compound could not be detected. The yields of 6, 9, and 10 initially increase, reach maximal levels, and then decrease. By contrast, yields of 13 and 16 increase

throughout the course of the reaction, although the latter compound was always a very minor product. Cyclic Voltammetry of N-Acetylcysteinyl Conjugates of DA. Cyclic voltammograms (ν ) 100 mV s-1) of 1.0 mM solutions of 5, 6, 9, and 13 at pH 7.4 are presented in Figure 4. The Ep values for each of the peaks observed in these cyclic voltammograms are shown in Figure 4. Neurobehavioral Effects Evoked by and LD50 Values for 5, 6, 9, and 13. Experiments using mice were employed to assess the toxicity and behavioral responses evoked when 5, 6, 9, or 13 were administered into the brain (insufficient amounts of 10 and 16 were available for biological testing). In all of these experiments solutions of each compound dissolved in 5 µL of isotonic saline were administered into the vicinity of the left lateral ventricle while mice were maintained under a light ether anesthetic. The LD50 values, used as a measure of toxicity and defined as the dose of the injected compound (expressed as free base) at which 50% of the treated animals died within 1 h, were determined using the statistical method of Dixon (34). Compounds 5, 6, and 9 were lethal, having experimental LD50 values of 14 ( 1 µg (45 ( 3 nmol) (mean ( SD), 25 ( 1 µg (80 ( 3 nmol), and 42 ( 1 µg (88 ( 2 nmol), respectively. Doses of 5 ranging from 10 to 100 µg, and of 6 and 9 ranging from 20 to 100 µg, were employed to determine the LD50 values. The neurobehavioral effects evoked by intracerebral injection of 5, 6, and 9 were very similar. Those described below were observed following administration of LD50 doses. Thus, after recovery from the anesthetic, animals either circled contralateral to the site of injection or walked backwards. Following this brief period (2-5 min) mice exhibited episodes of extreme hyperactivity, characterized by very rapid running, jumping, rolling along the head-tail axis, in addition to squeaking and severe trembling. At doses up to 100 µg, 13 evoked no unusual behavioral responses. Control animals treated with 5 µL of vehicle alone exhibited none of the above behavioral effects and all survived. Intracerebroventricular (icv) Administration of 5-S-Glu-DA and 5-S-CyS-DA. Rats that received icv infusions of 5-S-Glu-DA (44 nmol) or 5-S-CyS-DA (44 nmol) were sacrificed 15 min after the infusion was completed. An HPLC-EC chromatogram of a homogenate of the rat midbrain prepared 15 min after infusion of 5-SGlu-DA is presented in Figure 5A. Thus, even after such a small time interval, 5-S-Glu-DA was almost totally eliminated from this region of the brain. However, significant peaks for 5-S-CyS-DA and particularly 5 were present. Figure 5B is a chromatogram of a homogenate of the midbrain of a rat sacrificed 15 min after icv

N-Acetylcysteinyl Conjugates of Dopamine

Chem. Res. Toxicol., Vol. 9, No. 7, 1996 1123

Figure 4. Cyclic voltammograms at the PGE of 1.0 mM solutions of (A) 5, (B) 6, (C) 9, and (D) 13 in phosphate buffer (pH 7.4; µ ) 1.0). Sweep rate: 100 mV s-1.

Figure 6. (A) Concentrations of 5-S-Glu-DA, 5-S-CyS-DA, and 5 in five regions of rat brain 15 min after icv administration of 5-S-Glu-DA (44 nmol), and (B) concentrations of 5-S-CyS-DA and 5 in the same brain regions 15 min after icv administration of 5-S-CyS-DA (44 nmol). Data obtained by HPLC-EC analysis and represent the mean values obtained using 4 rats.

Figure 5. HPLC-EC chromatograms of supernatants of midbrain homogenates prepared from rats 15 min after icv administration of (A) 44 nmol of 5-S-Glu-DA and (B) 44 nmol of 5-SCyS-DA.

infusion of 5-S-CyS-DA and reveals that a significant fraction of this conjugate had been N-acetylated to 5. A summary of the concentrations of 5-S-Glu-DA, 5-S-CySDA, and 5 measured in five regions of rat brain 15 min after icv infusion of 5-S-Glu-DA (44 nmol) is presented in Figure 6A. These data indicate that 5-S-Glu-DA has a short lifetime in all brain regions. Furthermore, in all brain regions, but most notably in the midbrain, 5-S-GluDA was metabolized to a significant extent to 5. Figure 6B shows the concentrations of 5-S-CyS-DA and 5 measured in five regions of rat brain 15 min following icv infusion of 5-S-CyS-DA (44 nmol). These data tend

to suggest that N-acetylation of 5-S-CyS-DA to 5 is somewhat slower than metabolism of 5-S-Glu-DA to 5. However, again, N-acetylation of 5-S-CyS-DA to 5 appears to occur most rapidly in the midbrain. This observation is in accord with the report that cysteine conjugate N-acetyltransferase activity is greatest in the rat midbrain whereas activities of N-acetyl-L-cysteine conjugate deacetylase are much lower and uniform throughout the brain (29). HPLC-EC analyses of homogenates of the brains of rats sacrificed 15 min after icv administration of artificial CSF exhibited no peaks corresponding to 5-S-Glu-DA, 5-S-CyS-DA, or 5 in any brain region at the detector sensitivity employed. Because rats were anesthetized prior to and after icv administration of 5-S-Glu-DA and 5-S-CyS-DA, it was not possible to observe any neurobehavioral effects evoked by 5 formed as a result of metabolism of these conjugates.

Discussion The appearance of oxidation peak IIIa at more negative potentials than peak Ia in cyclic voltammograms of DA in the presence of NAc-CySH (Figure 1B-E) results from several processes. Thus, at potentials corresponding to the rising segment of oxidation peak Ia, DA is oxidized (2e, 2H+) to DA-o-quinone that is rapidly scavenged by

1124 Chem. Res. Toxicol., Vol. 9, No. 7, 1996

NAc-CySH to give 5 (major product) and 6 (minor product) (Scheme 1). Thus the concentration of DA-oquinone at the electrode surface is greatly reduced. Nernstian considerations therefore demand that additional DA is oxidized in order to reestablish the surface concentration of DA-o-quinone with a resultant increase in the current at potentials less positive than peak Ia; i.e., NAc-CySH facilitates the oxidation of DA leading, in part, to the appearance of peak IIIa. However, the Ep values for the primary oxidation peaks observed on the initial anodic sweep in cyclic voltammograms (ν ) 100 mV s-1) of 5 and 6 are 106 and 124 mV, respectively. Thus, both of these N-acetylcysteinyl conjugates, particularly 5, are more easily oxidized than DA (Ep ) 130 mV). Accordingly, at peak IIIa potentials both 5 and, to a lesser extent, 6 must be further oxidized (2e, 2H+) to o-quinones 7 and 8, respectively. The major reactions of 7 and 8 involve nucleophilic addition of NAc-CySH to give the 2,5-di-S-(N-acetylcysteinyl) conjugate 9 (Scheme 1). A very minor reaction of 7 involves nucleophilic addition of NAc-CySH to give 10. The Ep for the primary oxidation peak of 9 is 95 mV. Thus, this compound must also be oxidized at peak IIIc potentials to give o-quinone 11 that is attacked by NAc-CySH to give 13. Chromatographic evidence (Figures 2 and 3) indicates that 5 and 9 represent the major initial products of oxidation of DA in the presence of NAc-CySH. Furthermore, the Ep values for 5 and 9 are 106 and 95 mV, respectively. By contrast, Ep values for 6, a minor product, and 13 are 124 and 138 mV, respectively. Accordingly, it can be concluded that the major process contributing to peak IIIa involves oxidation of 5 and, subsequently, 9. The very low yields of 10 precluded measurement of its cyclic voltammetric behavior. Nevertheless, it is probable that this compound also oxidized (2e, 2H+) to o-quinone 12 that is scavenged by NAc-CySH, thus serving as an additional pathway to 13 (Scheme 1). A very minor product of controlled potential electro-oxidation of DA in the presence of NAc-CySH at 50 mV is 16. Experiments revealed that this compound is formed by oxidation of 13 and its yields increased significantly when more positive potentials were employed in controlled potential electrolyses. Although details of this reaction remain to be elucidated, it is probable that 13 is initially oxidized (2e, 2H+) to o-quinone 14 that exists in equilibrium with the p-hemithioquinone 15a which rearranges to its tautomer 15b followed by decarboxylation to give 16 (Scheme 1). Cyclic voltammograms of 5, 6, 9, and 13 all exhibit reduction peaks reversibly coupled to their primarily oxidation peaks corresponding to reduction of o-quinones 7, 8, 11, and 14, respectively. The Ep values for the reduction peaks of 7, 8, 11, and 14 wee 81, 96, 78, and 116 mV, respectively (Figure 4). The fact that Ep for peak IIIc, observed in cyclic voltammograms of DA (1 mM) in the presence of NAc-CySH (Figure 1B-D), is 88 mV (0.5 mM NAc-CySH), 82 mV (1.0 mM NAc-CySH), and 78 mV (2.0 mM NAc-CySH) suggests that this peak predominantly represents reduction of o-quinones 7 and 11. This, again, provides further support for the conclusion that peak IIIa corresponds primarily to oxidation of 5 and 9 and peak IIIc to reduction of their proximate products, 7 and 11, respectively. Cyclic voltammograms of 5, 6, 9, and 13 exhibit reversible followup couples at E°′ values of -307, -307, -312, and -314 mV, respectively (Figure 4), which are all similar to that for the peak IVc/peak IVa couple (E°′ ) -311 mV) observed in cyclic voltam-

Shen et al.

mograms of DA in the presence of NAc-CySH (Figure 1B-E). These couples appear in the same potential region as peak IIc/peak IIa (E°′ ) -295 mV) observed in cyclic voltammograms of DA corresponding to the 3/2 couple (Figure 1A). This suggests that o-quinones 7, 8, 11, and 14, the proximate oxidation products of 5, 6, 9, and 13, respectively, can undergo intramolecular cyclizations to indolines analogous to 2. Further chemical oxidations of such indolines by o-quinones 7, 8, 11, and 14 presumably then yield p-quinone imines analogous to 3 (Figure 1A) that are responsible for the peak IVc/peak IVa couple. Cyclic voltammograms of 5, 9, and 13 (but not 6) exhibit an irreversible oxidation peak at Ep ) 571 ( 7 mV. This peak corresponds to oxidation peak Va observed in cyclic voltammograms of DA in the presence of NAc-CySH (Figure 1B-E). However, the oxidation reaction responsible for peak Va remains to be elucidated. Electrochemical oxidation of DA in the presence of NAc-CySH thus provides the basis for a simple method to synthesize several N-acetylcysteinyl conjugates of DA, particularly 5, 9, and 13 along with lower yields of 6, 10, and 16. Voltammetric data reveal that 5 and 9 are significantly more easily oxidized than DA but are less easily oxidized than 5-S-CyS-DA and all other known cysteinyldopamines (18). The nature of the products of oxidation of 5, 9, and other N-acetylcysteinyl conjugates of DA remains to be determined. However, the cyclic voltammetric behaviors of these compounds suggest that their o-quinone proximate oxidation products undergo intramolecular cyclization of the ethylamino side chain similar to that which occurs with DA-o-quinone (Figure 1A). The results summarized in Figures 5A and 6A indicate that 5-S-Glu-DA is rapidly (ca. 15 min) cleared from all regions of rat brain with concomitant appearance of 5-SCyS-DA and 5. Based on previous studies on a structurally-related glutathionyl conjugate (29), it can be concluded that the first step in the metabolism of 5-S-GluDA involves the extraneuronal γ-GT-mediated cleavage of the γ-glutamyl bond of its glutathionyl residue to give the corresponding 5-S-cysteinylglycine conjugate. Further hydrolysis of this conjugate by the action of a cysteinyl glycine dipeptidase then leads to 5-S-CyS-DA. However, within 15 min after administration of 5-S-GluDA, significant levels of 5 were measured in all regions of rat brain. Notably, the largest 5/5-S-CyS-DA ratio was observed in the midbrain, a region having the highest brain cysteine conjugate N-acetyltransferase activity (29). However, 15 min following icv administration of 5-S-CySDA, the 5/5-S-CyS-DA ratio was much smaller in all brain regions, compared to that measured after administration of the same dose (44 nmol) of 5-S-Glu-DA, although the highest ratio was again observed in the midbrain. Preliminary studies have also demonstrated that 2,5-di-S-Glu-DA and 2,5,6-tri-S-Glu-DA are also rapidly cleared from rat brain and that the corresponding cysteinyl and N-acetylcysteinyl conjugates of DA are formed. However, HPLC-EC chromatograms of brain homogenates prepared shortly after icv infusion of these compounds contained additional peaks corresponding to metabolites that remain to be identified (data not shown). It has recently been proposed that an early event in the pathogenesis of idiopathic Parkinson’s disease is an elevated translocation of CySH and/or GSH into pigmented dopaminergic SN cell bodies (15, 18) in which cytoplasmic DA is normally oxidized to DA-o-quinone and thence neuromelanin (13). The major initial products of

N-Acetylcysteinyl Conjugates of Dopamine

the in vitro oxidation of DA in the presence of CySH or GSH at pH 7.4 are 5-S-CyS-DA (15, 18) and 5-S-Glu-DA (14), respectively. The same initial metabolites would be expected to be formed in the cytoplasm of pigmented dopaminergic neurons in the Parkinsonian SN. However, the subsequent fates of 5-S-CyS-DA and 5-S-GluDA probably depend on several factors. To illustrate, at their intraneuronal site of formation additional oxidation chemistry would be expected to occur. This is so because 5-S-Glu-DA (14) and particularly 5-S-CyS-DA (15, 18, 30) are more easily oxidized than DA. Oxidation of 5-S-GluDA in the presence of free GSH generates an o-quinone intermediate that is rapidly scavenged by the tripeptide to give 2,5-di-S-Glu-DA that is further oxidized to 2,5,6tri-S-Glu-DA (14). By contrast, oxidation of DA in the presence of free CySH gives primarily 2,5-di-S-CyS-DA, 2,5,6-tri-S-CyS-DA, and a number of dihydrobenzothiazines (DHBTs), (15, 18). These DHBTs are formed by intramolecular cyclizations of o-quinone intermediates generated by oxidations of cysteinyldopamines (18). Several of these DHBTs and 2,5-di-S-CyS-DA are toxic when administered into mouse and rat brains (15, 18). Furthermore, DHBTs are also easily oxidized to benzothiazine (BT) derivatives, some of which are even more lethal than their parent compounds.2 While it remains to be established whether these putative intraneuronal metabolites, formed by oxidative reactions between DA and CySH, are dopaminergic neurotoxins, these observations have led us to speculate that one or more of these compounds might be the endotoxins formed in the cytoplasm of neuromelanin-pigmented SN cells that contribute to the degeneration of these neurons in PD (15, 18). The detection of very low levels of 5-S-CyS-DA in the CSF of PD patients (35) provides evidence to suggest that cysteinyldopamines and glutathionyldopamines are either released or leak from dopaminergic neurons in which they are formed. Extraneuronally, these metabolites would be exposed to γ-GT, cysteinyl glycinyl dipeptidases, and cysteine conjugate N-acetyltransferase. The results of this investigation indicate that 5-S-Glu-DA, other glutathionyldopamines, and 5-SCyS-DA are rapidly metabolized to N-acetylcysteinyl conjugates of DA including 5, 6, and 9, all of which are toxic compounds. It is of particular interest that both 5-S-Glu-DA and 5-S-CyS-DA are metabolized to 5 most rapidly in the midbrain of the rat, a region that includes the SN. These observations, therefore, raise the possibility that 5 and other N-acetylcysteinyl conjugates of DA might play roles in the neuropathological processes that occur in the Parkinsonian SN. The mechanisms that underlie the neurobehavioral effects evoked by 5, 6, and 9 when administered into the mouse brain are unknown. However, it is unlikely that these effects are related to the redox properties of these compounds since similar doses of 5-S-CyS-DA, an easily oxidized compound, evoke no unusual neurobehavioral responses in the mouse. It is also important to emphasize that at this time there is no experimental evidence that 5, 6, 9, or other N-acetylcysteinyl conjugates of DA are toxic toward dopaminergic or other neurons. Fornstedt et al. (7) have reported that the 5-S-CySDA/DA ratio increases with increasing degeneration and depigmentation of the SN and that this ratio reaches its maximal value in the Parkinsonian brain. Similarly, an elevated 5-S-CyS-DA/HVA ratio has been measured in 2

Unpublished observations.

Chem. Res. Toxicol., Vol. 9, No. 7, 1996 1125

the CSF of PD patients (35). However, within experimental error, the absolute concentrations of 5-S-CyS-DA measured both in Parkinsonian SN tissue (7) and CSF (35) were not only extremely low but virtually identical to those measured in age-matched controls. If the large decrements of nigral GSH that occur in the Parkinsonian SN (5, 6), which also occur at early presymptomatic stages of the disease (36), and apparent depigmentation of dopaminergic cell bodies prior to degeneration (7, 9-12) do indeed reflect the consequences of an elevated influx of CySH and/or GSH into these neurons, it might be expected that levels of 5-S-CyS-DA would be higher than those measured experimentally in both brain tissue and CSF. However, the results of this and previous investigations suggest that these levels probably represent only a remnant of the 5-S-CyS-DA that is actually formed. Thus, intraneuronally, a significant fraction of 5-S-CyS-DA and 5-S-Glu-DA would be expected to undergo further facile oxidation to more complex cysteinyl and glutathionyl conjugates of DA, respectively. Further oxidation of cysteinyldopamines would lead to a number of DHBTs and thence BTs (18, 30). Extraneuronally, 5-SCyS-DA and 5-S-Glu-DA and other more complex cysteinyl- and glutathionyldopamines would be rapidly metabolized to N-acetylcysteinyl conjugates of DA. Together, such intraneuronal and extraneuronal processesd might account not only for the very low levels of 5-S-CySDA measured in Parkinsonian SN tissue and CSF but also for formation of a cascade of endotoxins that contribute to the death of not only nigrostriatal dopaminergic neurons but several other neuronal systems that degenerate in PD.

Acknowledgment. This work was supported by National Institutes of Health Grant NS-29886. Additional support was provided by the Vice President for Research and Research Council at the University of Oklahoma.

References (1) Hornykiewicz, O., and Kish, S. J. (1986) Biochemical pathophysiology of Parkinson’s Disease. Adv. Neurol. 45, 19-34. (2) Hornykiewicz, O. (1989) Aging and neurotoxins as causitive factors in idiopathic Parkinson’s Disease: A critical analysisd of neurochemical evidence. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 13, 319-328. (3) Dexter, D. T., Carter, C. J., Wells, F. R., Javoy-Agid, F., Lees, A. J., Jenner, P., and Marsden, C. D. (1989) Basal lipid peroxidation in substantia nigra is increased in Parkinson’s Disease. J. Neurochem. 52, 381-389. (4) Schapira, A. H. V., Cooper, J. M., Dexter, D., Clark, J. B., Jenner, P., and Marsden, C. D. (1990) Mitochondrial complex I deficiency in Parkinson’s Disease. J. Neurochem. 54, 323-327. (5) Riederer, P., Sofic, E., Rausch, W. D., Schmidt, B., Reynolds, G., Jellinger, K., and Youdim, M. B. H. (1989) Transition metals, ferritin, glutathione and ascorbic acid in Parkinsonian brain. J. Neurochem. 52, 515-520. (6) Jenner, P. (1993) Altered mitochondrial function, iron metabolism and glutathione levels in Parkinson’s disease. Acta Neurol. Scand. 87 (Suppl. 146), 6-13. (7) Fornstedt, B., Brun, A., Rosengren, E., and Carlsson, A. (1989) The apparent autoxidation rate of catechols in dopamine-rich regions of human brains increases with the degree of depigmentation of substantia nigra. J. Neural Transm.: Parkinson’s Dis. Dementia Sect. 1, 279-295. (8) Sian, J., Dexter, D. T., Jenner, P., and Marsden, C. D. (1992) Glutathione-related enzymes in brain in basal ganglia degenerative disorders. Br. J. Pharmacol. 107 (Suppl.), 428P. (9) Mann, D. M. A., and Yates, P. O. (1983) Possible role of neuromelanin in the pathogenesis of Parkinson’s disease. Mech. Ageing Dev. 21, 193-203.

1126 Chem. Res. Toxicol., Vol. 9, No. 7, 1996 (10) Hirsch, E. C., Graybiel, A. M., and Agid, Y. (1988) Melanized dopaminergic neurons are differentially susceptible to degeneration in Parkinson’s disease. Nature 334, 345-348. (11) Kastner, A., Hirsch, E. C., Lejeune, O., Javoy-Agid, F., Rascol, O., and Agid, Y. (1992) Is the vulnerability of neurons in the substantia nigra of patients with Parkinson’s disease related to their neuromelanin content? J. Neurochem. 59, 1080-1089. (12) Hirsch, E. C. (1992) Why are nigral catecholaminergic neurons more vulnerable than other cells in Parkinson’s disease? Ann. Neurol. 32, S88-S93. (13) Graham, D. G. (1978) Oxidation pathways for catecholamines in the genesis of neuromelanin and cytotoxic quinones. Mol. Pharmacol. 14, 633-643. (14) Zhang, F., and Dryhurst, G. (1995) Influence of glutathione on the oxidation chemistry of the catecholaminergic neurotransmitter dopamine. J. Electroanal. Chem. Interfacial Electrochem. 398, 117-128. (15) 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. (16) Carstam, R., Brinck, C., Hindemith-Angustsson, A., Rorsman, H., and Rosengren, E. (1991) The neuromelanin of the substantia nigra. Biochim. Biophys. Acta 1097, 152-160. (17) Slivka, A., Mytilineou, C., and Cohen, G. (1987) Histochemical evaluation of glutathione in brain. Brain Res. 409, 275-284. (18) Shen, X.-M., and Dryhurst, G. (1996) Further insights into the influence of L-cysteine on the oxidation chemistry of dopamine: Reaction pathways of potential relevance to Parkinson’s Disease. Chem. Res. Toxicol. 9, 751-763. (19) Steventon, G. B., Heafield, M. T. E., Waring, R. H., and Williams, A. C. (1989) Xenobiotic metabolism in Parkinson’s Disease. Neurology 39, 883-887. (20) Waring, R. H., Sturman, S. G., Smith, M. C. G., Steventon, G. B., Heafield, M. T. E., and Williams, A. C. (1989) S-Methylation in motoneuron and Parkinson’s Disease. Lancet, 356-357. (21) Yong, V. W., Perry, T. L., and Krisman, A. A. (1985) Depletion of glutathione in brainstem of mice caused by N-methyl-4-phenyl1,2,3,6-tetrahydropyridine is prevented by antioxidant treatment. Neurosci. Lett. 63, 56-60. (22) Raps, S. P., Lai, J. C. K., Hertz, L., and Cooper, A. J. L. (1989) Glutathione is present in high concentrations in cultured astrocytes but not in cultured neurons. Brain Res. 493, 398-401. (23) Sagara, J.-I., Miura, K., and Bannai, S. (1993) Maintenance of neuronal glutathione by glial cells. J. Neurochem. 61, 1672-1676. (24) Liccione, J. L., and Maines, M. D. (1988) Selective vulnerability of glutathione metabolism and cellular defense mechanisms in rat striatum to manganese. J. Pharmacol. Exp. Ther. 247, 156161.

Shen et al. (25) Orlowski, M., and Meister, A. (1979) The γ-glutamyl cycle. A possible transport system for amino acids. Proc. Nat. Acad. Sci. U.S.A. 67, 1248-1255. (26) Meister, A. (1974) Glutathione metabolism and function via the γ-glutamyl cycle. Life Sci. 15, 177-190. (27) Hiramatsu, M., Kumagai, Y., Unger, S. E., and Cho, A. K. (1990) Metabolism of methylenedioxymethamphetamine: formation of dihydroxymethamphetamine and a quinone identified as its glutathione adduct. J. Pharmacol. Exp. Ther. 254, 521-527. (28) Patel, N., Kumagai, Y., Unger, S. E., and Cho, A. K. (1991) Transformation of dopamine and R-methyldopamine by NG-10815 cells: Formation of thiol adducts. Chem. Res. Toxicol. 4, 421426. (29) Miller, R. T., Lau, S. S., and Monks, T. J. (1995) Metabolism of 5-(glutathion-S-yl)-R-methyldiopamine following intracerebroventricular administration to male Sprague-Dawley rats. Chem. Res. Toxicol. 8, 634-641. (30) Zhang, F., and Dryhurst, G. (1995) Reactions of cysteine and cysteinyl derivatives with dopamine-o-quinone and further insights into the oxidation chemistry of 5-S-cysteinyldopamine: Potential relevance to idiopathic Parkinson’s Disease. Bioorg. Chem. 23, 193-216. (31) Wrona, M. Z., Goyal, R. N., Turk, D., Blank, C. L., and Dryhurst, G. (1992) 5,5′-Dihydroxy-4,4′-bitryptamine: A potentially aberrant neurotoxic metabolite of serotonin. J. Neurochem. 59, 13921398. (32) Tse, D. C. S., McCreery, R. L., and Adams, R. N. (1976) Potential oxidative pathways of brain catecholamines. J. Med. Chem. 19, 37-40. (33) Zhang, F., and Dryhurst, G. (1993) Oxidation chemistry of dopamine: Possible insights into the age-dependent loss of dopaminergic nigrostriatal neurons. Bioorg. Chem. 21, 392-410. (34) Dixon, W. J. (1965) The up-and-down method for small samples. J. Am. Stat. Assoc. 60, 967-978. (35) Cheng, F.-C., Kuo, J.-S., Ghia, L.-G., and Dryhurst, G. (1996) Elevated 5-S-Cysteinyldopamine/homovanillic acid ratio and reduced homovanillic acid in cerebrospinal fluid: Possible markers for and potential insights into the pathoetiology of Parkinson’s disease. J. Neural Transm. 103, 433-446. (36) Jenner, P., Dexter, D. T., Sian, J., Schapira, A. H. V., and Marsden, D. (1992) Oxidative stress as a cause of nigral cell death in Parkinson’s Disease and Incidental Lewy Body Disease. Ann. Neurol. 32 (Suppl.), S82-S87.

TX960052V