A novel hydrogen peroxide-dependent oxidation pathway of dopamine via 6-hydroxydopamine

Article abstract of DOI:10.1016/S0040-4020(03)00242-4

In the presence of excess H₂O₂, oxidation of dopamine was diverted from the usual pigment-forming pathway to afford 6-hydroxydopamine and then a colorless reaction mixture comprising a polar non-extractable product. The latter was obtained in 20% yield by oxidation of 6-hydroxydopamine and was tentatively formulated as the novel 5-(2-aminoethyl)-2-hydroxy-5-(3-hydroxy-2-oxotetrahydro-1aH-oxireno[2,3] cyclopenta[1,2-b]pyrrol-3a(4H)-yl)cyclohex-2-ene-1,4-dione by extensive spectral analysis and conversion to a tetraacetyl derivative. Mechanistic experiments suggested that formation of the product proceeds via 6-hydroxydopamine by H₂O₂-dependent epoxidation and cyclization steps followed by dimerization and ring contraction with decarboxylation.

Full text of DOI:10.1016/S0040-4020(03)00242-4

Tetrahedron 59 (2003) 2215–2221
A novel hydrogen peroxide-dependent oxidation pathway
of dopamine via 6-hydroxydopamine
*
Paola Manini, Lucia Panzella, Alessandra Napolitano and Marco d’Ischia
Department of Organic Chemistry and Biochemistry, University of Naples ‘Federico II’, Complesso Universitario Monte S. Angelo,
Via Cinthia 4, I-80126 Naples, Italy
Received 14 October 2002; revised 17 January 2003; accepted 13 February 2003
Dedicated in the memory of Professor Giuseppe Prota, an extraordinary teacher and scientist
Abstract—In the presence of excess H
2 2
O , oxidation of dopamine was diverted from the usual pigment-forming pathway to afford
6-hydroxydopamine and then a colorless reaction mixture comprising a polar non-extractable product. The latter was obtained in 20% yield
by oxidation of 6-hydroxydopamine and was tentatively formulated as the novel 5-(2-aminoethyl)-2-hydroxy-5-(3-hydroxy-2-
oxotetrahydro-1aH-oxireno[2,3]cyclopenta[1,2-b]pyrrol-3a(4H)-yl)cyclohex-2-ene-1,4-dione by extensive spectral analysis and conversion
to a tetraacetyl derivative. Mechanistic experiments suggested that formation of the product proceeds via 6-hydroxydopamine by H O -
2 2
dependent epoxidation and cyclization steps followed by dimerization and ring contraction with decarboxylation. q 2003 Elsevier Science
Ltd. All rights reserved.
1
. Introduction
Recently, in a study of the effects of nitrite ions on the
1
4
hydrogen peroxide-promoted oxidation of 1, as a possible
route to the putative dopamine metabolite 6-nitrodopa-
The oxidation chemistry of dopamine (1) and related
catecholamines has been an active research subject over
the past decades because of its central relevance to
nigrostriatal neuronal degeneration in Parkinson’s dis-
and the etiopathogenesis of vitiligo.
incentives for these studies have derived from chemical
issues relating to the biomimetic synthesis of indole
1
5
mine, we noticed the generation at high substrate
concentrations of a product whose chromatographic and
spectral features did not match with those of the known
1
–5
6–8
4,5,16,17
ease,
Further
oxidation products of 1. We report herein the
isolation and spectral characterization of this new product,
9
,10
systems
in food technology.
and the control of oxidative browning processes
The first insight into dopamine oxidation was provided by
the elucidation of the classic indoline-aminochrome-5,6-
1
1–13
dihydroxyindole pathway depicted in Scheme 1.
,5
In the
†
6-hydroxydopamine (2)
4
presence of hydroperoxides,
quinone may be formed, which affords likewise 5,6-di-
hydroxyindole, the ultimate monomer precursor of the
dark polymeric pigments that constitute the characteristic
2
,9
visible outcome of dopamine oxidation. Little is currently
known about alternative oxidation pathways of 1, nor
has any intermediate product between the monomer
and pigment stage ever been isolated and structurally
characterized.
Keywords: dopamine; 6-hydroxydopamine; oxidation; hydrogen peroxide;
dimerization.
*
Corresponding author. Tel.: þ39-81-674133; fax: þ39-81-674393;
e-mail: alesnapo@unina.it
The common catecholamine nomenclature system was adopted which
assigns numbers 3, 4 and 6 to the OH bearing carbons.
†
Scheme 1.
0040–4020/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4020(03)00242-4

2216
P. Manini et al. / Tetrahedron 59 (2003) 2215–2221
which features an unusual heterocyclic ring system arising
by a remarkable H O -dependent oxidative dimerization
route of 2.
for preparative purposes. After several attempts, the product
was eventually isolated by ion exchange chromatography of
the reaction mixture followed by preparative HPLC
fractionation. The glassy oil thus obtained contained a
single product homogeneous on HPLC and sufficiently,
though not indefinitely, stable to be subjected to extensive
NMR analysis.
2
2
1
A noticeable feature of the H NMR spectrum in D O/DCl
2
was the lack of signals in the aromatic region. The most
deshielded resonance was a doublet (J¼1.8 Hz) at d 4.99,
coupled to a similar doublet at d 3.36. In addition, a series of
multiplets (8H overall) were apparent in the region between
d 2.0–4.5, suggesting two aminoethyl chains. A singlet (1H)
at d 3.32 was also observed. In DMSO-d the two doublets
6
were shifted upfield at d 4.53 and 3.00, and additional
signals at d 5.00 (1H, singlet), 3.18 and 2.68 (doublets,
J¼16.8 Hz) were present, the latter indicating diastereo-
topic CH protons. Addition of D O resulted in the partial
2
. Results and discussion
2
2
Oxidation of 1 at 1 mM concentration with horseradish
peroxidase (HRP)/H O in 0.1 M phosphate buffer, pH 7.4,
disappearance of the doublet at d 2.68, suggesting different
rates of D-exchange of the CH protons. COSY analysis of
2
2
2
proceeded smoothly to give in the early stages a main
1
the spin systems revealed moreover, marked differences in
both chemical shifts and coupling constants of the protons
of the two ethylamine chains.
8,19
transient species coinciding with 2 quinone,
which
gradually changed into 5,6-dihydroxyindole and then into a
dark brown pigment. A quite different pattern of reactivity
was observed when a large excess of H O and/or higher
1
3
Main features of the C NMR spectrum (DMSO-d ) were
6
2
2
concentrations of dopamine were used. Under these latter
conditions, pigment formation was inhibited and a pale
yellow, clear solution developed. HPLC analysis showed
small amounts of a species eluted just before 2 quinone and
with an absorption maximum centred at 293 nm in acidic
media. The product resisted treatment with dithionite or
ascorbate or with acids (HPLC evidence), but was smoothly
four quaternary carbons at d 211.5, 189.6, 187.3, 170.9,
three CH signals at d 88.9 (showing C–D coupling
following D O exchange), 80.4 and 60.4, three signals for
2
quaternary carbons at d 75.6, 69.3 and 48.8, and five CH2
resonances, one of which (d 34.4) proved sensitive to D O
exchange.
2
converted by NaBH to a slower eluting species with similar
4
absorption spectrum. It could not be recovered on standing
at pH.10.
Overall, these data suggested that: (a) the number of carbons
was that expected for two molecules of starting material,
minus one carbon; (b) an additional CH2 was formed;
(c) the aromatic units were profoundly modified; (d) the two
The same product was formed in much higher yields by
oxidation of 2 under similar conditions, suggesting that this
catecholamine was the actual precursor. Accordingly, in
subsequent experiments 2 was preferably used as substrate
aminoethyl chains experienced different chemical environ-
ments, reflecting adjacent stereocenters and presumably,
in one case, cyclization; (e) three carbonyl groups and a
2
deshielded sp carbon were present; (f) an epoxide ring was
6
Table 1. Spectral data of compound 3 (DMSO-d )
13
C
1
d H (mult, J Hz)
1
1
H, H COSY
1 13
H, C HMBC
d
C-1
C-2
C-3
C-4
C-5
C-6
187.3
170.9
88.9
189.6
48.8
34.4
–
–
–
–
–
5.00 (s)
–
–
2.68 (d, 16.8)
–
–
–
–
3.18
2.68
2.25, 3.47
2.17, 3.54
3.54, 2.17
3.47, 2.25
4.53
48.8, 75.6
–
–
–
3
.18 (d, 16.8)
2.17 (m)
.25 (m)
3.47 (m)
.54 (m)
60.4, 69.3, 170.9, 189.6
34.4, 48.8, 170.9
–
CH
2
CH
2
N
32.5
45.2
2
CH
2
N
–
3
–
34.4, 48.8, 69.3, 75.6, 80.4, 211.5
–
–
21.0, 69.3, 189.6, 211.5
–
–
–
–
–
0
C-1
C-1a
C-2
C-3
60.4
69.3
211.5
80.4
75.6
53.3
3.00
–
–
4.53
–
3.85 (m)
0
–
–
3.00
0
0
0
C-3a
0
–
C-5
4.07, 2.72, 2.32
3.85, 2.72, 2.32
2.72, 4.07
2.32, 3.85
4
.07 (m)
2.32 (m)
.72 (m)
0
C-6
21.0
2

P. Manini et al. / Tetrahedron 59 (2003) 2215–2221
2217
present (1H signal at d 3.00 correlating with carbon
resonances at d 60.4 (HMQC) and 69.3 (HMBC)).
degradation of the sample or tendency to separate as oil or
amorphous solid.
Based on correlation data, a most likely structure for the
compound was 3 in which a unit of 2 quinone was linked
through the quaternary C-1 carbon to a tetrahydrooxireno-
cyclopenta[1,2-b]pyrrole ring system. Resonances were
Reaction of 3 with acetic anhydride/pyridine at room
temperature for 16 h resulted in the quantitative conversion
to a main acetylated derivative, which could be purified to
give eventually a pale yellow oil homogeneous on TLC and
HPLC analysis.
1
1
1
13
assigned on the basis of the H, H COSY, H, C HMQC
and HMBC spectra (Table 1). Long range correlations
between high field CH protons at d 2.17/2.25, and the
1
13
Both the H and C NMR spectra indicated a tetraacetyl
derivative of 3 in which the methylene group and one of the
carbonyl groups on the cyclohexenedione ring of 3 were
missing and were replaced by two CH groups at d
6.51/106.2 and 6.68/108.2. Three quaternary carbons
resonating at d 52.3, 138.6 and 166.2, and a carbonyl
resonance at d 190.2 were also apparent. The epoxide ring
carbons on the bicyclic moiety resonated markedly up-field
at d 55.2 and 58.1, whereas the epoxide proton was
deshielded (d 3.53), possibly due to the acetylation of the
2
carbon resonances at d 34.4, 48.8 and 170.9 indicated that
the uncyclized ethylamine chain was linked to the
cyclohexenedione ring. Distinguishable correlations
0
between proton signal at d 4.53 (C-3 ) and the carbon
resonance at d 21.0 (C-6 ), the quaternary epoxide carbon (d
0
9.3) and the ketone carbon (d 211.5) resonances; and
6
between the epoxide proton signal at d 3.00 (C-1 ) with the
0
carbon C-3a (d 75.6), the adjacent epoxide carbon at d 69.3,
0
and the carbon resonances of the hydroxy ketone moiety (d
0
8
0.4, 211.5) were supportive of the tetrahydrooxireno-
OH group at the C-3 position, affecting also the geminal
1
13
cyclopentapyrrole ring system. The high field chemical
shift of the C-3 carbon was consistent with a vinylogous
a-oxocarboxylic acid moiety which is present in a
proton (d 5.98). H, C HMBC experiments showed
correlations between the proton at d 6.51 and resonances
at d 52.3 and 80.5, assigned to the carbons at the junction of
the two ring systems, and the adjacent b-carbonyl carbon (d
2
0–22
substantially ionized form in neutral aqueous medium.
166.2) bearing an acetoxy group. Intense cross peaks in the
HMBC spectrum were observed between the b-CH proton
The FAB-MS spectrum showed pseudomolecular MþH
2
(
fragmentation peaks at m/z 306 and 289 were also apparent
m/z 323) and MþNa (m/z 345) ion peaks. Intense
signal of the uncyclized ethylamine chain (d 3.84), the
carbon resonance of the adjacent CH2 (d 49.2) and a
quaternary carbon at d 171.5, due to the acetamido group. In
addition, the signal at d 5.98 correlated with a carbonyl
carbon at d 204.9, and with the carbon resonances (d 169.5
and 24.9) of the acetoxy function. These data, coupled
with the correlations between the epoxide proton signal
(d 3.53) and the carbon resonances at d 204.9, 80.5, and
55.2 (quaternary epoxide carbon), corroborated the
bicyclic tetrahydrooxirenocyclopentapyrrole ring system.
A summary of spectral data with correlations is given in
Table 2.
due to consecutive losses of two OH groups.
Structure 3 can exist in eight diastereoisomeric pairs of
enantiomers. Careful inspection of the NMR spectra,
however, did not provide evidence for a mixture of
stereoisomers. Moreover, the product should be formed in
different stereoisomers, at least some of these would
conceivably elute on HPLC at similar, but not identical,
retention times, giving rise to a group or a cluster of peaks,
at variance with experimental evidence.
Brief inspection of geometry optimized structures for all
diastereomers of 3 predicted lower energy values for those
structures featuring a cis-fused [3.3.0]azabicyclooctane
system. Unfortunately, NOESY and ROESY experiments
did not provide conclusive evidence to settle the stereo-
chemical issues.
The EI-MS spectrum of the acetylated derivative displayed
a weak molecular ion peak at m/z 490 and significant peaks
6
Table 2. Spectral data of compound 4 (DMSO-d )
13
C
1
d H (mult, J (Hz))
a
1
13
H, C HMBC
d
Attempts to grow crystals suitable for X-ray diffraction
analysis either from 3 or its bromide, hydrogensulfate and
perchlorate salts, using a variety of solvents (THF, acetic
acid, acetone, alcohols) or mixture of solvents and
temperature conditions proved unsuccessful, due to partial
C-1
C-2
C-3
C-4
C-5
C-6
190.2
106.2
166.2
138.6
108.2
52.3
–
6.51 (s)
–
–
52.3, 80.5, 166.2
–
–
–
–
6.68 (s)
–
–
CH
2
CH
N
2
N
30.7
49.2
58.1
55.2
204.9
84.6
80.5
48.0
24.2
3.84 (m)
4.16 (m)
49.2, 171.5
–
CH
2
0
C-1
C-1a
C-2
C-3
3– .53 (d, J¼1.3 Hz) 55.2, 80.5, 204.9
–
5– .98 (d, J¼1.3 Hz) 24.9, 169.5, 204.9
3.59 (m), 4.06 (m)
2.28 (m), 2.52 (m)
0
–
–
0
0
0
C-3a
0
–
C-5
C-6
–
–
–
–
–
–
0
OCOCH
OCOCH
NHCOCH
NHCOCH
3
169.2, 169.5, 169.7
20.5, 20.9, 24.9
171.5
–
3
2.11–2.30 (s)
–
2.11–2.30 (s)
3
3
24.3
a
3
Chemical shift values of the spectrum run in CD OD.

2218
P. Manini et al. / Tetrahedron 59 (2003) 2215–2221
at m/z 456, 414, 372, 330, denoting formation of a primary
fragment (M22O–2H), as observed in the case of 3, from
which three ketene molecules are sequentially lost.
Overall, these data allowed straightforward formulation of
the tetraacetyl derivative as 4. In this structure, the
acetylated OH group at C-3 exerted a lower shielding effect
on the adjacent CH carbon compared to the free ionized
2
0–22
group in 3.
Unfortunately, compound 4 resisted all attempts at crystal-
lization, affording invariably oily or amorphous materials.
Efforts to obtain other derivatives of 3 amenable to
crystallization for X-ray analysis proved unrewarding.
To get some insight into the mechanism of formation of 3
the oxidation of 2 was carried out under a variety of
experimental conditions summarized in Table 3. Formation
of 3 occurs exclusively in the presence of H O or under
2
2
conditions where H O could be generated, e.g. by
2
2
autoxidation of 2. Conversely, in the absence of oxygen or
in the presence of catalase as H O scavenger, formation of
2
2
3
was clearly inhibited, 2 quinone being the only product in
the early phases of the oxidation. Consistent with the
dimeric nature of 3, a high concentration of the catechol-
amine also proved an important requisite for its formation.
1
Figure 1. H NMR spectrum of the oxidation mixture of 2 in deuterated
.1 M phosphate buffer (pH 7.4) at the addition of ferricyanide
0.5 mol equiv.) (plot A), at 5 min (plot B), and at 15 min (plot C) reaction
0
(
Kinetic analysis revealed a relatively slow rate of formation
of 3, compared to the decay of 2, in line with a complex
mechanism in which several intermediates are produced.
tq i um i ne o. nS ey .mbols mark resonances due to the same species: ( p ): 2; (þ): 2
Table 3. Product formation from 2 under different oxidation conditions and
at varying substrate concentrations
Structure 3 arises apparently by reaction of one unit of 2
with a highly oxygenated intermediate derived from 2.
This species could be latter conceivably derived by
H O -induced epoxidation of 2 quinone. To probe this
a
[
2] (mM)
Reaction conditions
Product(s)
2
2
2
2
2
þ
þ
þ
b
1
0
1
1
0
1
1
0
1
1
1
1
0
1
0
0
0
.5
0
0
.5
0
0
.5
0
0
0
0
.5
0
.5
.5
50 mM Fe
50 mM Fe
50 mM Fe , 150 U/mL catalase
3
route, oxidation of 2 with ferricyanide was run in phosphate
1
b
2 quinone
2 quinone
2 quinone
2 quinone
3
3 (traces)
2 quinone
3
2 quinone
3
3
2 quinone
3
2 quinone
3 (traces)
buffered D O at neutral pH and was monitored by H NMR
2
at 400 MHz. (Fig. 1). When the oxidant was added, the
resonances of 2 were gradually replaced by a new set of
signals including two singlets at d 6.39 and 5.53 and two 2H
triplets (one of which partially overlapped to that of 2)
2
þ
3.0 mM Fe /EDTA,10 mM H
2
O
O
2
2
2
þ
3.0 mM Fe /EDTA,10 mM H
2
5 mM NaIO
10 mM
0.5 mM NaIO
4
19
4
which were attributed to 2 quinone (plot A). Proton–
deuterium exchange accounted for the low integrated area
10 mM K
10 mM K
10 mM K
3
3
3
Fe(CN)
Fe(CN)
Fe(CN)
c
6
6
6
, O
, 10 mM H O , O excluded
2
excluded
2
of the high field sp proton signal. The identity of the species
2 2 2
Tyrosinase
Tyrosinase
as 2 quinone was confirmed by HPLC analysis of the
mixture showing the presence of a main component at
R ¼20 min which was converted to 2 (R ¼16 min) on
d
Peroxidase , 10 mM H
2
O
2
d
Peroxidase , 0.5 mM H
t
t
2
O
2
d
treatment with NaBH . At 5 min reaction time, the
4
2 2
Peroxidase , 5 mM H O
resonances of 2 quinone disappeared, whereas those of 2
persisted (plot B). The aliphatic region of the spectrum
displayed a series of multiplets and spread over the region
between d 2.1–3.4 ppm which were attributed to the spin
systems of two aminoethyl chains (COSY analysis). In
a
Product analysis was carried out at 3 h reaction time unless otherwise
stated.
At 12 h reaction time.
b
c
d
7
1
2 U/mmol of 2.
.5 U/mmol of 2.

P. Manini et al. / Tetrahedron 59 (2003) 2215–2221
2219
Scheme 2.
4
addition, singlets at d 6.65 and 3.54, with comparable
integrated areas became apparent (plot B). HPLC analysis
of the mixture confirmed the decay of the quinone with
formation of a species eluting at around 7 min. After
of indoline do not seem to undergo facile epoxidation.
Second, the spreading of the proton resonances of the
ethanamine unit observed in the early stages of the reaction
is indicative of a stereogenic center adjacent to the
ethylamine chain, e.g. due to epoxidation at the C-1 carbon
1
0 min, following nearly complete decay of 2, the pattern of
2
3,24
aliphatic resonances was simplified, disclosing the presence
of an additional singlet at d 3.22 (plot C) ascribable to an
epoxide species. In the later phases of the reaction (2 h) the
smooth formation of 3 became apparent with concomitant
decay of the less retained components (HPLC analysis).
of the catecholamine.
Overall, these pieces of information allowed to propose a
simplified mechanistic scheme for the H O -promoted
oxidation of dopamine via 2, which is outlined in Scheme 2.
2
2
To determine whether epoxidation occurred prior to or
subsequent to cyclization, in another experiment the
aminochrome was generated in the NMR tube by ferri-
cyanide oxidation of 5,6-dihydroxindoline (see Scheme 1)
in deuterated phosphate buffer at pH 7.4. Formation of the
aminochrome was evidenced by complete loss of the
indoline resonances and the appearance of two singlets at
d 6.48 and 5.78, the latter showing a lower integrated area
with respect to the former due to rapid D-exchange, as well
as of two broad 2H signals at d 3.94 and 3.11. Addition of
H O resulted in partial conversion of the red aminochrome
This scheme involves the H O -promoted conversion of
2
2
4
dopamine to 2 quinone, followed by epoxidation and
cyclization leading to an a-hydroxyimine intermediate. This
would undergo nucleophilic attack by 2 through the
quaternary C-1 carbon. Brief inspection of the HOMO of
2 by semiempirical methods (AM1 and PM3) (Table 4)
consistently indicated a relatively large coefficient at the
C-1 carbon.
The peculiar pattern of oxygenation of the adduct would
ensure rapid equilibration with its tautomeric forms as well
as a marked facility to oxidation to give a 1,2,3-trioxo
species. This latter, in the hydrated form, would undergo a
decarboxylative ring contraction in which the driving force
would be provided by relief of the steric tension contributed,
e.g. by the epoxide ring and the encumbering substituent.
This rearrangement would share several features in common
with other known rearrangements, such as the benzil–
2
2
to dark materials without detectable epoxidic proton signals
in the range of d 3–3.5.
Although the inherent complexity of the chemistry pre-
cluded detailed characterization of the transient inter-
mediates, some conclusion can be safely drawn based on
the NMR monitoring experiments. First, aminochrome
1
8
25
species derived by cyclization of 2 quinone or oxidation
benzilic acid rearrangement. It is known that 2 quinone
cyclizes quite slowly compared to dopamine or other related
1
8,26
catecholamines,
competing routes depending on concentration and H O
so it can well partition among different
Table 4. HOMO coefficients for 2 as determined by semiempirical methods
2
2
a
levels.
Carbon
AM1
PM3
1
2
3
4
5
6
0.43
0.11
0.45
0.38
0.02
0.41
0.44
0.11
0.45
0.39
0.02
0.41
3. Experimental
3
.1. General methods
a
†
Numbering refers to catecholamine nomenclature (see footnote ).
For (HR) EI-MS spectra samples were ionized with a 70 eV

2220
P. Manini et al. / Tetrahedron 59 (2003) 2215–2221
beam and the source was taken at 2308C. Main fragmen-
tation peaks are reported with their relative intensities
3.3.1. Isolation of 5-(2-aminoethyl)-2-hydroxy-5-(3-
hydroxy-2-oxotetrahydro-1aH-oxireno[2,3]cyclopenta-
[1,2-b]pyrrol-3a(4H)-yl)cyclohex-2-ene-1,4-dione (3). A
solution of 2 hydrochloride (100 mg, 0.49 mmol) in 0.1 M
phosphate buffer, pH 7.4 (32 mL) was treated with
K Fe(CN) (104 mg, 0.32 mmol) under vigorous stirring
(
percent values are in brackets). (HR) FAB-MS spectra were
1
obtained using glycerol as the matrix. H spectra were
recorded at 400 MHz using a Bruker WM 400 spectrometer;
1
3
t-butyl alcohol (d 1.23) was added as internal standard.
1
C
NMR spectra were recorded at 100 MHz. H, H COSY,
3
6
1
and after 5 min fluxed with a stream of oxygen. After 3 h the
reaction was stopped by addition of 6 M HCl (70 mL) and
the mixture fractionated by ion-exchange chromatography
1
13
1
13
H, C HMQC, H, C HMBC, NOESY and ROESY
experiments were run at 400 MHz using standard pulse
programs from the Bruker library.
þ
(Dowex 50W-X2(H ) 2£40 cm column) using water
(1.4 L), 0.1 M HCl (340 mL), 0.5 M HCl (1 L), 1 M HCl
UV spectra were performed with a Beckmann DU 640
spectrophotometer. Analytical and preparative HPLC was
carried out on a Gilson apparatus equipped with a UV
detector set at 280 nm using a Sphereclone ODS (5m,
(400 mL), 2 M HCl (100 mL) as the eluants. Fractions
eluted with 0.5 M HCl (lmax 293 nm) were collected,
evaporated to dryness under reduced pressure and subjected
to preparative HPLC to afford 3 (31 mg, 20% yield) as a
glassy oil.
2
2
4
.6£250 mm ) or Econosil (10m, 22£250 mm ) column,
respectively. For analytical runs, citrate (0.04 M), phos-
phate (0.04 M), EDTA (0.8 mM), octane-1-sulfonic acid
Compound 3. lmax 293 nm (pH 1); dH (DMSO-d ), see
6
(
7.5 mM)-pH 3.1 containing 11% methanol and 5% aceto-
Table 1; d (D O/DCl) 2.28 (1H, m), 2.41–2.57 (2H, m),
2
H
nitrile was used as the eluant, at a flow rate of 1 mL/min. In
preparative runs elution conditions were: 0.1 M HCOOH-
acetonitrile 98:2 (v/v), flow rate 15 mL/min.
2.91 (1H, m), 3.32 (1H, s), 3.36 (1H, d, J¼1.8 Hz), 3.82
(2H, m), 4.12 (1H, m), 4.26 (1H, m), 4.99 (1H, d, J¼
1.8 Hz); d (DMSO-d ), see Table 1; d (D O/DCl) 20.7
C
6
C
2
(CH ), 31.6 (CH ), 47.4 (CH ), 50.2 (C), 53.7 (CH ), 59.7
(CH), 68.6 (C), 75.3 (C), 80.4 (CH), 174.9 (C), 188.7 (C),
2 2 2 2
TLC and PLC was carried out on silica gel plates (0.25 and
.50 mm, respectively) from Merck. Dopamine, 6-hydroxy-
þ
þ
0
211.1 (C); FAB-MS: m/z 345 [MþNa] , 323 [MþH] .
þ
dopamine hydrochloride, hydrogen peroxide (30% solution
in water), NaIO4 and K Fe(CN) , were from Aldrich
HRFABMS calcd for C H N O [MþH] 323.1243,
1
5 19 2 6
found m/z 323.1239.
3
6
Chemie; mushroom tyrosinase (EC 1.14.18.1), horseradish
peroxidase type II (HRP) (EC 1.11.1.7) and catalase from
bovine liver (EC 1.11.1.6) were from Sigma. 5,6 dihydroxy-
indole, and 5,6-dihydroxyindoline hydrobromide were
prepared as reported. Molecular mechanics and semi-
empirical (AM1/PM3) calculations were carried out with
Hyperchem 5.0 package produced by Hypercube, Inc.
3.3.2. Preparation of 6-(3-acetoxy-2-oxotetrahydro-1aH-
oxireno[2,3]cyclopenta[1,2-b]pyrrol-3a(4H)-yl)-6-(2-
acetylamino)-3,4-diacetoxycyclohexa-2,4-dien-1-one (4).
Compound 3 (50 mg) was treated with acetic anhydride
(2 mL) and pyridine (50 mL) at room temperature for 16 h.
The mixture was taken to dryness and fractionated by TLC
(benzene–acetone 1:1) to give 4 (15 mg, 20% yield) as a
pale yellow oil.
2
7
4
(
Waterloo, Ontario, Canada) 1997.
3
.2. Oxidation of dopamine (1) and 6-hydroxydopamine
(
2)
Compound 4. lmax 318 nm (ethanol); FT-IR (CHCl ) n
3
max
3
689, 3652, 1769, 1752, 1701, 1680, 1674, 1656, 1636,
1602, 1399, 1383, 1307, 1132; d (CD OD), see Table 2; d
C
A solution of 1 or 2 hydrochloride (0.5–10 mM) in 0.1 M
phosphate buffer, pH 7.4, was treated with HRP
H
3
(DMSO-d ), see Table 2; EI MS m/z 490 (5), 456 (48), 414
6
(
1.5 U/mmol of substrate) and hydrogen peroxide (10 mol
(72), 372 (76), 370 (100), 330 (79). HREIMS calcd for
C H N O 490.1587, found m/z 490.1582.
equiv.). In the case of 2, different reaction conditions,
oxidants and additives were used as detailed in Table 2.
When required, the solution of 2 was purged with a stream
of nitrogen for at least 30 min prior to addition of the
solution of the oxidant thoroughly purged with nitrogen.
The reaction course was followed by periodical HPLC
analysis of aliquots of the mixture.
2
3 26 2 10
Acknowledgements
This work was carried out in the frame of the project
Vitiligine: studio sui meccanismi patogenetici e sulle
‘
3
.3. NMR analyses
modalit a` di approccio terapeutico’ (IFO convenzione
n.121, Italian Ministry of Health). We thank the ‘Centro
Interdipartimentale di Metodologie Chimico-Fisiche’
(CIMCF, University of Naples Federico II) for NMR and
mass facilities. We thank Mrs Silvana Corsani for technical
assistance.
A solution of 2 (73 mM) in deuterated 0.1 M phosphate buffer
pH 7.4 was treated with K Fe(CN) (0.5 mol equiv.) and
spectra were run soon after addition and at 5 min time intervals.
The mixture was eventually analyzed at 2 h reaction time.
3
6
In other experiments, a solution of 5,6-dihydroxyindoline
hydrobromide (73 mM) in deuterated 0.1 M phosphate
buffer pH 7.4 was treated with K Fe(CN) (2.0 mol equiv.)
3
6
References
and with H O portionwise up to 2 mol equiv. Spectra were
2
2
run after each addition and at 10 min intervals over 30 min
reaction time.
1. Berman, S. B.; Hastings, T. G. J. Neurochem. 1999, 73,
1127–1137.

P. Manini et al. / Tetrahedron 59 (2003) 2215–2221
2221
2
. Prota, G. Melanins and Melanogenesis; Academic: San Diego,
992.
15. d’Ischia, M.; Costantini, C. Bioorg. Med. Chem. 1995, 3,
923–927.
1
3
. Zhang, F.; Dryhurst, G. J. Med. Chem. 1994, 37, 1084–1098.
. Napolitano, A.; Crescenzi, O.; Pezzella, A.; Prota, G. J. Med.
Chem. 1995, 38, 917–922.
16. Napolitano, A.; Pezzella, A.; Prota, G. Tetrahedron Lett. 1999,
40, 2833–2836.
4
17. Napolitano, A.; Pezzella, A.; Prota, G. Chem. Res. Toxicol.
1999, 12, 1090–1097.
5
6
7
8
9
. Pezzella, A.; d’Ischia, M.; Napolitano, A.; Misuraca, G.; Prota,
G. J. Med. Chem. 1997, 40, 2211–2216.
. Schallreuter Wood, K. U.; Wood, J. M. Dermatology 1996,
1
1
8. Swan, G. A. J. Chem. Soc., Perkin Trans. 1 1976, 339–341.
9. Wehrli, P. A.; Pigott, F.; Fischer, U.; Kaiser, A. Helv. Chim.
Acta 1972, 55, 3057–3061.
1
92, 191–192.
2
2
2
2
2
2
2
2
0. Floris, B. In The Chemistry of Enols; Rappoport, Z., Ed.;
Wiley: New York, 1990; pp 147–305, Chapter 4.
1. Fellmann, P.; Dubois, J.-E. Tetrahedron Lett. 1977, 3,
. Morrone, A.; Picardo, M.; De Luca, C.; Terminali, O.; Passi,
S.; Ippolito, F. Pigment Cell Res. 1992, 5, 65–69.
. Cucchi, M. L.; Frattini, P.; Santagostino, G.; Orecchia, G.
Pigment Cell Res. 2000, 13, 28–32.
2
47–250.
2. Wen, J. Q.; Grutzner, J. B. J. Org. Chem. 1986, 51,
220–4224.
. Prota, G. Forschr. Chem. Org. Naturst. 1995, 64, 94–148.
4
1
0. Swan, G. A. Forschr. Chem. Org. Naturst. 1974, 31, 521–582.
1. Tse, D. C. S.; McCreery, R. L.; Adams, R. N. J. Med. Chem.
3. Napolitano, A.; d’Ischia, M.; Costantini, C.; Prota, G.
Tetrahedron 1992, 48, 8515–8522.
1
1
976, 19, 37–40.
4. Crescenzi, O.; Napolitano, A.; Prota, G.; Peter, M. G.
Tetrahedron 1991, 47, 6243–6250.
5. March, J. Advanced Organic Chemistry; Wiley: New York,
1
1
2. Herlinger, E.; Jameson, R. F.; Linert, W. J. Chem. Soc., Perkin
Trans. 2 1995, 259–263.
3. Garcia-Moreno, M.; Rodriguez-Lopez, J. N.; Martinez-Ortiz,
F.; Tudela, J.; Varon, R.; Garcia-Canovas, F. Arch. Biochem.
Biophys. 1991, 288, 427–434.
1
992; p 1080 and refs. therein.
6. Blank, C. L.; McCreery, R. L.; Wightman, R. M.; Chey, W.;
Adams, R. N. J. Med. Chem. 1976, 19, 178–180.
7. Benigni, J. D.; Minnis, R. L. J. Heterocycl. Chem. 1965, 2,
1
4. Palumbo, A.; Napolitano, A.; Barone, P.; d’Ischia, M. Chem.
Res. Toxicol. 1999, 12, 1213–1222.
3
87–392.