A New Insight in the Biosynthesis of Pheomelanins: Characterization of a Labile 1,4-Benzothiazine Intermediate

Article abstract of DOI:10.1021/jo981018j

It has long been known that pheomelanins, the distinctive pigments of red hair and Celtic skin, arise by the oxidative cyclization of cysteinyldopas, mainly the 5-S-isomer 1, via 1,4-benzothiazines. However, the nature and reactivity of these intermediates have remained poorly defined. In an reexamination of the oxidation of 1, in aqueous buffers at physiological pHs, a hitherto unknown labile intermediate was identified and formulated as the 3-hydroxy-3,4-dihydro-1,4-benzothiazine 5 based on direct NMR analysis of the reaction mixture and conversion to the stable benzothiazines 3 and 4 under different conditions. Structure 5 was further supported by oxidation of the model compound 6 leading to the analogous more stable 2,2-dimethyldihydro-1,4-benzothiazine 9.

Full text of DOI:10.1021/jo981018j

J . Org. Chem. 1999, 64, 3009-3011
3009
A New In sigh t in th e Biosyn th esis of P h eom ela n in s:
Ch a r a cter iza tion of a La bile 1,4-Ben zoth ia zin e In ter m ed ia te
Alessandra Napolitano, Sofia Memoli, and Giuseppe Prota*,†
Department of Organic and Biological Chemistry, University of Naples “Federico II”,
Via Mezzocannone 16, I-80134 Naples, Italy
Received May 29, 1998 (Revised Manuscript Received March 1, 1999)
It has long been known that pheomelanins, the distinctive pigments of red hair and celtic skin,
arise by the oxidative cyclization of cysteinyldopas, mainly the 5-S-isomer 1, via 1,4-benzothiazines.
However, the nature and reactivity of these intermediates have remained poorly defined. In an
reexamination of the oxidation of 1, in aqueous buffers at physiological pHs, a hitherto unknown
labile intermediate was identified and formulated as the 3-hydroxy-3,4-dihydro-1,4-benzothiazine
5 based on direct NMR analysis of the reaction mixture and conversion to the stable benzothiazines
3 and 4 under different conditions. Structure 5 was further supported by oxidation of the model
compound 6 leading to the analogous more stable 2,2-dimethyldihydro-1,4-benzothiazine 9.
In tr od u ction
Pheomelanins, the distinctive pigments of red human
hair, 1-3 have become in recent years the focus of renewed
interest because of their implication as major contribu-
tory factors of the abnormal susceptibility of red haired,
fair skinned Caucasians to acute and chronic UV dam-
age.4 Interpretation of the postulated effects on molecular
basis has greatly been hampered by the poor knowledge
of the chemistry of these pigments.5
Biosynthetic studies provided evidence that pigment
formation involves oxidative polymerization of cysteinyl-
dopas, mainly the 5-S-isomer 1.3,6 Under biomimetic
conditions, oxidation of 1 was shown to proceed through
the 3,4-dihydro-1,4-benzothiazine-3-carboxylic acid 2.7
Further investigations led to formulation of more complex
oxidation pathways involving the intermediacy of a 1,4-
benzothiazine non-carboxylated at the 2-position, isolated
after reduction as 3.8-10
In a reexamination of the oxidation of 1 in aqueous
buffer at physiological pHs, we have now identified in
the early stages of the reaction a new transient 1,4-
benzothiazine intermediate which has been characterized
by analysis of the spectral features and chemical reactiv-
ity, as well as by use of model compounds.
† Phone: +39-81-7041249. Fax: +39-81-5521217. E-mail: prota@
cds.unina.it.
(1) Prota, G. Melanins and Melanogenesis; Academic Press: San
Diego, 1992.
(2) Thomson, R. H. Angew. Chem., Int. Ed. Engl. 1974, 13, 305-
312.
Resu lts a n d Discu ssion
Oxidation of 1 was initially performed by potassium
ferricyanide at pH 7.2, and the mixture was examined
by HPLC with or without reductive treatment in the first
few minutes of the reaction. The dihydrobenzothiazine
acid 2 was present in both mixtures, but the untreated
one contained another significant species whose chro-
matographic properties did not match those of any of the
known intermediates of in vitro pheomelanogenesis.3,9
Moreover, it was not present in the reduced mixture
which showed the dihydrobenzothiazine 3 as the major
component. Similar reaction patterns were obtained
using other oxidizing agents including sodium periodate,
the Fenton reagent,11 and the enzymatic system peroxi-
dase/H2O2. The HPLC fraction of the untreated oxidation
mixture corresponding to the unknown peak showed an
absorption maximum at 302 nm, which, on acidification,
was rapidly converted to that of tricochrome F (587 nm),
exhibiting the characteristic reversible pH-dependent
shift.12 Alkalinization resulted in a slow batochromic shift
of the absorption maximum to 320 nm, corresponding to
(3) Prota, G. Fortsch. Chem. Org. Naturst. 1995, 64, 94-148.
(4) (a) Thody, A. J .; Higgins, E. M.; Wakamatsu, K.; Ito, S.; Burchill,
S. A. Marks, J . M. J . Invest. Dermatol. 1991, 97, 340-344. (b) Vincensi,
M. R.; d’Ischia, M.; Napolitano, A.; Procaccini, E. M.; Riccio, G.;
Monfrecola, G.; Santoianni, P.; Prota, G. Melanoma Res. 1998, 8, 55-
58.
(5) Prota, G.; d’Ischia, M.; Napolitano, A. In The Pigmentary
System: Its Physiology and Pathophysiology; Nordlund, J . J ., Boissy,
R. E., Hearing, V. J ., King, R. A., Ortonne, J . P., Eds.; Oxford
University Press: New York, 1998; pp 307-333.
(6) (a) Prota, G.; Nicolaus, R. A. Gazz. Chim. Ital. 1967, 97, 665-
684. (b) Ito, S.; Prota, G. Experientia 1977, 33, 1118-1119.
(7) Prota, G.; Crescenzi, S.; Misuraca, G.; Nicolaus, R. A. Experientia
1970, 26, 1508-1509.
(8) Napolitano, A.; Costantini, C.; Crescenzi, O.; Prota, G. Tetrahe-
dron Lett. 1994, 35, 6365-6368.
(9) Napolitano, A.; Memoli, S.; Crescenzi, O.; Prota, G. J . Org. Chem.
1996, 61, 598-604.
(10) Costantini, C.; Crescenzi, O.; Prota, G.; Palumbo, A. Tetrahe-
dron 1990, 46, 6831-6838.
(11) (a) Maskos, Z.; Rush, J . D.; Koppenol, W. H. Arch. Biochim.
Biophys. 1992, 296, 521-529. (b) Napolitano, A.; Memoli, S.; Nappi,
A. J .; d’Ischia, M.; Prota, G. Biochim. Biophys. Acta 1996, 1291, 75-
82.
(12) Prota, G.; Scherillo, G.; Petrillo, O.; Nicolaus, R. A. Gazz. Chim.
Ital. 1969, 90, 1193-1207.
10.1021/jo981018j CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/03/1999

3010 J . Org. Chem., Vol. 64, No. 9, 1999
Napolitano et al.
Sch em e 1
another species identified as the 7-(2-amino-2-carboxy-
ethyl)-5-hydroxy-3-oxo-3,4-dihydro-2H-1,4-benzothiazine
(4),13 while sodium borohydride reduction led to the
dihydrobenzothiazine 3.8
Attempts to isolate the major component of the non-
reduced mixture met invariably with failure despite very
careful experiments, owing to its marked instability. To
circumvent these difficulties, oxidation of 1 was run in
phosphate buffer in D2O at neutral pHs and monitored
1
by H NMR at 400 MHz. At the addition of the oxidant,
the resonances of 1 progressively diminished, being
replaced by new sets of signals, among which those
attributable to a major component could be readily
recognized. The most prominent feature of the spectrum
was a downfield multiplet at δ 5.44 which showed
coupling with CH2 alanyl chain resonances at around 3
ppm (COSY evidence) and direct CH bond correlation
with a signal at δ 73.20. Coupled with consideration of
the reactivity, spectral analysis led to formulation of the
intermediate as the 3-hydroxy-3,4-dihydro-1,4-benzothi-
azine 5. Unfortunately, all efforts toward a more complete
characterization were thwarted by the low concentrations
which could be generated in the NMR tube and by its
relatively short half-life (about 30 min).
Further support of structure 5 for the labile intermedi-
ate was gained by oxidation of the model thioalkylcat-
echol 6. The major component of the oxidation mixture
exhibited an absorption maximum at 302 nm closely
resembling that of 5. On reduction with NaBH4, it was
rapidly converted into the dihydro-1,4-benzothiazine 7,
while exposure to basic or acidic media led to smooth
formation of the 3-oxo-dihydro-1,4-benzothiazine 8.
catechol 6, which readily undergoes reduction to the
dihydrobenzothiazine 7. The route depicted in Scheme 1
to the 1,4-benzothiazin-3-one 4/8 by aerial oxidation in
acidic or alkaline media of the 3-hydroxy-3,4-dihydro-1,4-
benzothiazine 5/9, followed by rearrangement with H-
shift and tautomerization, accounts for the reactivity
observed for this labile intermediate and represents an
alternative to that proposed previously in the literature.14
In the case of the benzothiazine intermediate 10 derived
from 1, such a reaction path is superseded by the acid
dimerization to trichochrome F.12
Exp er im en ta l Section
Gen er a l. Low- and high-resolution fast atom bombardment
mass spectrometry (FAB MS) was carried out with a double
focusing spectrometer equipped with a cesium gun operating
at 25 kV using a glycerol/thioglycerol matrix. 1H (13C) NMR
spectra were recorded at 400.1 (100.6) or 270.1 (67.9) MHz,
using tert-butyl alcohol or dioxane as internal standards.
COSY experiments were run using a standard Bruker pulse
program. 2D proton-carbon shift correlation experiments were
performed with gradient pulse using a delay corresponding to
J values of 140 Hz. HPLC analyses were run under conditions
previously described.15 Ion exchange chromatography was
carried out using a Dowex 50W-X2 (H+ form, 200-400 mesh)
resin. Horseradish peroxidase (donor: H2O2 oxidoreductase,
EC 1.11.1.7) type II (167 U/mg, RZ E430/E275 ) 2.0) was used.
3-[(R)-2-Amino-2-carboxyethylthio]-5-[(S)-2-amino-2-carboxy-
ethyl]-1,2-dihydroxybenzene (1) was prepared as previously
described.16
HPLC purification of the mixture obtained by oxidation
of 6, carefully avoiding strong acid or alkaline conditions,
afforded eventually a fraction corresponding to the major
component, which was characterized as the carbinola-
mine 9, analogous to 5 obtained from 1, on the basis of
spectrometric and NMR analysis.
The several facets of the reactivity of the benzothiazine
intermediate evidenced by HPLC analysis of the oxida-
tion mixture of 1 at the early stages may be reconciled
in terms of an equilibrium between the 3-hydroxy-3,4-
dihydro-1,4-benzothiazine 5 and the 2H-1,4-benzothia-
zine 10 as a result of reversible addition of water to the
imine function (Scheme 1). A similar behavior is observed
for the benzothiazine 11 generated from the model
3-[(S)-2-Am in o-2-car boxy-1,1-dim eth yleth ylth io]-5-[(S)-
2-a m in o-2-ca r boxyeth yl]-1,2-d ih yd r oxyben zen e (6). A so-
lution of L-dopa (1.97 g, 10 mmol) in 2 M sulfuric acid (50 mL)
(14) Crescenzi, S.; Misuraca, G.; Novellino, E.; Prota, G. La Chimica
e L’Industria 1975, 57, 392-393.
(15) Napolitano, A.; Crescenzi, O.; Pezzella, A.; Prota, G. J . Med.
Chem. 1995, 38, 917-922.
(13) Prota, G.; Scherillo, G.; Nicolaus, R. A. Gazz. Chim. Ital. 1968,
98, 495-510.
(16) Chioccara, F.; Novellino, E. Synth. Commun. 1986, 16, 967-
971.

New Insight in the Biosynthesis of Pheomelanins
J . Org. Chem., Vol. 64, No. 9, 1999 3011
was treated with cerium ammonium nitrate (10.96 g, 20 mmol)
in 2 M sulfuric acid (100 mL) under stirring. The resultant
yellow-orange mixture was added to a solution of D-penicil-
lamine (6.0 g, 40 mmol) in 2 M sulfuric acid (50 mL). After 10
min, the reaction mixture was purified by ion-exchange
chromatography (2 × 60 cm column), using water (200 mL),
0.5 M HCl (1.5 L), and 3 M HCl (2.0 L) as the eluant. Fractions
from 3 M HCl showing an absorption maximum at 295 nm
were collected and evaporated to dryness under reduced
pressure to afford 6 dihydrochloride (2.10 g, 50% yield) as a
yellow glassy oil, homogeneous to HPLC (0.1 M phosphoric
acid, 0.1 M methanesulfonic acid, 0.1 mM EDTA, pH 3.1,
eluant A; 0.05 M sodium citrate pH 4.0/methanol 85:15, eluant
B): UV λmax (0.1 M HCl) 295 nm; 1H NMR (D2O) δ 1.32 (s,
3H), 1.36 (s, 3H), 3.08 (dd, 2H, J ) 12.2, 6.6 Hz), 3.68 (s,
1H),4.20 (t, 1H, J ) 6.6 Hz), 6.88 (s, 1H × 2); 13C NMR (D2O)
δ 18.76 (CH3), 23.26 (CH3), 31.32 (CH2), 46.56 (C), 50.63 (CH),
56.85 (CH), 111.94 (C), 115.85 (CH), 122.83 (C), 127.92 (CH),
141.27 (C), 143.65 (C), 166.18 (C), 167.84 (C); FAB MS m/z
345 (MH+, 100); exact mass calcd for C14H21N2O6S 345.1120,
found 345.1115.
7-(2-Am in o-2-car boxyeth yl)-3,5-dih ydr oxy-2,2-dim eth yl-
3,4-d ih yd r o-2H-1,4-ben zoth ia zin e (9). A solution of 6 di-
hydrochloride (100 mg, 0.24 mmol) in 0.05 M phosphate buffer
pH 7.2 (72 mL), at 310 K, was treated with K3Fe(CN)6 (0.8
molar equiv). HPLC analysis of the reaction mixture after 2
min was found to consist of a predominant product eluted at
tR 16 min (eluant D). Preparative HPLC fractionation (0.05
M formic acid/methanol 55:45, eluant E, flow rate 10 mL/min)
of the mixture after lyophilization afforded 9 (15 mg, 21% yield)
1
as a glassy oil: UV λmax (H2O) 302 nm; H NMR (D2O) δ 1.33
(s, 3H), 1.45 (s, 3H), 3.02 (m, 1H), 3.17 (m, 1H), 3.98 (m, 1H),
4.85 (s, 1H), 6.65 (s, 2H);13C NMR (D2O) δ 19.58 (CH3), 22.85
(CH3), 31.55 (CH2), 40.05 (C) 51.96 (CH), 73.78 (CH), 108.10
(CH), 115.77 (CH), 120.90 (C), 140.06 (C), 149.01 (C), 169.94
(C); FAB MS m/z 299 (MH+, 100); exact mass calcd for
C
13H19N2O4S 299.1065, found 299.1072.
7-(2-Am in o-2-ca r b oxyet h yl)-5-h yd r oxy-2,2-d im et h yl-
3,4-d ih yd r o-2H-1,4-ben zoth ia zin e (7). For isolation of 7, the
oxidation of 1 was repeated under the above conditions, and
the reaction mixture was treated with sodium borohydride (50
mg) and, after additional 3 min, acidified to pH 2 with 2 M
HCl. HPLC fractionation of the reduced mixture (eluant E)
yielded 7 (40 mg, 59% yield) as a glassy oil, homogeneous to
HPLC (eluant D and E): UV λmax (0.1 M HCl) 288, 292 nm;
1H NMR (0.1 M DCl) δ 1.45 (s, 6H), 3.08 (dd, 1H, J ) 6.8, 14.6
Hz), 3.16 (dd, 1H, J ) 14.6, 5.8 Hz), 3.52 (s, 2H), 4.22 (dd, 1H,
J ) 6.8, 5.8 Hz,), 6.65 (s, 1H), 6.66 (s, 1H); 13C NMR (0.1 M
DCl) δ 27.42 (CH3), 36.58 (CH2), 40.55 (C), 53.87 (CH2), 55.28
(CH), 114.13 (CH), 115.32 (C), 120.97 (CH), 130.07 (C), 137.39
(C), 152.32 (C), 172.75 (C); FAB MS m/z 283 (MH+, 100); exact
mass calcd for C13H19N2O3S 283.1116, found 283.1110.
Oxid a tion of 1 or 6. To a solution of 1 or 6 (1 mM) in
phosphate buffer (0.05 M) pH 7.2 at 310 K was added K3Fe-
(CN)6 (0.8 molar equiv) under vigorous stirring. The reaction
course was monitored by HPLC analysis (eluant B in the case
of 1, 0.05 M sodium citrate pH 4.0/methanol 80:20, eluant D
in the case of 6). Aliquots of the reaction mixture were
periodically withdrawn, treated with NaBH4 (0.5 mg/mg
substrate) when necessary, and analyzed. Oxidation of 1 by
11b
peroxidase/H2O2,9 sodium periodate, and Fe2+/EDTA-H2O2
was run under similar conditions.
7-(2-Am in o-2-ca r boxyeth yl)-5-h yd r oxy-3-oxo-3,4-d ih y-
d r o-2H-1,4-ben zoth ia zin e (4) a n d 7-(2-Am in o-2-ca r boxy-
eth yl)-3,5-dih ydr oxy-3,4-dih ydr o-2H-1,4-ben zoth iazin e (5).
The mixture obtained by K3Fe(CN)6 oxidation of 1 as above
was fractionated at 2 min reaction time by preparative HPLC
(eluant B, flow rate 10 mL/min). The peak corresponding to
the species at tR 8 min was collected, and the UV spectrum
was taken immediately (λmax, HPLC eluant, pH 4: 302 nm).
Aliquots of the HPLC fraction were treated with (a) NaBH4,
after careful neutralization with 0.1 M NaOH, (b) 1 M NaOH
to pH 13, and (c) 1 M HCl to pH 1. The mixture thus obtained
was then subjected to HPLC, spectrophotometric, and in the
case of the acidified mixture paper chromatographic (eluant:
butanol-acetic acid-water-concentrated HCl, 20:30:50:1 or
2-propanol-formic acid-concentrated HCl, 70:30:1) analysis.
Identification of 4 in the oxidation mixture after alkalinization
as in (b) followed from HPLC analysis (eluants A and B).
Preparative HPLC (eluant B) and subsequent desalting by ion-
exchange chromatography (eluant HCl gradient) afforded 4 as
a glassy oil: UV λmax (0.1 M HCl) 300 nm; (0.1 M NaOH) 320
nm; 1H NMR (D2O) δ 3.06 (dd, 1H, J ) 15.7, 8.0 Hz), 3.19 (dd,
1H, J ) 15.7, 6.0 Hz), 3.41 (s, 2H), 4.26 (dd, 1H, J ) 8.1, 6.0
Hz), 6.65 (s, 1H), 6.75 (s, 1H), 7.90 (bs, 1H); 13C NMR (D2O) δ
32.19 (CH2), 37.65 (CH2), 56.56 (CH), 116.86 (CH), 122.09 (CH),
124.27 (C), 126.54 (C) 132.98 (C) 147.51 (C), 170.32 (C), 174.01
(C). In other experiments, the oxidation of 1 was repeated in
a NMR cuvette: to a solution of 1 (3 mM) in deuterated 0.2 M
phosphate buffer pH 7.2 was added K3Fe(CN)6 (0.8 equiv), and
7-(2-Am in o-2-ca r boxyeth yl)-5-h yd r oxy-2,2-d im eth yl-3-
oxo-3,4-d ih yd r o-2H-1,4-ben zoth ia zin e (8). The oxidation
mixture of 6 obtained as for isolation of 9 was acidified to pH
1 with 2 M HCl at 2 min reaction time and allowed to stand
at room temperature for additional 15 min. After removal of
the solvents under reduced pressure, the mixture was frac-
tionated by preparative HPLC (eluant: 0.05 M formic acid/
acetonitrile 75:25, eluant F, flow: 10 mL/min) to give 8 (22
mg, 31% yield) as a glassy oil, homogeneous by HPLC (eluant
F and D): UV λmax (0.1 M HCl) 300 nm, (0.1 M NaOH) 320
nm; 1H NMR (0.1 M DCl) δ 1.58 (s, 6H), 3.18 (dd, 1H, J )
14.7, 7.8 Hz), 3.28 (dd, 1H, J ) 14.7, 7.4 Hz), 4.35 (t, 1H, J )
7.4 Hz), 6.84 (s, 1H), 6.91 (s, 1H), 7.90 (bs, 1H); 13C NMR (DCl
0.1 M) δ 25.37 (CH3), 37.75 (CH2), 41.51 (C), 55.80 (CH), 117.03
(CH), 122.05 (CH), 131.75 (C), 138.76 (C), 141.44 (C), 152.89
(C), 173.15 (C); 175.65 (C); FAB MS m/z 297 (MH+, 100); exact
mass calcd for C13H17N2O4S 297.0909, found 297.0897.
Ack n ow led gm en t. This work was supported by
grants from Ministero della Pubblica Istruzione (Rome)
and Consiglio Nazionale delle Ricerche (CNR, Rome).
We thank the Centro Interdipartimentale Metodologie
Chimico-Fisiche (CIMCF, University of Naples) for
NMR facilities and Miss Silvana Corsani for technical
assistance.
1
spectra were recorded at 2 min intervals. H NMR (D2O) of 5:
δ 3.00 (m, 4H), 3.88 (m, 1H), 5.44 (m, 1H), 6.56 (s, 1H), 6.59
(s, 1H).
J O981018J