Synthesis of cycleanine mono-N-oxides

Article abstract of DOI:10.1021/np970362t

Oxidation of cycleanine (3) with m-chloroperbenzoic acid gave two diastereomeric N-oxides (1 and 2), and their stereochemistry was unambiguously determined on the basis of spectroscopic evidence. The NMR spectra of synthetic cycleanine mono-N-oxides 1 and 2 were significantly different from those of the natural product previously reported to be cycleanine N-oxide.

Full text of DOI:10.1021/np970362t

J . Nat. Prod. 1998, 61, 253-255
253
Syn th esis of Cyclea n in e Mon o-N-oxid es
Noriaki Kashiwaba,*,† Minoru Ono,† J un Toda,‡ Hideki Suzuki,‡ and Takehiro Sano‡
Research Laboratories, Kaken Shoyaku Co., Ltd., 3-37-10, Shimorenjaku, Mitaka-shi, Tokyo 181, J apan, and Showa College of
Pharmaceutical Sciences, 3-3165, Higashi-Tamagawagakuen, Machida-shi, Tokyo 194, J apan
Received J uly 29, 1997
Oxidation of cycleanine (3) with m-chloroperbenzoic acid gave two diastereomeric N-oxides (1
and 2), and their stereochemistry was unambiguously determined on the basis of spectroscopic
evidence. The NMR spectra of synthetic cycleanine mono-N-oxides 1 and 2 were significantly
different from those of the natural product previously reported to be cycleanine N-oxide.
In connection with the metabolic study of bisbenzyl-
isoquinoline alkaloids, we prepared cycleanine N-oxides
(1 and 2) as standard compounds. These synthetic
compounds, however, gave significantly different NMR
spectra from those of the natural product reported as
cycleanine mono-N-oxide isolated from Synclisia sca-
brida by Ohiri et al.1 and Epinetrum villosum by Parvez
et al.2 In this paper, we describe the preparation and
structure determination of 1 and 2 and present evidence
that the structure of the natural product previously
reported to be cycleanine N-oxide may be in error.
Oxidation of cycleanine (3) with m-chloroperbenzoic
acid (m-CPBA) gave two N-oxides (1 and 2) in yields of
8 and 9%, respectively. The separation of these com-
F igu r e 1. Structures of 1, 2, and 3.
pounds was readily achieved by column chromatography
and preparative TLC. Reduction of 1 and 2 with
sulfurous acid gave only 3, indicating that skeletal
change did not occur during the oxidation process (See
Figure 1).
Ta ble 1. 1H-NMR Data for 1, 2, and 3a
position
1
2
3
1
3
4.73 d (11.0)
3.64 m
4.99 d (10.7)
3.56 m
4.26 d (10.4)
2.90 m
The molecular formula of N-oxide 1 was determined
to be C38H42N2O7 by HRMS. The EIMS showed the
molecular ion peak at m/z 638 (M+, 51%), base ion peak
at m/z 622 (M+ -16), and strong ion peaks at m/z 312
(89%), 311 (36%), 204 (23%), and 190 (16%), suggesting
that 1 is a compound with one oxygen atom (16 mass
3.86 m
3.19 m
3.39 m
6.62 s
3.76 m
2.93 m
3.58 m
6.63 s
2.59 dd
(13.1, 10.7)
3.38 d (13.1)
3.25 m
2.91 m
3.02 m
6.57 s
4
5
R
2.38 dd
(12.2, 11.0)
4.60 d (12.2)
6.27 dd (8.5, 2.1) 6.30 dd (8.5, 2.1)
5.85 dd (8.5, 2.8) 5.91 dd (8.5, 2.8)
6.63 dd (8.5, 2.8) 6.65 dd (8.5, 2.1)
7.27 dd (8.5, 2.1) 7.06 dd (8.5, 2.1)
3.18 s
3.85 s
3.44 s
4.26 d (11.0)
2.92 m
2.52 dd
(12.8, 10.4)
3.22 d (12.8)
6.28 dd (8.2, 2.1)
5.83 dd (8.2, 2.8)
6.60 dd (8.5, 2.8)
7.04 dd (8.5, 2.1)
2.53 s
10
11
13
14
N-CH3
6-OCH3
7-OCH3
1′
1
units) more than that of cycleanine (3). The H-NMR
spectrum (Table 1) was similar to that of 3,3 except that
the signals for H-1, H-3, and an N-methyl proton were
shifted downfield compared to that of 3. The same
tendency was shown in the 13C-NMR spectrum (Table
2), in which the signals for C-1, C-3, and an N-methyl
carbon were observed at a lower magnetic field than
that of 3.3 The downfield shifts of these specific protons
and carbons suggested the presence of an N-oxide group
having electron-withdrawing properties4 and indicated
that 1 is an N-oxide of 3. This structure was also
supported by COLOC experiments.
3.67 s
3.83 s
3.37 s
4.24 d (10.7)
2.92 m
3.25 m
2.91 m
3.03 m
6.59 s
3.81 s
3.40 s
4.26 d (10.4)
2.90 m
3.25 m
2.91 m
3.02 m
6.57 s
2.52 dd
(12.8, 10.4)
3.22 d (12.8)
6.28 dd (8.2, 2.1)
5.83 dd (8.2, 2.8)
6.60 dd (8.5, 2.8)
7.04 dd (8.5, 2.1)
2.53 s
3′
3.27 m
2.90 m
3.02 m
6.58 s
4′
5′
R′
2.53 dd
(12.5, 11.0)
3.23 d (12.5)
6.30 dd (8.5, 2.1) 6.28 dd (8.2, 2.1)
5.81 dd (8.5,2.8) 5.83 dd (8.2, 2.8)
6.60 dd (8.5, 2.8) 6.62 dd (8.2, 2.8)
7.07 dd (8.5, 2.1) 7.04 dd (8.2, 2.1)
2.51 s
2.50 dd
(13.1, 10.7)
3.22 dd (13.1)
10′
11′
13′
14′
The stereochemistry of N-oxide 1 was determined as
follows: in the NOESY experiments, the H-14 (δH 7.27)
signal showed cross peaks to the H-1 (δH 4.73) and H-R
(δH 4.60) signals, indicating that these protons were in
close proximity as shown in Figure 2. Furthermore, in
N′-CH3
2.53 s
3.82 s
3.40 s
6′-OCH3 3.82 s
7′-OCH3 3.40 s
3.81 s
3.40 s
1
the H-NMR spectrum, the H-R and H-14 signals were
a
Values in parentheses are coupling constants (Hz).
* To whom correspondence should be addressed. Phone: 81-422-
44-0108. Fax: 81-422-43-9744.
shifted downfield due to the anisotropic effect of N-oxide
oxygen compared to those (δH 3.22, 7.04) of 3, suggesting
that the N-oxide should have an equatorial orientation
† Kaken Shoyaku Co., Ltd.
‡ Showa College of Pharmaceutical Sciences.
S0163-3864(97)00362-5 CCC: $15.00
© 1998 American Chemical Society and American Society of Pharmacognosy
Published on Web 01/16/1998

254 J ournal of Natural Products, 1998, Vol. 61, No. 2
Ta ble 2. 13C-NMR Data for 1, 2, and 3
Notes
Ta ble 3. 1H-NMR Data for Cycleanine N-Oxide Reported by
Ohiri et al.1
position
1
2
3
position
chemical shifta
1
3
74.61
59.23
73.88
57.51
59.48
44.62
N-CH3
3.32 s
N′-CH3
-OCH3
-OCH3
-OCH3
-OCH3
2.60 s
4
26.51
24.77
24.79
3.84 s
4a
5
6
7
8
8a
R
9
10
11
12
13
14
N-CH3
125.26
108.70
153.30
139.96
143.60
121.99
37.45
128.14
128.88
114.06
153.42
117.77
128.48
56.51
126.48
108.41
152.55
139.13
144.03
121.35
41.73
126.36
128.87
114.64
154.79
118.02
127.60
58.45
129.73
109.18
151.80
138.87
143.64
125.59
37.72
130.43
128.68
113.98
154.10
117.36
128.10
42.37
3.84 s
3.42 s
3.42 s
4.31 d (10)
4.31 d (10)
Ar-H
Ar-H
Ar-H
Ar-H
Ar-H
Ar-H
Ar-H
Ar-H
5.81 dd (8.5, 2.7)
5.81 dd (8.5, 2.7)
6.27 dd (8.5, 2.0)
6.27 dd (8.5, 2.0)
6.58 dd (8.5, 2.7)
6.58 dd (8.5, 2.7)
7.10 dd (8.5, 2.0)
7.10 dd (8.5, 2.0)
a
Values in parentheses are coupling constants (Hz).
6-OCH3
56.02
55.94
55.99
7-OCH3
1′
3′
4′
4a′
5′
6′
7′
8′
8a′
R′
9′
10′
11′
12′
13′
60.26
59.46
44.62
24.72
60.09
59.59
44.63
24.91
60.01
59.48
44.62
24.79
related to the H-R (δH 3.38) signal, and furthermore,
the chemical shift (δH 3.67) of the N-methyl proton of 2
was observed at a lower magnetic field than that (δH
3.18) of cycleanine â-N-oxide (1), indicating that the
N-methyl group of 2 should have an equatorial orienta-
tion. These facts suggested that the N-oxide has an
axial orientation (R-configuration). Therefore, 2 was
confirmed to be cycleanine R-mono-N-oxide.
129.73
109.24
151.89
138.87
143.76
125.28
37.95
131.23
128.64
113.80
154.43
117.45
128.47
42.36
130.08
109.38
151.81
138.73
143.37
125.61
37.56
130.89
128.55
114.26
153.38
117.77
128.14
42.41
129.73
109.18
151.80
138.87
143.64
125.59
37.72
130.43
128.68
113.98
154.10
117.36
128.10
42.37
Although cycleanine (3) possesses two basic nitrogen
atoms in the molecule, only two mono-N-oxides are
theoretically possible, inasmuch as 3 is a symmetrical
molecule consisting of the same two benzylisoquinoline
moieties. The 1H-NMR data (Table 3) of natural cy-
cleanine mono-N-oxide reported by Ohiri et al. and
Parvez et al.5 were not identical with that of either
synthetic cycleanine â-N-oxide (1) or R-N-oxide (2). In
addition, their reported 1H-NMR data seem to be
inconsistent with the structure of cycleanine N-oxide.
14′
N′-CH3
6′-OCH3
7′-OCH3
56.02
60.06
56.01
60.09
55.99
60.01
1
The H-NMR data suggested that the compound has a
symmetrical structure such as 3, since the signals due
to methoxy and aromatic protons, except for two differ-
ent N-methyl proton signals, were observed as two sets
of signal patterns (Table 3). Probably, on the basis of
this symmetrical character, Ohiri et al. assigned this
compound as cycleanine mono-N-oxide. However, cy-
cleanine mono-N-oxide is not a symmetrical molecule,
and therefore, it is not necessarily required that each
pair of protons have the same chemical shift. Moreover,
Ohiri et al. did not describe the signal corresponding to
the H-1 proton, which should shift downfield from that
of the H-1′ proton due to the electron-withdrawing effect
of the N-oxide. The two proton signals at δH 4.31,
though they were not assigned, are not attributable to
either the H-1 or H-1′ proton because it is unreasonable
that these protons have the same chemical shift. Ohiri
et al. reported that cycleanine N-oxide was prepared by
oxidation of cycleanine with hydrogen peroxide, and
identification of the synthetic cycleanine N-oxide was
performed only by comparison of TLC behavior and color
F igu r e 2. Stereochemistry of N-oxides 1 and 2.
or â-configuration as shown in Figure 2. Therefore, 1
was determined to be cycleanine â-mono-N-oxide.
The molecular formula of the other N-oxide (2) was
established as C38H42N2O7 by HRMS, and the EIMS
showed a similar fragmentation pattern to that of
cycleanine â-N-oxide (1). The 1H- and 13C-NMR spectra
(Tables 1 and 2) were similar to those of cycleanine (3),
except for the signals for H-1, H-3, an N-methyl proton,
C-1, C-3, and an N-methyl carbon. These data and the
COLOC experiments suggested that the structure of 2
is also consistent with an mono-N-oxide derivative of
3.
1
reaction with the natural product.1 However, the H-
1
As to the stereochemistry of N-oxide 2, the H-NMR
NMR data (Table 3) of the natural product, if they are
correct, are not consistent with the structure of cyclea-
nine N-oxide as described above.
spectrum showed that the H-4 (δH 3.58) signal was
shifted downfield from that (δH 3.02) of cycleanine (3),
indicating that this downfield shift should be due to the
anisotropic effect of the N-oxide oxygen. Thus, the H-4
and oxygen are in quasi 1,3-diaxial positions as shown
in Figure 2. On the other hand, in the NOESY experi-
ments, the N-methyl proton (δH 3.67) signal was cor-
Exp er im en ta l Section
Gen er a l Exp er im en ta l P r oced u r es. Optical rota-
tions were determined on a DIP-1000 (J ASCO) spec-

Notes
J ournal of Natural Products, 1998, Vol. 61, No. 2 255
trometer in CHCl3. IR spectra were recorded on an FT/
IR-5000 (J ASCO) spectrometer as KBr pellets, and data
are given in cm-1. UV spectra were measured on a
Ubest-35 (J ASCO) spectrometer in MeOH, and data are
given as λmax nm (log ꢀ). NMR spectra were taken on a
J NM-R500 (J EOL) (500 MHz for 1H and 125 MHz for
13C) spectrometer in CDCl3 with TMS as an internal
standard, and the chemical shifts are given in δ values.
MS were performed on J MS-D300 (J EOL) spectrometers
at 30 eV, and data are given in m/z (rel int). Column
chromatography was performed on Wakogel C-200
(Wako Pure Chemical Industries, Ltd.). Preparative
TLC was done on precoated Si gel 60 F254 (0.25-mm
thick) plates (Merck).
Cyclean in e r-N-oxide (2): amorphous powder; [R]24
D
-38° (c 0.15); IR 1607, 1584, 1510, 1454, 1421, 1352,
1299, 1220, 1172, 1118; UV 283 sh (3.48), 276 (3.53);
EIMS 638 (M+, 96), 622 (100), 607 (18), 312 (85), 311
(41), 204 (28), 190 (15); HRMS 638.2997 (C38H42N2O7
requires 638.2992).
Red u ction s of 1 a n d 2 w ith H2SO3. To 5% H2SO3
solution (20 mL), 1 (10 mg) was added and stirred for 8
h at room temperature. The solution was diluted with
H2O, basified with NH4OH, and extracted with CHCl3.
The extract was washed H2O and dried over Na2SO4.
After removal of the solvent in vacuo, the residue was
subjected to preparative TLC [with EtOAc-Et2NH (9:
1)] to afford 3 (7 mg, 70%), which is identical with an
authentic sample of 3 by comparison of TLC behavior
and IR spectrum. The reduction of 2 was performed in
the same manner as that of 1.
Oxid a tion of 3 w ith m -CP BA. A mixture of 3 (500
mg) and m-CPBA (220 mg as 80% purity) in dry CH2-
Cl2 (20 mL) was stirred for 3 h at room temperature.
The solution was extracted with CH2Cl2, washed with
NaHCO3 following H2O, and dried over Na2SO4. After
removal of the solvent in vacuo, the residue was
chromatographed on Si gel using 5% MeOH-CHCl3 as
eluent to give 3 (295 mg, 59%). The material eluted
with 10% MeOH-CHCl3 was subjected to preparative
TLC [with MeOH-CHCl3 (2:3)] to yield 1 (42 mg, 8%)
and 2 (48 mg, 9%).
Refer en ces a n d Notes
(1) Ohiri, F. C.; Verpoorte, R.; Svendsen, A. B. Planta Med. 1983,
47, 87-89.
(2) Parvez, M.; Rahman, A.; Rahman, S.; Ogbeide, O. N. J . Chem.
Soc. Pak. 1994, 16, 257-260.
(3) Kashiwaba, N.; Morooka, S.; Kimura, M.; Murakoshi, Y.; Ono,
M.; Toda, J .; Suzuki, H.; Sano, T. Chem. Pharm. Bull. 1997, 45,
470-475.
Cyclea n in e â-N-oxid e (1): amorphous powder; [R]24
D
(4) Phillipson, J . D.; Handa, S. S. Phytochemistry 1975, 14, 2683-
2690.
-17° (c 0.13); IR 1605, 1582, 1510, 1454, 1421, 1346,
1303, 1218, 1172, 1122; UV 283 sh (3.52), 276 (3.58);
EIMS 638 (M+, 52), 622 (100), 607 (29), 312 (89), 311
(36), 204 (23), 190 (16); HRMS 638.2993 (C38H42N2O7
requires 638.2992).
(5) The spectroscopic data ([R], UV, and 1H-NMR) reported by
Parvez et al.2 were completely in agreement with those of Ohiri
et al.1
NP970362T