Novel cytotoxic, alkylated hydroquinones from Lannea welwitschii

Article abstract of DOI:10.1021/np960435t

Two novel natural products, lanneaquinol (1) and 2'(R)- hydroxylanneaquinol (2), were isolated from the organic extract of the plant Lannea welwitschii (Hiern) Engl. Their structures were solved by spectroanalytical methods and confirmed by comparison to

Full text of DOI:10.1021/np960435t

+
+
116
J . Nat. Prod. 1997, 60, 116-121
Novel Cytotoxic, Alk yla ted Hyd r oqu in on es fr om La n n ea welwitsch ii
Amiram Groweiss,†,‡ J ohn H. Cardellina II,‡ Lewis K. Pannell,§ Duangchan Uyakul,§ Yoel Kashman,‡,| and
Michael R. Boyd*,‡
Laboratory of Drug Discovery Research and Development, Developmental Therapeutics Program, Division of Cancer Treatment,
Diagnosis, and Centers, National Cancer Institute-Frederick Cancer Research and Development Center, P.O. Box B, Building
1052, Room 121, Frederick, Maryland 21702-1201, and Laboratory of Analytical Chemistry, National Institute of Diabetes and
Digestive Kidney Disease, 9000 Rockville Pike, Building 8, Room B2A23, Bethesda, Maryland 20892
Received May 6, 1996X
Two novel natural products, lanneaquinol (1) and 2′(R)-hydroxylanneaquinol (2), were isolated
from the organic extract of the plant Lannea welwitschii (Hiern) Engl. Their structures were
solved by spectroanalytical methods and confirmed by comparison to synthetic models. The
absolute configuration of 2 was determined by the modified Mosher method. Both compounds
exhibited modest cytotoxicity against the NCI panel of 60 human tumor cell lines. The
structures of two isomeric 4,5-dihydroxy-5-alkyl-2-cyclohexenones (7 and 8), which appear to
be biogenetic precursors of 1 and 2, were also elucidated.
An organic extract of the plant Lannea welwitschii
(Hiern) Engl. (family Anacardiaceae) was selected for
bioassay-guided fractionation based upon modest po-
tency and preliminary indications of an unusual pattern
of differential cytotoxicity in the NCI 60-cell line, human
disease oriented, in vitro tumor screen.1,2 The crude
CH2Cl2-CH3OH (1:1) extract was subjected to a solvent-
solvent partitioning protocol, which concentrated the
antitumor activity in the CCl4-soluble fraction. Sequen-
tial bioassay-guided chromatography on Sephadex LH-
20 and Si gel gave two novel metabolites, lanneaquinol
(1) and 2′(R)-hydroxylanneaquinol (2), in 1.0% and 0.2%
of the crude organic extract, respectively.
The UV absorption at λmax 294 nm was indicative of
a phenolic moiety. Examination of the 13C NMR re-
vealed an aromatic ring containing three protons, while
the other substituents were two oxygen atoms (δC 149.2
and 147.3 ppm) and a carbon atom (δc 130.1 ppm).
Further interpretation led to the conclusion that the
ring was substituted by two phenol groups and by a
linear alkyl chain of 17 carbons. This chain contained
the remaining (disubstituted) double bond. In order to
complete the structure elucidation, two points had to
be solved: the substitution pattern of the aromatic ring
and the location and geometry of the double bond in the
alkyl side chain.
The 1H NMR spectrum revealed that the three
aromatic protons (a proton at δ 6.53 coupled ortho and
meta, respectively, to protons resonating at δ 6.62 and
6.60) comprised a 1,2,4-trisubstituted system, which
could be presented by three possible geometrical isomers
(1a -c). A comparison of the spectral data of lan-
neaquinol and a synthetic analogue, 4-n-hexyl-1,3-
dihydroxybenzene, quickly eliminated the possibility of
1b. There were many chemical shift differences, espe-
cially in the 13C-NMR spectrum, between the natural
product and this 1,3-dihydroxybenzene (“resorcinol”)
derivative, for example, the resonances at δC 107.6 and
102.8 ppm, which had no corresponding signals in
lanneaquinol (Table 1). On the other hand, spectral
data alone could not readily differentiate between
2-alkyl-hydroquinone 1a and the catechol derivative 1c
(4-alkyl-1,2-dihydroxybenzene).
Resu lts a n d Discu ssion
The molecular composition of lanneaquinol, C23H38O2,
was determined by HREIMS (m/z 346.2867), a composi-
tion also confirmed by CIMS and in accord with five
degrees of unsaturation. The structure was elucidated
mainly by interpretation of the spectral data, especially
the 1H- and 13C-NMR and 2D-NMR spectra, and by
spectral comparisons with synthetic analogues.
The first evidence that favored isomer 1a over 1c
came from the HMBC experiment. All the correlations
in isomer 1a lay within the expected two and three
bonds, while for 1c some crucial connectivities could
only be explained by less acceptable four-bond correla-
tions. Two important examples were the correlations
between one of the phenolic carbons (δC 149.2) and the
* To whom correspondence should be addressed.
† On sabbatical leave from TAMI (IMI) Institute for Research and
Development, Haifa Bay 26111, Israel.
‡ Laboratory of Drug Discovery Research and Development.
§ Laboratory of Analytical Chemistry, NIDDK, NIH, Bethesda, MD
20892.
| On sabbatical leave from Department of Chemistry, Tel Aviv
University, Tel Aviv 69978, Israel.
X Abstract published in Advance ACS Abstracts, February 1, 1997.
S0163-3864(96)00435-1 This article not subject to U.S. Copyright. Published 1997 by the Am. Chem. Soc. and the Am. Soc. of Pharmacogn.

+
+
Cytotoxic Hydroquinones from Lannea
J ournal of Natural Products, 1997, Vol. 60, No. 2 117
Ta ble 1. NMR Data for Compounds 1 and 2a
13C-NMR data
1H-NMR data
J (Hz)
C#
δ (ppm)
149.2
147.3
130.1
129.9, 129.8
116.8
116.0
113.3
30.1
mult.
HMBC
δ (ppm)
mult.
COSY
NOE
Lanneaquinol (1)
4
1
2
C
C
C
6.62, 6.60, 6.53
6.62, 6.53, 2.52
6.62, 6.60, 2.52
8′ & 9′
CH
CH
CH
CH
CH2
2, CH2
CH2
CH3
5.33
6.60
6.62
6.53
2.52
1.99
1.25
0.86
m, 2H
d
d
1.99
6.53
6.53
3
6
5
1′
6.53, 2.52
2.9
8.3
8.3, 2.9
7.8
2.52
6.60
6.60
5.33
dd
6.62, 6.60
1.56
5.33, 1.2
m, 2H
m, 4H
br s, 22H
t, 3H
6.60, 1.56
7′ & 10′
16′
17′
27.3
22.7
14.2
7.3
1.2
2′-Hydroxy-lanneaquinol (2)
1
4
149.2
148.9
130.2, 129.4
126.5
118.0
117.8
114.7
74.5
C
C
CH
C
CH
CH
CH
CH
CH2
6.59, 6.52, 2.75, 2.69
6.75
8′ & 9′
2
3
6
5
2′
1′
5.33
m, 2H
1.99
6.75, 3.94, 2.75, 2.69
6.53, 2.52
6.59
6.75, 6.52
2.75, 2.69, 1.49
6.52, 3.94
6.52
6.75
6.59
3.94
d
d
2.9
8.8
6.59
6.59
6.75, 6.52
2.75, 2.69 1.49
3.94
2.75, 2.69
6.75
dd
dddd
dd
8.8, 2.9
10.4, 7.3 6.5, 2.9
14.5, 2.9
14.5, 7.3
38.9
2.75, 2.69
6.52, 3.94
1.94
dd
3′
36.9
27.3, 27.1
22.7
CH2
2, CH2
CH2
2.75, 2.69
5.33
0.86
1.49
1.99, 1.2
1.25
m, 2H
t, 4H
br s, 18H
t, 3H
3.94, 1.2
5.33, 1.2
7′ & 10′
16′
17′
7.8
7.3
14.1
CH3
0.86
1.2
a
Recorded in CDCl3.
Sch em e 1a
a
Key: (a) n-BuLi, THF; (b) n-C16H33Br; (c) BBr3, CH2Cl2; (d) Mg, n-C16H33Br, THF; (e) NH2NH2, KOH.
benzylic protons (δH 2.52) and the aromatic proton at
δH 6.53. These correlations could best be explained by
1a .
benzene was alkylated under standard conditions (n-
BuLi, n-C16H33Br). The resulting intermediate (4a ) was
then treated with BBr3 to cleave the methyl ethers to
provide the alkylhydroquinone 4b. The catechol ana-
logue 5b was prepared in an analogous manner. For
the synthesis of 6b, however, a different approach was
required. The starting material, 3,4-dimethoxybenzal-
dehyde, was treated with hexadecyl magnesium bro-
mide under the conditions of the Grignard reaction in
an attempt to produce the appropriate benzyl alcohol.3
Quite unexpectedly, the main product was the alkyl aryl
ketone 6 rather than the expected alcohol. The mass
spectrum of 6, which had a molecular ion at m/z 390
(C25H42O3), also showed strong fragment ions at m/z 180
(100%) and 225 (C16H33), which could be attributed to a
McLafferty rearrangement and cleavage R to the car-
Additional evidence favoring the hydroquinone came
from difference NOE experiments. Irradiation of the
signal of the benzylic protons at δ 2.52 enhanced only
the signal of the proton at δ 6.60. This correlation would
be expected for structure 1a , while in 1c it could be
argued that an NOE to the aromatic proton at δ 6.53
might also be observed. However, this argument pro-
vided no discriminating evidence; therefore, in order to
distinguish between the two possible isomers, the
synthesis of analogues was undertaken.
Compound 4b, the model for 1a , was prepared using
methodology developed for the synthesis of urushiol-
related compounds3 (see Scheme 1). 1,4-Dimethoxy-

+
+
118 J ournal of Natural Products, 1997, Vol. 60, No. 2
Groweiss et al.
A
B
C
F igu r e 1. Comparison of the 1H-NMR spectra (aromatic region) of (A) 6b, (B) 4b, and (C) natural 1.
bonyl, respectively. The IR spectrum and the 13C NMR
(δC 199.2) were also indicative of the presence of a
ketone. Therefore, Wolff-Kishner reduction of the
carbonyl to 6a and cleavage of the methoxy groups by
reaction with BBr3 completed the synthesis of the
desired 4-heptadecyl-1,2-dihydroxybenzene (6b) (Scheme
1). The NMR spectra of 4b and 6b showed many similar
chemical shifts, and only careful comparison between
the 1H- and the 13C-NMR spectra of both compounds
with lanneaquinol (see Figure 1 and Experimental
Section) finally confirmed that the natural product
indeed had the hydroquinone regiochemistry, as in
structure 1a .
1) showed distinct correlations between the signal at δ
3.94 and the two benzylic protons (δ 2.75 and 2.69), and,
most importantly, the HMBC spectrum showed a three-
bond correlation to one of the aromatic carbons (δC
126.5), thereby confirming that the OH group was
located in the homobenzylic position (C2′). The absolute
configuration at this carbon was determined to be R by
a modified Mosher method.7
The natural products 1 and 2 and their synthetic
analogues were all evaluated in the NCI primary
2
antitumor screen.1,2,8 The mean panel GI50 value for
1 and 2 was approximately 1 µM. The range of
differential sensitivity of the panel cell lines was 5-10-
fold; the mean-graph profiles were otherwise unremark-
able. None of the O-methyl compounds (4a , 5a , 6, and
6a ) had any significant differential cytotoxicity.
The location of the extra double bond on the linear
8′,9′
alkyl side chain was determined as
by mass
spectrometric techniques. The FABMS (negative ion)
of 1 gave intense peaks at m/z 345 (M - H), 259 (M -
H - C6H14), and 191 (M - H - C11H22). These latter
two fragmentations, namely the cleavage of a saturated
six-carbon unit and an unsaturated C11 fragment,
revealed that the double bond was located between C7′
and C11′ of the side chain.
Lanneaquinol was then acetylated and treated with
dimethyl disulfide, which added to the double bond to
give compound 3.4-6 The EIMS of the latter molecule
(m/z 524, M+, C29H48O4S2) showed an intense fragment
ion corresponding to cleavage of C10H21S. This clearly
revealed that the double bond was originally located
between the tenth and eleventh carbons from the
terminal methyl group, namely at C8′-C9′.
The most active compound among the three geo-
metrical isomers 4b, 5b, and 6b was indeed the 2-alkyl-
1,4-dihydroxybenzene (4b), and the activity of 4b was
virtually identical to that of lanneaquinol. The biologi-
cal testing suggested that the activities of 1 and 2
resulted from the 2-alkyl-1,4-dihydroxybenzene frame-
work and that the actual length of the side chain or the
presence of a double bond therein was less important.
Alkylated phenols have previously been shown to be
cytotoxic. In a study of natural antimutagenic agents,
Wall reported that cymopol (2-bromo-5-geranyl hydro-
quinone) was quite toxic,9 while the Hecht group has
reported the DNA-cleaving activity of 5-alkyl resorci-
nols.10,11
The second isolated cytotoxic compound was slightly
more polar than lanneaquinol; its composition, C23H38O3,
was established by HREIMS. Comparing the spectral
data of 2 with 1 clearly showed that both compounds
were very similar with respect to the aromatic ring and
the alkyl moiety with only one exception, the presence
of secondary hydroxyl group on the alkyl side chain. The
data for the latter group included an IR absorption at
3592 cm-1 a proton chemical shift at δ 3.94 and a D2O
exchangeable proton (δ 7.75), and the observed δC 74.9
in the 13C-NMR spectrum. The COSY spectrum (Table
Another Sephadex LH-20 fraction from the CCl4-
soluble portion, which showed moderate activity in the
antitumor screen, was further separated by several
consecutive column chromatographies on Si gel and
flash chromatography on cyano-bonded phase bulk
packing. This separation led to the isolation of two
diastereomers, 7 and 8, which were similar in several
aspects to lanneaquinol and 2′-hydroxylanneaquinol and
may be their biosynthetic precursors. The molecular
composition of both compounds, C23H40O3, was deter-
mined by high-resolution mass spectrometry. The four

+
+
Cytotoxic Hydroquinones from Lannea
J ournal of Natural Products, 1997, Vol. 60, No. 2 119
Ta ble 2. NMR Data for Compounds 7 and 8
13C-NMR data
1H-NMR data
C#
δ (ppm)
197.7
mult.
HMBC
δ (ppm)
mult.
J (Hz)
COSY
NOE
Dihydroxyketone 7
1
3
2
C
6.82, 2.78, 2.41
149.2
129.7
CH
CH
4.50
4.50
6.82
6.01
dd
ddd
10.3, 2.9
10.3, 2.2
0.9
6.01, 4.50
6.82, 4.50
6.01, 4.50
6.82
8′ & 9′
5
4
130.0, 129.7
77.1
74.2
CH
C
CH
1.98
5.31
m, 2H
br s
1.98
6.82, 4.50, 2.78, 2.41, 1.66, 1.52
6.01, 2.78, 2.41, 1.66, 1.52
4.50
6.62, 6.01
2.41
6.82, 2.41
6
47.6
32.9
CH2
CH2
1.66
2.78
2.41
1.66
1.52
1.26
0.86
d
dd
m
m
m, 22H
t, 3H
15.9
2.41
2.41
4.50, 2.78
4.50
15.9,0.9
16, 12, 6
16, 10, 7
4.50, 2.78
1.52, 1.26
1.66, 1.26
1′
4.50, 2.41, 1.66, 1.52
0.86
no NOE
16′
17′
31.9
14.1
CH2
CH3
7
1.26
Dihydroxyketone 8
1
3
2
197.3
149.0
129.3
130.0, 129.7
76.5
C
6.72, 2.66, 2.47
4.34
4.34, 2.66
1.98
6.72, 2.66, 2.47 w
6.03, 2.66, 2.47 w
6.03
CH
CH
CH
C
CH
CH2
6.72
6.03
5.32
dd
dd
m, 2H
10.3, 2.4
10.3, 1
6.03, 4.34
6.72, 4.34
1.98
6.03, 4.34
6.72
8′ & 9′
5
4
6
70.5
46.8
4.34
2.66
2.47
1.65
1.50
1.26
0.86
t
dd
d
m
2.4
16.3, 1
16.3
6.72, 6.03
2.47
2.66
1.26
1.26
6.72, 2.47
2.47
4.34, 2.66
4.34, 2.66, 2.47
no NOE
1′
38.8
CH2
4.34, 2.66
0.86
m
16′
17′
31.9
14.1
CH2
CH3
br m, 22H
t, 3H
7
1.26
a
Recorded in CDCl3.
degrees of unsaturation were in agreement with the 1H-
and 13C-NMR spectra, which showed only two double
bonds and a carbonyl group, and, therefore, lacked the
aromatic ring of 1 and 2. The structure elucidation was
based on the interpretation of the spectral data, espe-
cially the 1D and 2D NMR. A comparison between
lanneaquinol and 7 revealed that the side chain was
similar in both compounds, and the difference lay only
in the six-membered ring. This ring in 7 is composed
of an R,â-unsaturated ketone (carbonyl at δC 197.7 ppm,
two coupled olefinic protons at δ 6.82 and 6.01), a
secondary hydroxy group (δ 4.50, coupled in the HMQC
experiment to the carbon at δ 74.2), and a tertiary
alcohol (δC 77.1). Besides these functional groups, there
was a ring methylene carbon (δC 47.6 ppm), which was
coupled (HMQC) to two protons at δ 2.78 and 2.41. The
structure and substitution of the cyclohexenone moiety
were established by COSY and HMBC experiments (see
Table 2). The former experiment established that the
secondary alcohol was allylic, while all the HMBC
connectivities, including the highly important one be-
tween the protons R to the carbonyl (δ 2.78 and 2.41)
and one of the side-chain methylenes (δC 32.9 ppm),
established the structure as 7. The position and geom-
etry of the olefinic bond in the side chain were deter-
mined as described for 1 and 2.
F igu r e 2. Relative stereochemistry of 7 and 8.
δC7 - δC8 ) 3.7 ppm), C5 ( δ ) 0.6 ppm), C6 ( δ ) 0.8
ppm), and C1′ ( δ ) -5.9 ppm), while the chemical shift
differences for all the other carbon pairs were e0.4 ppm.
The relative stereochemistry of the two isomers was
established by difference NOE experiments. In both 7
and 8, an NOE was observed between H-4 and one of
the H-6 protons (e.g., δ 2.41 in 7), indicating a pseudo-
axial disposition of H-4. In 8, an additional NOE was
detected between one of the C-1′ methylene protons (δ
1.65) and both H-6 protons (δ 2.66 and 2.47), revealing
that the alkyl ring substituent was equatorial. Thus,
the hydroxyl groups were cis-oriented in 8 (4S*, 5R*
relative stereochemistry) and, by inference, trans-
oriented in 7 (4S*, 5S*), as shown in Figure 2. The
absolute configurations of 7 and 8 have not been defined.
Dihydroxy ketone 8 is a potential biogenetic precursor
of lanneaquinol (1). The trans-anti-parallel configura-
tion of H-4 and OH-5 in 8 suggests an easy elimination
of water; subsequent keto-enol tautomerism at C-1
would lead to the aromatic product 1. This phenomenon
was indeed observed when compound 8 was left for
several weeks in an NMR tube (in CDCl3); 8 was
partially transformed to lanneaquinol. This transfor-
mation also occurred, but to a much lesser extent, with
7; after 1 year, only a small fraction of the compound
was transformed to 1. This observation may explain
Compound 8 was a diastereomer of 7 and differed only
in the stereochemistry at C5. The main changes in the
13C-NMR spectra were found at the region of C4 ( δ )

+
+
120 J ournal of Natural Products, 1997, Vol. 60, No. 2
Groweiss et al.
the moderate cytotoxicity of 7 and 8, the result of partial
in situ transformation into the more potent lan-
neaquinol.
Si gel and subsequent column chromatography on
cyano-bonded phase (elution with hexane-i-PrOH),
leading to the isolation of two additional compounds, 7
(16 mg) and 8 (9 mg). Both compounds were obtained
as colorless, viscous oils.
Exp er im en ta l Section
Com pou n d 7: [R]D -51.5° (c 0.54,CHCl3); UV (MeOH)
Gen er a l. Exp er im en ta l P r oced u r es: NMR spec-
tra were recorded on a Varian VXR-500 spectrometer
using CDCl3 and CD3OD as solvents and internal
standards. IR spectra were measured on a Perkin-
Elmer 267 and Nicolet 5MX spectrometers; UV spectra
were obtained with a Beckman 34 spectrophotometer.
Optical rotations were measured on a Perkin-Elmer 241
polarimeter. MS were recorded on VG Micromass ZAB
2F and J EOL SX-102 mass spectrometers. Elemental
analyses were carried out on a Carlo-Erba NA-1500
analyzer.
P la n t Ma ter ia l. Fruits, leaves, stems, and twigs of
L. welwitschii (Hiern) Engl. were collected by D. Tho-
mas, under contract to the National Cancer Institute,
near Mundemba, Cameroon, in March 1987. The plant
was identified by L. Ake´ Assi; a voucher specimen
(Q66P-6785) is maintained at the Missouri Botanical
Garden. Extraction of 558 g of dried, ground plant
material with CH2Cl2-MeOH (1:1) and MeOH yielded
18.22 g of crude extract.
λ
ν
max nm (log ꢀ) 294 (2.81), 215 (3.82), 204 (3.88); IR (neat)
max 3418 (br), 2924, 2854, 1674, 1456, 1259, 1075 cm-1
;
HREIMS m/z 364.2930 (M+, calcd for C23H40O3,
364.2977); EIMS m/z (rel abundance) 364 (M+, 20), 346
(44), 328 (M+- 2 H2O, 5), 265 (13), 123 (25), 111 (37),
84 (100), 69 (22), 55 (33); 13C NMR (CDCl3) δ 197.7,
149.2, 130.0, 129.73, 129.66, 77.1, 74.2, 47.6, 32.9, 31.9,
30.0, 29.75, 29.70, 29.50, 29.44, 29.30, 29.29, 29.20,
1
27.21, 27.16, 22.7, 22.4, 14.1; for H NMR, see Table 2.
Com p ou n d 8: [R]D +23° (c 0.9,CHCl3); UV(MeOH)
max nm (log ꢀ) 290 (3.21), 217 (3.82), 204 (3.83); IR (neat)
max 3408 (br), 2928, 2854, 1674, 1466, 1378, 1258, 1157,
λ
ν
1069, 833, 723 cm-1; HREIMS m/z 364.2978 (M+, calcd
for C23H40O3, 364.2977); EIMS m/z (rel abundance) 364
(M+, 19), 346 (41), 328 (9), 265 (13), 123 (22), 111 (41),
84 (100), 69 (22), 55 (34); 13C NMR (CDCl3) δ 197.3,
149.0, 130.0, 129.7, 129.3, 76.5, 70.5, 46.8, 38.8, 31.9,
29.9, 29.8, 29.7, 29.5, 29.4, 29.3 (× 2), 29.2, 27.20, 27.18,
1
23.6, 22.7, 14.1; for H NMR, see Table 2.
2-Hexa d ecyl-1,4-d im eth oxyben zen e (4a ). All the
reactions were done solely to prepare small samples of
the products and therefore were not optimized. The
reactions were run under N2 with magnetic stirring.
Isola tion of Com p ou n d s 1 a n d 2. The crude
organic extract (14.04 g) of L. welwitschii was parti-
tioned by the following protocol: distribution between
90% aqueous MeOH and hexane (affording 5.89 g), 80%
aqueous MeOH and CCl4 (1.91 g), and 70% MeOH-H2O
and CHCl3 (1.00 g). The MeOH was then removed from
the aqueous phase, which was subsequently extracted
with EtOAc to give 1.11 g; finally, lyophilization of the
H2O phase gave a residue of 2.07 g. Cytotoxic activity
was concentrated primarily in the CCl4 fraction. This
fraction was subjected to gel permeation through a
Sephadex LH-20 column (elution with 2:1:1 CH2Cl2-
MeOH-hexane), followed by Si gel column chromatog-
raphy of the most active fraction (109 mg; elution with
a mixture of hexane-EtOAc). This chromatography
gave pure 1 (59 mg) and 2 (19 mg) as amorphous solids.
La n n ea qu in ol (1): UV (MeOH) λmax nm (log ꢀ) 294
(3.52), 218 (3.71), and 205 (4.00); IR (CH2Cl2) νmax 3589,
2928, 1503, 1260, 1174, 750 cm-1; HREIMS m/z 346.2867
(M+, calcd for C23H38O2, 346.2872); EIMS m/z (rel
abundance): 346 (M+, 96), 163 (10), 151 (10), 123 (100);
13C NMR (CDCl3) δ 149.2, 147.3, 130.1, 129.9, 129.8,
116.8, 116.0, 113.3, 31.9, 30.1, 29.8 (× 2), 29.7, 29.55,
29.50, 29.46, 29.36, 29.34, 29.28, 27.3 (× 2), 22.7, 14.2;
A
100-mL three-necked, round-bottomed flask,
equipped with a reflux condenser and a funnel for
addition of reagents, was charged with 2.33 g (16.9
mmol) of 1,4-dimethoxybenzene and 30 mL of dry THF.
The solution was cooled to 0 °C, and 8 mL of 1.6 M
n-BuLi (in hexane, 12.8 mmol) was added dropwise over
10 min. The reaction mixture was stirred for 30 min
and then refluxed for 90 min. Then the flask was cooled
to room temperature, a mixture of n-hexadecyl bromide
(2.585 g, 8.5 mmol) in dry THF (7 mL) was added slowly
over 15 min, and the reaction mixture was refluxed for
another 90 min. The solvents were evaporated, and the
white solid residue was dissolved in Et2O-H2O (1:1, 100
mL). The phases were separated, the aqueous residue
was extracted with Et2O (2 × 80 mL), and the organic
fractions were combined, washed with H2O, dried over
anhydrous Na2SO4, and evaporated to dryness. The
resulting product (2.93 g) contained a mixture of 4a and
1,4-dimethoxybenzene. Crystallization from 95% EtOH
(with a few drops of Et2O) gave 2.46 g of 4a (> 95%
purity) in the first crop of crystals (80% yield). Com-
pound 4a : mp 45 °C; HREIMS m/ z 362.3175 (calcd for
1
for H NMR, see Table 1.
2′-R-Hyd r oxyla n n ea qu in ol (2): [R]D + 0.8° (c 1.24,
CHCl3); UV (MeOH) λmax nm (log ꢀ) 295 (3.46), 227
(3.57), 218 (3.68), 204 (4.10); IR (CH2Cl2) νmax 3592,
2928, 2856, 1499, 1239, 1174, 750 cm-1; HREIMS m/z
362.2824 (M+, calcd for C23H38O3, 362.2821); EIMS m/z
(rel abundance) 362 (M+, 20), 344 (22), 124 (96), 123
(40), 69 (100); 13C NMR (CDCl3) δ 149.2, 148.9, 130.2,
129.4, 126.5, 118.0, 117.8, 114.7, 74.5, 38.9, 36.9, 31.9,
29.8, 29.7, 29.6, 29.4, 29.3, 29.1, 27.3, 27.1, 25.6, 22.7,
1
C27H42O2, 362.3185); H NMR (CDCl3) δ 6.74 (d, J )
8.8 Hz), 6.71 (d, 2.9), 6.66 (dd, 8.8,2.9), 3.76 (s, 3H), 3.74
(s, 3H), 2.54 (m, 2H), 1.52 (m, 2H), 1.26 (m, 26H), 0.86
(t, 3H, 7); 13C NMR (CDCl3) δ 153.4 (× 2), 132.7, 116.2,
111.2, 110.5, 56.0, 55.6, 31.9, 30.2-29.4 (13C), 22.7, 14.1.
2-Hexa d ecyl-1,4-h yd r oqu in on e (4b). Compound
4a (1.45 g, 4.0 mmol) was dissolved in CH2Cl2 (30 mL)
in a three-necked flask under N2. The solution was
cooled to 0 °C and 8 mL of BBr3 (in CH2Cl2) was added
dropwise over 15 min. The reaction mixture was stirred
for 45 min at 0 °C and 4 h at room temperature. The
reaction was quenched with H2O, the phases were
separated, and the organic portion was washed with
H2O (2 × 100 mL), then with 10% Na2CO3 (100 mL),
1
14.1; for H NMR, see Table 1.
Isola tion of Com p ou n d s 7 a n d 8. Several fractions
from the Sephadex LH-20 chromatography of the CCl4
fraction showed moderate in vitro activity in the anti-
tumor screen. These fractions (a total of 1.03 g) were
combined and subjected to a flash chromatography over

+
+
Cytotoxic Hydroquinones from Lannea
J ournal of Natural Products, 1997, Vol. 60, No. 2 121
and finally with H2O. The CH2Cl2 phase was dried over
anhydrous Na2SO4 and evaporated to dryness. Only
300 mg of 4b were obtained in a pure form as a tan
solid: mp 114 °C; HREIMS m/ z 334.2866 (calcd for
199.2, 153.0, 148.9, 130.3, 122.6, 110.1, 109.9, 56.0, 55.9,
38.1, 31.9, 29.6-29.3 (11C), 24.7, 22.6, 14.1; EIMS m/ z
390 (M+, C25H42O3), 180 (100), 165.
Wolff-Kish n er Red u ction of 6. Compound 6 (450
mg) and KOH (250 mg) were dissolved in diethylene
glycol (10 mL) in a 100-mL flask. The temperature was
raised to 80 °C, and hydrazine hydrate (400 µL) was
added. Then the temperature was raised to 175 °C.
After 75 min, when TLC indicated total consumption
of 6, the reaction was cooled to room temperature, H2O
(10 mL) was added, and the organic materials were
extracted with CHCl3. The CHCl3 fraction, containing
most of the diethylene glycol, was dried over anhydrous
Na2SO4 and evaporated to give 1.43 g of oily product.
This material was subjected to flash chromatography
on Si gel. Elution with hexane-EtOAc (19:1) afforded
123 mg (28% yield) of pure 4-heptadecyl-1,2-dimethoxy-
benzene (6a ): 1H NMR(CDCl3) δ 6.80 (d J ) 8 Hz), 6.71
(d, 8), 6.70 (s), 3.87 (s, 3H), 3.86 (s, 3H), 2.55 (m, 2H),
1.60 (m, 2H), 1.25 (m, 28H), 0.88 (t, 3H,7); 13C NMR
(CDCl3) δ 148.7, 146.9, 135.6, 120.1, 111.7, 111.1, 55.9,
55.7, 35.6, 31.9, 31.7, 29.7-29.3 (12C), 22.7, 14.1.
1
C22H38O2, 334.2872); H NMR (1:1 CDCl3-CD3OD) δ
6.28 (d, J ) 8.3 Hz), 6.24 (d, 2.9), 6.15 (dd, 8.3,2.9), 2.19
(m, 2H), 1.23 (m, 2H), 0.92 (m, 26H), 0.54 (t, 3H, J )
7); 13C NMR (1:1 CDCl3-CD3OD) δ 149.0, 147.1, 129.9,
116.0, 115.0, 112.2, 31.3, 29.6-28.7 (13C), 22.0, 13.2;
anal. C 79.10%, H 11.40% (calcd C 78.99%, H 11.45%).
3-Hexadecyl-1,2-dih ydr oxyben zen e (5b). The same
procedures used for preparation of 4a and 4b were
applied. 1,2-Dimethoxybenzene (veratrole, 2.36 g) was
reacted with n-BuLi and then with 1.80 g of n-hexade-
cylbromide to give a mixture of 5a and starting material
(3.12 g). This fraction was separated by column chro-
matography to give 1.1 g of 3-hexadecyl-1,2-dimethoxy-
benzene (5a ): mp 36 °C; 1H NMR (CDCl3) δ 7.00 (t, J )
8 Hz), 6.80 (d, 8), 6.79 (d, 8), 3.88 (s, 3H), 3.86 (s, 3H),
2.66 (t, 8, 2H), 1.62 (m, 2H), 1.28 (m, 22H), 0.92 (t, 3H,
7); 13C NMR (CDCl3) δ 151.6, 146.0, 135.7, 122.6, 120.8,
108.8, 59.5, 54.5, 30.9, 29.8, 28.8-28.3 (12C), 21.7, 13.1;
anal. found C 79.65%, H 11.03% (calcd C 79.50%, H
11.68%).
4-Hep ta d ecyl-1,2-d ih yd r oxyben zen e (6b). Com-
pound 6a (112 mg, 0.3 mmol) was converted to 6b by
reaction with excess of BBr3 (1 mmol) in CH2Cl2 as
described above for 4b. After the usual workup, 98 mg
(95% yield) of 6b were obtained as a white crystalline
material. Crude 6b was recrystallized from EtOAc-
hexane to give 82 mg of highly purified product (6b):
mp 95-96 °C; IR (KBr) νmax 3452, 3349 (br), 2945, 2847,
1518, 1470, 1276, 979, 931, 813 and 716 cm-1; HREIMS
3-Hexadecyl-1,2-dimethoxybenzene (226 mg, 0.62
mmol) was then dissolved in CH2Cl2 and treated with
BBr3 as described for 4b. 3-Hexadecylcatechol (5b, 207
1
mg) was obtained as a solid: mp 44-45 °C; H NMR
(CDCl3) δ 6.72 (3H, m), 5.37 and 5.26 (br s, OH), 2.61
(t, 8, 2H), 1.62 (m, 2H), 1.27 (m, 26H), 0.90 (t, 3H, J )
7); 13C NMR (CDCl3) δ 143.0, 141.8, 129.4, 122.1, 120.1,
112.9, 31.9, 29.8-29.4 (13C), 22.7, 14.1; IR (KBr) νmax
3368 (br), 2917, 2845, 1468, 1282, 1192, 959, and 726
cm-1; anal. found C 79.01%, H 11.13%, (calcd for C
78.99%, H 11.45%).
1
m/ z 348.3016 (M+, calcd for C23H40O2, 348.3028); H
NMR (1:1 CDCl3-CD3OD) δ 6.66 (d, J ) 8 Hz), 6.60 (d,
2), 6.47 (dd, 8, 2), 2.42 (m, 2H), 1.50 (m, 2H), 1.25 (m,
28H), 0.83 (t, 3H, 7); 13C NMR (1:1 CDCl3-CD3OD,
partial) δ 145.1, 143.1, 135.5, 120.3, 116.0, 115.7, 112.2;
anal. found C 78.72%, H 11.69%, calcd C 79.25%, H
11.57%).
3,4-Dim eth oxyp h en yl n -Hep ta d ecyl Keton e (6).
A 250-mL three-necked flask, equipped with a reflux
condenser and a funnel for the addition of reagents, was
charged with 0.24 g of magnesium turnings, 15 mL of
dry THF, a few crystals of I2. and a magnetic stirring
bar. A solution of n-hexadecyl bromide (3.05 g, 10
mmol) in dry THF (15 mL) was introduced into the
funnel, and about one quarter of this solution was added
into the reactor under a stream of N2. The reaction was
initiated by heating the mixture gently with a heat gun.
The rest of the alkyl bromide solution was added
dropwise, and the reaction advanced very slowly, as
observed by the consumption of the Mg turnings (ca. 4
h). Then a mixture of 3,4-dimethoxybenzaldehyde (1.66
g, 10.0 mmol) in dry THF (15 mL) was added gradually
over 20 min, and the reaction mixture was refluxed
gently for 20 h. The reaction was quenched by the
addition of 1 M HCl (5 mL); the THF was evaporated,
CHCl3 was added, and the phases were separated. The
organic phase was washed with 1 M HCl and then H2O,
dried over anhydrous Na2SO4, and then evaporated to
dryness. The resulting product, which solidified upon
standing, was separated by column chromatography to
give 1.52 g (39%) of the unexpected product 6. All the
fractions were examined by NMR, and none contained
the expected Grignard product. The only other product
isolated was 3,4-dimethoxybenzyl alcohol (6): 1H NMR
(CDCl3) δ 7.57 (dd, J ) 8.3,2 Hz), 7.52 (d, 2), 6.87 (d,
8.3), 3.93 (s, 3H), 3.92 (s, 3H), 2.90 (m, 2H), 1.70 (m,
2H), 1.24 (m, 26H), 0.86 (t, 3H, 7); 13C NMR (CDCl3) δ
Ack n ow led gm en t. We thank D. Thomas, the Mis-
souri Botanical Garden, and G. Cragg for the plant
collection; L. Ake´ Assi for taxonomic identification; T.
McCloud for extraction; and A. Monks, D. Scudiero, T.
Prather, and R. Shoemaker for antitumor screening
assays.
Refer en ces a n d Notes
(1) Boyd, M. R.; Paull, K. D.; Rubinstein, L. R. In Cytotoxic
Anticancer Drugs: Models and Concepts for Drug Discovery and
Development. Developments in Oncology; Valeriote, F. A., Cor-
bett, T. H., Baker, L. H., Eds. Kluwer Academic Publishers:
Amsterdam, 1992; pp 11-34.
(2) Boyd, M. R.; Paull, K. D. Drug Dev. Res. 1995, 34, 91-109.
(3) ElSohly, M. A.; Adawadkar, P. D.; Benigni, D. A.; Watson, E.
S.; Little, T. L., J r. J . Med. Chem. 1986, 29, 606-611.
(4) Mason, R. T.; Fales, H. M.; J ones, T. H.; Pannell, L. K.; Chinn,
J . W.; Crews, D. Science 1989, 245, 290-293.
(5) Buser, H.-R.; Arn, H.; Guerin, P.; Rauscher, S. Anal. Chem. 1983,
55, 818-822.
(6) Francis, G. W.; Veland, K. J . Chromatogr. 1981, 219, 379-384.
(7) Ohtani, I.; Kusumi, T.; Ishitsuka, M. O.; Kakisawa, H. Tetra-
hedron Lett. 1989, 30, 3147-3150.
(8) Boyd, M. R. In Current Therapy in Oncology; Niederhuber, J .
E., Ed.; B. C. Decker: Philadelphia, 1993; pp 11-22.
(9) Wall, M. E. J . Nat. Prod. 1992, 55, 1561-1568.
(10) Lytollis, W.; Scannell, R. T.; An, H.; Murty, V. S.; Reddy, K. S.;
Barr, J . R.; Hecht, S. M. J . Am. Chem. Soc. 1995, 117, 12 682-
12 690.
(11) Singh, U. S.; Scannell, R. T.; An, H.; Carter, B. J .; Hecht, S. M.
J . Am. Chem. Soc. 117, 12 691-12 699.
NP960435T