Artepillin C isoprenomics: Design and synthesis of artepillin C isoprene analogues as lipid peroxidation inhibitor having low mitochondrial toxicity

Article abstract of DOI:10.1016/j.bmc.2006.04.015

We designed and synthesized isoprene analogues of artepillin C, a major component of Brazilian propolis, and investigated the inhibitory activity on lipid peroxidation of rat liver mitochondria (RLM) and RLM toxicity based on isoprenomics. We succeeded in

Full text of DOI:10.1016/j.bmc.2006.04.015

Bioorganic & Medicinal Chemistry 14 (2006) 5721–5728
Artepillin C isoprenomics: Design and synthesis of artepillin C
isoprene analogues as lipid peroxidation inhibitor
having low mitochondrial toxicity
Yoshihiro Uto,a Shutaro Ae,a Daisuke Koyama,a Mitsutoshi Sakakibara,a Naoki Otomo,a
Mamoru Otsuki,a Hideko Nagasawa,a Kenneth L. Kirkb and Hitoshi Horia,*
aDepartment of Biological Science and Technology, Faculty of Engineering, The University of Tokushima,
Minamijosanjimacho-2, Tokushima 770-8506, Japan
bLaboratory of Bioorganic Chemistry, National Institutes of Diabetes and Digestive and Kidney Diseases,
National Institute of Health, Bethesda, MD 20892, USA
Received 17 March 2006; revised 5 April 2006; accepted 6 April 2006
Available online 11 May 2006
Abstract—We designed and synthesized isoprene analogues of artepillin C, a major component of Brazilian propolis, and investi-
gated the inhibitory activity on lipid peroxidation of rat liver mitochondria (RLM) and RLM toxicity based on isoprenomics. We
succeeded in the synthesis of artepillin C isoprene analogues using regioselective prenylation within the range from 22% to 53% total
yield. Reactivity of artepillin C and its isoprene analogues with ABTS (2,20-Azinobis(3-ethylbenzothiazoline-6-sulfonate)) radical
cations showed only a slight difference among the molecules. The isoprene side-chain elongation analogues of artepillin C showed
almost the same inhibitory activity against RLM lipid peroxidation as artepillin C. Artepillin C and its isoprene analogues had very
weak RLM uncoupling activity. Moreover, artepillin C and its isoprene analogues exhibited a lower inhibitory activity against aden-
osine 50-triphosphate (ATP) synthesis by about two orders of magnitude than the effective inhibitory activity against RLM lipid
peroxidation. From these results we conclude that artepillin C isoprene analogues could be potent lipid peroxidation inhibitors hav-
ing low mitochondrial toxicity. We also conclude that elongation of the isoprene side chain of artepillin C to increase lipophilicity
had little influence on the inhibitory activity toward RLM lipid peroxidation.
Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction
(coenzyme Q) is a component of the mitochondrial
respiratory chain.5 A structural feature of ubiquinone
is the presence of from 1 to 12 of consecutive prenyl side
chains that makes ubiquinone a highly hydrophobic
structure with high affinity for membrane. In addition,
a semiquinone radical of ubiquinol, as a reduced form
of ubiquinone, is relatively stable so it functions as an
effective antioxidant in the mitochondria.6 Paradoxical-
ly, coenzyme Q is also involved in superoxide produc-
tion by the respiratory chain.7 The development of
additional effective antioxidants targeted to the mito-
chondria is an important strategy for the prevention
and treatment of mitochondrial dysfunction.
Mitochondria are indispensable to the maintenance of
human life as the site of organization of energy produc-
tion. However, it is well known that potentially toxic
reactive oxygen species (ROS) are produced by the elec-
tron transport chain (ETC) in the mitochondrial inner
membrane.1 In addition, mitochondria are very suscep-
tible to lipid peroxidation by ROS because of the pres-
ence in the mitochondrion of high concentrations of
unsaturated lipids.2 It has been pointed out that such
ROS-mediated peroxidation is related to carcinogenesis
and aging.3,4 Natural anti-oxidants are present to
provide protection against this. Thus, ubiquinone
a-Tocopherol (vitamin E) is an endogenous antioxidant
and potent lipid peroxidation inhibitor, possessing the
phenolic hydroxyl group and hydrophobic alkyl side
chain present in ubiquinone. Exogenous antioxidants
also are known Artepillin C (Fig. 1) is a diprenyl-p-
hydroxycinnamic acid derivative first isolated from Bac-
Keywords: Artepillin
peroxidation inhibitor; Rat liver mitochondria.
C isoprene analogue; Isoprenomics; Lipid
*
Corresponding author. Tel.: +81 88 656 7514; fax: +81 88 656
9164; e-mail: hori@bio.tokushima-u.ac.jp
0968-0896/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.04.015

5722
Y. Uto et al. / Bioorg. Med. Chem. 14 (2006) 5721–5728
charis species,8 as a major constituent (>5%) in Brazilian
propolis.9 Artepillin C is a prenyl-containing compound
that displays a variety of medicinal effects such as anti-
bacterial,10 antitumor,11 apoptosis-inducing,12 immuno-
modulating,13 and antioxidative activities.14 It is
interesting that, despite being less hydrophobic than a-
tocopherol, artepillin C has a three times higher inhibi-
tory activity on lipid peroxidation than a-tocopherol.15
Though the reason for the higher activity is unclear, this
suggested to us that there exists an optimal hydropho-
bicity and molecular structure for the design of this class
of lipid peroxidation inhibitors. In fact, drupanin, a
mono-prenyl analogue of artepillin C, exhibited weaker
inhibition of peroxidation of linoleic acid than did arte-
pillin C.14 In contrast to prenylated and non-prenylated
flavonoids, an increase in the number of prenyl substit-
uents decreased antioxidant activity in the lipid peroxi-
dation system.16 Therefore, we felt it was important to
elucidate a relationship between the number and length
of prenyl groups of isoprenoids (optimal hydrophobici-
ty) and the potency of inhibition of lipid peroxidation
based on what we have termed ‘isoprenomics’15,17
(medicinal chemistry of isoprenoid involved in structur-
al analysis, biosynthesis, biological function, and
chemotype18).
chain such as mono-farnesyl (TX-2101), prenyl and ger-
anyl (TX-2013), and di-geranyl (TX-2007) (Fig. 1). The
highest occupied molecular orbital (HOMO) and the
lowest unoccupied molecular orbital (LUMO) energy
of artepillin C and these analogues were calculated to
have almost the same values (Table 1).
2.2. Synthesis
As shown in Scheme 1, the preparation of the target
compounds was based on regioselective C-prenylation
of p-iodophenol according to our established proce-
dure.15,21 In the C-prenylation reaction, increasing the
length of isoprene units tends to increase the reactivity
(the yield of compounds 1, 7, and 13 was 58%, 62%,
and 75%, respectively). The yield of 4 was relatively
low (46%) due to competing O-prenylation. Acetylation
of prenylated compounds 1, 4, 7, 10, and 13 produced
compound 2, 5, 8, 11, and 14 in good yields. These
underwent Mizoroki–Heck coupling with methyl acry-
late in the presence of Pd(OAc)2, (o-tol)3P, and Et3N
in dry toluene to give a, b-unsaturated carboxylates 3,
6, 9, 12, and 15 in reasonable yield, except for com-
pound 6. Finally, these compounds were deprotected
in methanolic potassium hydroxide to give analytically
pure TX-1959, TX-2007, TX-2012, TX-2013, and TX-
2101 in 38%, 22%, 45%, 26%, and 53% total yield,
respectively.
In this paper, we report the design and synthesis of nat-
ural and synthetic artepillin C analogues having various
isoprene side chains. We then studied the inhibitory
activity on rat liver mitochondria (RLM) lipid peroxida-
tion of these analogues. In addition, we evaluated their
mitochondrial toxicity with respect to RLM uncoupling
activity and inhibitory activity of adenosine 50-triphos-
phate (ATP) synthesis.
2.3. Antioxidant activity
The antioxidant activities of artepillin C and its isoprene
analogues were evaluated by spectroscopic analysis with
ABTS
(2,20-Azinobis(3-ethylbenzothiazoline-6-sulfo-
nate)) radical cations. The IC50 values of artepillin C
and its isoprene analogues were almost the same in the
2. Results
Table 1. HOMO and LUMO energy of artepillin C and its isoprene
analogues
2.1. Design and molecular modeling
Compound
EHOMO (eV)
ELUMO (eV)
Naturally occurring p-hydroxycinnamic acid analogues
having isoprene side chains are limited in number. Three
such compounds, including artepillin C, drupanin19 (we
have designated as ‘TX-1959’), and the mono-geranyl
compound20 that we have designated ‘TX-2012’ have
been identified. For this study, we designed additional
p-hydroxycinnamic acid analogues having isoprene side
Artepillin C
TX-1959
TX-2007
TX-2012
TX-2013
TX-2101
ꢀ8.910
ꢀ8.977
ꢀ8.917
ꢀ9.005
ꢀ8.907
ꢀ9.063
ꢀ0.693
ꢀ0.732
ꢀ0.702
ꢀ0.728
ꢀ0.695
ꢀ0.745
Figure 1. Structure of artepillin C and our designed artepillin C isoprene analogues based on isoprenomics.

Y. Uto et al. / Bioorg. Med. Chem. 14 (2006) 5721–5728
5723
Scheme 1. Reagents and conditions: (a) prenyl or geranyl or farnesyl bromide/NaH/toluene, rt, 1 h; (b) AcCl/DMAP/CH2Cl2, reflux, 4 h;
(c) CH2CHCOOMe/Et3N/(o-tol)3P/Pd(OAc)2/toluene, reflux, 20 h; (d) KOH/MeOH/H2O, reflux, 3 h.
range of 6.84–8.37 lM. The mono-isoprene analogues
such as TX-1959, TX-2012, and TX-2101 consistently
showed a slightly higher reactivity than the di-isoprene
analogues such as TX-2007 and TX-2013 (Table 2).
lin C (IC50 = 0.72 0.03). In addition, the mono-prenyl
analogue TX-1959 showed a weaker inhibitory activity
(IC50 = 37.80 3.04) than artepillin C. The mono-gera-
nyl analogue TX-2012 (IC50 = 2.61 0.65) and mono-
farnesyl analogue TX-2101 (IC50 = 2.32
0.26) have
2.4. Inhibitory activity on lipid peroxidation
similar inhibitory activities, and these compounds were
more potent than TX-1959 (IC50 = 37.80 3.04).
We next investigated the activities of the artepillin C iso-
prene analogues as inhibitors of RLM lipid peroxida-
tion. ADP and Fe2+ were used as initiators of a
radical reaction, and then IC50 values were estimated
from O2 consumption in the presence or absence of arte-
pillin C isoprene analogues (Table 2). We found that the
inhibitory activities of artepillin C isoprene elongation
analogues TX-2013 (IC50 = 0.84 0.08) and TX-2007
(IC50 = 0.83 0.09) were nearly equal to that of artepil-
2.5. Mitochondrial uncoupling activity
Phenolic compounds having a weakly acidic proton sur-
rounded by a hydrophobic side chain act as proton
uncouplers in mitochondria. An example is 3,5-di-tert-
butyl-4-hydroxybenzylidenemalononitrile (SF6847).22
If effective uncoupling and antioxidant activity are
similar, these activities would compete so the net antiox-
idant activity would be weakened. Therefore, we
examined the uncoupling activity of artepillin C and
its analogues, all of which showed very weak or no
uncoupling activity (Table 3). Interestingly, we observed
an inhibition of mitochondrial ETC activities with arte-
pillin C and its isoprene analogues at high dose (data not
shown).
Table 2. Antioxidant activity of artepillin C and its isoprene analogues
Compound
ABTS (IC50, lM)
RLM lipid (IC50, lM)
Artepillin C
TX-1959
TX-2007
TX-2012
TX-2013
TX-2101
7.81 0.15
6.92 0.39
8.21 0.09
7.33 0.81
8.37 0.39
6.84 0.29
0.72 0.03
37.80 3.04
0.83 0.09
2.61 0.65
0.84 0.08
2.32 0.26
2.6. Inhibitory activity of ATP synthesis
To examine mitochondrial toxicity further, we investi-
gated the activity of artepillin C and its isoprene ana-
Values represent means SD of three experiments.

5724
Y. Uto et al. / Bioorg. Med. Chem. 14 (2006) 5721–5728
Table 3. Mitochondrial toxicity of artepillin C and its isoprene
analogues
ment of the prenyl side chain of TX-1959 with the gera-
nyl side chain (TX-2012), whereas the activity was not
changed in the replacement of the geranyl side chain
of TX-2012 with the farnesyl side chain (TX-2101). In
general, we have shown that a more hydrophobic anti-
oxidant has a higher inhibitory activity toward lipid per-
oxidation,23 although we suggested that the relationship
may be applicable to a limited range of hydrophobicity.
In the case of the artepillin C isoprene analogues, the
upper limit of the logP value of artepillin C (5.66) or
TX-2012 (5.60) may be taken for the hydrophobicity–
activity relationship.
Compound
Uncoupling activity
ATP synthesis
(IC50, lM)
Cmax
(lM)
Vmax
(natomO/mg/min)
Artepillin C
TX-1959
TX-2007
TX-2012
TX-2013
TX-2101
SF6847
125
100
20
180
98
47
110
13
86
88
50
25
34
23
108
119
280
40
18
0.01
0.007
The relatively hydrophobic artepillin C isoprene ana-
logues act as potent RLM lipid peroxidation inhibitors.
It seemed possible that these molecules could also affect
mitochondrial functions such as mitochondrial uncou-
pling or ATP synthesis. Despite this expectation, mito-
chondrial uncoupling activities of artepillin C isoprene
analogues were lower compared to their activities as
inhibitors of lipid peroxidation. Artepillin C (pKa =
9.64) is much less acidic than the potent mitochondrial
uncoupler SF6847 (pKa = 6.86). Thus, the phenolic
hydroxyl group of artepillin C is hardly dissociated under
physiological pH. For this reason, we felt that artepillin C
and its isoprene analogues would not function as proton-
ophore uncouplers on the mitochondrial membrane.
However, the question of inhibition of mitochondrial
ETC activities with artepillin C and its isoprene analogues
at high dose remains open. On the other hand, inhibition
of ATP synthesis by artepillin C and its isoprene ana-
logues was observed. However, the IC50 values for inhibi-
tion of ATP synthesis were higher by about one to two
orders of magnitude than those for inhibition of lipid per-
oxidation. Interestingly, we found a good correlation
(r = 0.973) between the magnitude of hydrophobic
parameter logP values and the potency of inhibition of
ATP synthesis (Fig. 2). Therefore, we propose that arte-
pillin C and its isoprene analogues may be a useful inhib-
itor of lipid peroxidation in particular because of its low
mitochondrial toxicity.
logues as inhibitors of ATP synthesis (Table 3). As
shown in Table 3, activities as inhibitors of ATP synthe-
sis were lower than the activities as inhibitors of RLM
lipid peroxidation by a factor of two orders of
magnitude.
3. Discussion
In this paper, we designed and synthesized isoprene ana-
logues of artepillin C and investigated their reactivity to-
wards free radicals and activities as inhibitors of RLM
lipid peroxidation, and as well as RLM toxicity. Our
molecular orbital calculations suggest that the conversion
of two-prenyl group on artepillin C into various isoprene
side chains does not influence the HOMO or LUMO ener-
gies and electron distribution on artepillin C. Thus we pre-
dicted the two-prenyl groups on artepillin C do not
contribute to the reactivity toward free radicals. In fact,
essentially no difference in the reactivity of all isoprene
analogues of artepillin C with the ABTS radical could
be detected. On the other hand, we have already revealed
that di-isoprene analogues, including artepillin C, have
higher reactivity toward the DPPH (1,1-diphenyl-2-pic-
rylhydrazyl) free radical than do mono-isoprene ana-
logues.17 These results may be explained in terms of
difference in stability of the two free radicals. From these
results we conclude that the reactivity of antioxidants to-
ward free radical dependson the nature of the free radicals
and the experimental conditions. Thus, care must be tak-
en in the design of radical scavenging experiments in eval-
uation of the potency of antioxidants.
Excluding post-modification products, at present only
three p-hydroxycinnamic acid derivatives having iso-
prene side chains have been isolated from natural sourc-
es, namely artepillin C (two prenyl side chain), drupanin
(one prenyl side chain), and mono-geranyl ‘TX2012’. It
is interesting to note that naturally occurring small mol-
ecules having isoprene side chains longer than the gera-
nyl group, such as the farnesyl group or larger, have not
found except for tocotrienol (vitamin E), vitamin K2,
and ubiquinone, all required by a living organism. Based
on this, we have suggested that these highly hydropho-
bic isoprene side chains became mandatory for these
molecules as the assumed important biological functions
in lipid membrane. For example, Ras proteins are guan-
ine nucleotide-binding proteins that undergo post-trans-
lational modifications,24 such as farnesylation, that lead
to their localization tin the plasma membrane. Further-
more, natural products having isoprene side chains iso-
lated from plants seem limited to those possessing the
prenyl side chain. With respect to this difference, we pro-
posed that an optimal hydrophobicity of essential bio-
It is well known that mitochondria are susceptible to lip-
id peroxidation by reactive free radicals because of the
presence in the mitochondrion of a high concentration
of unsaturated lipid.2 Consequently, RLM are widely
used as a preparation to evaluate the activities of inhib-
itors of lipid peroxidation. In this study, we have shown
that artepillin C, TX-2007, and TX-2013 have essentially
the same activities as inhibitors of RLM lipid peroxida-
tion. This result suggests that isoprene chain elongation
of artepillin C has no effect on activity with respect to
inhibition of RLM lipid peroxidation. However, the
inhibition of RLM lipid peroxidation by mono-prenyl
analogue TX-1959 was greatly decreased compared to
artepillin C, probably because of its lower hydrophobic-
ity (logP = 3.53). Interestingly, the potency of inhibition
of RLM lipid peroxidation was increased by the replace-

Y. Uto et al. / Bioorg. Med. Chem. 14 (2006) 5721–5728
5725
on Kanto Chemical silica gel 60N (230–400 mesh). All
the chemicals were purchased from Wako Pure Chemi-
cal Industries, Ltd (Osaka, Japan), Kanto Chemical
Co., Inc. (Tokyo, Japan), Tokyo Chemical Industry
Co., Ltd (Tokyo, Japan), and Sigma–Aldrich Japan
(Tokyo, Japan).
5.2. Molecular modeling methods
The semi-empirical MO calculation was applied by
the AM1 methods of Stewart and the program
MOPAC2000 (Fujitsu, Japan).25 The orbital energy
(eV) of the HOMO and LUMO was calculated.
5.3. General procedure for C-prenylation: synthesis
of 1, 4, 7, 10, and 13
p-Iodophenol was dissolved in dry toluene, and then
NaH (1.1 mol equiv) was added. After the reaction mix-
ture was stirred at room temperature for 5 min, prenyl
or geranyl or farnesyl bromide (1.0 or 2.0 per mol of
p-iodophenol) was added. After the reaction mixture
was stirred at room temperature for 1 h, the mixture
was poured into ice water and then acidified with 2 M
CH3COOH. The mixture was extracted with Et2O and
washed with saturated aqueous NaHCO3 followed by
saturated aqueous NaCl. The Et2O layer was dried
(anhydrous MgSO4) and evaporated under reduced
pressure. Residues were purified by silica gel column
chromatography with hexanes and Et2O to give the
product as a colorless oil.
Figure 2. Hydrophobic parameter logP versus inhibitory activity of
ATP synthesis (1/IC50).
logical molecules exists. Thus, such natural products are
prenylated to protect them from being too strongly
bound to membrane but they have sufficient hydropho-
bicity to express membrane associated biological
activities. If this hypothesis were true, novel mem-
brane-bound biologically active small molecules might
be designed by artificial modification of various plant
materials with farnesyl side chains.
5.3.1. 4-Iodo-2-(3-methyl-2-butenyl)phenol (1). Yield
1
4. Conclusion
58%. H NMR (CDCl3) d 7.35 (m, 2H), 6.55 (d, 1H,
J = 8.4 Hz), 5.92 (s, 1H), 5.26 (t, 1H, J = 7.2 Hz), 3.27
(d, 2H, J = 7.2 Hz), 1.75 (s, 6H).
In this report, we report the design and synthesis of nov-
el artepillin C isoprene analogues, including natural and
synthetic molecules, based on isoprenomics. We further
demonstrate that these isoprene analogues may function
as potent lipid peroxidation inhibitors having low mito-
chondrial toxicity. We also show that elongation of iso-
prene side chain of artepillin C has little influence on
potency of inhibition RLM lipid peroxidation but leads
to an increase in RLM toxicity.
5.3.2.
2,6-Di[(2E)-3,7-dimethyl-2,6-octadienyl]-4-iod-
1
ophenol (4). Yield 46%. H NMR (CDCl3) d 7.26 (s,
2H), 5.37 (s, 1H), 5.25 (td, 2H, J = 7.3 Hz, 1.2 Hz),
5.07 (m, 2H), 3.29 (d, 4H, J = 7.3 Hz), 2.09 (m, 8H),
1.73 (s, 6H), 1.69 (s, 6H), 1.60 (s, 6H).
5.3.3. 2-[(2E)-3,7-Dimethyl-2,6-octadienyl]-4-iodophenol
1
(7). Yield 62%. H NMR (CDCl3) d 7.38 (m, 2H), 6.58
(dd, 1H, J = 6.3 Hz, 2.7 Hz), 5.27 (td, 1H, J = 7.1 Hz,
1.2 Hz), 5.16 (s, 1H), 5.06 (m, 1H), 3.30 (d, 2H,
J = 7.1 Hz), 2.10 (m, 4H), 1.76 (s, 3H), 1.69 (s, 3H),
1.60 (s, 3H); EIMS m/e: 356 (M+).
5. Experimental
5.1. General procedures
1H NMR spectra were recorded on a JEOL JNM-
EX400 spectrometer (400 MHz) with tetramethylsilane
as the internal standard. Chemical shifts are reported
in ppm. Coupling constants are reported in Hz. Mass
5.3.4. 2-[(2E)-3,7-Dimethyl-2,6-octadienyl]-4-iodo-6-(3-
methyl-2-butenyl)phenol (10). Yield 57%. 1H NMR
(CDCl3) d 7.26 (s, 2H), 5.38 (s, 1H), 5.26 (t, 2H,
J = 7.2 Hz), 5.07 (t, 1H, J = 6.8 Hz), 3.28 (t, 4H,
J = 6.6 Hz), 2.11 (m, 4H), 1.77 (s, 3H), 1.74 (s, 6H),
1.69 (s, 3H), 1.60 (s, 3H); EIMS m/e: 424 (M+).
spectra were measured on
a Shimadzu GC-MS
QP-1000 mass spectrometer using an electron ionization
(EI) method. All melting points were determined by
Yanagimoto micro melting point apparatus (MP-S3
model) and are uncorrected. Elemental analyses were
performed with a Yanako CHN recorder MT-5. Reac-
tions were monitored by analytical thin-layer chroma-
tography (TLC) with use of Merck silica gel 60F254
glass plates. Column chromatography was performed
5.3.5. 4-Iodo-2-[(2E,6E)-3,7,11-trimethyl-2,6,10-dodecat-
1
rienyl]phenol (13). Yield 75%. H NMR (CDCl3) d 7.38
(m, 2H), 6.57 (dd, 1H, J = 8.8 Hz, 1.7 Hz), 5.27 (t, 1H,
J = 7.2 Hz), 5.14 (s, 1H), 5.08 (t, 2H, J = 6.1 Hz), 3.31
(d, 2H, J = 7.2 Hz), 2.05 (m, 8H), 1.76 (s, 3H), 1.67
(s, 3H), 1.60 (s, 6H); EIMS m/e: 424 (M+).

5726
Y. Uto et al. / Bioorg. Med. Chem. 14 (2006) 5721–5728
5.4. General procedure for acetylation: synthesis of 2, 5, 8,
11, and 14
5.5.2. Methyl (E)-3-4-(acetyloxy)-3,5-di[(2E)-3,7-dimeth-
yl-2,6-octadienyl]phenyl-2-propenoate (6). Yield 65%. H
1
NMR (CDCl3) d 7.62 (d, 1H, J = 16.0 Hz), 7.24 (s, 2H),
6.34 (d, 1H, J = 16.0 Hz), 5.23 (t, 2H, J = 7.1 Hz), 5.10
(m, 2H), 3.79 (s, 3H), 3.21 (d, 4H, J = 7.1 Hz), 2.32 (s,
3H), 2.08 (m, 8H), 1.68 (s, 12H), 1.60 (s, 6H); EIMS
m/e: 450 ([MꢀAc]+).
Prenylated p-iodophenol (1, 4, 7, 10, 13) and DMAP
(1.2 mol equiv) were dissolved in dry CH2Cl2, and then
AcCl (2.0 mol equiv) was added dropwise. The mixture
was refluxed for 4 h, the mixture was cooled to room
temperature and evaporated. The residue was chromato-
graphed over silica gel eluated by hexanes/Et2O (10:1) to
afford acetylated compound as colorless oil.
5.5.3. Methyl (E)-3-4-(acetyloxy)-3-[(2E)-3,7-dimethyl-
2,6-octadienyl]phenyl-2-propenoate (9). Yield 91%. 1H
NMR (CDCl3) d 7.65 (d, 1H, J = 16.1 Hz), 7.39 (m,
2H), 7.05 (dd, 1H, J = 7.6 Hz, 1.0 Hz), 6.37 (d, 1H,
J = 16.1 Hz), 5.23 (td, 1H, J = 7.3 Hz, 1.2 Hz), 5.10
(td, 1H, J = 6.6 Hz, 1.2 Hz), 3.80 (s, 3H), 3.25 (d, 2H,
J = 7.3 Hz), 2.32 (s, 3H), 2.08 (m, 4H), 1.70 (s, 3H),
1.68 (s, 3H), 1.60 (s, 3H); EIMS m/e: 356 (M+).
5.4.1. 4-Iodo-2-(3-methyl-2-butenyl)phenyl acetate (2).
1
Yield 96%. H NMR (CDCl3) d 7.52 (m, 2H), 6.77 (d,
1H, J = 8.0 Hz), 5.17 (t, 1H, J = 6.8 Hz), 3.17 (d, 2H,
J = 6.8 Hz), 2.29 (s, 3H), 1.76 (s, 6H).
5.4.2.
2,6-Di[(2E)-3,7-dimethyl-2,6-octadienyl]-4-iod-
ophenyl acetate (5). Yield 88%. 1H NMR (CDCl3) d
7.37 (s, 2H), 5.19 (t, 2H, J = 7.1 Hz), 5.09 (m, 2H),
3.14 (d, 4H, J = 7.1 Hz), 2.29 (s, 3H), 2.06 (m, 8H),
1.69 (s, 6H), 1.66 (s, 6H), 1.60 (s, 6H); EIMS m/e: 491
([MꢀAc]+).
5.5.4. Methyl (E)-3-{4-(acetyloxy)-3-[(2E)-3,7-dimethyl-
2,6-octadienyl]-5-(3-methyl-2-butenyl)phenyl}-2-propeno-
1
ate (12). Yield 91%. H NMR (CDCl3) d 7.62 (d, 1H,
J = 16.0 Hz), 7.23 (s, 2H), 6.35 (d, 1H, J = 16.0 Hz),
5.23 (m, 2H), 5.11 (t, 1H, J = 6.8 Hz), 3.80 (s, 3H),
3.20 (t, 4H, J = 6.3 Hz), 2.32 (s, 3H), 2.09 (m, 4H),
1.76 (s, 3H), 1.69 (s, 6H), 1.68 (s, 3H), 1.61 (s, 3H);
EIMS m/e: 424 (M+).
5.4.3. 2-[(2E)-3,7-Dimethyl-2,6-octadienyl]-4-iodophenyl
acetate (8). Yield 89%. H NMR (CDCl3) d 7.50 (m,
1
2H), 6.77 (d, 1H, J = 8.3 Hz), 5.19 (td, 1H, J = 7.3 Hz,
1.1 Hz), 5.10 (t, 1H, J = 6.7 Hz), 3.19 (d, 2H,
J = 7.3 Hz), 2.30 (s, 3H), 2.07 (m, 4H), 1.69 (s, 3H),
1.67 (s, 3H), 1.61 (s, 3H); EIMS m/e: 398 (M+).
5.5.5. Methyl (E)-3-4-(acetyloxy)-3-[(2E,6E)-3,7,11-tri-
methyl-2,6,10-dodecatrienyl]phenyl-2-propenoate
(15).
7.65 (d, 1H,
Yield 77%. 1H NMR (CDCl3)
d
J = 16.1 Hz), 7.39 (m, 2H), 7.05 (dd, 1H, J = 7.3 Hz,
1.7 Hz), 6.37 (d, 1H, J = 16.1 Hz), 5.23 (td, 1H,
J = 7.3 Hz, 1.2 Hz), 5.10 (m, 2H), 3.80 (s, 3H), 3.26 (d,
2H, J = 7.3 Hz), 2.31 (s, 3H), 2.03 (m, 8H), 1.70 (s,
3H), 1.67 (s, 3H), 1.60 (s, 3H), 1.59 (s, 3H); EIMS m/
e: 424 (M+).
5.4.4. 2-[(2E)-3,7-Dimethyl-2,6-octadienyl]-4-iodo-6-(3-
methyl-2-butenyl)phenyl acetate (11). Yield 99%. 1H
NMR (CDCl3) d 7.37 (s, 2H), 5.17 (m, 2H), 5.10 (m,
1H), 3.13 (t, 4H, J = 6.3 Hz), 2.30 (s, 3H), 2.08 (m,
4H), 1.74 (s, 3H), 1.69 (s, 3H), 1.67 (s, 3H), 1.66 (s,
3H), 1.61 (s, 3H); EIMS m/e: 466 (M+).
5.6. General procedure for deprotection: synthesis of
TX-1959, TX-2007, TX-2012, TX-2013, and TX-2101
5.4.5. 4-Iodo-2-[(2E,6E)-3,7,11-trimethyl-2,6,10-dodecat-
rienyl]phenyl acetate (14). Yield 98%. 1H NMR (CDCl3)
d 7.52 (m, 2H), 6.77 (d, 1H, J = 8.3 Hz), 5.19 (td, 1H,
J = 7.3 Hz, 1.2 Hz), 5.10 (m, 2H), 3.19 (d, 2H,
J = 7.3 Hz), 2.29 (s, 3H), 2.05 (m, 8H), 1.76 (s, 3H),
1.67 (s, 3H), 1.60 (s, 6H); EIMS m/e: 466 (M+).
KOH (5% aqueous solution) was added to a solution of
the a, b-unsaturated carboxylate (3, 6, 9, 12, and 15) in
MeOH. The mixture was refluxed for 3 h, cooled to
0 °C, and acidified with 1 N HCl. The MeOH was evap-
orated under reduced pressure, and the aqueous residue
was extracted with Et2O. Extracts were washed with sat-
urated aqueous NaHCO3 and followed saturated aque-
ous NaCl. The Et2O layer was dried (anhydrous
MgSO4) and evaporated under reduced pressure. Resi-
dues were purified by column chromatography on silica
gel (CH2Cl2/MeOH) to give artepillin C isoprene
analogues.
5.5. General procedure for Mizoroki–Heck coupling:
synthesis of 3, 6, 9, 12, 15
A solution of acetylated compound (2, 5, 8, 11, and 14),
methyl acrylate (5.0 mol equiv), Et3N (2.0 mol equiv),
(o-tol)3P (10% mol equiv), and Pd(OAc)2 (5% mol e-
quiv) in dry toluene was refluxed for 20 h. The mixture
was cooled to room temperature, diluted with Et2O,
and filtered through celite. The filtrate was evaporated
under reduced pressure. Residues were purified by col-
umn chromatography on silica gel (hexanes/Et2O) to
give a,b-unsaturated carboxylate as a light yellow oil.
5.6.1. TX-1959 (drupanin19). (E)-3-[4-Hydroxy-3-(3-
methyl-2-butenyl)phenyl]-2-propenoic acid. White solid.
Yield 74%. 1H NMR (CDCl3)
d 7.71 (d, 1H,
J = 15.6 Hz), 7.34 (m, 2H), 6.82 (d, 1H, J = 8.4 Hz),
6.30 (d, 1H, J = 15.6 Hz), 5.31 (t, 1H, J = 6.8 Hz), 3.37
(d, 2H, J = 6.8 Hz), 1.79 (s, 6H). Anal. Calcd for
C14H16O3: C, 72.39; H, 6.94. Found: C, 72.14; H, 6.86.
5.5.1. Methyl (E)-3-[4-(acetyloxy)-3-(3-methyl-2-bute-
nyl)phenyl]-2-propenoate (3). Yield 93%. 1H NMR
(CDCl3) d 7.65 (d, 1H, J = 15.6 Hz), 7.28 (m, 2H),
7.05 (d, 1H, J = 8.8 Hz), 6.38 (d, 1H, J = 15.6 Hz),
5.21 (t, 1H, J = 6.8 Hz), 3.80 (s, 3H), 3.24 (d, 2H,
J = 6.8 Hz), 2.32 (s, 3H), 1.76 (s, 3H), 1.72 (s, 3H).
5.6.2. TX-2007. (E)-3-{3,5-Di[(2E)-3,7-dimethyl-2,6-
octadienyl]-4-hydroxyphenyl}-2-propenoic acid. Light
yellow oil. Yield 85%. 1H NMR (CDCl3) d 7.70 (d,

Y. Uto et al. / Bioorg. Med. Chem. 14 (2006) 5721–5728
5727
1H, J = 15.9 Hz), 7.21 (s, 2H), 6.28 (d, 1H, J = 15.9 Hz),
5.9. Inhibition activity of RLM lipid peroxidation
5.74 (br s, 1H), 5.32 (t, 2H, J = 6.4 Hz), 5.09 (t, 2H,
J = 6.8 Hz), 3.36 (d, 4H, J = 7.0 Hz), 2.11 (m, 8H),
1.77 (s, 6H), 1.69 (s, 6H), 1.61 (s, 6H); EIMS m/e: 436
(M+). Anal. Calcd for C29H40O3: C, 79.77; H, 9.23.
Found: C, 79.93; H, 9.43.
Lipid peroxidation in RLM was performed at 25 °C and
monitored as oxygen consumption with a Clark-type
oxygen electrode in a total volume of 2.53 ml. Mito-
chondria were added at concentrations of 0.7 mg pro-
tein/ml. Peroxidation was started by addition of final
concentrations of 1 mM ADP and 100 lM FeSO4 in
medium consisting of 175 mM KCl and 10 mM Tris–
HCl buffer, pH 7.4. The amount of O2 consumed during
peroxidation was calculated assuming that the satura-
tion concentration of O2 at 25 °C is 258 lM. The percent
inhibition with drug on RLM lipid peroxidation was
calculated by the following equation:
5.6.3. TX-2012. (E)-3-{3-[(2E)-3,7-Dimethyl-2,6-octa-
dienyl]-4-hydroxyphenyl}-2-propenoic acid. Light yel-
1
low oil. Yield 89%. H NMR (CDCl3) d 7.72 (d, 1H,
J = 15.9 Hz), 7.33 (m, 2H), 6.83 (d, 1H, J = 8.1 Hz),
6.30 (d, 1H, J = 15.9 Hz), 5.32 (td, 1H, J = 7.2 Hz,
1.2 Hz), 5.07 (t, 1H, J = 6.6 Hz), 3.38 (d, 2H,
J = 7.2 Hz), 2.12 (m, 4H), 1.78 (s, 3H), 1.69 (s, 3H),
1.60 (s, 3H); EIMS m/e: 300 (M+). Anal. Calcd for
C19H24O3: C, 75.97; H, 8.05. Found: C, 75.67; H, 8.06.
Inhibitionð%Þ ¼ f1 ꢀ ðRpt=tinhÞ ꢁ kg ꢁ 100%;
k ¼ tinh0=Rp0t0;
where Rp, rate of lipid peroxidation (nmol O/min); Rp0,
rate of lipid peroxidation of control (nmol O/min); t, to-
tal time (min); t0, total time of control (min); tinh, induc-
tion time (min); tinh0, induction time of control (min).
The mean effective concentration of antioxidant tested
required for its 50% inhibition was defined as IC50 in
the RLM lipid peroxidation reaction.
5.6.4. TX-2013. (E)-3-{3-[(2E)-3,7-Dimethyl-2,6-octa-
dienyl]-4-hydroxy-5-(3-methyl-2-butenyl)phenyl}-2-
1
propenoic acid. Light yellow oil. Yield 88%. H NMR
(CDCl3) d 7.70 (d, 1H, J = 15.9 Hz), 7.20 (s, 2H), 6.28
(d, 1H, J = 15.9 Hz), 5.73 (brs, 1H), 5.31 (t, 2H,
J = 7.1 Hz), 5.08 (m, 1H), 3.36 (t, 4H, J = 7.1 Hz),
2.11 (m, 4H), 1.79 (s, 3H), 1.77 (s, 6H), 1.69 (s, 3H),
1.61 (s, 3H); EIMS m/e: 368 (M+). Anal. Calcd for
C24H32O3: C, 78.22; H, 8.75. Found: C, 78.03; H, 8.59.
5.10. Mitochondrial uncoupling activity
5.6.5. TX-2101. (E)-3-{4-Hydroxy-3-[(2E,6E)-3,7,11-tri-
methyl-2,6,10-dodecatrienyl]phenyl}-2-propenoic acid.
Light yellow oil. Yield 94%. H NMR (CDCl3) d 7.72
(d, 1H, J = 15.9 Hz), 7.33 (m, 2H), 6.82 (d, 1H,
J = 8.3 Hz), 6.30 (d, 1H, J = 15.9 Hz), 5.32 (td, 1H,
J = 7.1 Hz, 1.1 Hz), 5.08 (m, 2H), 3.38 (d, 2H,
J = 7.1 Hz), 2.06 (m, 8H), 1.79 (s, 3H), 1.67 (s, 3H),
1.60 (s, 6H); EIMS m/e: 368 (M+). Anal. Calcd for
C24H32O3: C, 78.22; H, 8.75. Found: C, 78.05; H, 8.73.
The respiration of mitochondria was monitored polaro-
graphically with a Clark-type oxygen electrode. The
incubation medium consisted of 200 mM sucrose,
2 mM MgCl2, and 1 mM EDTA in 10 mM potassium
phosphate buffer (pH 7.4). Mitochondria were added
at 0.7 mg protein/ml in a total volume of 2.53 ml succi-
nate (10 mM, plus rotenone at 1 lg/mg protein) was the
respiratory substrate. The activity was determined by
measuring changes in the rate of state 4 respiration upon
addition of a test compound.
1
5.7. Antioxidant activity
5.11. Inhibitory activity of ATP synthesis
Reactivity of artepillin C isoprene analogues with ABTS
(2,20-Azinobis(3-ethylbenzothiazoline-6-sulfonate)) rad-
ical cations was measured according to the method of
Re et al.26 A stock solution of ABTS radical cations
was prepared one day before the assay by 5 ml of
7 mM ABTS and 2.45 mM potassium persulfate, fol-
lowed by storage in dark at room temperature. The
stock solution of ABTS radical cations was diluted with
water or ethanol to an absorbance of about 0.70 at
734 nm. The decolorization assay was started by mixing
2 ml of the diluted ABTS solution with 20 ll of a test
compound solution in a cuvette. Test compounds were
dissolved with ethanol. At 6 min after mixing, the absor-
bance was measured at 734 nm. In a control experiment,
a test compound solution was replaced with ethanol,
and the radical scavenging activity of the test compound
was expressed as IC50 value.
ATP synthesis in mitochondria was determined as
reported by Nishimura et al.28 by measuring the increase
in pH of the medium associated with ATP synthesis at
25 °C in medium consisting of 200 mM sucrose,
20 mM KCl, 3 mM MgCl2, and 3 mM potassium phos-
phate buffer, pH 7.4, in a total volume of 1.68 ml. The
reaction was started by adding 160 lM ADP to the
mitochondrial suspension (0.7 mg protein/ml) energized
with 5 mM succinate (plus rotenone at 1 lg/mg protein).
The pH change was monitored using a pH meter.
Acknowledgment
This work was supported in part by the Intramural
Research Program of the NIH, NIDDK.
5.8. Preparation of rat liver mitochondria (RLM)
RLM were isolated from male Wistar rats (140–220 g)
as reported by Myers and Slater.27 The mitochondrial
protein content was determined by the biuret method
using bovine serum albumin as a standard.
References and notes
1. Raha, S.; Robinson, B. H. Trends Biochem. Sci 2000, 25,
502.

5728
Y. Uto et al. / Bioorg. Med. Chem. 14 (2006) 5721–5728
2. Richardson, T.; Tappel, A. L.; Gruger, E. H., Jr. Arch.
Biochem. Biophys. 1961, 94, 1.
15. Uto, Y.; Ae, S.; Nagasawa, H.; Hori, H. ACS Symp. Ser.
2005, 909, 176.
3. Penta, J. S.; Johnson, F. M.; Wachsman, J. T.; Copeland,
W. C. Mutat. Res. 2001, 488, 119.
4. Harman, D. J. Am. Geriatr. Soc. 1972, 20, 145.
5. Crane, F. L.; Hatefi, Y.; Lester, R. L.; Widmer, C.
Biochim. Biophys. Acta 1957, 25, 220.
6. Mellors, A.; Tappel, A. L. J. Biol. Chem. 1966, 241, 4353.
7. James, A. M.; Smith, R. A.; Murphy, M. P. Arch.
Biochem. Biophys. 2004, 423, 47.
16. Rodriguez, R. J.; Miranda, C. L.; Stevens, J. F.; Deinzer,
M. L.; Buhler, D. R. Food Chem. Toxicol. 2001, 39, 437.
17. Uto, Y.; Ae, S.; Hotta, A.; Terao, J.; Nagasawa, H.; Hori,
H. Adv. Exp. Med. Biol. 2006, 578, in press.
18. Simmons, D. A.; Luderitz, O.; Westphal, O. Biochem. J.
1965, 97, 807.
19. Bohlmann, F.; Jakupovic, J. Phytochemistry 1979, 18,
1189.
8. Huneck, S.; Zdero, C.; Bohlmann, F. Phytochemistry
1986, 25, 883.
20. Jakupovic, J.; Warning, U.; Bohlmann, F.; King, R. M.
Rev. Latinoamer. Quim. 1987, 18, 75.
9. Aga, H.; Shibuya, T.; Sugimoto, T.; Nakajima, S.;
Kurimoto, M. Biosci. Biotechnol. Biochem. 1994, 58, 945.
10. Marcucci, M. C.; Ferreres, F.; Garcia-Viguera, C.; Bank-
ova, V. S.; De Castro, S. L.; Dantas, A. P.; Valente, P. H.
M.; Paulino, N. J. Ethnopharmacol. 2001, 74, 105.
11. Matsuno, T.; Jung, S. K.; Matsumoto, Y.; Saito, M.;
Morikawa, J. Anticancer Res. 1997, 17, 3565.
12. Kimoto, T.; Arai, S.; Aga, M.; Hanaya, T.; Kohguchi, M.;
Nomura, Y.; Kurimoto, M. Gan To Kagaku Ryoho 1996,
23, 1855.
13. Kimoto, T.; Arai, S.; Kohguchi, M.; Aga, M.; Nomura,
Y.; Micallef, M. J.; Kurimoto, M.; Mito, K. Cancer
Detect. Prev. 1998, 22, 506.
14. Hayashi, K.; Komura, S.; Isaji, N.; Ohishi, N.; Yagi, K.
Chem. Pharm. Bull. 1999, 47, 1521.
21. Uto, Y.; Hirata, A.; Fujita, T.; Takubo, S.; Nagasawa, H.;
Hori, H. J. Org. Chem. 2002, 67, 2355.
22. Muraoka, S.; Terada, H. Biochim. Biophys. Acta 1972,
275, 271.
23. Takahashi, N.; Tamagawa, K.; Kubo, Y.; Fukui, T.;
Wakabayashi, H.; Honda, T. Bioorg. Med. Chem. 2003,
11, 3255.
24. Bos, J. L. Cancer Res. 1989, 49, 4682.
25. Stewart, J. J. J. Comput. Aided Mol. Des. 1990, 4, 1.
26. Re, R.; Pellegrini, N.; Proteggente, A.; Pannala, A.; Yang,
M.; Rice-Evans, C. A. Free Radic. Biol. Med. 1999, 26,
1231.
27. Myers, D. K.; Slater, E. C. Biochem. J. 1957, 67, 558.
28. Nishimura, M.; Ito, T.; Chance, B. Biochim. Biophys. Acta
1962, 59, 177.