Chemo-enzymatic synthesis and cell-growth inhibition activity of resveratrol analogues

Article abstract of DOI:10.1016/j.bioorg.2004.08.003

The stilbenoid resveratrol (1) was subjected to regioselective acetylation catalysed by Candida antarctica lipase (CAL) to obtain 4′- acetylresveratrol (2). CAL biocatalysed regioselective alcoholysis of 3,5,4′-triacetylresveratrol (3), 3,5,4′-tributanoylresveratrol (6), and 3, 4, 5′-trioctanoylresveratrol (9) afforded derivatives 4, 5, 7, 8, 10, and 11. Further resveratrol analogues (12-18) were obtained through methylation and hydrogenation reactions, whereas the 3,4,4′- trimethoxystilbene (19) was obtained by complete synthesis. Resveratrol and its lipophylic analogues were subjected to cell-growth inhibition bioassays towards DU-145 human prostate cancer cells. Compounds 2-19 showed cell-growth inhibition activity comparable to or higher than resveratrol (GI₅₀ = 24.09 μM), displaying low or very low toxicity against non-tumorigenic human fibroblast cells. Comparison of the trimethoxy stilbenes 12 (GI₅₀ = 2.92 μM) and 19 (GI₅₀ = 25.39 μM) indicates that the position of the substituents is important for the activity. The marked activity of methyl ethers 12, 13, and 18 in comparison with that of the corresponding esters suggests that the different chemical reactivity, rather than steric factors, strongly influences the activity.

Full text of DOI:10.1016/j.bioorg.2004.08.003

Bioorganic Chemistry 33 (2005) 22–33
www.elsevier.com/locate/bioorg
Chemo-enzymatic synthesis and cell-growth
inhibition activity of resveratrol analogues
Venera Cardilea, Laura Lombardoa, Carmela Spataforab,
Corrado Tringalib,*
a
`
Dipartimento di Scienze Fisiologiche, Universita di Catania, Viale A. Doria 6, 95125 Catania, Italy
b
`
Dipartimento Scienze Chimiche, Universita di Catania, Viale A. Doria 6, 95125 Catania, Italy
Received 14 June 2004
Available online 1 October 2004
Abstract
The stilbenoid resveratrol (1) was subjected to regioselective acetylation catalysed by Can-
dida antarctica lipase (CAL) to obtain 40-acetylresveratrol (2). CAL biocatalysed regioselective
alcoholysis of 3,5,40-triacetylresveratrol (3), 3,5,40-tributanoylresveratrol (6), and 3, 4, 50-trioc-
tanoylresveratrol (9) afforded derivatives 4, 5, 7, 8, 10, and 11. Further resveratrol analogues
(12–18) were obtained through methylation and hydrogenation reactions, whereas the 3,4,40-
trimethoxystilbene (19) was obtained by complete synthesis. Resveratrol and its lipophylic
analogues were subjected to cell-growth inhibition bioassays towards DU-145 human prostate
cancer cells. Compounds 2–19 showed cell-growth inhibition activity comparable to or higher
than resveratrol (GI50 = 24.09 lM), displaying low or very low toxicity against non-tumori-
genic human fibroblast cells. Comparison of the trimethoxy stilbenes 12 (GI50 = 2.92 lM)
and 19 (GI50 = 25.39 lM) indicates that the position of the substituents is important for the
activity. The marked activity of methyl ethers 12, 13, and 18 in comparison with that of the
corresponding esters suggests that the different chemical reactivity, rather than steric factors,
strongly influences the activity.
ꢀ 2004 Elsevier Inc. All rights reserved.
Keywords: Resveratrol; Candida antarctica lipase; Synthesis; Cell-growth inhibition activity; DU-145 cells;
Prostate cancer
*
Corresponding author. Fax: +39 095 580138.
E-mail address: ctringali@unict.it (C. Tringali).
0045-2068/$ - see front matter ꢀ 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.bioorg.2004.08.003

V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
23
1. Introduction
Resveratrol (trans-3,5,40-trihydroxystilbene, 1), found in grapes as well as in other
plants, is a natural phytoalexin biosynthesized in response to pathogenic attack or
other stress conditions [1]. In recent times, it has been intensively studied for a vari-
ety of promising biological properties [2,3], among them antioxidant activity [4],
inhibition of cyclooxygenase [5], inhibition of platelet aggregation [6], antiestrogenic
activity [7]. Most of the popularity of resveratrol is probably due to its relatively high
concentration in red wines and its possible role in the prevention of heart diseases, as
evidenced by the so-called ÔFrench paradoxÕ [8]. More generally, to resveratrol have
been ascribed chemopreventive activities against human degenerative diseases such
as atherosclerosis [9] and cancer [10]. In particular, resveratrol showed cancer
chemopreventive activity in assays representing anti-initiation, anti-promotion and
antiprogression activity, inhibiting the development of preneoplastic lesions and
tumorigenesis [10]. Further evidence showed that 1 is able to inhibit cell growth
and to induce apoptosis (programmed cell death) in various human cancer cell lines
[11–15]. In particular, a number of studies have been reported on the chemopreven-
tive activity of 1 against prostate cancer [16]. More specifically, 1 showed differential
effects on growth, cell cycle arrest, and induction of apoptosis in human prostate
cancer cell lines [17], and resulted acutely toxic against DU-145 prostate cancer cells
[18]. Thus, resveratrol may also play a role in cancer therapy and may be considered
as a lead compound for preparation of analogues with enhanced bioactivity. A num-
ber of synthetic analogues of resveratrol have already been synthesized and submit-
ted to bioassays. However, the structural requirements for the anti-tumour
properties of resveratrol have not been completely established, and some conflicting
results are reported in the literature [18–23]. Minor attention has been devoted to
lipophylic analogues, and in particular to esters bearing acyl chains of different
lengths, notwithstanding that the study of this kind of derivatives may be important
in view of their enhanced affinity with lipophylic membrane constituents, and conse-
quently of the probable increase of their cell uptake.
On this basis, we decided to carry out a study of lipophylic resveratrol derivatives, as
part of our ongoing interest in bioactive natural products and their analogues. We have
recently reported a chemo-enzymatic methodology for the preparation of regioselec-
tively acylated resveratrol analogues in the presence of Candida antarctica lipase
(CAL) [24]. This method allows preparation of resveratrol esters with different substi-
tution patterns, a goal not trivial using traditional organic reagents in the presence of
functional groups with very similar reactivity, such as the phenol groups of resveratrol.
We planned also the preparation of methylated derivatives of 1, in view of the recently
reported data on the anti-tumour properties of these lipophylic analogues [19]. The pre-
pared compounds were tested for cell-growth inhibition activity towards the above
cited human prostate cancer DU-145 cell line, a model that some of us have recently
used in a study on resveratrol-induced apoptosis of human prostate cancer [25]. Thus,
we report here the results on the DU-145 cell-growth inhibition of a number of resve-
ratrol analogues. To verify the possible cytotoxicity on normal cells, we performed also
parallel experiments using primary lines of non-tumorigenic human fibroblasts.

24
V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
2. Experimental
2.1. Materials
All reagents were of commercial quality and were used as received (Merck and
Sigma-Aldrich); solvents were distilled and dried using standard techniques.
Resveratrol (1) was purchased from Sigma; the CAL (Chirazyme L-2, c. –f., C2,
Lyo) was a gift from Roche.
The 1H NMR spectra were recorded on a Varian Unity Inova spectrometer at 500
MHz and performed at constant temperature (27ꢁC). Electron impact (EI) and fast
atom bombardment (FAB) mass spectra were recorded on a Kratos MS 50 instru-
ment using 3-nitrobenzylalcohol (NBA) as matrix.
The enzymatic reactions were incubated at 40ꢁ under shaking at 400 rpm; control
reactions without enzyme were carried out under the same conditions.
The progress of the chemo-enzymatic reactions was monitored, at regular time
intervals, by thin-layer chromatography (TLC). Reactions were quenched on com-
pletion by filtering off the enzyme. Solvents were removed under reduced pressure
and the products were purified by flash chromatography on silica gel DIOL 40–
63 m (Merck) or using the Buchi-Biotage systems FLASH 12 or FLASH 40,
equipped with prepacked silica gel cartridges 12S, 12M or 40S.
Analytical TLC was performed on silica gel (Merck 60 F254) plates using cerium
sulphate as developing reagents.
2.2. Acylation reactions
The peracylated derivatives [3,5,40-tri-O-acetylresveratrol (3), 3,5,40-tri-O-buta-
noylresveratrol (6), and 3,5,40-tri-O-octanoylresveratrol (9)] were prepared under
standard conditions using acetic and butyric anhydride in pyridine (1:1) and octa-
noyl chloride in pyridine (1:1) respectively.
Mass Spectroscopy (MS)-FAB and 1H NMR spectra of 3,5,40-tri-O-acetylresvera-
trol (3) were in perfect agreement with those reported in the literature [26].
3,5,40-tri-O-Butanoylresveratrol (6): white amorphous powder; EI-MS m/z: 438
1
[M+]; H NMR (CDCl3): d 7.48 (d, 2H, J = 8.5 Hz, H-20, H-60), 7.10 (d, 2H,
J = 2.0 Hz, H-2, H-6), 7.08 (d, 2H, J = 8.5 Hz, H-30, H-50), 7.06 (d, 1H,
J = 16.0 Hz, H-b), 6.97 (d, 1H, J = 16.0 Hz, H-a), 6.81 (t, 1H, J = 2.0 Hz, H-4),
2.54 (t, 6H, J = 7.5 Hz, 3,5,40-OCOCH2CH2CH3), 1.79 (sest., 6H, J = 7.5 Hz,
3,5,40-OCOCH2CH2CH3), 1.05 (t, 9H, J = 7.5 Hz, 3,5,40-OCOCH2CH2CH3). Anal.
Calcd for C26H30O6: C, 71.21; H, 6.90; O, 21.89. Found: C, 71.92; H, 6.83; O, 21.25.
3,5,40-tri-O-Octanoylresveratrol (9): white amorphous powder; FAB-MS m/z 607
1
[M+H]+; H NMR (CDCl3): d 7.47 (d, 2H, J = 8.5 Hz, H-20, H-60), 7.10 (d, 2H,
J = 2.0 Hz, H-2, H-6), 7.07 (d, 2H, J = 8.5 Hz, H-30, H-50), 7.05 (d, 1H,
J = 16.0 Hz, H-b), 6.98 (d, 1H, J = 16.0 Hz, H-a), 6.79 (t, 1H, J = 2.0 Hz, H-4),
2.55 [t, 6H, J = 7.5 Hz, 3,5,40-COCH2(CH2)5CH3], 1.75 [pent, 6H, J = 7.5 Hz,
3,5,40- OCOCH2CH2(CH2)4CH3], 1.43 [pent, 6H, J = 7.5 Hz, 3,5,40-OCOCH2
(CH2)4CH2CH3], 1.36 [bs, 18H, 3,5,40-OCOCH2CH2(CH2)3CH2CH3], 0.90 [t, 9H,

V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
25
J = 7.5 Hz, 3,5,40-OCOCH2(CH2)5CH3]. Anal. Calcd for C38H54O6: C, 75.21; H,
8.97; O, 15.82. Found: C, 75.09; H, 8.83; O, 16.08.
2.3. Methylation reaction
Sample (1, 90 mg, 0.39 mmol) was placed into a boiling flask and dispersed with
40 mL of acetone and 60 mg of potassium carbonate; 60 ll of dimethyl sulphate was
added to this suspension, then heated for 24 h under a reflux condenser. Acetone was
removed from the mixture by a rotary evaporator. The resulting mixture was purified
by LC (silica gel, CH2Cl2 in n-hexane from 20 to 100% ), thus affording 80 mg (89%
yield) of 12 and 7.5 mg (8% yield) of 13.
1
MS-FAB and H NMR data of 3,5,40-trimethoxystilbene (12) and 3,40-dimeth-
oxystilbene (13) are in agreement with those reported in the literature [27,28].
2.4. Hydrogenation reactions
A solution of the each substrate (1, 2, 3, 4 or 12, 100 mg) in anhydrous EtOH
(10 mL) was stirred at room temperature under 2 atm H2 pressure in the presence
of Pd/C (10 mg); the reactions were monitored by TLC and stopped after 12 h, by
filtering off the catalyst. The filtrate was evaporated in vacuo and the product (14,
15, 16, 17, or 18) was obtained in quantitative yield.
MS-FAB and 1H NMR data of dihydroresveratrol (14), 3,5,40-tri-O-acetyldihydr-
oresveratrol (15), and 3,5,40-trimethoxydihydrostilbene (18) are in agreement with
those reported in the literature [27,29].
3,5-di-O-Acetyl-dihydroresveratrol (16): white amorphous powder; EI-MS m/z
1
314 [M+]; H NMR (CDCl3): d 7.02 (d, 2H, J = 8.0 Hz, H-30, H-50), 6.79 (d, 2H,
J = 2.0 Hz, H-2, H-6), 6.76 (bt, 1H, H-4), 6.74 (d, 2H, J = 8.0 Hz, H-20, H-60),
2.85 (s, 4H), 2.28 (s, 6H). Anal. Calcd for C18H18O5: C, 68.78; H, 5.77; O, 25.45.
Found: C, 69.02.; H, 5.93; O, 25.05.
40-O-Acetyl-dihydroresveratrol (17): white amorphous powder. EI-MS m/z 272
1
[M+]; H NMR (CD3OD): d 7.17 (d, 2H, J = 8.5 Hz, H-20, H-60), 6.96 (d, 2H,
J = 8.5 Hz, H-30, H-50), 6.14 (bd, 2H, H-2, H-6), 6.10 (bt, 1H, H-4), 2.85 (t, 2H,
J = 7.2 Hz, H-a), 2.73 (t, J = 7.2 Hz, 2H, H-b), 2.24 (s, 3H, 40-OCOCH3). Anal.
Calcd for C16H16O4: C, 70.57; H, 5.92; O, 23.5. Found: C, 70.81; H, 5.84; O, 23.35.
2.5. Enzymatic acetylation of 1
The esterification reaction was carried out in sealed flasks at 40ꢁC. Resveratrol (1,
500 mg, 2.19 mmol) was dissolved in t-amyl alcohol (100 mL) and vinyl acetate
(2 mL, 21.6 mmol). CAL (500 mg) was added to a solution of 1. After 90 h incuba-
tion, the reactions were quenched and the residues purified by flash chromatography
on prepacked silica gel cartridge (from 1 to 4% di-MeOH/CH2Cl2) to give 218 mg
(43 %) of 2.
1
40-O-Acetylresveratrol (2): MS-FAB and H NMR spectra of 2 were identical to
those reported in the literature [24].

26
V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
2.6. Enzymatic alcoholysis
Each substrate (3, 6, or 9, 500 mg) was dissolved in t-butylmethylether (50 mL)
and n-BuOH (2 mL). CAL (500 mg) was added to this solution and the mixture
was maintained under shaking (400 rpm) at 40ꢁC for 0.5 h (3) or 48 h (6 and 9).
The reactions were quenched by filtering off the enzyme and the filtrate was evapo-
rated in vacuo. The residue was subjected to flash chromatography to obtain the
deacylated analogues.
1
MS-FAB and H NMR spectra of 3,5-di-O-acetylresveratrol (4) and 3-O-acetyl-
resveratrol (5) are in perfect agreement with those reported in the literature [24].
3,5-di-O-Butanoylresveratrol (7): white amorphous powder, 256 mg (61% yield)
purified by LC (silica gel, from CH2Cl2 to 3% EtOAc–CH2Cl2); EI-MS m/z 368
1
[M+]; H NMR (CDCl3): d 7.34 (d, 2H, J = 8.5 Hz, H-20, H-60), 7.06 (d, 2H,
J = 2.0 Hz, H-2, H-6), 6.98 (d, 1H, J = 16.0 Hz, H-b), 6.82 (d, 1H, J = 16.0 Hz,
H-a), 6.78 (d, 2H, J = 8.5 Hz, H-30, H-50), 6.76 (bt, 1H, H-4), 2.55 (t, 4H, J = 7.5
Hz, 3,5-OCOCH2CH2CH3), 1.79 (sest., 4H, J = 7.5 Hz, 3,5-OCOCH2CH2CH3),
1.05 (t, 6H, J = 7.5 Hz, 3,5-OCOCH2CH2CH3). Anal. Calcd for C22H24O5: C,
71.72; H, 6.57; O, 21.71. Found: C, 71.58.; H, 6.45; O, 21.97.
3-O-Butanoylresveratrol (8): white amorphous powder, 41 mg (12% yield) puri-
fied by LC (silica gel, from CH2Cl2 to 5% EtOAc–CH2Cl2); EI-MS m/z 298 [M+];
1H NMR (CD3OD): d 7.38 (d, 2H, J = 8.5 Hz, H-20, H-60), 7.03 (d, 1H,
J = 16.5 Hz, H-b), 6.87 (d, 1H, J = 16.5 Hz, H-a), 6.81 (bt, 1H, H-2), 6.77 (d, 2H,
J = 8.5 Hz, H-30, H-50), 6.71 (bt, 1H, H-6), 6.38 (t, 1H, J = 2.0 Hz, H-4), 2.55 (t,
2H, J = 7.0 Hz, 3-OCOCH2CH2CH3), 1.76 (sest., 2H, 3-OCOCH2CH2CH3), 1.05
(t, 3H, 3-OCOCH2CH2CH3). Anal. Calcd for C18H18O4: C, 72.47; H, 6.08; O,
21.45. Found: C, 72.22; H, 5.98; O, 21.8.
3,5-di-O-Octanoylresveratrol (10): white amorphous powder, 166 mg (42% yield)
purified by LC (silica gel, EtOAc-n-hexane from 10 to 25%); FAB-MS m/z 481
1
[M+H]+; H NMR (CDCl3): d 7.36 (d, 2H, J = 8.5 Hz, H-20, H-60), 7.06 (d, 2H,
J = 2.0 Hz, H-2, H-6), 6.80 (d, 2H, J = 8.5 Hz, H-30, H-50), 7.01 (d, 1H,
J = 16.0 Hz, H-b), 6.85 (d, 1H, J = 16.0 Hz, H-a), 6.76 (t, 1H, J = 2.0 Hz, H-4),
2.55 [t, 4H, J = 7.5 Hz, 3,5-COCH2(CH2)5CH3], 1.75 [pent, 4H, J = 7.5 Hz,
3,5-OCOCH2CH2(CH2)4CH3], 1.41 [pent., 4H, J = 7.5 Hz, 3,5-OCOCH2(CH2)4
CH2CH3], 1.36 [bs, 12H, 3,5-OCOCH2CH2(CH2)3CH2CH3], 0.90 [t, 6H,
J = 7.5 Hz, 3,5-OCOCH2(CH2)5CH3]. Anal. Calcd for C30H40O5: C, 74.97; H,
8.39; O, 16.64. Found: C, 75.01; H, 8.27; O, 16.72.
3-O-Octanoylresveratrol (11): white amorphous powder, 32 mg (11% yield) purified
by LC (silica gel, EtOAc-n-hexane from 10% to 30% ); FAB-MS m/z 355 [M+H]+; 1H
NMR (CDCl3): d 7.38 (d, 2H, J = 8.5 Hz, H-20, H-60), 7.06 (s, 1H, H-2), 6.79 (s, 1H, H-
6), 6.81(d, 2H, J = 8.5 Hz, H-30, H-50), 7.01 (d, 1H, J = 16.0 Hz, H-b), 6.85 (d, 1H,
J = 16.0 Hz, H-a), 6.48 (t, 1H, J = 2.0 Hz, H-4), 2.54 [t, 2H, J = 7.5 Hz, 3-
COCH2(CH2)5CH3], 1.74 [pent, 2H, J = 7.5 Hz, 3-OCOCH2CH2(CH2)4CH3], 1.41
[pent., 2H, J = 7.5 Hz, 3-OCOCH2(CH2)4CH2CH3], 1.36 [bs, 6H, 3-OCOCH2CH2
(CH2)3CH2CH3], 0.90 [t, 3H, J = 7.5 Hz, 3-OCOCH2(CH2)5CH3]. Anal. Calcd for
C22H26O4.: C, 74.55 H, 7.39; O, 18.06. Found: C, 75.01; H, 7.27; O, 17.72.

V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
27
2.7. Synthesis of 3,4,40-trimethoxystilbene (19)
4-Methoxybenzylchloride (1.1 mL, 8.13 mmol) was heated with excess of triethyl
phosphite (1.85 mL, 10.6 mmol) to 130 ꢁC to give a 4-methoxy-diethylbenzylphosph-
onate (1.82 g, 7.1 mmol, yield 87%), which was cooled to 0ꢁC, and dry DMF (10
mL) and 0.41 g (7.6 mmol) of sodium methoxide were added. To this solution,
1.18 g (7.1 mmol) of 3,4-dimethoxybenzaldeide was added to stand at room temper-
ature for 1 h. The mixture was heated to 100ꢁC, allowed to stand at this temperature
for 1 h. After cooling, the reaction mixture was stirred overnight and subsequently
quenched with water and extracted with diethylether. The combined organic layers
were washed with water and dried over Na2SO4, affording 1.57 g of the product
19 (yield 83%). EI-MS m/z 270 [M+]; 1H NMR (CDCl3): d 7.44 (d, 2H,
J = 8.0 Hz, H-30, H-50), 7.05 (d, 1H, J = 2.0 Hz, H-2), 7.02 (dd, 1H, J = 8.0,
J = 2.0 Hz, H-6), 6.92 (bs, 2H, H-a e b), 6.89 (d, 2H, J = 8.0 Hz, H-20, H-60), 6.86
(d, 1H, J = 8.0 Hz, H-5, 3.94 (s, 3H, CH3), 3.90 (s, 3H, CH3), 3.83 (s, 3H, CH3).
Anal. Calcd for C17H18O3: C, 75.53; H, 6.71; O, 17.76. Found: C, 75.37; H, 6.52;
O, 18.11.
2.8. Cell cultures
Androgen-non-responsive DU-145 human prostate cancer cells were cultured in
Earle minimal essential medium (EMEM), containing 10% fetal calf serum, 1 mM
L
-glutamine, antibiotics (50 IU/mL penicillin and 50 lg/mL streptomycin), and
1% non-essential amino acids (Invitrogen Life Technologies, Inc.).
Primary lines of non-tumorigenic, non-immortalized, human adult fibroblasts
were grown in DulbeccoÕs modified eagleÕs medium containing 10% fetal calf serum,
100 U/mL penicillin, 100 g/mL streptomycin, and 25 g/mL fungizone (Invitrogen
Life Technologies, Inc.).
The cells were incubated at 37 ꢁC in humidified 5% CO2/95% atmosphere and
transferred to subcultures every 3 days following treatment with trypsin–EDTA.
2.9. Cell-growth inhibition assay (MTT test)
The MTT assay, used to evaluate cell viability, measures the cellular capacity to
reduce 3-(4,5-dimethylthiazol-2-yl)diphenyltetrazolium bromide (MTT) to blue for-
mazan products by various mitochondrial dehydrogenase enzymes. The method de-
scribed by Mosmann and Chung [30] was employed. Briefly, DU-145 and human
adult fibroblast cells (1 · 104) were set up in flat-bottomed 200 L microplates, incu-
bated at 37ꢁC in a humidified 5% CO2/95% air mixture and, 24 h later, treated with
the resveratrol compounds at concentrations of 50, 25, 12.5, and 6.25 lM for 72 h,
except for the compound 12, which was tested also at concentrations of 3.12 and
1.56 lM for 72 h, before cell harvesting. Four hours before the end of the culture,
20 L of 0.5% MTT in phosphate buffer saline was added to each microwell. After
incubation with the reagent, the supernatant was removed and replaced with
100 lL of dimethyl sulfoxide. The optical density of each sample was measured with

28
V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
a microplate spectrophotometer reader (Titertek Multiskan, Flow Laboratories) at k
570 nm.
2.10. Statistical analysis
Each experiment was repeated at least three times in quadruplicate and the mean
SEM for each value was calculated. Statistical analysis of results was performed
using StudentÕs t test and one-way ANOVA by the statistical software package SY-
STAT, version 9 (Systat Inc., Evanston, IL, USA). A difference was considered sig-
nificant at P < 0.01.
3. Results and discussion
We report here the preparation of 17 resveratrol derivatives (2–18) obtained
through acylation, methylation, and hydrogenation of 1, as well as the complete
synthesis of the 3,4,40-trimethoxystilbene (19). These compounds were assayed for
cell-growth inhibition activity towards DU-145 human prostate cancer cells and
non-tumorigenic human fibroblasts. For the latter, the values of percent viability
of treated samples (at the maximum concentration, 50 M) vs. untreated controls
were from 80 to 100% except for the compound 19 (74%), indicating that resveratrol
analogues are not cytotoxic against normal cells. Thus, we report here only the GI50
values on DU-145 cells (see Table 1). We planned to investigate the effect on the cell-
growth inhibition activity of a regioselective substitution of the –OH groups in 1
with more lipophylic groups, like –OMe or –OCOR, which should enhance the affin-
ity with biological membranes and consequently increase the cell uptake. In addition,
we prepared also the dihydroderivatives of 1 and its acetylated (2–4) and permethy-
lated (12) analogues in order to evaluate the possible role of the central double bond.
To this purpose, we have employed both classical chemical reactions and enzymatic
methods, according to our previous work [24]. The 40-acetylresveratrol (2) was more
conveniently obtained in high yield through direct CAL-catalysed esterification with
vinyl acetate employing t-amyl alcohol as solvent (see Scheme 1A): in these condi-
tions a high selectivity of CAL was observed, so that after 90 h only the 40-acetyl
derivative was obtained. The high selectivity of the direct esterification was counter-
balanced by a low reactivity, so that the 40-acylated derivatives with longer acyl
chain could not be obtained with satisfactory yields. The majority of regioselectively
acylated resveratrol analogues were obtained through alcoholysis of a peracylated
resveratrol derivative dissolved in organic solvent and in the presence of CAL [24].
The peracylated analogues, namely 3,5,40-triacetylresveratrol (3), 3,5,40-tributa-
noylresveratrol (6), and trioctanoylresveratrol (9), were prepared from 1 according
to standard acylation procedures (see Section 2). The biocatalysed deacylation reac-
tions were carried out in t-butylmethylether, using an eccess of n-butanol as nucleo-
file (see Scheme 1B). Due to the preferential discrimination of the 40 position by
CAL, the 3,5-diacylderivative is normally obtained as the main product. This then
becomes a substrate in a further deacylation step, thus giving the 3-acyl derivative.

V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
29
Table 1
DU-145 growth inhibitiona of resveratrol (1) and analogues (2–19)
GI50 (lM)b
1
2
3
4
5
24.09 0.4
32.18 0.7
23.34 0.5
25.21 0.9
25.42 0.4
6
7
8
9
10
30.82 0.7
23.87
0.5
19.07 0.9
21.63 1.1
25.07 0.6
11
25.20 0.8
12
2.92 0.9
13
12.24 1.4
14
22.85 0.6
15
29.76 0.8
16
21.91 1.3
17
23.66 0.6
18
12.37 0.5
19
25.39 0.5
a
Human prostate tumour cell line.
b
GI50 calculated after 72 h of continuous exposure relative to untreated controls. Values are the mean
SD) of three experiments.
(
Further progression of the reaction would have afforded 1, so we stopped the reac-
tion when a convenient amount of the diacylated and monoacylated derivatives had
been obtained. It is worth noting here that the little difference in the length of the
acyl chain from acetyl to butanoyl causes a dramatic decrease in the reaction rate:
the 3,5-diacetylresveratrol is obtained with good yield in 30 min, while a comparable
amount of the 3,5-dibutanoyl can be recovered only after 30 h. Also, the diacetate 3
is more reactive than the dibutanoate 6: the latter, after 48 h, was still present in the
reaction mixture to an extent of 60%, affording also 12% of the monobutanoate 8
and only trace amounts of resveratrol. In contrast, 3 is rapidly converted into 5
and the reaction had to be stopped after 30–40 min to avoid a massive conversion
to resveratrol. A standard duration of 48 h was used for the conversion of trioctano-

30
V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
Scheme 1.
ate 9, in order to have a satisfactory yield of both the main product 10 and the mono-
acylated analogue 11.
The 3,5,40-trimethoxystilbene (12) and the related 3,40-dimethoxy-5-hydroxystil-
bene (13) were obtained as the main product and as a minor product, respectively,
following standard methylation of 1 (see Section 2). Some of our preliminary results
from bioassays as well as recent literature data prompted us to investigate the effect
of a different substitution pattern of the stilbenoids nucleus bearing three methoxy
group. Thus, we carried out a complete synthesis of trans-3,4,40-trimethoxystilbene
(19), according to Scheme 2. We employed a classical method [31] for the synthesis
of stilbenoids, through a Wittig-like reaction of 4-methoxy-diethylbenzylphospho-
nate with 3,4-dimethoxybenzaldehyde to obtain the stilbenoid 19.
The dihydroderivatives, namely dihydroresveratrol (14), 3,5,40-triacetyldihydr-
oresveratrol (15), 3,5-diacetyldihydroresveratrol (16), 40-acetyldihydroresveratrol
(17), and 3,5,40-trimethoxydihydrostilbene (18), were obtained by catalytic hydroge-
nation of the corresponding unsaturated compounds (see Section 2).
The results of the cell-growth inhibition assays on a total of 19 compounds,
including resveratrol (1), are reported as GI50 (lM) and summarised in Table 1. It
is apparent at a glance that none of the analogues shows absent or negligible activity
with respect to 1. Conversely, nine compounds show an activity higher than that of
resveratrol. In particular, the most active analogue is the 3,5,40-trimethoxystilbene
Scheme 2.

V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
31
12, showing GI50 = 2.92 lM. Also, the 3,40-dimethoxystilbene 13 (GI50 = 12.24 lM)
and the 3,4,40-dihydrotrimethoxystilbene 18 (GI50 = 12.37 lM) are significantly
more active than resveratrol, while all the other compounds show cell-growth
inhibition activity comparable to that of resveratrol, and in some cases (3,5,40-tri-
acetylresveratrol 3, 3,5,40-tributanoylresveratrol 6, 3,5-dibutanoylresveratrol 7, 3-
butanoylresveratrol 8, dihydroresveratrol 14, 3,5-dihydrodiacetylresveratrol, 16),
slightly higher. Two hydrogenated derivatives are more active than the parent com-
pound, namely 14 and 16, but small differences are observed. Conversely, the dihy-
droderivative 18 is less active than the parent compound 12, although it maintains an
activity higher than resveratrol. These results suggest that the central double bond is
not essential for the activity but may have a role, for instance, in blocking the con-
formational motion of the trimethoxy structure 12 in the highly active configuration
19. Of course, this point requires further evidence to be established. Interestingly, the
3,4,40-trimethoxystilbene 19 (GI50 = 25.39 lM) results in dramatically less activity
than the permethylated resveratrol 12, nothwithstanding the minimal structural dif-
ference between these compounds.
4. Conclusions
The present study confirmed the usefulness of biocatalysed reactions carried out
in organic solvents for the regioselective modification of bioactive polyphenols. This
methodology — employing CAL — allowed us to obtain resveratrol analogues with
a different substitution pattern, that is, 40-acetylresveratrol (2) as well as 3-acyl, 3,5-
diacyl, and triacyl derivatives with C-2, C-4, and C-8 acyl chain (3–11). Conventional
chemical conversions and synthesis afforded also compounds 12–19, and cell-growth
inhibition activity towards DU-145 prostate cancer cells was evaluated for this panel
of 19 compounds. The first conclusion we can draw is that lipophylic analogues of
resveratrol maintain cell-growth inhibition activity, which is enhanced in particular
in methyl ethers with 3,5,40 substitution pattern. A second important observation is
that the activity is maintained or increased even in the absence of free hydroxyl
groups. In fact, we report here nine derivatives showing activity comparable or high-
er than resveratrol and lacking free hydroxyl groups. This is worth noting in connec-
tion with previous literature data reporting that chemically available hydroxyl
groups are required for DU-145 cell cycle arrest and decreased cell viability [18],
and at least the 40-OH group is required for inhibition of cell proliferation [23]. Com-
parison of the two trimethoxy stilbenes 12 and 19 also indicates that the position of
the substituents is important for cell-growth inhibition activity, at least towards DU-
145 cells. The marked activity of methyl ethers 12, 13, and 18 in comparison with
that of the corresponding esters suggests that the different chemical reactivity, rather
then steric factors, strongly influences the activity. Resveratrol and its esters might be
ÔprodrugsÕ of the metabolite active in vivo, that is, a methoxystilbene formed through
biological methylation. In this hypothesis, lipophylic analogues would be better ab-
sorbed into the cell, but esters would require hydrolysis by cellular esterases to be
converted into the active O-methylated form. Of course, further studies are needed

32
V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
for better elucidation of the structural determinants and the mechanism of action of
these cell-growth inhibition stilbenoids. However, the results reported here confirm
the promising biological properties of resveratrol analogues and may offer a further
impulse to the development of stilbenoids-based anti-tumour drugs with low or ab-
sent toxicity against non-tumorigenic cells.
Acknowledgments
This work was financially supported by Ministero della Pubblica Istruzione, Uni-
`
versita e Ricerca Scientifica (PRIN 2003, Rome, Italy) and by University of Catania
(Progetti di Ricerca di Ateneo, Catania, Italy). Thanks are due to Dr. Giovanni
Nicolosi (Istituto di Chimica Biomolecolare, Sez. di Catania, Valverde, Catania,
Italy) for helpful suggestions and to Ms Ute Hill (Roche Diagnostics GmbH, Mann-
heim, Germany) for the generous gift of the CAL.
References
[1] P. Sarig, Y. Zutkhi, A. Monjauze, N. Lisker, R. Ben-Arie, Physio. Mol. Plant Pathol. 50 (1997) 337–
347.
[2] J. Gusman, H. Malonne, G. Atassi, Carcinogenesis 22 (2001) 1111–1117.
[3] L. Fremont, Life Sci. 66 (2000) 633–673.
[4] L. Belguendouz, L. Fremont, A. Linard, Biochem. Pharmacol. 53 (1997) 1347–1355.
´
[5] M. Maccarrone, T. Lorenzon, P. Guerrieri, A. Finazzi, Agro, Eur. J. Biochem. 265 (1999) 27–34.
[6] C.R. Pace-Asciak, O. Rounova, S.E. Hahn, E.P. Diamandis, D.M. Goldberg, Clin. Chimi. Acta 246
(1996) 163–182.
[7] B.D. Gehm, J.M. McAndrews, P.-Y. Chien, L. Jameson, Proc. Natl. Acad. Sci. USA 94 (1997)
14138–14143.
[8] A.Y. Sun, A. Simonyi, G. Sun, Free Radic. Biol. Med. 32 (2002) 314–318.
[9] C.R. Pace-Asciak, S. Hahn, E.P. Diamandis, G. Soleas, D.M. Goldberg, Clin. Chim. Acta 235 (1995)
207–219.
[10] M. Jang, L. Cai, G.O. Udeani, K.V. Slowing, C.F. Thomas, C.W.W. Beecher, H.H.S. Fong, N.R.
Farnsworth, A.D. Kinghorn, R.G. Mehta, R.C. Moon, J.M. Pezzuto, Science 275 (1997) 218–220.
[11] Y.-J. Surh, Y.-J. Hurh, J.-Y. Kang, E. Lee, G. Kong, S.J. Lee, Cancer Lett. 140 (1999) 1–10.
´
[12] Y. Schneider, V. Florence, D. Benoit, B. Lassina, F. Gosse, C. Bergmann, N. Seilers, F. Raoul,
Cancer Lett. 158 (2000) 85–91.
[13] K. Joe, H. Liu, M. Suzui, M.E. Vural, D. Xiao, I.B. Weinstein, Clin. Cancer Res. 8 (2002) 893–903.
[14] P.-L. Kuo, L.-C. Chiang, C.-C. Lin, Life Sci. 72 (2002) 23–34.
[15] O.P. Mgbonyebi, J. Russo, I.H. Russo, Int. J. Oncol. 12 (1998) 865–869.
[16] H.L. Ratan, W.P. Steward, A.J. Gesher, J.K. Mellon, Urol. Oncol. 7 (2002) 223–227.
[17] T.-C. Hsieh, J.M. Wu, Exp. Cell Res. 249 (1999) 109–115.
[18] G.Z. Morris, R.L. Williams, M.S. Elliot, S.J. Beebe, Prostate 52 (2002) 319–329.
[19] G.R. Pettit, M.P. Grealish, M.K. Jung, E. Hamel, R.K. Pettit, J.-C. Chapuis, J.M. Schidt, J. Med.
Chem. 45 (2002) 2534–2542.
[20] M. Roberti, D. Pizzirani, D. Simoni, R. Rondanin, R. Baruchello, C. Bonora, F. Buscemi, S.
Grimaudo, M. Tolomeo, J. Med. Chem. 46 (2002) 3546–3554.
´
[21] Y. Schneider, P. Chabert, J. Stutzamann, D. Coelho, A. Fougerousse, F. Gosse, J.-F. Launay, R.
Brouillard, F. Raul, Int. J. Cancer 107 (2003) 189–196.
[22] F. Raul, Y. Schneider, R. Brouillard, A. Fougerousse, P. Chabert, E. Gonzalez, PCT Int. Appl.
(2003) 38 (CODEN: PIXXD2 WO 2003031381).

V. Cardile et al. / Bioorganic Chemistry 33 (2005) 22–33
33
[23] L.A. Stivala, M. Savio, F. Carafoli, P. Perucca, L. Bianchi, G. Maga, L. Forti, U.M. Pagnoni, A.
Albini, E. Prosperi, V. Vannini, J. Biol. Chem. 276 (2001) 22586–22594.
[24] G. Nicolosi, C. Spatafora, C. Tringali, J. Mol. Catal., B: Enzym. 16 (2002) 223–229.
[25] V. Cardile, C. Scifo, A. Russo, M. Falsaperla, G. Morgia, M. Motta, M. Renis, E. Imbriani, G.
Silvestre, Anticancer Res. 23 (2003) 4921–4926.
[26] G.S. Jayatilake, B.J. Baker, J.B. McClintock, J. Nat. Prod. 58 (1995) 1958–1960.
[27] E. Mannila, A. Talvitie, E. Kolehmainen, Phytochemistry 33 (1993) 813–816.
[28] M.J.T.G. Gonzales, M.M.M. Pinto, A. Kijjoa, C. Anantachoke, W. Herz, Phytochemistry 32 (1993)
433–438.
[29] El-Feraly, S. Farouk, J. Nat. Prod. 47 (1984) 89–92.
[30] B.T. Mossman, A. Churg, Am. J. Respir. Crit. Care Med. 157 (1998) 1666–1680.
[31] F.W. Bachelor, A.A. Loman, L.R. Snowdon, Can. J. of Chem. 48 (1970) 1554–1557.