Antioxidant potential of natural and synthesised polyprenylated hydroquinones

Article abstract of DOI:10.1016/S0968-0896(01)00346-7

Metabolites 2-octaprenyl-1,4-hydroquinone (1) and 2-(24-hydroxy)-octaprenyl-1,4-hydroquinone (2), isolated from the sponge Ircinia spinosula, along with eight synthetic derivatives, were evaluated for their antioxidant capacity. The antioxidant potential

Full text of DOI:10.1016/S0968-0896(01)00346-7

Bioorganic & Medicinal Chemistry 10 (2002) 935–939
Antioxidant Potential of Natural and Synthesised
Polyprenylated Hydroquinones
Leto-A. Tziveleka,a Angeliki P. Kourounakis,b Panos N. Kourounakis,b
Vassilios Roussisa and Constantinos Vagiasa,*
aSchool of Pharmacy, Department of Pharmacognosy, University of Athens,
Panepistimiopolis Zografou, 157 71 Athens, Greece
bSchool of Pharmacy, Department of Medicinal Chemistry, Aristotles University of Thessaloniki,
GR-540 06 Thessaloniki, Greece
Received 9 July 2001; accepted 27 September 2001
Abstract—The metabolites 2-octaprenyl-1,4-hydroquinone (1) and 2-(24-hydroxy)-octaprenyl-1,4-hydroquinone (2), isolated from
the sponge Ircinia spinosula, along with a series of synthetic derivatives, were evaluated for their antioxidant capacity, in order to
establish a potential relationship between structural characteristics and antioxidant activity. The antioxidant potential of both
natural and synthesised compounds was evaluated in vitro by their ability: (1) to interact with the stable free 1,1-diphenyl-2-
picrylhydrazyl radical (DPPH) and (2) to inhibit the peroxidation, induced by the Fe++/ascorbate system, of heat inactivated
hepatic microsomal membrane lipids. Metabolite 1 presented a strong interaction with DPPH and had a moderate effect on lipid
peroxidation, while metabolite 2 interacted extensively with DPPH and exhibited a significant effect against lipid peroxidation. All
derivatives retaining the free 1,4-hydroquinone system maintained fully or partly the free radical scavenging capacity. # 2002
Elsevier Science Ltd. All rights reserved.
Introduction
sidered to be the most abundant representatives. Biolo-
Oxidative damage is considered to contribute to ather-
osclerosis, cardiovascular diseases, and cancer, patho-
logical conditions that are the major causes of mortality
in industrialised countries. Certain metabolites derived
from natural sources could play a preventive role due to
their antioxidant properties. Within the context of our
investigations towards the isolation of bioactive natural
products with prospects for therapeutic use, we recently
studied the black massive sponge Ircinia spinosula found
in shallow Mediterranean marine ecosystems.
gical studies conducted on several polyprenyl-
hydroquinones showed them: (1) to be effective inhibi-
tors of phospholipase A2;3 (2) to exert an interesting anti-
inflammatory action through inhibition of 5-lipoxy-
genase activity;9 (3) to have a moderate anti-microbial
action2,4 and (4) to act as general inhibitors of retroviral
reverse transcriptases as well as of cellular DNA poly-
merases.10
In addition, terpenoid 1,4-benzoquinones, such as toco-
pherols, ubiquinones and plastoquinones, along with
their reduced forms, often major metabolites in the
plants, participate in electron transport, photosynthesis,
and play an important role as endogenous cellular anti-
oxidants.11,12
The extracts of several Ircinia species have been found
to exert analgesic, muscle relaxant, anti-inflammatory,
and anti-microbial effects.1À4 In particular, a number of
biological activities of I. spinosula have been attributed
to polyprenyl-benzoquinones and polyprenyl-hydro-
quinones,5À8 which constitute the majority of its chemi-
cal content. Among this class of metabolites,
triterpenyl- and tetraterpenyl-hydroquinones are con-
In continuation of our studies on the biological activ-
ities of hydroquinones, intrigued by the aforementioned
distinctive activities, we decided to evaluate the anti-
oxidant potency of two natural hydroquinones and a
series of synthesised derivatives, in an attempt to relate
the structural features of the metabolites with the exer-
ted activity (Fig. 1).
*Corresponding author. Tel.: +30-1-7274590; fax: +30-1-7274592;
e-mail: vagias@pharm.uoa.gr
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00346-7

936
L.-A. Tziveleka et al. / Bioorg. Med. Chem. 10 (2002) 935–939
Results
activity (Fig. 2B, Table 1). The derivative that retained
most of the lipid peroxidation activity (IC50 of 350 mM)
possessed a hydrogenated side chain, a free chain
hydroxyl group and an acetylated aromatic ring (7),
(Fig. 3C, Table 1). However, the same derivative (7)
does not retain the initial DPPH activity exhibited by
compound 2 (Fig. 2B). The oxidised derivative (10)
exhibits a moderate DPPH activity (Fig. 2B) and is
completely inactive on lipid peroxidation (Table 1).
The metabolites 2-octaprenyl-1,4-hydroquinone (1) and
2-(24-hydroxy)-octaprenyl-1,4-hydroquinone (2) along
with a series of oxidised, hydrogenated, and acetylated
derivatives, afforded via chemical modifications, were
evaluated for their antioxidant activity in order to
establish the relationship between structural character-
istics and their antioxidant potential.
In the assays employed, natural product 1 presented a
strong interaction with DPPH (Fig. 2A) and a moderate
effect on lipid peroxidation (Fig. 3A, Table 1). It was
found that acetylation of the phenolic hydroxyl groups
(3) as well as the combination of acetylation and
hydrogenation of the octaprenyl side chain (6) resulted
in significant reduction of its antioxidant activity (Fig.
2A, Table 1). On the other hand, derivatives 5 and 9,
that maintain the free aromatic hydroxyl groups and the
semi-hydroquinone system, respectively, retain fully or
partly their free radical scavenging activity (Fig. 2A).
Discussion
The results obtained from the DPPH as well as the lipid
peroxidation assay employed to test the antioxidant
activity of the hydroquinones, demonstrated that nat-
ural product 2 is the best antioxidant among the com-
pounds tested (Figs 2 and 3, Table 1). Our results are
analogous to earlier data that showed enhanced anti-
inflammatory, anti-microbial and cytotoxic activity for
the hydroquinones bearing an additional hydroxyl
group on the side chain.3,4 In all above mentioned cases,
the presence of the hydroxyl group on the side chain
enhances remarkably the activity although no satisfac-
tory explanation has yet been found. In an attempt to
explain this, we have considered and estimated (by the-
oretical calculations) the lipophilicity of the molecules
and the influence of the absence or presence of the
hydroxyl group on the side chain. Although in general
Natural product 2 interacted extensively with DPPH
(Fig. 2B) and exhibited a significant effect against lipid
peroxidation with an IC50 of 220 mM (Fig. 3B, Table 1).
Its antioxidant action was greater than that of com-
pound 1. Similar to metabolite 1, acetylation of the
aromatic hydroxyl groups of compound 2 (4 and 8)
resulted in a dramatic reduction of the antioxidant
Table 1. IC50 values of the different compounds tested for their lipid
peroxidation activity
Compounds
IC50 (mM)a
1
2
0.99
0.22
7
0.35
0.99
> > 1
9
3, 4, 5, 6, 8, 10
aValues are means of three independent experiments, which did not
differ more than 2–8%, and are estimated after incubation of the
reaction mixture at 37 ꢀC for 45 min.
Figure 2. Interaction at 30 min of compounds 1 (A) and 2 (B) along
with their derivatives, with 0.2 mM of the stable free radical DPPH
(each value is the mean of three independent measurements, which did
not differ more than 5–12%).
Figure 1. Chemical structures of the natural and synthesised hydro-
quinones.

L.-A. Tziveleka et al. / Bioorg. Med. Chem. 10 (2002) 935–939
937
we cannot find any significant correlation of activity
with the calculated lipophilicity of the molecules, per-
haps this aspect is worth further exploring in a longer
series of derivatives.
delocalisation, exerting in that way radical scavenging
capacity as has been suggested for flavonoids.13
Our findings also indicate that the combined presence of
phenolic hydroxyls and the side-chain hydroxyl group
enhances the protective activity against lipid peroxida-
tion (1, 2) (Figs 3A and 2B, Table 1). It is well known
that p-hydroquinone systems participate in oxido-
reductive cycles in the presence of Fe++(+). In parti-
cular, there are indications14 pointing to either the
hydroquinone or semiquinone as the active inter-
mediates in scavenging mechanisms involved in lipid
peroxidation.
Acetylation of the phenolic hydroxyl groups resulted in
reduction of the antioxidant activity of both natural
compounds (Fig. 2, Table 1), exhibiting the same lack of
action as in earlier studies regarding the anti-microbial
action of polyprenylated hydroquinones and their syn-
thetic derivatives.4
A series of anti-microbial and cytotoxic activity studies
concerning the biological action of polyprenyl hydro-
quinones revealed that hydrogenation of the side chain
double bonds enhanced the pharmacological potency of
the molecules,2,4 while, we found that hydrogenation
per se does not affect the antioxidant activity of the
compounds tested, since derivative 5 exerts the same
action as compound 1 (Fig. 2A).
Hydroquinones are probably able to reduce highly oxi-
dising free radicals by hydrogen atom donation. Some
activity is also observed for the quinone system (9),
which may be attributed to the above mentioned prop-
erties. Finally, although the activity of 7 is unexpected,
it could be explained by partial hydrolysis of the acet-
oxy groups during the incubation period of the assay.
Despite previous studies showing that quinone deriva-
tives exert a better anti-microbial effect,2,4 the oxidised
derivative (10) exhibits a moderate DPPH activity (Fig.
2B) and is completely inactive on lipid peroxidation
(Table 1). Our results suggest that the necessary and
sufficient condition for the interaction of the com-
pounds with the stable free radical DPPH is the pre-
sence of the semi-hydroquinone system. These results
also point to the involvement of the double bonds con-
jugation with the ring oxo-groups, to facilitate electron
Conclusion
In addition to the significant antioxidant capacity of the
investigated natural products, this study revealed some
interesting structural requirements for this activity.
From the results obtained, we can attribute the proper-
ties of the active compounds to the free 1,4-hydro-
quinone system in combination with the presence of a
Figure 3. Representative graphs of the time course of lipid peroxidation as affected by various concentrations of compounds 1 (A), 2 (B) and 7 (C)
(each value is the mean of three independent measurements, which did not differ more than 2–8%).

938
L.-A. Tziveleka et al. / Bioorg. Med. Chem. 10 (2002) 935–939
hydroxy group on the octaprenyl side chain as an
enhancing moiety. It was found that changes in the
prenyl chain result in important alterations of their
antioxidant activity. These structural requirements pro-
mote the antioxidant potency of these compounds that
may be potentially interesting antioxidant supplement
candidates.
none (10) and the aldehyde 2-[24-oxy]-octaisopentyl-1,4-
diacetoxy-benzene (8), respectively. The structures of
the derivatives were confirmed by their spectral char-
acteristics.
Derivatives 3–5 and 10 were prepared as earlier descri-
bed, and their spectral features were in accordance with
previously reported data.4 Derivatives 4, 7 and 8 are
reported for first time and their spectroscopic data are
shown below.
Experimental
Chemistry
2-[24-hydroxy]-octaprenyl-1,4-diacetoxy-benzene
(4).
Ac2O (200 mL) was added to a solution of 2 (669.4 mg,
1.0 mmol) in pyridine (5 mL) in a molar ratio of 2.1:1,
and the mixture was stirred overnight at room tem-
perature. In the resulting mixture EtOAc and H2O were
added, and the organic layer was separated, washed
with water, dried over anhydrous Na2SO4, and con-
centrated in vacuo. The residue (706.2 mg) was chro-
matographed on silica gel (cyclohexane/EtOAc) to
afford 538.7 mg of 4 (71% yield), as a colorless oil; IR
(in CHCl3): 3611, 2927, 1759, 1490, 1447, 1371, 1224,
1170, 1013 cmÀ1; UV (C6H14): lmax (e)=267 (803) nm;
1H NMR (CDCl3): d 6.89–7.02 (3H, m), 5.28 (1H, t,
J=7.3 Hz), 5.20 (1H, t, J=7.3 Hz), 5.09 (6H, m), 4.08
(2H, s), 3.20 (2H, d, J=6.9 Hz), 2.27 (3H, s), 2.26
(3H, s), 1.98–2.11 (28H, m), 1.65 (6H, s), 1.57 (18H, br);
13C NMR (CDCl3): d 169.4 (s), 169.3 (s), 148.2 (s), 146.3
(s), 138.3 (s), 137.6 (s), 135.4 (s), 135.2 (s, Â2), 134.9 (s,
Â2), 134.4 (s), 131.3 (s), 128.6 (d), 124.9 (d), 124.3 (d),
124.2 (d), 124.0 (d), 123.9 (d, Â2), 122.9 (d), 122.6 (d),
120.6 (d), 119,9 (d), 60.3 (t), 39.9 (t), 39.7 (t, Â4), 35.2 (t,
Â2), 28.5 (t), 26.9 (t), 26.7 (t, Â2), 26.6 (t, Â2), 26.5 (t),
26.2 (t), 25.7 (q), 21.1 (q), 20.8 (q), 17.7 (q), 16.2 (q,),
16.0 (q).
General details. UV spectra were determined in spectro-
scopic grade C6H14 on a Shimadzu UV model 160A. IR
spectra were obtained using a Paragon 500 Perkin-
Elmer spectrophotometer. 1H and 13C NMR spectra
were recorded using a Bruker AC 200 and DRX 400
spectrometers. Chemical shifts are given on a d (ppm)
scale using TMS as internal standard (s, singlet; d,
doublet; t, triplet; m, multiplet; br, broad). High reso-
lution FAB Mass Spectra data were recorded on a
JEOL AX505HA Mass Selective Detector and were
provided by the University of Notre Dame, Department
of Chemistry and Biochemistry, Notre Dame, Indiana.
Column chromatography was performed with Kieselgel
60 (Merk); TLCs were performed with Kieselgel 60 F254
(Merk aluminum support plates).
Sponge material. I. spinosula was collected by Scuba (2–
15 m) from Saronicos Gulf, Greece and kept frozen
until analysed. A voucher specimen is deposited at the
Herbarium of the Laboratory of Pharmacognosy
(ATPH/MO/35).
Extraction and isolation. The organism was initially
freeze dried (452.0 g) and then exhaustively extracted at
room temperature with mixtures of CH2Cl2/MeOH (3:1,
v/v). The organic extract (21.3 g) after evaporation of
the solvents was subjected to vacuum column chroma-
tography using silica gel and a step gradient system
ranging from 100% CH2Cl2 to 100% EtOAc. Metabo-
lites 1 (4.28 g) and 2 (3.36 g) were isolated as oils from
the non polar fractions, and their structural confirma-
tion was based on previously reported spectral data.4,5
2-[24-hydroxy]-octaisopentyl-1,4-diacetoxy-benzene (7).
A solution of 4 (480.6 mg, 0.64 mmol) in EtOH (10
mL) was hydrogenated using 10% Pd/C under H2
atmospheric pressure. The mixture was stirred overnight
at 50 ꢀC. Then the catalyst was removed by filtration,
and the solvent was evaporated to give 489.8 mg of pure
7 (99% yield), as a colorless oil; IR (in CHCl3): 3660,
2928, 1759, 1463, 1371, 1220, 1209, 1171 cmÀ1; UV
(C6H14): lmax (e)=267 (1046) nm; FABHRMS m/z
771.6874 [M+H]+ (calcd. for C50H90O5 770.6792); m/z
(% rel. int.) 123 (100), 165 (78), 207 (9), 263 (6), 333 (4),
403 (2), 473 (1) 557 (1), 668 (20), 686 (6), 711 (3), 728
Chemical modifications and spectral data. Hydro-
quinone derivatives were prepared by simple chemical
manipulations, such as oxidation, acetylation of the
hydroxyl groups, and hydrogenation of the chain dou-
ble bonds (Fig. 1). Polyprenylated hydroquinones 1 and
2 were treated with Ac2O in dry pyridine to afford the
corresponding acetylated products, 2-octaprenyl-1,4-
diacetoxy-benzene (3) and 2-[24-hydroxy]-octaprenyl-
1,4-diacetoxy-benzene (4). Reduction of the metabolite
1, as well as of the previously referred products 3 and 4,
with H2 using 10% Pd/C as a catalyst produced the
saturated analogues 2-octaisopentyl-1,4-hydroquinone
(5), 2-octaisopentyl-1,4-diacetoxy-benzene (6) and 2-[24-
1
(14), 753 (1); H NMR (CDCl3): d 6.89–7.02 (3H, m),
3.52 (2H, d, J=4.75), 2.44 (2H, m), 2.29 (3H, s), 2.26
(3H, s), 1.05–1.43 (52H, m), 0.86 (3H, d, J=6.5 Hz),
0.84 (18H, br), 0.81 (3H, d, J=6.5 Hz); 13C NMR
(CDCl3): d 169.4 (s), 169.3 (s), 148.2 (s), 146.2 (s),
136.2 (s), 123.0 (d, Â2), 122.8 (d, Â4), 119.7 (d, Â2),
65.7 (t), 39.3 (t), 37.4 (t, Â4), 37.3 (t), 37.2 (t), 37.1 (t),
32.8 (t, Â4), 32.7 (t, Â3), 28.0 (t), 27.8 (t), 24.8 (t, Â2),
24.5 (t, Â2), 24.4 (t), 24.3 (t), 22.7 (q), 22.6 (q), 21.1 (q),
20.9 (q), 19.7 (q, Â4), 19.6 (q), 19.5 (q).
hydroxy]-octaisopentyl-1,4-diacetoxy-benzene
(7),
2-[24-oxy]-octaisopentyl-1,4-diacetoxy-benzene (8).
A
respectively. Oxidation of 1, 2 and 7 with CrO3 in 70%
acetic acid led to the formation of quinones 2-octapre-
nyl-1,4-quinone (9), 2-[24-hydroxy]-octaprenyl-1,4-qui-
quantity of 7 (360.2 mg, 0.47 mmol), dissolved in ace-
tone (3 mL) was added to a solution of CrO3 (100 mg,
1.0 mmol) in 70% HOAc (10 mL). The resulting mix-

L.-A. Tziveleka et al. / Bioorg. Med. Chem. 10 (2002) 935–939
939
ture was stirred at 50 ꢀC overnight, and after completion
of the reaction CH2Cl2 and H2O were added, the
organic layer was separated, washed with water, dried
over anhydrous Na2SO4, and concentrated in vacuo.
The residue (358.9 mg) was chromatographed on silica
gel (cyclohexane/EtOAc) to afford 67.8 mg of 8 (21%
yield) as a colorless oil; (8): IR (in CHCl3): 3660, 2928,
1759, 1491, 1463, 1371, 1232, 1199, 1171 cmÀ1; UV
(C6H14): lmax (e)=271 (535) nm; FABHRMS: m/z
769.6673 [M+H]+ (calcd for C50H88O5 768.66354); m/z
(% rel. int.) 123 (100), 165 (86), 207 (11), 263 (5), 333
Aliquots (0.3 mL) from the incubation mixture were
taken at various time intervals. Lipid peroxidation was
assessed spectrophotometrically (535 against 600 nm)
by the determination of the 2-thiobarbituric acid (TBA)
reactive material.18 Under the above experimental con-
ditions, all compounds, as well as DMSO, were tested
and found not to interfere with the assay. TBA was
purchased from Sigma Chemical Co. (St. Louis, MO,
USA).
1
(4), 403 (2), 473 (1) 557 (1), 755 (5); H NMR (CDCl3):
References and Notes
d 9.53 (1H, d, J=3.4), 6.87–7.02 (3H, m), 2.46 (2H, m),
2.29 (3H, s), 2.26 (3H, s), 1.05–1.46 (52H, m), 0.86 (3H,
d, J=6.5 Hz), 0.83, (18H, br), 0.81 (3H, d, J=6.5 Hz);
13C NMR (CDCl3): d 189.2 (d), 169.4 (s), 169.3 (s),
148.2 (s), 146.2 (s), 136.2 (s), 123.0 (d, Â2), 122.8 (d,
Â3), 119.7 (d, Â2), 39.4 (t), 37.5 (t, Â2), 37.4 (t, Â2),
37.3 (t), 37.2 (t), 37.1 (t), 32.8 (t, Â5), 32.6 (t, Â2), 28.0
(t), 27.8 (t), 24.8 (t), 24.7 (t), 24.6 (t), 24.5 (t, Â3), 22.7
(q), 22.6 (q), 21.1 (q), 20.9 (q), 19.7 (q, Â3), 19.6 (q, Â2),
19.5 (q).
1. De Pasquale, R.; Circosta, C.; Occhiuto, F.; De Rosa, S.;
De Stefano, S. Phytother. Res. 1991, 5, 49.
2. De Rosa, S.; Crispino, A.; De Guilio, A.; Iodice, C. J. Nat.
Prod. 1994, 57, 1711.
3. Gil, B.; Sanz, M. J.; Terencio, M. C.; De Guilio, A.; De Rosa,
S.; Alcaraz, M. J.; Paya, M. Eur. J. Pharmacol. 1995, 285, 281.
4. Mihopoulos, N.; Vagias, C.; Chinou, I.; Roussakis, C.;
Scoullos, M.; Harvala, C.; Roussis, V. Z. Naturforsch 1999,
54c, 417.
5. Cimino, G.; De Stefano, S.; Minale, L. Tetrahedron 1972,
28, 1315.
6. Minale, L. In Marine Natural Products; Scheuer, P. J., Ed.;
Academic: New York, 1978; Vol. 1, p. 175.
7. De Rosa, S.; Crispino, A.; De Guilio, A.; Iodice, C.;
Milone, A. J. Nat. Prod. 1995, 58, 1450.
8. Erdogan, I.; Tanaka, J.; Higa, T.; Sener, B. Nat. Prod. Sci.
1999, 5, 177.
Activity
Interaction of the test compounds with the stable radical
DPPH. The test compounds were dissolved in absolute
ethanol [analytical grade, iron content was less than
10À5% (w/v)], added to an equal volume of an ethanolic
solution of DPPH (final concentration 0.2 mmol/L) at
various concentrations (0.1, 0.2 and 0.4 mmol/L) and
kept at room temperature (22Æ2 ꢀC). Absorbance (517
nm) was recorded at 45 min.15 DPPH was purchased
from Aldrich-Chemie (Steinheim, Germany).
9. Terencio, M. C.; Ferra
De Rosa, S.; De Guilo, A.; Paya
Pharmacol. 1998, 357, 565.
´
ndiz, M. L.; Posadas, I.; Roig, E.;
´
, M.; Alcaraz, M. J. Arch.
10. Loya, S.; Rudi, A.; Kashman, Y.; Hizi, A. Biochem. J.
1997, 324, 721.
11. Ernster, L., In Oxidative Processes and Antioxidants; Pao-
letti, R.; Samuelsson, B.; Catapano, A.L.; Poli, A.; Rinetti,
M., Eds.; Raven: New York, 1994; p. 185.
In vitro lipid peroxidation. Hepatic microsomal fraction
from untreated female Fischer-344 rats (180–220 g) was
prepared as described earlier.16 The incubation mixture
contained heat inactivated (90 ꢀC for 90 s) hepatic
microsomal fraction, corresponding to 2.5 mg protein/
mL (final concentration) or 4 mmol/L fatty acid resi-
dues,17 ascorbic acid (0.2 mmol/L) in Tris–HCl/KCl
buffer (50 mmol/L/150 mmol/L, pH 7.4) and various
concentrations (1–0.05 mmol/L) of the test compounds
dissolved in dimethylsulphoxide (DMSO) or DMSO
alone (Control). The reaction was started by the addi-
tion of a freshly prepared FeSO4 solution (10 mmol/L),
and the mixture was incubated at 37 ꢀC for 45 min.
12. Hundal, T.; Forsmark-Andree, P.; Ernster, L.; Andersson,
´
B. Arch. Biochem. Biophys. 1995, 324, 117.
13. Pietta, P.-G. J. Nat. Prod. 2000, 63, 1035.
14. Foti, M.; Piattelli, M.; Amico, V.; Ruberto, G. J. Photo-
chem. Photobiol 1994, 26, 159.
15. Kourounakis, A.; Rekka, E.; Kourounakis, P. J. Pharm.
Pharmacol. 1997, 49, 938.
16. Rekka, E.; Kolstee, J.; Timmerman, H.; Bast, A. Eur. J.
Med. Chem. 1989, 24, 43.
17. Eichenberger, K.; Bohni, P.; Winterhalter, K. H.; Kawato,
S.; Richter, C. FEBS Lett. 1982, 142, 59.
18. Kourounakis, A.; Galanakis, D.; Rekka, E.; Kour-
ounakis, P. Drug Dev. Res. 1999, 46, 9.