Bioconjugates of curcumin display improved protection against glutathione depletion mediated oxidative stress in a dopaminergic neuronal cell line: Implications for Parkinson's disease

Article abstract of DOI:10.1016/j.bmc.2010.02.029

Oxidative stress is implicated in mitochondrial dysfunction associated with neurodegeneration in Parkinson's disease (PD). Depletion of the cellular antioxidant glutathione (GSH) resulting in oxidative stress is considered as an early event in neurodegeneration. We previously showed that curcumin, a dietary polyphenol from turmeric induced GSH synthesis in experimental models and protected against oxidative stress. Here we tested the effect of three bioconjugates of curcumin (involving diesters of demethylenated piperic acid, valine and glutamic acid) against GSH depletion mediated oxidative stress in dopaminergic neuronal cells and found that the glutamic acid derivative displayed improved neuroprotection compared to curcumin.

Full text of DOI:10.1016/j.bmc.2010.02.029

Bioorganic & Medicinal Chemistry 18 (2010) 2631–2638
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Bioconjugates of curcumin display improved protection against glutathione
depletion mediated oxidative stress in a dopaminergic neuronal cell line:
Implications for Parkinson’s disease
G. Harish a, C. Venkateshappa a, Rajeswara Babu Mythri a, Shiv Kumar Dubey b, , Krishna Mishra c,
Neetu Singh d, Shireen Vali d, M. M. Srinivas Bharath a,
*
a Department of Neurochemistry, National Institute of Mental Health and Neurosciences, #2900, Hosur Road, Bangalore 560 029, Karnataka, India
b Center for Biotechnology, University of Allahabad, Allahabad, India
c Indo-Russian Center for Biotechnology, Indian Institute of Information Technology-Allahabad, Jhalwa Campus, Allahabad 211 002, India
d Cellworks Group Inc., 3rd Floor, West-wing, ‘Neil-Rao Tower’, #118, Road #3, EPIP, White Field, Bangalore 560 066, India
a r t i c l e i n f o
a b s t r a c t
Article history:
Oxidative stress is implicated in mitochondrial dysfunction associated with neurodegeneration in Parkin-
son’s disease (PD). Depletion of the cellular antioxidant glutathione (GSH) resulting in oxidative stress is
considered as an early event in neurodegeneration. We previously showed that curcumin, a dietary pol-
yphenol from turmeric induced GSH synthesis in experimental models and protected against oxidative
stress. Here we tested the effect of three bioconjugates of curcumin (involving diesters of demethylenat-
ed piperic acid, valine and glutamic acid) against GSH depletion mediated oxidative stress in dopaminer-
gic neuronal cells and found that the glutamic acid derivative displayed improved neuroprotection
compared to curcumin.
Received 8 January 2010
Revised 16 February 2010
Accepted 17 February 2010
Available online 20 February 2010
Keywords:
Neurodegeneration
Oxidative stress
Glutathione
Ó 2010 Elsevier Ltd. All rights reserved.
Curcumin
Bioconjugates
Dopaminergic cell line
1. Introduction
gest that the early GSH loss in the SN of PD patients could be linked
to subsequent mitochondrial dysfunction ultimately leading to
Parkinson’s disease (PD) is a neurodegenerative disorder char-
acterized by loss of dopaminergic neurons of the substantia nigra
(SN) associated with dopamine loss in the basal ganglia. Among
the early events causing neurodegeneration are oxidative stress
and mitochondrial dysfunction.1–4 Oxidative stress during early
PD is elicited by a significant depletion of the cellular antioxidant
glutathione (GSH+GSSG) in the SN.5 GSH depletion precedes both
mitochondrial dysfunction and dopamine loss and is therefore con-
sidered as the earliest triggering factor of neurodegeneration.6,7
GSH depletion in dopaminergic cells in culture increased oxidative
stress and decreased mitochondrial function.8 These results sug-
neurodegeneration.1,8 We recently demonstrated the dynamic
relationship between GSH metabolism and mitochondrial dysfunc-
tion associated with PD.9 It could be surmised that therapeutic res-
toration of intracellular GSH levels might prevent triggering of the
early events of pre-symptomatic PD. Further, such therapeutic
molecules are efficacious only if they are non-toxic with the ability
to cross the blood brain barrier (BBB).
Turmeric (Curcuma longa) is a dietary spice used in Indian cui-
sine and traditional medicine. Curcumin (diferuloylmethane), the
most active and non-toxic component of turmeric is a polyphe-
nol,10,11 with antioxidant, anti-cancer properties and therapeutic
potential in neurological disorders and ability to cross the blood
brain barrier (BBB).12–15 Curcumin is neuroprotective against 1-
methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mediated
neurotoxicity in vivo.16,17 We earlier showed that curcumin pro-
tects mitochondria against nitrosative stress and oxidative stress
both in vitro and in vivo.18,19 Through systems biology, we have
also explained the antioxidant mechanism of curcumin action
and its effect on GSH metabolism.19
Abbreviations: PD, Parkinson’s disease; SN, substantia nigra; GSH, glutathione
(reduced); GSSG, glutathione (oxidized); DTNB, 5,5-dithio-bis-2-nitro benzoic acid;
ROS, reactive oxygen species; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyri-
dine; c-GCL, gamma-glutamyl cysteine ligase; BSO, buthionine sulfoximine; H2O2,
hydrogen peroxide; glu, glutamate; cys, cysteine; D1, di-demethylenated piperoyl
curcumin; D2, di-valinoyl curcumin; D3, di-glutamoyl curcumin.
* Corresponding author. Tel.: +91 080 26995167; fax: +91 080 26564830.
E-mail addresses: bharath@nimhans.kar.nic.in, thathachar@rediffmail.com
(M.M. Srinivas Bharath).
However, the bioavailability of curcumin is very limited due to
poor absorption, rapid metabolism and quick systemic elimination.
Present address: Post doctoral fellow, Division of Hematology
University of Michigan Cancer Center, Ann Arbor, MI 48109, USA.
& Oncology,
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.02.029

2632
G. Harish et al. / Bioorg. Med. Chem. 18 (2010) 2631–2638
To improve its bioavailability, we have reported the synthesis of
three curcumin derivatives involving diesters at the phenolic
groups: curcumin D1 (di-piperoyl), D2 (di-valinoyl), and D3 (di-
glutamoyl). These bioconjugates have improved bioavailability than
curcumin, due to improved solubility, enhanced adsorption and
slow metabolism.20–22 Unpublished data from our laboratory indi-
cated that D3 exhibited increased detoxification of PN in vitro thus
protecting brain mitochondria and preventing protein nitration
and against 1-methyl-4-phenylpyridinium (MPP+) neurotoxicity
in cell culture. In the current study, we have tested the neuroprotec-
tive ability of these derivatives against GSH depletion mediated oxi-
dative stress in dopaminergic neurons in culture. Our overall aim is
to obtain superior curcumin derivatives possessing both improved
antioxidant capacity and increased bioavailability.
the corresponding derivatives were denoted as D1, D2, and D3,
respectively. The ester bonds in these compounds are biodegrad-
able, that is, hydrolysable with common cellular esterases making
these derivatives potent pro-drugs in the target cells. The following
is the design, synthesis and analysis of these curcumin derivatives
as reported earlier (Fig. 1):20
2.1.1. 1,7-Bis-[4-O-demethylenated piperoyl-3-
methoxyphenyl]-1,6 heptadiene-3,5-dione (D1)
Curcumin (0.736 g, 2 mmol) was taken in dry pyridine an-
d mixed with demethylenated piperoyl chloride (1.008 g,
4.5 mmol) and stirred at room temperature (rt) for 6 h. After the
completion of reaction as indicated by thin layer chromatography
(TLC), the reaction mixture was poured onto crushed ice and
repeatedly extracted with EtOAc. The organic layer was concen-
trated and purified on a silica gel column using a DCM/methanol
(MeOH) gradient. Yield: 45% (0.667 g); Rf: 0.74 (DCM/MeOH
9.5:0.5); UV kmax (MeOH): 395 and 295 nm. Anal. Calcd for
C43H36O12: C, 69.29; H, 4.87; O, 25.78. Found: C, 69.34; H, 4.83;
O, 25.80. 1H NMR (CDCl3) d ppm = 3.65 (s, 6H, –OCH3), 4.15 (s,
2H, C4–H), 4.65 (d, 2H, O@C–CH2), 5.03 (s, 4H, aromatic C–OH),
5.65 (d, 2H, –CH@CH–C@O), 6.38 (d, 2H, –CH@C), 6.72–7.80 (m,
8H, olefinic and aromatic), 6.90 (d, 1H, Ar–H), 7.18–7.48 (d, 2H,
CH@CH@C@O), 7.48 (d, 2H, C1–H and C7–H).
2. Materials and methods
All the chemicals and solvents were of analytical grade. Routine
chemicals and bulk solvents were obtained from Sisco research
Laboratories (Mumbai, Maharashtra, India) and Merck (White-
house Station, NJ, USA). Fine chemicals for biochemical assays were
obtained from Sigma (St. Louis, MO, USA). Curcumin, valine, glu-
tamic acid and piperine for synthesis were purchased from Merck
(Schuchardt, Germany).
2.1. Preparation of curcumin derivatives
2.1.2. 1,7-Bis-(4-O-valinoyl-3-methoxy phenyl) 1,6-heptadiene-
3,5-dione (D2)
The bioconjugates of curcumin were synthesized by covalent
linking of different moieties through the two phenolic groups of
curcumin resulting in diesters. The moieties selected were dem-
ethylenated piperic acid, valine (val) and glutamic acid (glu) and
Curcumin (0.368 g, 1 mmol) was taken in dry pyridine and
mixed with N-phthaloyl-valinoyl chloride (0.53 g, 2 mmol) and
stirred at rt for 8 h. After completion of the reaction as indicated
by TLC, the reaction mixture was poured into crushed ice and
A
O
O
O
O
B
OCH3
OH
OCH3
OH
H3CO
HO
H3CO
HO
(Curcumin)
(Curcumin)
Pyridine/ demethylenated piperoyl
chloride (Ia) / 6h / r.t.
(i) Pyridine/ N-Phthaloylvalinoyl
Chloride (IIa) / N-Phthaloylglutamoyl
Chloride (IIb) / 6h / r.t.
(ii) NH3 : Pyridine (9:1 v/v)
O
O
O
O
OCH3
OR
OCH3
OR1
H3CO
RO
H3CO
R2O
D1
D2/D3
O
O
D1: R= A
(Ia)= A-Cl
A=
HO
HO
D2: R1=R2= -CO-CH-NH2
Cl-C-CH-HN
CH
H3C CH3
(IIa) =
CH=CH-CH=CH-C-
C
O
O
H3C CH3
(Demethylenated piperic acid)
(Valine)
O
O
D3: R1=R2= -CO-CH-NH2
(IIb) =
Cl-C-CH-HN
CH2
CH2
O
HOOC
HOOC
(Glutamic acid)
Figure 1. Chemical structure and scheme of synthesis of the curcumin derivatives used in this study. Scheme of synthesis of D1 (di-demethylenated piperoyl curcumin) (A),
D2 (di-valinoyl curcumin), and D3 (di-glutamoyl curcumin) (B) (see Section 2).

G. Harish et al. / Bioorg. Med. Chem. 18 (2010) 2631–2638
2633
repeatedly extracted with ethyl acetate. The organic layer was con-
centrated and treated with ammonia/pyridine (9:1 v/v) to remove
phthaloyl group. The organic layer was concentrated and purified
by silica gel column chromatography using DCM/methanol gradi-
ent. Yield: 45% (0.254 g); Rf: 0.9 (DCM/MeOH 9:1); UV kmax
(MeOH): 295 nm. Anal. Calcd for C31H38O8N2: C, 65.71; H, 6.76;
O, 22.59; N, 4.94. Found: C, 65.74; H, 6.71; O, 22.60; N, 4.91. 1H
NMR (CDCl3) d ppm = 1.03 (d, 12H, CH–(CH3)2), 2.65 (m, 2H, 30-C
of valine), 3.70 (s, 6H, –OCH3), 4.09 (s, 2H, C4–H), 4.19 (m, 2H,
CH–NH2), 6.53 (d, 2H, C2–H and C6–H), 6.82–7.13 (m, 6H, Ar–H),
7.56 (d, 2H, C1–H and C7–H).
suspension were sonicated (5 s ꢀ 4 on ice; pulsed setting) and cen-
trifuged at 13,500g (10 min) at 4 °C. Thirty microliters of the super-
natant (corresponding to ꢁ0.2 mg protein) was added to 0.95 ml of
FOX1 reagent and incubated at room temperature for 30 min. The
reaction mixture was centrifuged at 800g (10 min) and the absor-
bance of the supernatant was measured at 560 nm. The amount of
H2O2 was calculated using the molar extinction coefficient of H2O2
(1.5 ꢀ 104 molꢂ1cmꢂ1) and the amount of H2O2 was expressed as
nM/mg of protein.
2.5. Estimation of lipid peroxidation
2.1.3. 1,7-Bis-(4-O-glutamoyl-3-methoxy phenyl) 1,6-
heptadiene-3,5-dione (D3)
Lipid peroxidation was measured by estimation of malondialde-
hyde (MDA) by thiobarbituric acid reaction method modified from
Ohkawa et al.26 Briefly, N27 cells were sonicated and centrifuged at
16,000g (10 min) at 4 °C. One hundred microliters of the superna-
tant was added to a mixture containing 0.75 ml of acetic acid (pH
3.5, 20% v/v), 0.1 ml SDS (8% w/v) and 0.75 ml of thiobarbituric acid
(0.8%, w/v). The reaction mixture was heated in boiling water bath
for 45 min. The adducts formed were extracted into 1.5 ml of 1-
butanol by vortexing and centrifugation at 2500 rpm (10 min)
and their absorbance was measured at 532 nm. The amount of
MDA formed was calculated using the molar extinction coefficient
(241 molꢂ1cmꢂ1).
Curcumin (0.368 g, 1 mmol) was taken in dry pyridine and
mixed with N-phthaloyl-glutamoyl chloride (0.55 g, 2 mmol) and
stirred at rt for 8 h. After completion of the reaction as indicated
by TLC, the reaction mixture was poured into crushed ice and
repeatedly extracted with ethyl acetate. The organic layer was con-
centrated and treated with ammonia/pyridine (9:1 v/v) to remove
phthaloyl group. The organic layer was concentrated and purified
by silica gel column chromatography using DCM/methanol gradi-
ent. Yield: 40% (0.250 g); Rf: 0.9 (DCM/MeOH 9:1); UV kmax
(MeOH): 290 nm. Anal. Calcd for C31H34O12N2: C, 59.42; H, 5.47;
O, 30.64; N, 4.47. Found: C, 59.39; H, 5.40; O, 30.67; N, 4.40. 1H
NMR (CDCl3) d ppm = 2.06–2.18 (m, 8H, 30,40-C of glutamate),
3.45 (m, 2H, CH–NH), 3.70 (s, 6H, –OCH3), 4.46 (s, 2H, C4–H),
4.09 (m, 2H, CH–NH2), 6.53 (d, 2H, C2–H and C6–H), 6.83–7.11
(m, 6H, Ar–H), 7.54 (d, 2H, C1–H and C7–H), 11.2 (s, 2H, HO–C–O).
2.6. Measurement of reactive oxygen species (ROS)
Total ROS in N27 cells were measured by a modified method of
Ohashi et al.27 Briefly, the medium from treated or untreated N27
cells was replaced with 1 ml Locke’s solution (154 mM NaCl,
5.6 mM KCl, 3.6 mM NaHCO3, 5 mM HEPES, 2 mM CaCl2 and
10 mM Glucose, pH 7.4). Ten micromoles of dihydrodichlorofluo-
rescein diacetate (DCFDA) was then added and the cells were incu-
bated at 37 °C (10 min) in a CO2 incubator. The Locke’s solution
was then removed and the cells were harvested. The cell pellet
was washed with 1X phosphate buffered saline (PBS), pH 7.4, twice
and reconstituted in lysis buffer (10 mM Tris–HCl containing 0.5%
Tween-20). The lysate was centrifuged at 1000g (10 min) and the
fluorescence of the supernatant was measured (Excitation—
480 nm; Emission—530 nm).
2.2. Cell culture
1RB3AN27 (N27) dopaminergic neuronal cell line has been used
throughout this study. N27 cell line was generated by SV40 LT-
antigen immortalization of rat mesencephalic cultures.23 N27 cells
possess all the physiological and biochemical properties of dopa-
minergic neurons and emulate those cells that are lost during
PD.24 N27 cells were grown as described earlier.9,18
2.3. Total glutathione (GSH + GSSG) estimation
Following different treatments, soluble N27 cell extracts were
subjected to total glutathione estimations by the 5,50-dithio-bis-
2-nitro benzoic acid (DTNB) recycling method as described ear-
lier.18 The cell pellet was resuspended in PE buffer (100 mM potas-
sium phosphate buffer, pH 7.4 containing 1 mM EDTA) and
homogenized by sonication (5 s ꢀ 6 on ice) and total protein was
2.7. Glutathione-S-transferase (GST) assay
GST was assayed by measuring the rate of enzyme catalyzed con-
jugation of GSH with 1-chloro 2-4-dinitro benzene (CDNB) accord-
ing to the method of Guthenberg et al.28 Briefly, to 1 ml reaction
mixture containing phosphate buffer (0.1 M, pH 6.5; 0.5 mM EDTA),
estimated. 80
ll of the homogenate was acid precipitated with
CDNB (1.5 mM) and 50 ll GSH (1 mM) and 50 ll sample was added
an equal volume of 2% sulfosalicylic acid (w/v). The mixture was
centrifuged at 15,000g (15 min) and the resultant supernatant
was used for the assay. 20 ll aliquot of the supernatant was incu-
and the increase in absorbance at 340 nm was monitored for 5 min.
The enzyme activity was expressed as nmoles of S-2,4, dinitrophenyl
glutathione formed /min/mg protein (M EC-9.6 mMꢂ1 cmꢂ1).
bated with assay buffer [PE buffer containing 0.8 mM DTNB and
0.32 U/ml glutathione reductase] in a final reaction volume of
2.8. Statistical analyses
450 ll. The reaction was initiated by addition of 0.6 mM NADPH.
The reaction kinetics of DTNB recycling which was dependent on
total glutathione levels was monitored at 412 nm for 3 min. The
absolute GSH level in each sample was calculated based on the
maximum reaction rate compared with GSSG standards (0–
250 ng). All estimations were conducted in triplicate and GSH con-
centrations were normalized per mg protein.
All quantitative data were accumulated from at least three
independent experiments. The final data are expressed as mean
SEM. Differences between mean values were analyzed by one-
way analysis of variance (ANOVA) and p value <0.05 was consid-
ered to be significant.
2.9. Predictive data from virtual experimental system
2.4. Estimation of H2O2
2.9.1. Model description
Water soluble hydroperoxide levels in N27 cells were deter-
mined following ferrous iron oxidation with xylenol orange
We recently reported predictive studies linking GSH metabolism
and mitochondrial dysfunction associated with PD.9 We also de-
scribed the effect of curcumin administration on GSH
(FOX1) as previously described.25 Briefly, N27 cells in 100–150
ll

2634
G. Harish et al. / Bioorg. Med. Chem. 18 (2010) 2631–2638
metabolism.19 The current experiment includes the comparison of
curcumin and its derivatives in terms of their effect on GSH metab-
olism and oxidative stress in a module of the dopaminergic neuronal
platform. This module evaluates the induction of gamma-glutamyl
to decelerate neurodegeneration requires molecules that restore
neuronal GSH levels and prevent oxidative damage. We showed
that curcumin protects against oxidative and nitrosative stress
in vivo either by direct detoxification of toxic species or via induc-
tion of GSH synthesis.18,19 However, the bioavailability of curcumin
poses a major obstacle for its therapeutic efficacy. The synthesis
and improved cellular uptake of curcumin bioconjugates have been
recently reported.20–22 In the current study, we have tested the
neuroprotective ability of three derivatives of curcumin (diester
with demethylenated piperic acid or D1; diester with val or D2;
diester with glu or D3) compared to curcumin against GSH deple-
tion mediated oxidative stress in dopaminergic neurons (see Fig. 1
for structures of the compounds).
cysteineligase(c-GCL)andgenerationofGSH, ATP, ROS, andreactive
nitrogen species (RNS). The nodes included in each of the individual
modules are qualitatively described here with the equations and ki-
netic parameters for each node described previously.9,19
Oxidative stress due to GSH depletion: The homeostasis of ROS
and RNS in the neuron under normal physiological conditions is
maintained mainly by GSH and associated enzymes.9 The decline
in the protective antioxidant system with lowered GSH levels im-
pacts by oxidative damage to the mitochondria.29
We utilized N27 cell line as a model for GSH depletion associ-
ated with PD. N27 cell line, generated by immortalization of dopa-
minergic neurons from rat fetal mesencephalon are comparable to
the neurons selectively lost in PD.23,24 Hence N27 cell line repre-
sents an excellent cell model to understand GSH depletion medi-
ated oxidative stress associated with PD. N27 neuronal cells
2.9.1.2. GSH synthesis and metabolism
This module incorporating GSH synthesis, its cellular distribu-
tion and antioxidant function have been described earlier.9,30 Cur-
cumin (and its derivatives) administered into the system induces
GSH synthesis via up regulation of c
-GCL.31
treated with BSO (0–25
crease in the levels of total cellular GSH (Fig. 2A) with 50% deple-
tion obtained at 1.5 M BSO and maximum decrease at 25 M.
For subsequent experiments, cells were exposed to 1.5 M BSO
lM; 24 h) showed a dose dependent de-
2.9.2. Model construction
l
l
The model is a dynamic representation of GSH metabolism
modulated by curcumin. The model construction and its assump-
tions have been described earlier.9 The reaction kinetics for all
nodes in the dynamic integrated model is summarized in the sup-
plementary data of previously reports.9,19
l
for 24 h to maintain ꢁ50% depletion of cellular GSH, similar to
the decreases observed in the PD SN.7
We confirmed the neuroprotective ability of curcumin and its
derivatives against GSH depletion using the BSO-N27 model.19
Treatment of N27 cells with curcumin or the derivatives alone
2.9.3. Simulation protocol
(0.5 lM, 24 h) significantly increased total GSH as shown ear-
lier18,19 with D3 showing maximum increase (ꢁ64% compared
with untreated control) compared to curcumin (30%), D1 (ꢁ25%),
and D2 (ꢁ25%) (Fig. 2B). However, higher concentrations of curcu-
min and the derivatives did not cause further increase in GSH lev-
Initial conditions: The initial conditions of the neuronal platform
represent normal mitochondrial energetics with the free radicals
generated being scavenged by GSH antioxidant system. In the neu-
ron, the ATP utilization is ꢁ80
l
M/s32 and the PMF is ꢁ90–200 mV.
els (data not shown) indicating that under these conditions, 0.5 lM
curcumin or its derivatives is sufficient to induce maximum GSH
levels in these neurons.
The normal cytosolic and mitochondrial concentrations of GSH are
1–3 mM33 and 10–50
lM, respectively. The steady-state concentra-
tions of H2O2 and ROS in the mitochondria are ꢁ10 nM and 0.1 nM,
respectively.34 The normal physiological concentrations of NO and
125
A
calcium in matrix are ꢁ0.02–0.1
lM and 0.2 l
M, respectively.35,36
100
Administration of curcumin and its derivatives: In this experi-
ment, the dopaminergic neuronal platform was exposed to the
conditions indicated below followed by estimation of cellular
GSH and ROS:
***
75
50
*** ***
***
***
***
25
0
***
(a) Curcumin alone (0.5 lM).
(b) Improved bioavailability of curcumin representing the
***
10 25
***
derivatives D1 and D2 and indicated by doubled curcumin
0
1
1.5
2
2.5
3
4
5
levels (1
lM).
BSO (µM)
(c) Improved bioavailability of curcumin (1
lM) in addition to
mild increase in glutamate levels (from 1
resenting D3.
l
M to 3 M) rep-
l
200
150
100
B
***
**
*
*
**
Curcumin was modeled as having the following dual characteris-
tics as described in Jagatha et al.:19
(a) GSH up regulation: Curcumin administration (1
lM) to a nor-
50
0
mal cell increases -GCL levels from 0.26 M to 0.33 lM
c
l
resulting in a ꢁ1.5-fold increase of GSH level.
C
Cur
D1
D2
D3
(b) Direct scavenging of ROS and RNS: Curcumin shows SOD
activity and PN scavenging activity.37 The cumulative effects
of both reactions (i) and (ii) have been modeled.
0.5µM
Figure 2. Effect of BSO and curcumin on total GSH levels in N27 dopaminergic
neurons. (A) Estimation of total GSH in N27 cells incubated with 0–25 M BSO for
24 h (n = 5). Values shown as percentage of GSH compared with untreated control
l
3. Results
(100% GSH = 23.3 3.0 nmol/mg protein). (B) Estimation of total GSH in N27 cells
Oxidative stress implicated in PD pathology involves GSH
depletion as the earliest event.7 Hence, therapeutic intervention
*
**
p
incubated with 0.5
***
l
M of curcumin, D1, D2, and D3 for 24 h (n = 3). p <0.05,
<0.01 and
p <0.001 compared to untreated control (in A and B).

G. Harish et al. / Bioorg. Med. Chem. 18 (2010) 2631–2638
2635
When N27 cells were treated with BSO (1.5
l
M for 24 h) fol-
ments) restored GST activity in BSO treated cells comparable to
untreated controls. However, a direct correlation between the
magnitude of restoration of GST activity and the corresponding
MDA values in respective treatment groups was not observed.
Estimation of H2O2 indicated ꢁ40% increase in BSO treated N27
cells which decreased by ꢁ25% following BSO withdrawal (Fig. 4C).
The H2O2 levels were further reduced (ꢁ50% reduction compared
to BSO alone) in cells either pre or post-treated with curcumin or
its derivatives. Except for D2 (post-treatment) which showed
ꢁ80% decrease in H2O2, the other derivatives showed similar
reduction compared with curcumin. Nevertheless, these levels
were much lower than the H2O2 levels observed in untreated cell
extracts.
Similarly, ROS estimation indicated that compared to untreated
controls, there was ꢁ65% increase in BSO treated N27 cells which
was decreased by ꢁ30% following BSO withdrawal (Fig. 4D). The
ROS levels were further decreased in cells pre-treated with curcu-
min or its derivatives (50% reduction in ROS with curcumin, 55%
with D1, 70% with D2, and 80% with D3 compared with BSO alone).
Post treatment was also able to abrogate BSO mediated ROS gener-
ation (45% reduction in ROS with curcumin, 65% with D1, 60% with
D2 and 90% with D3 compared with BSO alone). As seen in the
MDA experiment, exposure to D3 caused maximum scavenging
of ROS compared with curcumin and other derivatives. However,
cells treated with either curcumin or the derivatives alone did
not show significant increase in ROS levels (data not shown).
Based on the data, we conclude that among the derivatives, D3
displayed significantly enhanced antioxidant potential against oxi-
dative stress in dopaminergic neurons compared with curcumin.
To explain the mechanism underlying this, we carried out a sys-
tems biology based predictive study of curcumin and its deriva-
lowed by BSO withdrawal for 24 h, it completely restored the cel-
lular GSH levels indicating that BSO mediated GSH depletion was
specific and reversible (Fig. 3). This regimen involving 24 h BSO
treatment followed by 24 h withdrawal was applied in all the sub-
sequent experiments. Pre treatment of BSO treated N27 cells with
curcumin and its derivatives (0.5
lM, 24 h) significantly restored
the GSH levels compared to BSO alone (1.5
l
M, 24 h) (curcu-
min = 78%, D1 = 78%, D2 = 72% and D3 = 83% compared to
BSO = ꢁ50% and control = 100%) (Fig. 3). However, post treatment
(0.5 lM, 24 h) caused tremendous increase in GSH compared with
the BSO alone and much higher compared to untreated control
(curcumin = 127%, D1 = 124%, D2 = 122% and D3 = 136% compared
to control = 100% and BSO alone = 50%) (Fig. 3). We observed that
D3 showed higher GSH levels both in the pre and post treatment
experiments compared to curcumin.
It is known that GSH depletion causes increased ROS produc-
tion.7 Since curcumin and its derivatives can restore GSH levels
in the BSO-N27 model, we tested whether these derivatives could
detoxify ROS more efficiently compared to curcumin. Accordingly,
we measured the cellular levels of MDA (marker of lipid peroxida-
tion), H2O2 and total ROS in N27 cells treated either with BSO
(0.5 lM) alone or in combination with curcumin or its derivatives
(both pre and post-treatment) compared to untreated N27 cells.
Figure 4A shows that BSO treatment resulted in ꢁ25% increase in
total MDA levels compared to untreated extracts, which was signif-
icantly decreased following BSO withdrawal for 24 h (104%). While
curcumin pre treatment resulted in ꢁ40% reduction in lipid perox-
idation compared to BSO alone, pretreatment with D1 and D3 re-
sulted in ꢁ50% reduction and D2 caused ꢁ30% reduction in lipid
peroxidation. On the other hand, post treatment with these com-
pounds caused higher reduction in MDA levels as shown in Figure
4A (65% reduction with curcumin, 75% with D1, 80% with D2, 90%
with D3 compared with BSO alone). These levels were much lower
than the MDA levels observed in untreated cell extracts. Such dras-
tic detoxification of lipid peroxidation might be possible due to sig-
nificantly higher GSH levels during pretreatment (as shown
previously in Fig. 3). We observed that the exposure to D3 resulted
in maximum reduction in lipid peroxidation compared with curcu-
min and other derivatives.
tives in
a dopaminergic neuronal virtual platform described
earlier.9,19 Figure 5 shows the dynamics of GSH metabolism in
the brain which represents a steady-state GSH concentration of
2.5 mM. Curcumin (0.5 lM) administration into the system in-
creased the expression of the enzyme
c-GCL by Nrf2-dependent
pathway as described earlier,19 resulting in elevation of GSH levels
up to ꢁ3 mM (Fig. 5A). Increase in curcumin concentration up to
1 lM mimicking improved bioavailability in derivatives did not
elevate the GSH levels further thus corroborating our cell culture
We surmised that the decrease in MDA levels in cells treated
with curcumin and derivatives might be due to the direct involve-
ment of cellular GST activity. Accordingly, GST assay in BSO treated
N27 extracts showed ꢁ20% reduction in activity compared with
untreated controls (Fig. 4B). Further, withdrawal of BSO or incuba-
tion with curcumin and its derivatives (both pre and post treat-
data. However, when curcumin at 0.5 M along with increased glu-
l
tamate levels mimicking the glutamoyl diester of curcumin (D3)
was administered, the GSH levels almost doubled to 4.75 mM. This
trend is similar to the increase in GSH levels in D3 treated N27 cells
compared to curcumin. Consequently, D3 showed improved detox-
ification of ROS compared with curcumin and other derivatives
(Fig. 5B). Figure 6 is a schematic representation of the effect of cur-
cumin and its derivatives on GSH metabolism in the dynamic
model.
150
***
*** *** ***
***
***
4. Discussion
100
**
**
**
*
There have been significant advances in PD therapy including
surgical and pharmacological interventions. Currently, there are
several options for pharmacological treatment of PD with Levo-
$
50
0
dopa (L-dopa) being the most popular drug. However, the occur-
rence of severe drug-induced side effects in chronic
L-dopa
therapy has allowed utilization of alternative molecules including
dopamine receptor agonists, anti-cholinergic drugs, monoamine
oxidase-B inhibitors, catechol-O-methyl transferase inhibitors etc.
These therapeutic strategies are mainly symptomatic and strive
to replenish striatal dopamine. But their ability to prevent or slow-
down neurodegeneration of SN neurons has not been completely
validated in humans. Therefore, there is scope for exploration of
novel therapeutic molecules that prevent neurodegeneration and
BSO(1.5µM) -
Pre (0.5µM) -
Post(0.5µm) -
+
-
-
+
-
-
(WD)
+
+
+
+
+
-
+
-
+
-
+
-
Cur D1 D2 D3
-
-
-
-
Cur D1 D2 D3
Figure 3. Neuroprotective effects of curcumin on total GSH levels GSH-depleted
dopaminergic neurons. Estimation of total GSH in BSO (1.5 M) treated N27 cells
incubated with or without 0.5 M curcumin or its derivatives (D1, D2, and D3) (pre
or post-treatment) (n = 6). WD = BSO withdrawal for 24 h. $p <0.001 compared to
** ***
l
l
*
untreated control; p <0.05, p <0.01, and
p <0.001 compared to BSO alone.

2636
G. Harish et al. / Bioorg. Med. Chem. 18 (2010) 2631–2638
A
150
100
50
B
150
***
***
***
***
***
***
100
*
$
**
***
***
***
***
***
50
0
***
0
BSO(1.5µM) -
Pre (0.5µM) -
Post(0.5µm) -
+
-
-
+(WD) +
-
-
+
+
+
+
-
+
+
-
+
-
BSO(1.5µM) -
Pre (0.5µM) -
Post(0.5µm) -
+
-
+(WD)+
+
+
+
+
-
+
-
+
-
+
-
Cur D1 D2 D3
-
-
-
Cur D1 D2 D3
-
***
+
-
-
-
Cur D1 D2 D3
-
-
-
***
+
-
-
Cur D1 D2 D3
C
D
200
150
100
50
$
$
150
*
***
**
**
***
***
***
100
50
0
***
***
***
***
***
***
***
***
***
0
BSO(1.5µM) -
Pre (0.5µM) -
Post(0.5µm) -
+
-
-
+
-
-
(WD)+
+
+
+
-
+
-
+
-
+
-
BSO(1.5µM) -
+
-
-
+
(WD)
+
+
+
+
-
+
-
+
-
+
-
Cur D1 D2 D3
Pre (0.5µM)
-
-
Cur D1 D2 D3
-
-
-
-
Cur D1 D2 D3
Post(0.5µm) -
-
-
-
-
-
Cur D1 D2 D3
Figure 4. Neuroprotective effects of curcumin on lipid peroxidation, GST activity, hydrogen peroxide production, and total ROS in GSH-depleted dopaminergic neurons. Total
cell extracts from untreated and BSO (1.5 M) treated N27 cells incubated with 0.5 M curcumin or its derivatives (D1, D2, and D3) (pre or post-treatment) were used to
estimate total cellular concentrations of (A) MDA (n = 6); values shown as percentage of MDA compared with untreated control (100% MDA = 115 5 nmol/mg protein);
l
l
*
p
**
<0.05, p <0.01, and
***
p <0.001 compared to BSO alone. (B) GST activity (n = 3); values shown as percentage of GST activity compared with untreated control (100% GST
activity = 0.407 0.02 IU/mg protein/min); $p <0.05 compared to untreated control; ***p <0.001 compared to BSO only. (C) H2O2 (n = 3); values shown as percentage of H2O2
compared with untreated control (100% H2O2 = 13 2 nmol/mg protein); $p <0.001 compared with untreated control; *p <0.05, **p <0.01, and ***p <0.001 compared to BSO
alone; and (D) Total ROS content (n = 6); values indicated by percentage arbitrary units (au) of DCF fluorescence; (100% Fluorescence = 264.4 1.2 au/mg protein); $p <0.001
***
compared with untreated control;
p <0.001 compared with BSO alone. WD = BSO withdrawal for 24 h in all the figures.
5
A
0.18
B
0.16
0.14
0.12
4
3
2
0.1
0.08
0.06
0.04
0.02
1
0
0
Con
cur
(0.5µM)
cur
cur + glu
Con
cur
(0.5µM)
cur
(1µM)
cur + glu
(0.5µM) (3µM)
(1µM) (0.5µM) (3µM)
Figure 5. In silico predictive data from the virtual experimental system on the effect of curcumin and its derivatives on cellular GSH synthesis and scavenging of ROS.
Curcumin (either at 0.5 and 1
estimation of ROS.
lM) and glutamate (at 3 lM) were introduced into the dopaminergic neuronal platform followed by (A) estimation of GSH synthesis and (B)
could be used as adjunctive therapies along with dopamine
replacement.38
binds to the redox-active metals, iron and copper thereby exerting
neuroprotection against oxidative damage.40 Interestingly, curcu-
min induces GSH synthesis in cells by activation of GCL activity
in vivo thereby enhancing the antioxidant potential.31 Dickinson
et al.31 demonstrated that curcumin-induced elevation of GCL
activity is probably by enhanced transcription of GCL genes via
Curcumin exhibits neuroprotective effect against MPTP16,17 and
6-hydroxy dopamine (6-OHDA) in vivo.39 We have recently re-
ported that curcumin protects mitochondria against PN and GSH
depletion mediated toxicity in vitro and in vivo.18,19 Curcumin also

G. Harish et al. / Bioorg. Med. Chem. 18 (2010) 2631–2638
2637
GSH
pool
Glu
Glutamine
Mitochondria
GGT
ATP
Cys
Gluglu
Cys
Gly
GR
GCL
Glucys
GSSG
GSH
Gly-cys
GPx
GS
ATP
H2O2
H2O
ANPEP
Neuron cytosol
Extracellular matrix
Figure 6. Schematic overview of the effect of curcumin or its derivatives on GSH metabolism. De novo synthesis of GSH from its constituent amino acids in the neuronal
cytoplasm occurs as a two-step reaction. Whereas curcumin and its derivatives D1, D2, and D3 contribute to GSH synthesis at the rate limiting step by increasing the gene
expression of c-GCL, D3 also contributes via a minor elevation of glu levels in the cytosol. Conversion of the oxidized form of GSH, GSSG also contributes to the total cellular
GSH pool. GSH released from the cell is hydrolyzed by the activity of GGT in the extracellular space to its constituent amino acids, which could later be taken up by the cells
for de novo GSH synthesis. The mitochondrial GSH pool is maintained by uptake of cytosolic GSH via an energy-dependent mechanism. GSH detoxifies ROS and RNS generated
in the mitochondria during different metabolic reactions. List of abbreviations: GSSG, GSH disulfide; GR, GSH reductase; ANPEP, alanyl (membrane) aminopeptidase; Gly,
glycine; gluCys, glutamate-cysteine; Cys, cysteine; Glu, glutamate; GGT, gamma-glutamyl transpeptidase; GS, GSH synthase; GPx, Glutathione peroxidase; H2O2, hydrogen
peroxide.
binding of specific transcription factor complexes to TRE and EpRE
elements. Further, the authors showed that curcumin modulates
the components of transcription factors that compose EpRE and
AP-1 complexes which bind to these cis elements, thereby impact-
ing the GCL expression. Lavoie et al.41 demonstrated both in astro-
cytes and neuronal cells that curcumin caused a specific and
significant increase in the expression of the gene coding for the
significant increase in GSH levels and showed maximum detoxifi-
cation of ROS. The predictive study carried out to delineate the
beneficial effects of D3 indicated the following: During GSH syn-
thesis, the rate limiting step is the formation of the dipeptide gam-
ma-glutamyl cysteine.46 This requires influx of glu (as glutamine)
and cys (as cystine) from the extracellular matrix as shown in
the schematic representation in Figure 6.9 Since it was not possible
to introduce the glutamoyl diester of curcumin in our neuronal
platform, we chose to increase the extracellular glu levels (thus
modifier subunit of
ulation of
c-GCL (GCLM) which is required for the up reg-
c
-GCL activity. Based on these data and our previous re-
ports,18,19 we surmise that curcumin could be a potential candidate
for adjunctive pharmacotherapy in PD.
increasing its influx into the neuronal cytosol) from 1
lM (control
condition) to 3 M along with increased levels of cytosolic curcu-
l
However, there are several reports indicating poor bioavailabil-
ity of curcumin due to the rapid systemic derivatization in the GI
tract suggesting that small doses of curcumin are necessary for
its neuroprotective effect.42 Limited bioavailability might be due
to poor absorption, rapid metabolism and quick systemic elimina-
tion. In order to improve its bioavailability, the use of adjuvants
like piperine to prevent glucuronidation, use of curcumin nanopar-
ticles, curcumin structural analogues, liposomal curcumin or cur-
cumin phospholipid complexes have been attempted.43
min. Consequently, the initial increase in GSH was basically due
to increased neuronal glu without any significant changes in cys-
tine uptake and intracellular cys levels. With increased glu in the
lM range in the system, we observed initial elevation in cellular
GSH. However, when the glutamate levels increased to 1–5 mM,
cystine uptake in astrocytes was inhibited leading to reduction in
neuronal cys thereby decreasing GSH levels (data not shown). De-
spite abundant glutamate levels, GSH levels dropped since cys here
became rate limiting for synthesis.
We have utilized diester derivatives of curcumin (di-piperoyl,
di-valinoyl and di-glutamoyl) that improve its bioavailability.20
These conjugates could be dissociated at the target site by cellular
esterases releasing curcumin thereby functioning on as pro-drugs.
Studies in animal models and human volunteers showed that pip-
erine, an active natural compound from black pepper enhances the
activity and bioavailability of curcumin several fold.44 Similarly,
valinoyl ester pro-drugs exhibit improved bioavailability.45 The
diesters of curcumin used in our study have: (i) enhanced meta-
bolic stability due to protection of phenolic hydroxyl groups thus
delaying their glucuronidation (ii) improved cellular uptake due
to transportation via amino acid carriers and (iii) better solubility
due to increased hydrophilicity.20
Our study indicated that administration of D3 probably en-
hanced cellular GSH levels not only via increased accumulation
of curcumin in the cell, but also by improved intracellular concen-
trations of glu in the non-toxic range. D3 not only exhibits im-
proved bioavailability but also independently enhances GSH
levels making it a specific pro-drug for diseases involving oxidative
stress. In conclusion, compounds such as D3 with improved up-
take, ROS-scavenging capacity and neuroprotective efficiency
in vivo compared with curcumin with a capacity to cross blood
brain barrier could serve as potential neuroprotective strategies
in disorders such as PD.
Acknowledgments
We observed that all the three derivatives tested were either
equal to or more potent compared to curcumin in countering the
GSH depletion mediated oxidative stress. Moreover, the derivatives
showed neuroprotection both during pre and post treatment sug-
gesting that they could reverse the oxidative damage caused dur-
ing GSH depletion. Among the three bioconjugates, D3 exhibited
This work was supported by a fast track grant from DST, India to
M.M.S.B. G.H. gratefully acknowledges the junior research
fellowship from ICMR, India. R.B.M. gratefully acknowledges the
senior research fellowship from CSIR.

2638
G. Harish et al. / Bioorg. Med. Chem. 18 (2010) 2631–2638
22. Mishra, S.; Kapoor, N.; Mubarak, A. A.; Pardharasaradhi, B. V. V.; Leela, K. A.;
References and notes
Khar, A.; Misra, K. Free Radical Biol. Med. 2005, 38, 1353.
23. Prasad, K. N.; Carvalho, E.; Kentroti, S.; Edwards-Prasad, J.; Freed, C.;
Vernadakis, A. In Vitro Cell Dev. Biol. Anim. 1994, 30A, 596.
24. Adams, F. S.; La Rosa, F. G.; Kumar, S.; Edwards-Prasad, J.; Kentroti, S.;
Vernadakis, A.; Freed, C. R.; Prasad, K. N. Neurochem. Res. 1996, 21, 619.
25. Wolff, S. P. Methods Enzymol. 1994, 233, 182.
26. Ohkawa, H.; Ohishi, U.; Yagi, K. Anal. Biochem. 1979, 95, 351.
27. Ohashi, T.; Mizutani, A.; Murakami, A.; Kojo, S.; Ishii, T.; Taketani, S. FEBS Lett.
2002, 511, 21.
28. Guthenberg, C.; Alin, P.; Mannervik, B. Methods Enzymol. 1985, 113, 507.
29. Muyderman, H.; Nilsson, M.; Sims, N. R. J. Neurosci. 2004, 24, 8019.
30. Chen, Y.; Shertzer, H. G.; Schneider, S. N.; Nebert, D. W.; Dalton, T. P. J. Biol.
Chem. 2005, 280, 33766.
31. Dickinson, D. A.; Iles, K. E.; Zhang, H.; Blank, V.; Forman, H. J. FASEB J. 2003, 17,
473.
32. Bernstein, B. W.; Bamburg, J. R. J. Neurosci. 2003, 23, 1.
33. Raichur, A.; Vali, S.; Gorin, F. Neuroscience 2006, 142, 859.
34. Cadenas, E.; Davies, K. J. Free Radical Biol. Med. 2000, 29, 222.
35. Jezek, P.; Hlavata, L. Int. J. Biochem. Cell Biol. 2005, 37, 2478.
36. Giulivi, C.; Poderoso, J. J.; Boveris, A. J. Biol. Chem. 1998, 273, 11038.
37. Iwunze, M. O.; McEwan, D. Cell Mol. Biol. (Noisy-le-grand) 2004, 50, 749.
38. Bharath, M. M. S. Recent Pat. Endocr. Metabol. Immune Drug Discov. 2008, 2, 135.
39. Zbarsky, V.; Datla, K. P.; Parkar, S.; Rai, D. K.; Aruoma, O. I.; Dexter, D. T. Free
Radical Res. 2005, 39, 1119.
40. Baum, L.; Ng, A. J. Alzheimers Dis. 2004, 6, 367.
41. Lavoie, S.; Chen, Y.; Dalton, T. P.; Gysin, R.; Cuénod, M.; Steullet, P.; Do, K. Q. J.
Neurochem. 2009, 108, 1410.
42. Ramassamy, C. Eur. J. Pharmacol. 2006, 545, 51.
43. Anand, P.; Kunnumakkara, A. B.; Newman, R. A.; Aggarwal, B. B. Mol. Pharmacol.
2007, 4, 807.
44. Shoba, G.; Joy, D.; Joseph, T.; Majeed, M.; Rajendran, R.; Srinivas, P. S. Planta
Med. 1998, 64, 353.
45. Sinko, P. J.; Balimane, P. V. Biopharm. Drug Dispos. 1998, 19, 209.
46. Griffith, O. W. Free Radical Biol. Med. 1999, 27, 922.
1. Beal, M. F. Ann. Neurol. 1992, 31, 119.
2. Burke, R. E. In Cell Death and Disease of The Nervous System; Koliatsos, V. E.,
Ratan, R. R., Eds.; Humana Press: Totowa, NJ, 1998; pp 459–475.
3. Adams, J. D., Jr.; Chang, M. L.; Klaidman, L. Curr. Med. Chem. 2001, 8, 809.
4. Sayre, L. M.; Smith, M. A.; Perry, G. Curr. Med. Chem. 2001, 8, 721.
5. Perry, T. L.; Yong, V. W. Neurosci. Lett. 1986, 67, 269.
6. Jenner, P. Acta Neurol. Scand. Suppl. 1993, 146, 6.
7. Bharath, S.; Hsu, M.; Kaur, D.; Rajagopalan, S.; Andersen, J. K. Biochem.
Pharmacol. 2002, 64, 1037.
8. Jha, N.; Jurma, O.; Lalli, G.; Liu, Y.; Pettusi, E. H.; Greenamyre, J. T.; Liu, R. M.;
Forman, H. J.; Andersen, J. K. J. Biol. Chem. 2000, 275, 26096.
9. Vali, S.; Mythri, R.; Jagatha, B.; Padiadpu, J.; Ramanujan, K. S.; Andersen, J. K.;
Gorin, F.; Bharath, M. M. Neuroscience 2007, 149, 904.
10. Aggarwal, B. B.; Kumar, A.; Bharti, A. C. Anticancer Res. 2003, 23, 363.
11. Aggarwal, B. B.; Sundaram, C.; Malani, N.; Ichikawa, H. Adv. Exp. Med. Biol. 2007,
595, 1.
12. DuVoix, A.; Blasius, R.; Delhalle, S.; Schnekenburger, M.; Morceau, F.; Henry, E.;
Dicato, M.; Diederich, M. Cancer Lett. 2005, 223, 181.
13. Lim, G. P.; Chu, T.; Yang, F.; Beech, W.; Frautschy, S. A.; Cole, G. M. J. Neurosci.
2001, 21, 8370.
14. Yang, F.; Lim, G. P.; Begum, A. N.; Ubeda, O. J.; Simmons, M. R.; Ambegaokar, S.
S.; Chen, P. P.; Kayed, R.; Glabe, C. G.; Frautschy, S. A.; Cole, G. M. J. Biol. Chem.
2005, 280, 5892.
15. Thiyagarajan, M.; Sharma, S. S. Life Sci. 2004, 74, 969.
16. Rajeswari, A.; Sabesan, M. Inflammopharmacology 2008, 16, 96.
17. Rajeswari, A. Eur. Rev. Med. Pharmacol. Sci. 2006, 10, 157.
18. Mythri, R. B.; Jagatha, B.; Pradhan, N.; Andersen, J.; Bharath, M. M. Antioxid.
Redox Signal. 2007, 9, 399.
19. Jagatha, B.; Mythri, R. B.; Vali, S.; Bharath, M. M. Free Radical Biol. Med. 2008, 44,
907.
20. Dubey, S. K.; Sharma, A. K.; Narain, U.; Misra, K.; Pati, U. Eur. J. Med. Chem. 2008,
43, 1837.
21. Mishra, S.; Narain, U.; Mishra, R.; Misra, K. Bioorg. Med. Chem. 2005, 13, 1477.