Synthesis and complexation properties of two new curcuminoid molecules bearing a diphenylmethane linkage

Article abstract of DOI:10.1016/S0022-2860(03)00050-4

Bis-curcuminoids 3 and 4, bearing a diphenylmethane bridge for both compounds and a crown ether chain adapted to complex Li⁺ for compound 4, were synthesized and characterized by mass and NMR spectroscopies. By their preorganized geometry, they represent a new class of curcuminoids able to complex transition metal cations. Their absorption spectra in DMF by comparison with those of curcumin 1 and dimethylcurcumin 2 have shown that the phenolate forms in the two halves are in interaction in compound 3 and are independent in compound 4. Complexation studies have revealed a poor selectivity of curcuminoids for transition metals. Nevertheless complexation of Cu (II), studied by UV-visible absorption spectroscopy, has shown subtle differences between curcumin 1 and bis-curcuminoid 3. These observations were supported by quantum mechanic calculations to establish the most probable structures of non- and complexed compounds.

Full text of DOI:10.1016/S0022-2860(03)00050-4

Journal of Molecular Structure 649 (2003) 177–190
www.elsevier.com/locate/molstruc
Synthesis and complexation properties of two new curcuminoid
molecules bearing a diphenylmethane linkage
Agus Sundaryonoa, Aziz Nourmamodea, Christian Gardrata, Alain Fritschb,
Alain Castellana,
*
a
´ ´ ´ ´
Laboratoire de Chimie des Substances Vegetales, UPRES EA-494, Universite Bordeaux 1, 351, cours de la Liberation, F-33405
Talence, France
b
´ ´ ´
Laboratoire de Physico-Chimie Moleculaire, CNRS UMR 5803, Universite Bordeaux 1, 351, cours de la Liberation, F-33405 Talence, France
Received 29 November 2002; revised 9 January 2003; accepted 9 January 2003
Abstract
Bis-curcuminoids 3 and 4, bearing a diphenylmethane bridge for both compounds and a crown ether chain adapted to
complex Liþ for compound 4, were synthesized and characterized by mass and NMR spectroscopies. By their preorganized
geometry, they represent a new class of curcuminoids able to complex transition metal cations. Their absorption spectra in DMF
by comparison with those of curcumin 1 and dimethylcurcumin 2 have shown that the phenolate forms in the two halves are in
interaction in compound 3 and are independent in compound 4. Complexation studies have revealed a poor selectivity of
curcuminoids for transition metals. Nevertheless complexation of Cu (II), studied by UV–visible absorption spectroscopy, has
shown subtle differences between curcumin 1 and bis-curcuminoid 3. These observations were supported by quantum mechanic
calculations to establish the most probable structures of non- and complexed compounds.
q 2003 Elsevier Science B.V. All rights reserved.
Keywords: Curcumin; Curcuminoids; UV–visible absorption; Complexation; Transition metals; Calculated structures
1. Introduction
its powerful antioxidant activity [2–8], which was
used in food system [9], curcumin has anti-inflamma-
tory properties [10], HIV antiproteases activity [11]
and cancer preventive properties [12].
Curcumin, 1,7-bis(4-hydroxy-3-methoxyphenyl)-
1,6-heptadiene-3,5-dione, 1, is the main yellow
compound of Curcuma longa rhizomes (turmeric)
and is widely used as food coloring additive [1]. The
medicinal activity of curcumin has been known since
ancient times and this molecule has been the object of
several investigations in the field of biology, medicine
and pharmacology over the last decades. In addition to
Curcumin belongs to a group of naturally occurring
1,3-diketones in which the carbonyl groups are
directly linked to olefinic carbons, conferring com-
plexing properties to the molecule. Curcumin has
been known for its ability to form colored complexes
with other molecules [13–20]. Curcumin reacts with
boric acid to form a red colored complex which is
used for boron determination by visible absorption
spectrometry [13,14]. Moreover, curcumin forms
*
Corresponding author.
E-mail address: a.castellan@lcsv.u-bordeaux1.fr (A. Castellan).
0022-2860/03/$ - see front matter q 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0022-2860(03)00050-4

178
A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
strong colored chelates with transition metals ions
Cu2þ, Ni2þ, Zn2þ, Pd2þ, Fe3þ [17–20].
2. Experimental
In our continuous effort in studying interactions
between aromatic chromophore systems [21,22],
curcumin offers an unique opportunity to design and
study systems with some specific complexation ability
in a restrained geometry. In this paper, complexation
of transition metals mainly copper(II) by curcumin 1,
dimethylcurcumin 2 and new curcuminoids 3 and 4
(Scheme 1) is described using molecular absorption
spectroscopy. Compound 3 displays a pre-organized
complexing system for transition metals whereas
compound 4 also incorporates a crown ether cavity
adapted to complex selectively lithium cation.
2.1. Materials and methods
The starting materials and solvents of appropriate
grade (for synthesis or for spectroscopy) were
obtained from Aldrich and used without further
purification. Melting points were measured on a
heating microscope Electrothermal 9100 Reichert.
Studies by 1H and 13C NMR were made using
Bruker WP400 and/or AC250 Fourier transform
spectrometers (solvents CDCl3 or DMF-d7). J values
are given in Hz. IR spectra were registered with a
Paragon 1000 PC Perkin–Elmer FT-IR spec-
trometer. UV/Vis spectra were recorded on a
Lambda 18 Perkin–Elmer spectrometer using just-
prepared curcuminoid solutions to avoid hydrolytic
degradation [15]. Low and high-resolution mass
spectra were obtained with a V.G. Micromass
AutoSpec Q operating with a positive LSIMS
ionization mode (Csþ, ion bombardment energy:
35 keV; matrix: 3-nitrobenzyl alcohol). The com-
pounds were purified using Merck silica gel 60
(eluent dichloromethane–diethyl ether mixture). The
purity of the synthesized compounds was confirmed
by TLC analysis carried out on Fluka silica gel F254
plates (thickness 0.20 mm). Also the purity of
curcuminoids 2 and 3 was verified by HPLC using
a Shimadzu liquid chromatography equipment (LC9
pump, SPD6A UV detector set at 280 nm and CR4A
data station) equipped with a silica gel phase
(Nucleosil 100, 5 mm) column (length 25 cm,
internal diameter 4.6 mm, eluent CH2Cl2–methyl t-
butyl ether 7v-3v). The amount of copper in the
complexes was determined by atomic absorption
spectroscopy (AAS) with an Instrument Laboratory
(Type IL151) spectrometer. The mineralization of
the complex was performed using persulfuric acid
produced by action of hydrogen peroxide on
concentrated sulfuric acid.
2.2. Syntheses
Curcumin 1 and dimethylcurcumin 2 were pre-
pared according to a procedure described in the
literature [23].
Scheme 1. Formulae of studied compounds.

A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
179
2.2.1. Bis{2-hydroxy-3-methoxy-5-[7-(4-hydroxy-3-
methoxyphenyl)-hepta-1E,6E-diene-3,5-
dioxo]}methane (3)
the value 2–3. The mixture was stirred for 1 h
and extracted with dichloromethane. The organic
layer was washed with water, dried (Na2SO4) and
evaporated under vacuum. The solid residue was
purified by column chromatography (eluent dichlor-
omethane–diethyl ether 9:1). Compound 3 was
isolated as an orange solid (360 mg, 34%), mp
149–155 8C. HPLC analysis of the isolated com-
pound revealed only one absorption peak. FT-IR
(KBr) (cm21): 3490, 3380, 2930, 2850, 1630, 1594,
6-(4-Hydroxy-3-methoxyphenyl)-hex-5E-ene-2,4-
dione ( feruloylacetone, 5). Acetylacetone (7.5 g,
75 mmol), boric anhydride (3.5 g, 54 mmol) and tri-
n-butyl borate (23 g, 100 mmol) were mixed in ethyl
acetate (50 ml) at 0 8C with stirring. A solution of
vanillin (3.75 g, 25 mmol) and n-butylamine (0.5 ml)
in ethyl acetate was slowly added for 1.5 h. The
reaction mixture was stirred at 0 8C for 90 min and
then overnight at room temperature. Hydrochloric
acid (0.4N) was then added to the solution, at 60 8C,
until the pH became acid. The reaction mixture was
stirred for an additional hour and then extracted with
ethyl acetate. The organic layer was washed with
water, dried (Na2SO4) and evaporated under vacuum.
Feruloylacetone was obtained by column chromatog-
raphy (eluent dichloromethane–diethyl ether 95:5) as
a pale yellow solid (2.14 g, 37%), mp 144–147 8C (lit.
1
1510, 1385, 1290, 1150, 1080, 970, 830. H NMR
(400 MHz; DMF-d7) d: 3.90 (s, 6H, H700 and H7000),
3.92 (s, 6H, H7 and H70), 4.02 (s, 2H, H8), 6.02 (s,
2H, Hd and Hd0), 6.77 (d, J ¼ 15.5 Hz, 2H, Hb and
Hb0), 6.81 (d, J ¼ 15.5 Hz, 2H, Hf and Hf0), 6.90 (d,
J ¼ 7.8 Hz, 2H, H500 and H5000), 7.09 (d, J ¼ 2.1 Hz,
2H, H6 and H60), 7.19 (dd, J ¼ 2.1, 7.8 Hz, 2H, H600
and H6000), 7.33 (d, J ¼ 2.1 Hz, 2H, H2 and H20),
7.40 (d, J ¼ 2.1 Hz, 2H, H200 and H2000), 7.55 (d,
J ¼ 15.5 Hz, 2H, Ha and Ha0), 7.60 (d, J ¼ 15.5 Hz,
2H, Hg and Hg0), 9.73 (br s, 4H, phenol), 16.50 (br
s, enol). 13C NMR (100 MHz; DMF-d7) d: see
Table 1. The assignment was based on DEPT,
COSY, HMQC and HMBC experiments (see
Table 1). LSIMS-MS m/z 771 (6%, MNaþ), 330
(8), 749 (6%, MHþ), 329 (30), 177 (38), 176 (100);
m/z 749.2630 (Found MHþ); C43H41O12 requires
749.2598.
1
[24] 146–147 8C). H NMR (250 MHz; CDCl3) d:
2.20 (s, 3H, CH3–), 3.95 (s, 3H, CH3O–), 5.63 (s, 1H,
–CHyC–OH), 6.5 (d, 1H, –CHyCH–CyO), 7.0 (m,
3H, H–Ar), 7.6 (d, 1H, –CHyCH–Ar).
4,40-Dihydroxy-5,50-dimethoxy-3,30-methanediyl-
dibenzaldehyde (6). Vanillin (178 g, 1.17 mol) and an
aqueous solution (35%) of formaldehyde (60 g,
0.7 mol) were heated under reflux and stirring. A
solution of NaOH (50 g, 1.25 mol) in 50 ml of H2O
was added and the reaction mixture was refluxed for
3 h, then poured into 2.5 l of H2O at 75–80 8C before
a dilute hydrochloric acid (10%) solution (0.5 l) was
added dropwise. The solid was filtered, washed with
water and acetone and dried under vacuum. TLC
analysis indicated the presence of pure 6 (17 g, 10%),
mp 255–262 8C (lit. [25] 274 8C). 1H NMR
(250 MHz; CDCl3) d: 3.90 (s, 6H, CH3O–Ar); 4.0
(s, 2H, Ar–CH2–Ar); 7.3 (m, 4H, –Ar); 9.75 (s, 2H,
–CHO).
2.2.2. 7,20-Dimethoxy-5,22-di-[7-(4-hydroxy-3-
methoxyphenyl)-hepta-1E,6E-diene 3,5-dioxo]
{[9,12,15,18-tetraoxatricyclo[17.4.0.03,8]tricosa-1
(23),3(8),4,6,19,21-hexaene} (4)
7,20-Dimethoxy-5,22-diformyl[9,12,15,18-tetraox-
atricyclo[17.4.0.03,8]tricosa-1 (23),3(8),4,6,19,21-
hexaene (7). Compound 6 (1 g, 3.16 mmol), sodium
carbonate (1.34 g, 12.64 mmol) and 1,8-diiodo-3,6-
dioxaoctane (1.17 g, 5.4 mmol) in DMF (150 ml)
were stirred at 100 8C, under a nitrogen atmosphere,
for 30 h. The solution was filtered then evaporated.
The solid residue was extracted with dichloro-
methane, washed by dilute hydrochloric acid and
water, dried (Na2SO4) and evaporated in vacuo. The
solid was purified by column chromatography (eluent
dichloromethane–diethyl ether 95:5) yielding the title
compound as a pale solid (650 mg, 48%), mp 171–
172 8C. FT-IR (KBr) (cm21): 3440, 2925, 2860, 2740,
1695, 1675, 1585, 1345, 1310, 1140, 1090, 1035, 745.
Curcuminoid 3. Feruloylacetone 5 (700 mg,
3.00 mmol), boric anhydride (294 mg, 4.20 mmol)
and tri-n-butyl borate (3.22 g, 14 mmol) were heated
at 80 8C, for 1 h, in DMF (30 ml). A solution of 6
(442 mg, 1.40 mmol) and n-butylamine (0.14 ml) in
DMF (5 ml) was slowly added, for 60 min, at the
same temperature. After 5 h, the reaction mixture
was cooled down to 60 8C and hydrochloric
acid (0.4N) was added until the pH reached

180
A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
Table 1
13C NMR data and correlated protons in HMQC and HMBC experiments of compound 3 (100 MHz, solvent: DMF-d7)
C-number
Chemical shift (ppm)
Correlated protons in HMQC
Correlated protons in HMBC
8
31.0
8
7, 70, 700 and 7000
d and d0
200and 20
2 and 2000
500 and 5000
b and b0
f and f0
57.2 and 57.4
102.6
110.3
112.8
117.4
122.7
123.0
124.9
126.4
127.5
128.6
142.3
142.9
149.9
150.1
150.1
151.5
151.5
185.2
185.6
7, 70, 700 and 7000
d and d0
200and 20
2 and 2000
500 and 5000
b and b0
f and f0
b and b0; f and f0
6 and 60; a and a0
600 and 6000, g and g0
d and d0
d and d0
600 and 6000
6 and 60
1 and 10
100 and 1000
g and g0
a and a0
5 and 50
4 and 40
400 and 4000
3 and 30
300 and 3000
e and e0
c and c0
600 and 6000
6 and 60
200 and 2000; g and g0
8; 2 and 20; a and a0
b and b0
f and f0; 500 and 5000
200 and 2000; 600 and 6000
2 and 20; 6 and 60
8; 6 and 60
2 and 20; 6 and 60
200 and 2000; 500 and 5000; 600 and 6000
2 and 20; 7 and 70; 8
200 and 2000; 500 and 5000; 700 and 7000
d and d0; f and f0; g and g0
a and a0; b and b0; d and d0
g and g0
a and a0
1H NMR (250 MHz; CDCl3) d: 3.6 (m, 8H, –CH2–
O–CH2–), 3.95 (s, 6H, CH3O–Ar), 4.3 (m, 4H, –
CH2–O–Ar), 4.4 (s, 2H, Ar–CH2–Ar), 7.0 (m, 4H, –
Ar), 9.7 (s, 2H, –CHO).
1 h and extracted with dichloromethane. The organic
layer was washed with water, dried (Na2SO4) and
evaporated under vacuum. The solid residue was
purified by column chromatography (eluent dichlor-
omethane–diethyl ether 9:1) giving compound 4 as an
orange solid (240 mg, 20%), mp 104–110 8C. As for
compound 3, HPLC analysis of the isolated com-
pound revealed only one absorption peak. FT-IR
(KBr) (cm21): 3400, 2930, 2860, 1625, 1580, 1510,
1420, 1300, 1130, 960, and 850. 1H NMR (400 MHz;
DMF-d7) d: 3.60 (br s, 4H, Hu and Hu0), 3.64 (br s, 4H,
Hv and Hv0), 3.90 (s, 6H, H7 and H70), 3.99 (s, 6H, H700
and H7000), 4.30 (br s, 4H, Hw and Hw0), 4.37 (s, 2H,
Curcuminoid 4. Feruloylacetone 5 (1.31 g,
5.6 mmol), boric anhydride (294 mg, 4.2 mmol) and
tri-n-butyl borate (2.57 g, 11.2 mmol) in DMF (30 ml)
were heated at 80 8C, for 1 h. A solution of 7 (600 mg,
1.40 mmol) and n-butylamine (0.15 ml) in DMF
(5 ml) was slowly added (60 min). After 5 h, the
reaction mixture was cooled down to 60 8C and
hydrochloric acid (0.4N) was added until the reaction
mixture reached pH 2–3. The mixture was stirred for

A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
181
H8), 6.05 (s, 2H, Hd and Hd0), 6.79 (d, J ¼ 15.7 Hz,
2H, Hf and Hf0), 6.80 (d, J ¼ 1.8 Hz, 2H, H6 and H60),
6.81 (d, J ¼ 15.7 Hz, 2H, Hb and Hb0), 6.89 (d,
J ¼ 7.3 Hz, 2H, H500 and H5000), 7.19 (dd, J ¼ 1.9,
7.3 Hz, 2H, H600 and H6000), 7.40 (d, J ¼ 1.8 Hz, 2H,
H200 and H2000), 7.42 (d, J ¼ 1.8 Hz, 2H, H2 and H20),
7.53 (d, J ¼ 15.7 Hz, 2H, Ha and Ha0), 7.62 (d,
J ¼ 15.7 Hz, 2H, Hg and Hg0), 9.73 (br s, 2H, phenol),
16.40 (br s, enol). 13C NMR (100 MHz; DMF-d7) d:
see Table 2. The assignment was based on DEPT,
COSY, HMQC and HMBC experiments (see Table 2);
LSIMS-MS m/z 885 (39%, MNaþ), 863 (10%, MHþ),
413 (100); m/z 885.3119 (Found MNaþ);
C49H50O14Na requires 885.3098.
2.2.3. Copper(II) complex with 3
The experimental procedure was adapted from
Krishnankutty’s work [19]. A solution of copper(II)
acetate (7.8 mg, 0.043 mmol) in methanol (20 ml) was
added within 2 h, under stirring, at room temperature,
to a solution of 3 (30 mg, 0.04 mmol) and sodium
acetate (9.6 mg, 0.12 mmol) in methanol (20 ml).
Table 2
13C NMR data and correlated protons in HMQC and HMBC experiments of compound 4 (100 MHz, solvent: DMF-d7)
C-number
Chemical shift (ppm)
Correlated protons in HMQC
Correlated protons in HMBC
8
31.4
57.3
8
7, 70, 700 and 7000
u and u0
v and v0
w and w0
d and d0
200and 20
2 and 2000
500 and 5000
f and f0
7, 70, 700 and 7000
u and u0
v and v0
w and w0
d and d0
200and 20
2 and 2000
500 and 5000
f and f0
71.6
71.6
74.6
102.8
111.4
112.9
117.4
123.0
124.0
124.0
124.0
128.6
132.2
136.6
141.8
142.8
150.1
150.3
151.7
154.4
184.6
186.2
b and b0; f and f0
6 and 60; a and a0
600 and 6000; g and g0
d and d0
8; 2 and 20; a and a0
d and d0
200 and 2000; g and g0
f and f0; 500 and 5000
b and b0
6 and 60
b00and b0
6 and 6000
100 and 1000
1 and 10
4 and 40
a and a0
g and g0
3 and 30
5 and 50
300 and 3000
400 and 4000
c and c0
e and e0
6 and 60
b and b0
600 and 6000
8
a and a0
g and g0
2 and 20; 6 and 60
200 and 2000; 600 and 6000
2 and 20; 7 and 70; 8
2 and 20; 8; 6 and 60
200 and 2000; 700 and 7000
200 and 2000; 600 and 6000; 500 and 5000
a and a0; b and b0; d and d0
d and d0; f and f0; g and g0

182
A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
After 15 h, filtration of the solution, washing the solid
with methanol and diethyl ether, allow isolation of the
complex in the form of a brown powder (20 mg, 65%),
mp .300 8C; FT-IR (KBr) (cm21): 3490, 3380, 2930,
2850, 1630, 1594, 1510, 1385, 1290, 1150, 1080, 970,
830; AAS Found: Cu, 9.05%.
complexation of b-diketone with transition metals
was performed in their presence [18–20]. The
position and intensity of the long-wavelength absorp-
tion bands indicate that ppp transitions of the
through-conjugated system dominate due to the
conjugation of their p-electrons in the enolic
configuration [28,29].
The UV/visible absorption of curcumin 1
(1.8 £ 1024 mol l21) in DMF, in absence of sodium
acetate or sodium hydroxide, displays an intense band
at 430 nm due to the neutral molecule and a smaller
one at 590 nm due to the deprotonated form [32]. In
presence of sodium acetate or sodium hydroxide, the
intensity of the band at 430 nm decreases and the one
at 590 nm increases. Different assignments of the acid
dissociation constants of curcumin can be found in the
literature [20,31,32]. Toniolo et al. [32], in accordance
with Dietze et al. [33] findings, indicate that curcumin
deprotonation occurs in two steps: the easiest proton
release involves both phenolic groups with little
difference in the relevant dissociation constants, while
the release of a proton from the enol system is more
difficult (3 units pKA difference depending on the
medium). Borsari et al. [20], by means of NMR,
potentiometric titrations and UV–visible absorption
measurements, indicate that the enolic proton is the
more acidic one. The presence of sodium acetate or
sodium hydroxide in dimethylcurcumin 2 solution
induces a decrease of the band intensity at 428 nm
assigned to the neutral form in curcumin. This shows
that formation of enolic anion does not create any new
observable ppp transitions and consequently, it
indicates that the band at 590 in curcumin spectrum
can be assigned to the phenolate anions. Conse-
quently, sodium acetate in large excess in DMF
solution is able, as does sodium hydroxide, to
deprotonate both enolic and phenolic protons, in
accordance with Borsari experiments [20].
3. Results and discussion
3.1. Synthesis of biscurcuminoids 3 and 4
Curcumin 1 and dimethylcurcumin 2 were syn-
thesized according to a procedure patented by
Krackov and Bellis [23] by reacting, in presence of
n-butylamine, the acetylacetone boron complex
(obtained by action of tributylborate and boric
anhydride) with vanillin and 3,4-dimethoxybenzalde-
hyde, respectively. Increasing the acetylacetone con-
tent, reacting with vanillin, allowed the isolation of
feruloylacetone 5 [24]. Compounds 3 and 4 were
synthesized by the general procedure described in
Scheme 2 using feruloylacetone 5, as 1,3-diketone,
and the diphenylmethanes 6 and 7, respectively.
The structuresof 3and 4were establishedmainlyby
massspectrometrywithCsþSIMSionizationatlowand
high resolution, by UV–visible and FT-IR absorption
spectroscopies in comparison with curcumin and by
two-dimensional NMR techniques (see Section 2). As
for curcumin and dimethylcurcumin, compounds 3 and
4 exist in their enol form [26]. Compounds 3 and 4
represent a new class of preorganized complexing
system based on b-diketone units able to selectively
complex transition metal cations.
3.2. UV–visible absorption spectroscopy
Electronic absorption spectroscopy is a very useful
technique to study complex formation [27]. Curcumin
and curcuminoid derivatives display absorption in the
ultraviolet and visible regions [28–31]. The absorp-
tion spectra of curcumin 1, dimethylcurcumin 2 and
the biscurcuminoids 3 and 4 in DMF solutions are
shown in Fig. 1; DMF, a polar non-protic solvent, was
selected because it dissolves easily complexes of
curcuminoids and transition metal cations. The
influence of the presence, in large excess, of sodium
acetate or sodium hydroxide is also presented because
The absorption behavior of 4 is very similar to
curcumin’s one: e.g. absorption maxima of neutral
molecule and enolate anion at 425 nm and of
phenolate anion at 580 nm. This indicates that each
halves of compound 4 are acting quite independently
in the ground electronic state.
Compound 3 displays UV–visible absorption
spectra which are quite different compared to those
of compounds 4 and 1. This is due to different behavior
of the phenolate anions. The maximum absorption of

A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
183
Scheme 2. Synthetic procedure used for the preparation of the curcuminoid compounds 3 and 4.
neutral and enolate forms of 3 (433 nm) compares well
with the one of biscurcuminoid 4 (425 nm), whereas
the absorption maximum assigned to phenolate species
in compound 3 is blue shift (538 nm). This peculiar
behavior of curcuminoid 3 has to be related to some
interactions between the two phenol anionic moieties
through the diphenylmethane bridge.
by elemental analysis and UV–visible absorption
spectroscopy, using Job’s method of continuous
variation [34].
The experimental procedure given by Krishnan-
kutty and Venugopalan [19], to obtain complexes
between curcuminoids and transition metals, was
applied. The procedure in this study, which consists in
refluxing a mixture of the curcuminoid with the metal
acetate in methanol, led to a solid material with no
well define crystalline structure for Cu(II), Co(II),
Hg(II), Cu(I) and Fe(III). The spectra of the isolated
complexes of compound 3 with a number of transition
metals, in DMF solution (where they are fully
soluble), are shown in Fig. 2. It is clear that complexes
with Cu(II), Fe(III) and Co(II) present very different
absorption curves compared to the non-complexed
compound 3. This is indicative of a strong pertur-
bation of curcumin chromophore by the metal cation.
3.3. Complexation of metal cations
3.3.1. Curcuminoid 3
Curcumin is known to complex transition metals
[17–20,34]. With Cu(II), Ni(II), Zn(II) and Pd(II)
acetates, curcumin 1 forms complexes, which were
isolated and characterized by elemental analysis
showing two ligands for one metal cation. In contrast,
with Zr(IV) oxychloride, curcumin forms a 1/1
stoichiometric complex. Its structure was established

184
A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
Fig. 1. UV–visible absorption spectra of DMF solutions of: (A) curcumin 1 (1.8 £ 1024 mol l21) in DMF (—1), in presence of sodium acetate
(12 £ 1022 mol l21) (—2) or sodium hydroxide (12 £ 1022 mol l21) (- - -3); (B) dimethylcurcumin 2 (2.2 £ 1024 mol l21) (—1), in presence
of sodium acetate (12 £ 1022 mol l21) (—2) or sodium hydroxide (12 £ 1022 mol l21) (- - -3); (C) biscurcuminoid 3 (1024 mol l21) (—1), in
presence of sodium acetate (12 £ 1022 mol l21) (—2) or sodium hydroxide (12 £ 1022 mol l21) (- - -3); (D) biscurcuminoid 4 (1024 mol l21
(—1), in presence of sodium acetate (12 £ 1022 mol l21) (—2) or sodium hydroxide (12 £ 1022 mol l21) (- - -3).
)
In contrast, with Hg(II) and Cu(I), the spectra are not
very perturbed. This is in favor of a week interaction
between the ligand and the metal.
In order to examine in more details complexation
of Cu(II) by curcuminoids, absorption spectra of
curcuminoids (concentration 1.3 £ 1024 mol l21
)
were recorded in presence of a large excess of sodium
acetate (1.2 £ 1022 mol l21) with variable amount of
cupric acetate, in DMF solution, at 20 8C. The spectra
are shown in Figs. 3–5 for dimethylcurcumin 2,
curcumin 1 and curcuminoid 3, respectively. Similar
figures were obtained when sodium hydroxide was
used instead of sodium acetate.
Fig. 2. UV–visible absorption spectra of isolated complexes formed
between biscurcuminoid 3 and transition metal cations (see Section
2 for Cu(II) as an example) in presence of AcONa (12 mmol l21).
(—): 2; (–A–): 2/Cu(II); (–K–): 2/Co(II); (–W–): 2/Hg(II);
(– £ –): 2/Cu(I); (– þ –): 2/Fe(III).
Kuhlwein et al. [17] already described the
formation of 1/2 complex (ML2) between dimethyl-
curcumin and Cu(II). Its maximum absorption, found
at 421 nm for CH2Cl2 solution, compares well with

A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
185
Fig. 3. UV–visible absorption of compound 2 (1.3 £ 1024 mol l21) in presence of AcONa (0.012 mol l21) with variable amounts of Cu(OAc)2
(1) 0 mol l21; (2) 2.8 £ 1025 mol l21; (3) 5.6 £ 1025 mol l21; (4) 8.3 £ 1025 mol l21; (5) 11 £ 1025 mol l21; (6) 14 £ 1025 mol l21; (7)
19 £ 1025 mol l21
.
425 nm observed in this work for DMF solution
(Fig. 3). The differences observed in the absorption
spectra of curcuminoids, when increasing amounts of
Cu(II) are added to the solution to form complexes,
are much more pronounced for curcumin 1 (Fig. 4)
and biscurcuminoid 3 (Fig. 5) than for dimethylcur-
cumin 2 (Fig. 3). This is related to the formation of
phenolate anions in compounds 1 and 3 when sodium
acetate or sodium hydroxide are used in large excess
(see vide infra and Fig. 1). Job’s method of continuous
variation was not applied to determine the stoichi-
ometry of the complex because different complexes
are formed and the curves could be analyzed by a
more efficient programme: the LETAGROP-SPEFO
software, designed for this type of study [35,36]. It
allows the determination of the stoichiometry of the
complex and its binding constants. The procedure was
applied to curcumin 1 and biscurcuminoid 3 due to the
large differences found in the UV–visible absorption
spectra compared to dimethylcurcumin 2.
The validity of the fit was verified by the good
repartition of the residuals (calculated 2 experimental
values) and the experimental x2 ¼ 10:92; compared to
the theoretical expected one ðx2 ¼ 12:60Þ: The equili-
brium constants at 20 8C are log K11 ¼ 6:53ð^0:23Þ
and log K2 ¼ 10:88ð^0:22Þ; corresponding to K11
3:4 £ 106 and K12 ¼ 2:2 £ 104:
¼
The binding constants K11 and K12 for complex
formation between curcumin 1 and Cu(II) are 7 order
of magnitude lower in DMF, than those determined by
Arrieta et al. [18] in water/dioxane solvent mixture by
conductimetry; the ratio K12=K11 being the same in
both conditions. These differences between Arrieta’s
data and the present study are mainly due to different
solvent effects with lesser dissociation of the ion-pairs
in DMF than in dioxane/water mixtures.
For bis-curcuminoid 3, the curves were correctly
fitted using the following 2/1 complexation scheme.
This scheme gives the best residual plot and the
closest experimental x2 ¼ 14:03 compared to the
theoretical expected one ðx2 ¼ 12:60Þ: The equili-
brium constants at 20 8C were log K11 ¼ 5:15ð^0:16Þ
A correct analysis of the absorption curves was
obtained according to the following 1/2 complexation
scheme:
and log K2 ¼ 8:7ð^0:17Þ; which correspond to K11
¼
1:4 £ 105 and K21 ¼ 3:6 £ 103:
Compound 1 þ CuðIIÞ Y LCu K11
LCu þ Compound 1 Y CuL2 K12
K2 ¼ K12 £ K11
Compound 3 þ CuðIIÞ Y LCu K11
LCu þ CuðIIÞ Y LCu2 K21

186
A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
Fig. 4. Electronic absorption of mixtures of curcumin
1
(1.3 £ 1024 mol l21) in presence of AcONa (0.012 mol l21) with
variable amounts of Cu (OAc)2 (A) (—): 0 mol l21; (–A–):
Fig. 5. UV–visible absorption of compound 3 (1.3 £ 1024 mol l21
)
1.43 £ 1025 mol l21
;
(–K–): 2.7 £ 1025 mol l21
;
(–W–):
in presence of AcONa (0.012 mol l21) with variable amounts of Cu
5.38 £ 1025 mol l21; (– £ –): 7.92 £ 1025 mol l21; (B) (—):
(OAc)2 (A) (—): 0 mol l21; (–A–): 1.43 £ 1025 mol l21; (–K–):
7.92 £ 1025 mol l21
;
;
(–A–): 10 £ 1025 mol l21
;
(–K–):
2.7 £ 1025 mol l21
;
(–W–): 5.38 £ 1025 mol l21
;
(– £ –):
12.7 £ 1025 mol l21
(–W–): 17 £ 1025 mol l21
;
(– £ –):
7.92 £ 1025 mol l21; (B) (—): 7.92 £ 1025 mol l21; (–A–):
21.3 £ 1025 mol l21
.
10 £ 1025 mol l21
;
(–K–): 12.7 £ 1025 mol l21
;
(–W–):
17 £ 1025 mol l21; (– £ –): 21.3 £ 1025 mol l21
.
The infrared spectrum of the complex (Cu(II)/3) is
similar to the spectrum of compound 3 except that the
intensity of the band at 1630 cm21, due to stretching
of intramolecular bonded carbonyl function, is
K2 ¼ K21 £ K11
Mass spectrometry analysis of prepared complexes of
curcuminoids 1 and 3 with Cu(II) did not show any
molecular peaks even using soft ionization technique
like LSIMS (see Section 2). This contrasts with our
observations on dimethylcurcumin 2 and 1,7-diphe-
reduced in the complex and the band at 1508 cm21
,
due to the coordinated carbonyl of the b-diketone, is
exalted. Similar observations were reported for
curcumin derivatives [14,17].
nyl-1,6-heptadiene-3,5-dione/Cu(II)
complexes
The binding constant K11 determined for the
complex of curcuminoid 3 with Cu (II) corresponds
either to the formation of a 1/1 complex between one
curcumin moiety of 3 and Cu (II) or to a 1/2 complex
between the two curcumin halves of 3 and Cu (II). The
value of K11 (1.4 £ 105), when compared to the value
K11 £ K12 ¼ 7:5 £ 1010 for a 1/2 complex between
curcumin 1 and Cu (II), indicates some restrain
geometry to form a 1/2 complex type between the two
(Cu(II)L2 stoichiometry) and to published results by
Krishnankutty and Venugopalan [19]. To confirm that
the solid material isolated from curcuminoid 3 and
cupric acetate is a copper complex, the copper content
was measured by flame AAS. The value 9.05% (on a
weight basis) obtained is between the expected value
for a 1/1 and 2/1 complexes (7.8 and 11.45%),
respectively. This is in favor of the formation of both
complexes.

A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
187
halves of compound 3 and Cu(II). Also, if the two
parts in compound 3 behave independently to form a
1/1 complex, the value K11 should be close for
compounds 1 and 3, respectively. There is one order
of magnitude difference. All these observations
probably indicate that the two curcumin halves in
compound 3 are involved in the binding process
between Cu(II) and their enol parts. Formation of the
1/2 complex involves some restrain geometry which
comes from torsion of the two curcumin moieties
through the diphenyl methane linkage (vide supra).
Moreover, if the two halves behaved indepen-
dently in compound 3, the value of K21
( ¼ 3.6 £ 103), which represents the binding con-
stant between the second curcumin half and Cu(II),
should be close to the constant K11 for curcumin 1
and Cu(II) ( ¼ 3.4 £ 106), though it is known that
stability constants for coordination of a second
ligand are generally smaller than for the first ligand.
The much lower value found for K21 supports the
hypothesis that the first atom of Cu(II) is localized
between the two halves. For binding, the second
Cu(II) atom should disconnect the two parts before
binding with the enolate. For that reason, in the
complex curcuminoid 3/Cu(II)2, the diphenyl-
methane structure likely adopts a head to tail
conformation with the two complexed enol parts
situated in a non-interacting position.
Complexation of compound 4 with cupric acetate
in presence of a large excess of sodium acetate was
also studied by UV–visible absorption spectroscopy.
Different evolution of the absorption curves than for
biscurcuminoid 3 was noted (results not shown). No
good fitting between the calculated and experimental
curves was obtained using either 1/1 or 2/1 complex
schemes. This indicates that the mechanism of
complex formation between compound 4 and Cu (II)
is more complex than for compound 3, probably due
to the involvement of Naþ by the crown ether part in
the complexation process.
3.4. Study of the geometry of complexes by quantum
molecular calculations
No X-rays structures of complexes involving only
curcumin derivatives and transition metal cations
were published. Our attempts to obtain good crystals
from curcuminoid 3 and Cu(II) were unsuccessful.
Nevertheless, in order to get some insight into the
structure of the complexes, we performed semi-
empirical quantum chemistry calculations using the
AMPAC programme (AMPAC 6.0, q 1997, Semi-
chem, 7204 Mullen, Shawnee, KS 66216 (USA)). The
molecular structures were fully optimized using
the SAM1d parametrization [37] that accounts for
the copper element. Harmonic frequencies were
determined to insure that the optimized structures
are indeed minima on the potential energy surfaces.
The SAM1d geometries for C–O and Cu–O bonds
were confirmed against more refined Density Func-
tional Theory [38], using a non-local B3LYP func-
tional [39] with a DZVP2 double zeta basis set [40].
We used GAUSSIAN 98 to perform the calculation [39].
The latter was performed on a simplified model
system involving a copper (II) atom complexed with
two bidentate monoanions formed from hepta-1,6-
dien-3,5-dione. The SAM1d results were found to
provide very similar geometries (bond lengths and
angles). They represent conformations corresponding
to the global minimum in the potential energy surface.
It was shown previously [17,19] that the complex
between curcumin and Cu (II) displays a 2/1
stoichiometry involving two enolate anions. The
phenol groups were chosen in their neutral form to
calculate the different geometries to facilitate the
calculations.
3.3.2. Curcuminoid 4
As indicated in Section 1, compound 4 was
synthesized to form, in the same molecule, complexes
with both Liþ and transition metal cations. The
presence of different alkaline cation acetates (Liþ,
Naþ, Kþ, Csþ), in large excess, modifies the relative
intensityoftheabsorptionbandofthephenolatespecies
(580 nm) compared to the band at 425 nm due to the
enolate and non-deprotonated molecule. With Liþ, the
bandat580 nmhasaverylowintensity.Theadditionof
Liþ inincreasingquantitiestoasolutionofcompound 4
in presence of cesium or potassium acetates induces a
decrease of the band at 580 nm. The same effects were
observed for curcumin 1. All these observations are
indicative of the absence of selective complexation of
Liþ by the crown ether chain in compound 4, even
though enough room is available to form the complex.
The formation of tight ion pairs between phenolate and
Liþ appears more favored in DMF.

188
A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
Fig. 6. (A) L2Cu complex of curcumin 1; (B) curcuminoid 3; (C) LCu complex of curcuminoid 3.

A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
189
The structure of the L2Cu complex where L is
the enolate form of compound 1 is shown in
Fig. 6A. The complex is centro-symmetrical, with
the copper atom at the center, in a square planar
systems involving transition metal complexes, for
example in the catalysis area.
˚
environment. The Cu–O distances are 2.02 A, very
Acknowledgements
close to the observed value found in many copper
oxides based crystals [41]. The C–O distances in
˚
the enolate group are stretched to 1.30 A compared
We thank the cooperative programme between
France and Indonesia for a thesis grant to A.S. We are
indebted to Dr J.P. Desvergne for the analysis of the
electronic absorption curves using the LETAGROP-
SPEFO software and to Dr P. Pardon for atomic
absorption measurements. The assistance of Mr
M. Petraud in the NMR work is acknowledged.
˚
to 1.29 A for the carbonyl group in curcumin [42]
and rhodium (I) curcuminoid complexes [17].
The calculated structure for compound 3, with a Cs
symmetry, is shown in Fig. 6B. The methylene bridge
does not show significant strain, with a 1088 C–C–C
valence angle. The curcumin halves are very similar
to the geometry determined for compound 1. The
structure of a LCu complex in which both deproto-
nated enol halves are bonded to the Cu(II) cation was
determined. This is shown in Fig. 6C. This complex
retains a Cs symmetry. The copper site shows a quasi
square planar environment and the Cu–O bond length
References
[1] V.S. Govindarajan, Crit. Rev. Food Sci. Nutr. 12 (1980) 199.
[2] T. Masuda, K. Hidaka, A. Shinohara, T. Maekawa, Y. Takeda,
H. Yamaguchi, J. Agric. Food Chem. 47 (1999) 71.
[3] T. Masuda, K. Hidaka, T. Toi, H. Bando, T. Maekawa, Y.
Takeda, H. Yamaguchi, J. Agric. Food Chem. 50 (2002) 2524.
[4] S.M. Khopde, K.I. Priyadarsini, P. Venkatesan, M.N.A. Rao,
Biophys. Chem. 80 (1999) 85.
˚
is 2.03 A on average, very close to that found for the
L2Cu complex of compound 1. Moreover, the
methylene bridge bonding angle is barely affected
(1078). Most of the necessary conformational modi-
fications therefore occurs in each curcumin halves,
with a large torsional effect at the rings junctions.
[5] S.V. Jovanovic, S. Steenken, C.W. Boone, G. Simic, J. Am.
Chem. Soc. 121 (1999) 9677.
[6] L. Ross, C. Barclay, M.R. Vinqvist, Org. Lett. 2 (2000) 2841.
[7] F. Ortica, M.A.J. Rodgers, Photochem. Photobiol. 74 (2001)
745.
[8] Y.M. Sun, H.Y. Zhang, D.Z. Chen, C.B. Liu, Org. Lett. 4
(2002) 2909.
4. Conclusion
[9] O.I. Aruoma, B. Halliewell, G. Williamson, in: O.L. Aruoma,
S.L. Cuppet (Eds.), Antioxidant Methodology in Vivo and in
Vitro Concepts, AOCS Press, Champaign, IL, 1997.
[10] M. Ali, A. Bagati, J. Gupta, Indian J. Chem. 34B (1995) 884.
[11] Z. Sui, R. Salto, J. Li, C. Craik, P.R.O. De Montellano, Bioorg.
Med. Chem. 1 (1993) 415.
Curcuminoids 3 and 4 were designed and syn-
thesized to selectively complex transition metal
cations. The binding force between the transition
metal cation and the central keto–enol group was
found sufficient to form stable complexes and the pre-
organized structure given by the diphenylmethane
structural element in compounds 3 and 4 does not
bring additional selectivity. A better selectivity is
related to higher organized systems as shown by
Wong et al. [43]. For that purpose, polyaza derivatives
appear more suitable than the keto–enol groups.
Nevertheless, a thorough investigation of the com-
plexation of curcumin 1 and bis-curcuminoid 3 with
Cu(II) using UV–visible absorption spectroscopy
combined with quantum molecular calculations led,
for the first time, to new important insight on the
complexation ability of curcuminoid derivatives.
Such fruitful approach might be used for others
[12] J.M. Kim, S. Araki, D.J. Kim, C.B. Park, N. Takasuka, H.
Baba-Toriyama, T. Ota, Z. Nir, F. Khachik, N. Shimidzu, Y.
Tanaka, T. Osawa, T. Uraji, M. Murakoshi, H. Nishino, H.
Tsuda, Carcinogenesis 19 (1998) 81.
[13] R.N. Sah, P.H. Brown, Microchem. J. 56 (1997) 285.
[14] G.S. Spicer, J.D.H. Strickl, J. Chem. Soc. (1952) 4644.
[15] H.H. Tonnesen, M. Masson, T. Loftsson, Intern. J. Pharm. 244
(2002) 127.
[16] B. Tang, L. Ma, H.Y. Wang, G.Y. Zhang, J. Agric. Food
Chem. 50 (2002) 1355.
[17] F. Kuhlwein, K. Polborn, W. Beck, Z. Anorg. Allg. Chem. 623
(1997) 1211.
[18] A. Arrieta, F. Dietze, G. Mann, L. Beyer, J. Hartung, J. Prakt.
Chem. 330 (1988) 111.
[19] K. Krishnankutty, P. Venugopalan, Synth. React. Inorg. Met.
Org. Chem. 28 (1998) 1313.

190
A. Sundaryono et al. / Journal of Molecular Structure 649 (2003) 177–190
[20] M. Borsari, E. Ferrari, R. Grandi, M. Saladini, Inorg. Chim.
Acta 328 (2002) 61.
[37] M.J.S. Dewar, C. Jie, G. Yu, Tetrahedron 23 (1993) 5003.
[38] R.G. Parr, W. Yang, Density-Functional Theory of Atoms and
Molecules, Oxford University Press, Oxford, 1989.
[21] A. Castellan, S. Grelier, L. Kessab, A. Nourmamode, Y.
Hannachi, J. Chem. Soc. Perkin Trans. 2 (1996) 1131.
[22] S. Guilardi-Ruggiero, A. Castellan, S. Grelier, A. Nourma-
mode, M. Cotrait, J. Mol. Struct. 484 (1999) 235.
[23] M.H. Krackov, H.E. Bellis, W.O. Patent 97 16403, 1997.
[24] P.J. Roughley, D.A. Whiting, J. Chem. Soc. Perkin Trans. 1
(1973) 2379.
[39] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A.
Robb, J.R. Cheeseman, V.G. Zakrzewski, J.A. Montgomery,
Jr., R.E. Stratmann, J.C. Burant, S. Dapprich, J.M. Millam,
A.D. Daniels, K.N. Kudin, M.C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C.
Adamo, S. Clifford, J. Ochterski, G.A. Petersson, P.Y. Ayala,
Q. Cui, K. Morokuma, D.K. Malick, A.D. Rabuck,
K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V. Ortiz,
A.G. Baboul, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz,
I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith,
M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez,
M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W.
Wong, J.L. Andres, C. Gonzalez, M. Head-Gordon, E.S.
Replogle, J.A. Pople, GAUSSIAN 98, Revision A.7. Edited by
Gaussian Inc., Pittsburgh, PA, 1998.
[25] J.D. Garber, US Patent, 2,775,613, 1956.
[26] U. Pedersen, P.B. Rasmussen, S.O. Lawesson, Liebigs Ann.
Chem. (1985) 1557.
[27] J.K. Hardy, The University of Akron, http://ull.chemistry.
uakron.edu/analytical/index.html, 2000.
[28] H.H. Tonnesen, A.F. Arrieta, D. Lerner, Pharmazie 50 (1995)
689.
[29] P.H. Bong, Bull. Kor. Chem. Soc. 21 (2000) 81.
[30] C.F. Chignell, K.J. Reszka, A.G. Motten, R.H. Sik, T.A. Dahl,
Photochem. Photobiol. 59 (1994) 295.
[40] N. Godbout, D.R. Salahub, J. Andzelm, E. Wimmer, Can.
J. Chem. 70 (1992) 560.
[31] S.M. Khopde, K.I. Priyadarsini, D.K. Palit, T. Mukherjee,
Photochem. Photobiol. 75 (2000) 625.
[41] I. Robertson, M. Truter, J. Chem. Soc. (1967) 309.
[42] Y. Ishigami, M. Goto, T. Matsuda, Y. Takizawa, S. Suzuki,
J. Jpn. Soc. Colour Mater. 72 (1999) 71.
[32] R. Toniolo, F. Di Narda, S. Susmel, M. Martelli, L. Martelli,
G. Bontempelli, Ann. Chim. 92 (2002) 281.
[33] F. Dietze, F. Arrieta, U. Zimmer, Pharmazie 52 (1997) 302.
[34] P. Umapathy, D.N. Sen, Ind. J. Chem. 2 (1964) 350.
¨
[35] B. Konig, M. Pelka, H. Zieg, T. Ritter, H. Bouas-Laurent, R.
[43] E.H. Wong, G.R. Weisman, D.C. Hill, D.P. Reed, M.E.
Rogers, J.S. Condon, M.A. Fagan, J.C. Calabrese, K.C. Lam,
I.A. Guzei, A. Rheingold, J. Am. Chem. Soc. 122 (2000)
10531.
Bonneau, J.P. Desvergne, J. Am. Chem. Soc. 121 (1999) 1681.
[36] L.G. Sillen, B. Warnquist, Ark. Kemi 31 (1968) 315.