Circular dichroism of anthocyanidin 3-glucoside self-aggregates Raquel Gavara, Vesselin Petrov, Alexandre Quintas, Fernando Pina*
The circular dichroism spectra of the six most common anthocyanidin 3-glucoside show the formation of left handed aggregates compatible with dimers. The absorption bands of the monomer split by increasing concentration according to the formation of H and J aggregates. The angle and distance between the transition moments of the two monomers in the dimer was calculated from the splitting of the 0–0 absorption band. While the angle is similar for the series the distance changes dramatically. The intensity of the CD signal is proportional to the inverse of the square of the distance.
Highlights
"The circular dichroism spectra of six common anthocyanins 3-glucosides was obtained. " Like 3,5-diglucoside analogous
Q10 they exhibit
left-handed CD signals. " J and H aggregates are formed by concentration increasing. " The distance of the transition moments correlate with the intensity of the CD signal.
2
Circular
dichroism
of
anthocyanidin
3-glucoside
self-aggregates
3Raquel
Gavara
a,
Vesselin
Petrov
a,
Alexandre
Quintas
b,
Fernando
Pina
a,⇑4 aREQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa,2829516Monte de Caparica,Portugal 5 bInstituto Superior de Ciencias da Saude Egas Moniz, Centro de Investigação Interdisciplinar Egas Moniz, P-2829511 Monte de Caparica, Caparica, Portugal 6 7 9
a r t i c l e
i n f o
10 11 12 13 14 Keywords: 15 Circulardichroism 16 Anthocyanins 17 J and H aggregates 18 1 9a b s t r a c t
20 Self-association constants for the flavylium cations of the six most common anthocyanidin 3-glucosides21 were determined by circular dichroism (CD) andUV–Visspectroscopy. Along with previous1H NMR
22 results, all measurements were consistent with amonomer–dimermodel. The CD spectra of the
antho-23 cyanidin 3-glucosides were similar to the analogues 3,5-diglucosides. All dimers of the anthocyanidin
24 3-glucosides exhibited left-handed CD signals, with petunidin-3-glucoside and myrtillin having the most
25 intense signals. In addition, the magnitude of the molar ellipticity,[h],was generally higher for the
3-glu-26 cosides than for the 3,5-diglucosides. For all six anthocyanins studied, the CD absorption spectra of their
27 dimers showed evidence of the splitting of the monomer absorption into lower (J) and higher (H) energy
28 bands. The angle and the distance between the dipolar moments of the two monomers comprising the
29 dimer were obtained from the lower energy absorption band.
30 While the angle was more or less similar in all six dimers, the separation distance between the
mono-31 mer dipole moments differed dramatically. The intensity of the CD signal displayed a linear dependence
32 with the inverse square of the dipole moment distances.
33 34 35
36 1. Introduction
37 Anthocyanins are the pigments responsible for the beautiful red
38 to blue colours of flowers and fruits. However, when their pH
39 dependent equilibrium is studied in water at low concentrations,
40 the red flavylium cation is only stable at very acidic pH values,
41 while the blue quinoidal base is a minor species (<5%) at
moder-42 ately acidic ones. Due to the fact that the pH of the vacuoles, where
43 anthocyanins are located, changes roughly from 3 to 6 (Steward
44 et al.,1975) some kind of interactions that permit to achieve colour
45 in these conditions must take place. One solution found by Nature
46 regarding the achievement of blue colour, is the formation of
47 supramolecular structures involving metals and flavonoids, which
48 are able to stabilize the blue quinoidal base (Yoshidaet al.,1995, 49 2009).
50 In the case of red colour it was reported for raspberry the
pres-51 ence of high concentrations of cyanidin 3-glucoside (kuromanin,
52 ca. 2.4mg/gof fresh fruit) and other derivatives bearing different
53 sugars in position 3 (Melo et al., 2000). Nature uses high
concentra-54 tions of the anthocyanin to compensate for the fact that is not
55 using the total colouring power of the flavylium cation (for pH 56 3.1found in raspberry extract, the mole fraction of flavylium cation
57 of the cyanidin 3-glucoside is only ca. 0.33). The use of high
58 concentration of anthocyanins (Wu et al., 2006), in particular
59
anthocyanidin 3-glucosides, raises the question of self-aggregation
60
and its influence on the colour definition.
61
In a series of papersHoshinoet al. (1982)demonstrated that
62
anthocyanidin 3,5-diglucosides self-aggregate by stacking in a
63
right-handed or left-handed screw axis. While quinoidal bases of
64
cyanin and pelargonin lead to right-handed adducts, peonin,
65
delphin and malvin form left-handedones.On the other hand, all
66
the respective flavylium cations lead to aggregates exhibiting
67
left-handed CD signals, due to a super-asymmetry imposed by
68
the oligomeric species (Rodgerand Nordén,1997).
69
In a recent paper we have carried out an extended studied of the
70
self-aggregation in the six most abundant anthocyanidin
3-gluco-71
sides (Leydet etal., in press), by means of1H NMR and UV–Vis
72
absorption. It was verified that the rate and equilibrium constants
73
of the network of chemical reactions involving the dyes is
dramat-74
ically dependent on the anthocyanin concentration. In this work
75
we report the circular dichroism spectra of these compounds,
76
Scheme 1, and correlate the magnitude of the CD signal with the
77
angle and distance between the dipolar moments of the monomers
78
in the dimer.
79
2. Experimental
80
Myrtillin chloride, oenin chloride, kuromanin chloride and
81
callistephin chloride were purchased from Extrasynthèse;
peoni-82
din-3-glucoside chloride and petunidin-3-glucoside chloride were
83
purchased from PhytoLab. All the reagents (P95%) were used ⇑Corresponding author.
E-mail address:[email protected](F. Pina).
Q2 Q1
Q4
dichroism
anthocyanidin
anthocyanidin
self-aggregates
self-aggregates
Raquel
Raquel
Vesselin
Alexandre
Fernando
2829516 Portugal
6 (
in these conditions must take place. One solution found by Nature
sugars in position 3 ( sugars in position 3 (
in raspberry extract, the mole fraction of flavylium cation
pH
in raspberry extract, the mole fraction of flavylium cation
Wu et al., 2006 Wu et al., 2006 et al. (1982) 3,5-diglucos right-hande Rodger
In a recent paper we have carried out an extended studied of the In a recent paper we have carried out an extended studied of the
al.,
It was verified that the rate and equilibrium constants
in press),
It was verified that the rate and equilibrium constants It was verified that the rate and equilibrium constants
84 without any further purification. The solutions were prepared in
85 Millipore water, and HCl 0.1 M was employed to acidify the
86 samples at pH 1.
87 UV–Visabsorption spectra were recorded on Varian Cary 100
88 Bio, Varian Cary 5000i and Jasco V-530 spectrophotometers.
89 2.1. Circular dichroism analysis
90 Myrtillin chloride, oenin chloride, kuromanin chloride,
calliste-91 phin chloride, peonidin-3-glucoside chloride and
petunidin-3-glu-92 coside solutions at pH 1 were prepared 12h before the
93 experiments and kept at 4 °C.
94 Circular dichroism were performed using near-UV and visible
95 (350–700nm) CD in a Jasco J810 spectropolarimeter equipped
96 with a temperature control unit Julabo F25 using a range of
con-97 centration of the different 3-glucoside anthocyanins from 10ÿ3to
98 10ÿ6M. Near-UV and visible CD spectra were recorded with0.01,
99 0.05,0.1 and 1 cm (linear) path length quartz cuvette at 20 °C,
100 according to the measured high tension (HT) voltage measured in
101 each sample. For each spectrum, three scans were averaged and
102 anthocyanidin 3-glucosides concentration was checked by
absor-103 bance at the maximum in the visible region using the molar
104 absorption coefficients of each flavylium cation under study
(myr-105
tillin chloride, oenin chloride, kuromanin chloride, callistephin
106
chloride, peonidin-3-glucoside chloride and petunidin-3-glucoside
107
chloride) on aUV–Visspectrophotometer Jasco V-530.
108
Molar ellipticity for each compound has been calculated using
109
theformulae:
110
½h ¼ h=ðC:l:10Þ deg cm2dmolÿ1 112112
113
where h is the ellipticity (mdeg), C the concentration (mol lÿ1) and l
114
the path length (cm).
115
CD spectra of the appropriate blanks (HCl 0.1 M) were recorded
116
and subtracted from the anthocyanidin 3-glucosides spectra.
117
3. Results anddiscussion
118
3.1. Circular dichroism
119
The circular dichroism spectra representing the ellipticity of
120
petunidin-3-glucoside at pH 1.0 is shown in Fig. 1A. The molar
121
ellipticity at the maximum of the first and second cottonbands
122
is represented inFig. 1B. The same is shown for oenin (malvidin
123
3-glucoside) inFig. 2A and B. Identical figures for the four
remain-124
ing anthocyanins of Scheme 1, are reported in Supplementary 125
material, Fig.S1–S4. At lower concentrations the molar ellipticity
O HO OH OH OGlc +2 3 4 5 6 7 8 6' 5' 4' 3' 2' Pelargonidin-3-glucoside (callistephin) O HO OH OH OGlc + Peonidin-3-glucoside O HO OH OGlc + OH Delphinidin-3-glucoside (myrtillin) OH A C B O HO OH OGlc + OCH3 OH Malvidin-3-glucoside (oenin) OCH3 O HO OH OGlc + OH O HO OH OGlc + OCH3 OH OH Cyanidin-3-glucoside (kuromanin) Petunidin-3-glucoside OH OH OCH3
Scheme 1.The six anthocyanidin 3-glucosides.
-400 -300 -200 -100 0 100 200 300 400 360 400 440 480 520 560 600 640 θ (m d e g ) Wavelength (nm) -4 104 -2 104 0 2 104 4 104 0 2 10-4 4 10-4 6 10-4 8 10-4 1 10-3 1.2 10-3 [θ ] (d e g c m 2 d m o l -1 ) Concentration (M) 539 nm 461 nm
A
B
Fig. 1.(A) Ellipticity of petunidin-3-glucoside as a function of concentration (up to 10ÿ3M); (B) molar ellipticity at the first (539 nm) and second (461 nm) cotton bands. KD= 900 Mÿ1.
(350–700
0.01,
pH 1.0
126 in the visible region is practically zero, indicating that the
127 monomerdoes notexhibit signal. Signal rises when concentration
128 increases due to the appearance of supramolecular chirality upon
129 aggregation.
130 The fittings of the molar ellipticity of all the six anthocyanins
131 were achieved using a monomer–dimer model (see below) and
132 the dimerization constants previously calculated by 1H NMR,
133 (Leydet et al.,in press).
134 As shown inFigs.1and2(as well asFig.S1–S4in Supplemen-135 tary material) the flavylium cation adducts of the six
anthocyani-136 din 3-glucosides of Scheme 1, self-associate in a left-handed
137 manner, as reported for the 3,5-diglucoside analogues (Hoshino,
138 1982). The CD strongest signal occurs for petunidin-3-glucoside
139 and delphinidin-3-glucoside (myrtillin) similarly to what was
140 previously observed for the 3,5-diglucoside analogous. However,
141 comparison with the molar ellipticity, [h], of 3,5-diglucosides
142 seems to indicate that 3-glucosides exhibit in general a higher
143 [h].This suggests that the crucial sugar to form the screw is the
144 one in position 3. The presence of a second sugar in position 5 of
145 the flavylium cation seems to have a small effect on the intensity
146 of the CD signal, and thus the screw shape and the respective
ellip-147 ticity is essentially modulated by the sugar in position 3.
148 It is worth of note the fact that the intensity of the CD signal is
149 not proportional to the value of the aggregation constant: for
150 example oenin possesses a high association constant but presents
151 a weak CD signal. Two main factors contribute to the appearance
152 of the CD signal:(i)the distance between the dipolar moments
153 of the two monomers, more properly the inverse of the square of
154 the distance;(ii)the angle formed between the dipolar moments
155
of the molecules during aggregation stacking. Additionally, the
156
CD signal is also proportional to the molar absorption coefficient
157
of the chromophore, in this case the dimer adduct (BerovaandDi 158
Bari,2007).
159
In order to normalize the effect of the chromophore absorption
160
on the CD signal, a ‘‘corrected’’ CD intensity was calculated by
con-161
sidering the sum of the molar ellipticity around the maximum of
162
the1st and2nd cotton bands (amplitude) divided by the molar
163
absorption coefficient of the dimer
e
dimer/Mÿ1cmÿ1 (see below)164
taken at the maximum (Table 1).
165
3.2.1H NMR
166
In a recent paper (Leydet et al.,in press) we reported that the1H
167
NMR peaks of the protons H4, H6and H8of the flavylium cation of
168
anthocyanidin 3-glucosides are dramatically shifted to high field
169
by increasing concentration. This is a result of the shielding of
170
these protons upon aggregation, a phenomenon previously
ob-171
served by other authors (Hoshino, 1992; Hoshino et al., 1982;
172
Houbiers et al., 1998), see Supplementary material for the 1H
173
NMR data. The difference between the chemical shift of the dimer
174
and monomer,Table 2, suggests that protons H4and H6are more
175
affected than proton H8by the interaction in the aggregate, and
176
this should be an important clue for the determination of the dimer
177
structure.
178
The1H NMR data of the six anthocyanidin 3-glucosides (Leydet
179
et al.,in press) was treated with amonomer–dimerand isodesmic
180
models (Dimicoli and Hélène, 1973) the last one considering the
181
existence of higher order aggregates with the same association -250 -200 -150 -100 -50 0 50 100 150 360 400 440 480 520 560 600 640 θ (m d e g ) Wavelength (nm) -1.5 104 -1 104 -5000 0 5000 1 104 0 4 10-4 8 10-4 1.2 10-3 1.6 10-3 Concentration (M) [θ ] (deg cm 2 dmol -1 ) 492 nm 550 nm
A
B
Fig. 2.(A) Ellipticity of oenin as a function of concentration (up to 1.5 10ÿ3M); (B) Molar ellipticity at the first (550 nm) and second (492 nm) cotton bands. K
D= 976 Mÿ1.
Table 1
Ellipticity data for the six anthocyanidin 3-glucosides.a
Anthocyanin [h]1st Cotton (deg cm2dmolÿ1) [h] 2nd Cotton (deg cm2dmolÿ1) e
dimer/Mÿ1cmÿ1(kmax) Corrected CD intensityb
Myrtillin ÿ76,610 (539 nm) 57,870 (452 nm) 28,600 (483 nm) 2.8 Oenin ÿ29,550 (550 nm) 16,830 (492 nm) 40,300 (514 nm) 3.3 Petunidin ÿ94,740 (539 nm) 69,820 (461 nm) 41,700 (497 nm) 1.1 Kuromanin ÿ55,130 (539 nm) 25,000 (452 nm) 22,700 (503 nm) 1.7 Peonidin ÿ26,300 (541 nm) 13,310 (466 nm) 41,100 (503 nm) 4.0 Callistephin ÿ19,080 (529 nm) 13,610 (414 nm) 56,220 (496 nm) 0.58
aObtained form the fitting ofFig. 1B and B and the equivalent figures in supplementary material.
bCalculated as the sum of the maximum of the absolute value of the 1st and 2nd Cotton bands divided by the mole absorption coefficient of the dimer, see below.
Q5 in press). As shown in in press). and Figs.
) the flavylium cation adducts of the six ) the flavylium cation adducts of the six anthocyani- Supplemen-) the flavylium cation adducts of the six
anthocyani-1982).
and delphinidi
[
to indicate that 3-glucosides exhibit in general a higher one
].
one
of the two monomers, more properly the inverse of the square of of the two monomers, more properly the inverse of the square of
absorption coefficien in press and between Supplement between models models 19731973
182 constant. In both cases good fittings have been obtained and the1H
183 NMR data was not able to discriminate between the two models.
184 3.3.Monomer–dimermodel
185 Similarly to the data of the1H NMR, the circular dichroism
spec-186 tra presented in this work was also treated with amonomer–dimer 187 model, as follows:
188 Themonomer (M)–dimer(D) equilibrium is described by Eq.
189 (1):
190
M þ M $ D ð1Þ
192 192
193 and the dimeric constant (KD) is defined byEq.(2):
194 KD¼ ½Deq: ð½Meq:Þ2¼
v
D eq: 2C0 eq: ðv
Meq:Þ 2 ð2Þ 196 196197 Considering the mass balance of the system given byEq.(3):
198
½Meq:þ 2½Deq:¼ C0eq: ð3Þ
200 200 201
v
M eq:þv
Deq:¼ 1 203 203204 where the molar fractions (
v
) are defined according theEqs.(4) and205 (5): 206
v
M eq:¼ ½Meq: C0eq: ¼ ÿ1 þ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1 þ 8C0eq:Kd q 4KdC0eq: ð4Þ 208 208 209v
D eq:¼ 2½Deq: C0eq: ð5Þ 211 211212 Finally, the molar ellipticity of the system is defined as:
213
½h ¼ ½hM
v
Mþ ½hDv
D ð6Þ215 215
216 where [h] is the1stor2nd cottonbands signal [h]Mthe CD signal of
217 the monomer (0) and [h]Dthe molar ellipticity of the dimer.
218 Fitting was achieved for the same set of constants calculated on
219 the basis of the1H NMR chemical shifts (Leydet et al.,in press), see
220 Table 3. However, good fittings are also obtained using the
isodes-221 mic model which implies the formation ofn-Merswith identical
222 successive association constants. Consequently, it cannot be
223 excluded the formation of higher order aggregates, but the
224 accuracy of the data does not allow discrimination between mono-225 mer–dimerand alternative models. It is worth to comment that
226 this conclusion cannot be extended, in principle, to the
self-aggre-227 gation of the quinoidal base, where indirect evidence to the
228 formation of higher order aggregates was obtained (Hoshino
229 et al.,1981a,b). Regarding to the flavylium cation it is possible that
230 its positive charge hinders the association of the dimer with
addi-231 tional monomers, because of the growing electrostatic repulsion.
232 The following discussion will be carried out using the
233 monomer–dimermodel.
234
3.4.UV–Visabsorption
235
The effect of the aggregation in theUV–Visabsorption spectra is
236
clearly observed when the molar absorption coefficients
e
(Mÿ1-237
cmÿ1), are represented as a function of the anthocyanin
concentra-238
tion, seeFig. 3A forpetunidin-3-O-glucosideandFig. 3B for oenin.
239
Inspection ofFig. 3shows the splitting of the monomer
absorp-240
tion in two new red and blue shifted absorptions upon aggregation,
241
see below. The individual spectra of the monomer and dimer in the
242
visible region were obtained as reported previously (Cruzet al., 243 2010). Thus, a set of absorption spectra at several concentrations
244
(the same range of concentrations employed in the CD
measure-245
ments) was treated by FiNAl algorithm (Antonov and Petrov,
246
2002; Antonov et al., 1999). Taking into account the dimerization
247
constant, this algorithm recovers the individual spectrum of the
248
monomer and dimer and gives their contribution (mole fraction)
249
for each concentration. The monomer and dimer absorption bands
250
were additionally deconvoluted using Gaussian functions, see
251
Fig. 4for oenin as an illustrative example.
252
The absorption spectrum of the monomer,Fig. 4A, shows the
253
existence of two transitions, the one of lower energy exhibiting
254
two vibrational bands (for example, in the case of oenin, the
dis-255
tance between the vibration maxima (532 and 505 nm)
256
corresponds to 1005 cmÿ1). Decomposition of the dimer of each
257
anthocyanin was achieved using six bands, Fig. 4B. Each of the
258
monomer absorption bands splits into new bands, corresponding
259
to H and J aggregates (Spano, 2010). 1The situation is illustrated
260
inScheme 2.
261
In the present paper calculations of the distance and angle were
262
performed on the 0–0 absorption band (532 nm for oenin in
263
Fig. 4A), following the Kasha theory (Kasha et al., 1965). It is
impor-264
tant to stress that the following calculations do not predict the
265
relative position of the two monomers in the dimer, but the
dis-266
tance and angle between the respective transition dipole moments.
267
Moreover, the position of the transition moment of the monomers
Table 2
Variation of the chemical shifts from the dimer to the monomer obtained by1H NMR (Leydet et al., 2011). Anthocyanin H4 H6 H8 Myrtillin 1.06 0.96 0.54 Oenin 0.73 0.78 0.49 Petunidin-3-glucoside 1.27 1.2 0.69 Kuromanin 0.75 0.7 0.4 Peonidin-3-glucoside 0.89 0.75 0.51 Callistephin 0.47 0.44 0.31 Table 3
Angle and distance between the dipole moments of monomers in dimers (calculated from the 0–0 electronic absorption band).f
Anthocyanin KD (Mÿ1)a 0–0 Band kM (nm)b kJ(nm)c kH(nm)d Angle e (°) Distancee (Å) Myrtillin 1240 529 572 497 ÿ100 2.9 Oenin 976 532 562 519 ÿ128 5.2 Petunidin-3-glucoside 900 531 569 498 ÿ97 2.5e Kuromanin 700 525 558 495 ÿ103 3.2 Peonidin-3-glucoside 661 526 557 495 ÿ114 3.2 Callistephin 404 508 529 475 ÿ110 4.5 aSame constants of the1H NMR studies (Leydet et al., in press). b Absorption maximum of the monomer.
c Absorption maximum of the J-band. d Absorption maximum of the H-band.
eEstimated error in the angle ± 10°; distance ± 0.2 Å: on the basis of the equa-tions involved, the distance error increases when the angle approaches 90° and this value could be underestimated.
f The sign of the anthocyanidin 3-glucoside self-aggregates overlay angle is negative because the front molecule covers the rear molecule in an anti-clockwise turn around an axis coinciding with the distance of the molecules (Klyne and Prelog, 1960).
1 H-aggregates are obtained when dye molecules aggregate following a sandwich-type arrangement and J aggregates when the monomer associates in a head-to-tail arrangement. H aggregates lead to a bathochromic shift of the absorption band of the dye, whereas J aggregates originate a hypsochromic shift.
Monomer–d monomer–dimer monomer Eq. Eq. Eq. Eqs. 0) and [ 0) and [ in press
. However, good fittings are also obtained using the isodes-s. Consequently, it cannot be s. Consequently, it cannot be
e models. It is worth to comment that e models. It is worth to comment that
positive positive petunidin-3-al., achieved Spano, the the
268 in each anthocyanin could be slightly different, due to the different
269 substitution pattern.
270 3.5. Calculations
271 The two H and J absorption bands of the dimer were obtained
272 by spectral decomposition and the angle and distance between
273 the transition moments can be interpreted through the exciton
274 model usingEqs.(7)and(8), where the angle(
a
)between thetran-275 sitions moments is defined by:
276
a
¼ 2: arctan ffiffiffiffiffiffiffiffiffiffiffiffim
H fJm
J fH s ð7Þ 278 278279 and the distance (R in Å) between the monomers in the dimer is
280 defined by the expression:
281 R ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2:14 1010 cos
a
fMD
m
m
M 3 s ð8Þ 283 283284
m
xis position of the band X in cmÿ1fXis oscillator straight defined as285 fX¼ 1:3 10ÿ8HðnÞ
R1
ÿ1
e
ðm
Þdm
,HðnÞ ¼ðn29:nþ2Þ2 is factor of theenvi-286
ronment, n is refraction index,D
m
~is the difference of the positions287
of the split bands(JandH)in cmÿ1
288
The control of the spectral decomposition can be made by the
289
coincidence, within experimental error, of the energy of the
290
absorption maximum of the monomer and the semi sum of the
291
identical energies of the H and J bands.
292
InTable 3, the results for the six anthocyanidin 3-glucosides are
293 shown. 0 1.5 104 3 104 200 300 400 500 600 700 ε (M -1 cm -1 ) Wavelength (nm) 0 1.5 104 3 104 200 300 400 500 600 700 ε (M -1 cm -1 ) Wavelength (nm)
A
B
Fig. 3.Representation of the molar absorption coefficient,e(Mÿ1cmÿ1), as a function of the anthocyanin concentration: (A) petunidin-3-glucoside 2.0 10ÿ5M (full line) and 7.0 10ÿ4M (pointed line); (B) the same for oenin 2.0 10ÿ5M (full line) and 9.0 10ÿ4M (pointed line).
Fig. 4.(A) Spectral decomposition of the monomer of oenin. Vibrational bands of the first absorption 0–0 (full line) and 0–1 (traced line); second singlet (pointed line): (B) spectral decomposition of the three bands upon aggregation (dimer). Each band of the monomer splits in two bands corresponding to J and H aggregates.
Scheme 2.Splitting of the lower energy band of the monomer upon aggregation in a dimer, including the vibronic bands.
moments
Eqs.
moments is defined by:
(8)
is defined by:
(
H Hð
of the spectral decomposition can be made by the
(J
of the spectral decomposition can be made by the of the spectral decomposition can be made by the
H)
of the spectral decomposition can be made by the of the spectral decomposition can be made by the of the spectral decomposition can be made by the
294 Inspection ofTable 3suggests that in the case of the0–0band
295 the angle between the transition moments of the monomers in
296 the dimer are similar and thus the distance should be the crucial
297 parameter to account for the CD signal. According to the theory,
298 the intensity of the CD signal is proportional to the inverse of the
299 square of the distance (Berova and Di Bari, 2007) and such a
300 relation was observed, as shown in Fig. 5A.
301 Calculations based on the exciton model have been questioned,
302 in particular when distances lower than Van Der Waals (3.5 Å) are
303 obtained (Arik and Onganer, 2003).2 However, the existence of
304 hydrogen bonds could lead to shorter values. In addition,
anthocya-305 nins are not planar molecules and by consequence their dipole
mo-306 ments are predicted to be out of the plane formed by ring A and C. It
307 is possible to conceive geometries where the two dipoles are closer
308 than the atoms of the whole molecule, if the dimers interact through
309 the same face.
310 InFig. 5B the plot of the difference between the chemical shift
311 of the monomer and dimer for protons H4, H6and H8(Table 2) as
312 a function of the corrected CD signal suggests, in spite of the great
313 dispersion, that the increasing of the CD signal and the magnitude
314 of the chemical shifts follow the same trend.
315 4. Conclusion
316 The self-aggregation of the flavylium cation of the six most
317 common anthocyanidin 3-glucosides is responsible for the
chemi-318 cal shift variations observed in the1H NMR spectra, appearance of
319 a CD signal characteristic of a left-handed aggregation and split of
320 the UV–Vis absorption band in two other bands (red and blue
321 shifted). In terms of the consequences on the colour expression
322 in plants, the splitting of the absorption bands upon aggregation
323 leads to a slight modification of the colour tonality. Moreover,
324 aggregation stabilizes the colour arising from the flavylium cation
325 because the nucleophilic attack of water occurs at higher pH
326 values. The three techniques (NMR, CD andUV–Vis)lead to
self-327 consistent results and are compatible with a monomer–dimer 328 aggregation type for the flavylium cation. One possible explanation
329 for the apparent lack of high order aggregates is the positive charge
330 of the flavylium cation, which prevents the addition of more
mono-331 mers to the dimer, due to the increasing electrostatic repulsion.
332
5. Uncited references
333
Goto etal. (1976),Hoshino(1991),Hoshino and Goto (1990) 334
andHoshinoetal. (1980).
335 Appendix A. Supplementary data
336
Supplementary data associated with this article can be found, in
337
the online version, athttp://dx.doi.org/10.1016/j.phytochem.2012.
338
12.011.
339
References
340 Antonov, L., Petrov, V., 2002. Quantitative analysis of undefined mixtures – ‘‘fishing
341 net’’ algorithm. Anal. Bioanal. Chem. 374, 1312–1317.
342 Antonov, L., Gergov, G., Petrov, V., Kubista, M., Nygren, J., 1999. UV–Vis
343 spectroscopic and chemometric study on the aggregation of ionic dyes in
344 water. Talanta 49, 99–106.
345 Arik, M., Onganer, Y., 2003. Molecular excitons of pyronin B and pyronin Y in
346 colloidal silica suspension. Chem. Phys. Lett. 375, 126–133.
347 Berova, N., Di Bari, L., 2007. Application of electronic circular dichroism in
348 configurational and conformational analysis of organic compounds. Chem.
349 Soc. Rev. 36, 914–931.
350 Cruz, L., Petrov, V., Teixeira, N., Mateus, N., Pina, F., Freitas, V., 2010. Establishment
351 of the chemical equilibria of different types of pyranoanthocyanins in aqueous
352 solutions. Evidence for the formation of aggregation in
pyranomalvidin-3-O-353 coumaroylglucoside-(+)-catechin. J. Phys. Chem. B 114, 13232–13240.
354 Dimicoli, J.-L., Hélène, C., 1973. Complex-formation between purine and
indole-355 derivatives in aqueous-solutions-proton magnetic-resonance studies. J. Am.
356 Chem. Soc. 95, 1036–1044.
357 Goto, T., Hoshino, T., Ohba, M., 1976. Stabilization effect of natural salts on
358 anthocyanins: flavylium salts, anhydrobases and genuine anthocyanins. Agric.
359 Biol. Chem. 40, 1593–1596.
360 Hoshino, T., 1991. Anthocyanin self-aggregates. 6. An approximation estimate of
361 self-association constants and the self stacking of malvin quinonoidal bases
362 studied by H-1-NMR. Phytochemistry 30, 2049–2055.
363 Hoshino, T., Goto, T., 1990. Effects of pH and concentration on the self-association of
364 malvidin quinonoidal base – electronic and circular dichroic studies.
365 Tetrahedron Lett. 31, 1593–1596.
366 Hoshino, T., Matsumoto, U., Goto, T., 1980. Evidence for self-association of
367 anthocyanins. 1. Circular dichroism of cyanin anhydrobase. Tetrahedron Lett.
368 21, 1751–1754.
369 Hoshino, T., Matsumoto, U., Goto, T., 1981a. Evidence for self-association of
370 anthocyanins. 2. Self-association of some anthocyanins in neutral
aqueous-371 solution. Phytochemistry 120, 1971–1976.
372 Hoshino, T., Matsumoto, U., Harada, N., Goto, T., 1981b. Chiral excitation couple
373 stacking of anthocyanins. Interpretation of the origin of anomalous CD induced
374 by anthocyanin association. Tetrahedron Lett. 22, 3621–3624.
375 Hoshino, T., Matsumoto, U., Goto, T., 1982. Evidence for self-association of
376 anthocyanins. 4. PMR spectroscopic evidence for the vertical stacking of
377 anthocyanin molecules. Tetrahedron Lett. 23, 433–436.
0 2 4 0 0 .1 0.2 Corrected CD signal 1 /d 2 (Å -2 ) 0 0.4 0.8 1.2 0 1 2 3 4 5 ∆ 1H NMR shift (ppm) Corrected CD signal
A
B
Fig. 5.(A) Representation of the inverse of the square of the distance as a function of the corrected CD signal leads to a relatively good linear relation; (B) representation of the difference between the chemical shifts of the monomer and dimer for protons H4 (d), H6 (h) and H8 (s) as a function of the corrected CD signal.
2We thank the referee who raised this question.
Q6 Onganer, shifted). with UV–Vis) with possible possible al. (1991), Goto al.
378 Houbiers, C., Lima, J.C., Maçanita, A.L., Santos, H., 1998. Color stabilization of 379 malvidin 3-glucoside: self-aggregation of the flavylium cation and 380 copigmentation with the Z-chalcone form. J. Phys. Chem. B 102, 3578–3585. 381 Kasha, M., Rawls, H.R., Ashraf El-Bayoumi, M., 1965. The exciton model in molecular 382 spectroscopy. Pure Appl. Chem. 11, 371–392.
383 Klyne, W., Prelog, V., 1960. Description of steric relationships across single bonds. 384 Experientia 16, 521–523.
385 Leydet, Y., Gavara, R., Petrov, V., Diniz, A. M., Parola, A. J., Lima, J. C., Pina, F., in press. 386 The effect of self-aggregation on the determination of the kinetic and 387 thermodynamic constants of the network of chemical reactions in 3-glucoside 388 anthocyanins. Phytochemistry.
389 Melo, M.J., Moncada, M., Pina, F., 2000. On the red colour of Raspberry (Rubus 390 idaeus). Tetrahedron Lett. 41, 1987–1991.
391 Rodger, A., Nordén, B., 1997. Circular Dichroism and Linear Dichroism. Oxford 392 University Press, Oxford.
393 Spano, F.C., 2010. The spectral signatures of Frenkel Polarons in H- and J-aggregates.
394 Acc. Chem. Res. 43, 429–439 (And references therein).
395 Steward, R.N., Norris, K.H., Asen, S., 1975. Microspectrophotometric measurements
396 of pH and pH effect on color of petal epidermal-cells. Phythochemistry 14, 937–
397 942.
398 Wu, X., Beecher, G.R., Holden, J.M., Haytowitz, D.B., Gebhart, S.E., Prior, R.L., 2006.
399 Concentrations of anthocyanins in common foods in the United States and
400 estimation of normal consumption. Agric. Food Chem. 54, 4069–4075.
401 Yoshida, K., Kondo, T., Okazaki, Y., Katou, K., 1995. Cause of blue petal color. Nature
402 373, 291.
403 Yoshida, K., Mori, M., Kondo, T., 2009. Blue flower color development by
404 anthocyanins: from chemical structure to cell physiology. Nat. Prod. Rep. 26,
405 884–915.
406
