Synthesis and equilibrium multistate of new pyrano-3-deoxyanthocyanin-type pigments in aqueous solutions

Article abstract of DOI:10.1016/j.tet.2017.08.051

Pyrano-flavylium compounds are flavylium derivatives with occurrence in processed foodstuffs such as wine and red fruit juices. In this work, the chemical equilibria of “bio-inspired” pyrano-3-deoxyanthocyanin dyes (4–8), presenting different substituent groups in D ring (10-methyl, 10-catechol, 10-dimethylaminophenyl, 10-carboxy) as well as in B ring (3′,4′-dihydroxy, 4′-carboxy), were studied for the first time in aqueous solutions by UV–Vis spectroscopy. The presence of a methyl group at C-10 (pigment 4) seems to stabilize the flavylium cation (AH⁺) shifting the deprotonation of 7-OH to a higher pK_a1 (pK_a1 = 5.0 ± 0.1). On the other hand, the quinoidal bases (A, A^?) were more stabilized for the pigments which undergo the first acid-base equilibrium at C-10 substituent, as could be observed by the pK_a2 constants (pK_a2 = 8.4 ± 0.1 and pK_a2 = 8.1 ± 0.1, pigments 6 and 8, respectively).

Full text of DOI:10.1016/j.tet.2017.08.051

Tetrahedron 73 (2017) 6021e6030
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Synthesis and equilibrium multistate of new
pyrano-3-deoxyanthocyanin-type pigments in aqueous solutions
a
a
a
b
a
^
Joana L.C. Sousa , Vania Gomes , Nuno Mateus , Fernando Pina , Victor de Freitas ,
,
*
Luís Cruz a
a
^
REQUIMTE/LAQV, Departamento de Química e Bioquímica, Faculdade de Ciencias, Universidade do Porto, Rua do Campo Alegre, 687, 4169-007 Porto,
Portugal
b
^
LAQV, REQUIMTE, Departamento de Química, Faculdade de Ciencias e Tecnologia, Universidade NOVA de Lisboa, 2829-516 Caparica, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history:
Pyrano-flavylium compounds are flavylium derivatives with occurrence in processed foodstuffs such as
wine and red fruit juices. In this work, the chemical equilibria of “bio-inspired” pyrano-3-
deoxyanthocyanin dyes (4e8), presenting different substituent groups in D ring (10-methyl, 10-
catechol, 10-dimethylaminophenyl, 10-carboxy) as well as in B ring (30,40-dihydroxy, 40-carboxy), were
studied for the first time in aqueous solutions by UVeVis spectroscopy. The presence of a methyl group at
C-10 (pigment 4) seems to stabilize the flavylium cation (AHþ) shifting the deprotonation of 7-OH to a
higher pKa1 (pKa1 ¼ 5.0 0.1). On the other hand, the quinoidal bases (A, Aꢀ) were more stabilized for the
pigments which undergo the first acid-base equilibrium at C-10 substituent, as could be observed by the
pKa2 constants (pKa2 ¼ 8.4 0.1 and pKa2 ¼ 8.1 0.1, pigments 6 and 8, respectively).
© 2017 Elsevier Ltd. All rights reserved.
Received 7 June 2017
Received in revised form
23 August 2017
Accepted 28 August 2017
Available online 1 September 2017
Keywords:
Pyrano-3-deoxyanthocyanins
Dyes
Chemical equilibria
UVeVis spectroscopy
pKa
1. Introduction
between C-4 and 5-OH of the flavylium cation and a double bond
from other molecule such as pyruvic acid,9 vinylcatechin10 and
Flavylium compounds represent a family of natural pigments
present in the plant kingdom, which include anthocyanins, 3-
deoxyanthocyanins and their metabolites.1e5 These compounds
are responsible for the coloration of many flowers and fruits (from
yellow to blue) and can be found in processed beverages such as red
wine. It is well known that the color of anthocyanins is greatly
vinylcatechol,11,12 resulting in the formation of an additional pyran
D ring. Over the last years, novel classes of pyrano-flavylium
compounds such as portisins (blue pigments),13 pyr-
anoanthocyanin dimers14 and pyrano-3-deoxyanthocyanins,15e17
have been reported as well as important insights about their for-
mation mechanism, physical-chemical and antioxidant proper-
ties.10,18,19 Furthermore, pyranoanthocyanins have been shown to
display higher color intensity and stability at a wider pH range
comparatively to their anthocyanin precursors.19e21
affected by pH. The sequence of chemical reactions was established
6e8
by Brouillard et al.
using temperature and pH jumps experi-
ments: at very acidic pH, the red flavylium cation (AHþ) is the
predominant species; when the pH is raised, the flavylium cation
immediately undergoes proton transfer reaction, giving rise to the
purple quinoidal base (A), and simultaneously, but more slowly, the
flavylium cation leads to the formation of the colorless hemiketal
(B) through a hydration reaction. The hemiketal further undergoes
a tautomerization reaction yielding the pale yellow cis-chalcone
(Cc) which isomerizes slowly to trans-chalcone (Ct) (Fig. 1).
This kind of pigments could undergo cycloaddition reactions
In fact, while anthocyanins practically lose their red color be-
tween pH 1 and 5, as a result of the formation of colorless hemiketal
B form (pKh 2e3), pyranoanthocyanins practically do not change
their color intensity.22 Equilibria network characterization of an-
thocyanins and several pyrano-anthocyanins are well documented
in the literature. On the other hand, the kinetic, thermodynamic
and photochromic features of oaklins (a family of 3-
deoxyanthocyanins present in oak aged red wines) were fully
studied.23
Conversely,
in
which
regards
pyrano-3-
deoxyanthocyanins, there is a lack of studies on their physical-
chemical parameters.
* Corresponding author.
E-mail address: luis.cruz@fc.up.pt (L. Cruz).
The achievement of the correct chemical equilibrium
http://dx.doi.org/10.1016/j.tet.2017.08.051
0040-4020/© 2017 Elsevier Ltd. All rights reserved.

6022
J.L.C. Sousa et al. / Tetrahedron 73 (2017) 6021e6030
Fig. 1. Equilibrium forms of natural anthocyanins in aqueous solutions.
Table 1
characterization of this kind of pigments in solution in a large pH
range will allow to know the stability of their species for further
potential applications such as photochromic and pH-sensors for
food packaging, dyes for hair and textile, and photosensitizers for
solar cells applications, among others.
The influence of different substituents in C-10 and B ring of five
“bio-inspired” pyrano-3-deoxyanthocyanin-type dyes (Scheme 1)
in their chemical equilibria was studied by UVeVis spectroscopy
and also compared with the natural pyrano-anthocyanin analogues
(with glucose moiety).
Mass and spectroscopic data of the synthesized pyrano-3-deoxyanthocyanins.
Pigments
lmax (nm)a
[M]þ m/z
PyLUT-10-Me 4
PyLUT-10-Catechol 5
PyLUT-10-N(Me)2Ph 6
460
496
550
492
439
309
403
414
339
351
PyLUT-10-Carboxy 7
Py-(40,10)-Dicarboxy-7-hydroxyflavylium 8
LUT: Luteolinidin.
a
Recorded from the LC-DAD.
2. Results and discussion
1H and 13C NMR characterization confirmed the structures of the
pyrano-3-deoxyanthocyanins 4e8.
2.1. Synthesis of the pyrano-3-deoxyanthocyanin pigments
Five different pyrano-3-deoxyanthocyanin-type pigments 4e8
were synthesized according to the conditions described in Scheme
1.
Firstly, 3-deoxyanthocyanins 3a,b were prepared through an
acidic aldol condensation between 2,4,6-trihydroxybenzaldehyde 1
and acetophenone derivatives 2a,b.23,24 The next step consisted in
the cycloaddition reactions between 3-deoxyanthocyanins and
cinnamic acids (or oxaloacetic acid in the case of pigments 7 and 8).
After the purification steps, the pigments were individually
analyzed by LC-DAD/ESI-MS in the positive ion mode (Table 1). Full
2.2. Thermodynamic equilibrium
In general, flavylium compounds display the same chemical
equilibrium network: flavylium cation (AHþ) is thermodynamically
stable at very acidic pH values (pH ¼ 1), being the dominant species
of the equilibrium; when pH increases, the flavylium cation could
be involved in two parallel reactions: i) acid-base reaction to form
the neutral quinonoidal base (A) and ii) hydration reaction to give
the colorless hemiketal (B). Proton transfer reactions take place
immediately after the pH jump in a micro-seconds time scale while
Scheme 1. Synthesis strategy to obtain the pyrano-3-deoxyanthocyanin pigments 4e8. Reagents and conditions: (3a) TMS-Cl (20 equiv.), AcOEt/MeOH (2:1), 0 ꢁC, 60 min, 72%; (3b)
AcOH/H2SO4 (4:1), RT, 24 h, 59%.

J.L.C. Sousa et al. / Tetrahedron 73 (2017) 6021e6030
6023
the hydration reaction is much slower (sub-seconds to seconds). On
ꢀ
ꢁ
ꢀ
ꢁ ꢀ
ꢁ
ꢀ
ꢁ ꢀ
ꢁ
½Aꢂ$ H3Oþ
Aꢀ $ H3Oþ
A2ꢀ $ H3Oþ
the other hand, at higher pH values (moderately acidic), A is not the
most stable species at the equilibrium but rather B, which means
that A is formed during the first instants of the process and dis-
appears partially later to give the equilibrium distribution at these
pH values, including also minor fractions of Cc and Ct (Fig. 1).
On the other hand, pyrano-flavyliuns such as pyrano-
anthocyanins display higher color intensity and stability at a
wider pH range comparatively to their anthocyanin precursors,
because the former do not undergo hydration reaction.19 It is
established that the substitution at C-4 should avoid the nucleo-
philic attack of water at position 2. Therefore, this means that only
ꢀ
ꢁ
ꢀ
ꢁ
Ka1
¼
Ka2
¼
Ka3
¼
AHþ
Aꢀ
½Aꢂ
(4)
The total concentration of individual species, C0, is given by
equation (5):
h
i
h
i
h
i
C0 ¼ AHþ þ ½Aꢂ þ Aꢀ þ A2ꢀ
(5)
and the mole fraction distribution of the different species, ci, is
calculated as follows:
ꢀ
ꢁ
ꢀ
ꢁ3
AHþ
C0
H3Oþ
cAHþ
¼
¼
(6)
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
H3Oþ 3 þ Ka1$ H3Oþ 2 þ Ka1$Ka2$ H3Oþ þ Ka1$Ka2$Ka3
ꢀ
ꢁ2
K
a1$ H3Oþ
½Aꢂ
cA
¼
¼
(7)
(8)
(9)
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
H3Oþ 3 þ Ka1$ H3Oþ 2 þ Ka1$Ka2$ H3Oþ þ Ka1$Ka2$Ka3
C0
ꢀ
ꢁ
ꢀ
ꢁ
Aꢀ
C0
K
a1$Ka2$ H3Oþ
cAꢀ
¼
¼
¼
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
H3Oþ 3 þ Ka1$ H3Oþ 2 þ Ka1$Ka2$ H3Oþ þ Ka1$Ka2$Ka3
ꢀ
ꢁ
A2
C0
-
K
a1$Ka2$Ka3
cA2
¼
-
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
H3Oþ 3 þ Ka1$ H3Oþ 2 þ Ka1$Ka2$ H3Oþ þ Ka1$Ka2$Ka3
acid-base reactions should take place, as recently reported for other
pyranoanthocyanin-type compounds by UVeVis and NMR
studies.25e27
These equations could be simplified to the following ones:
cAHþ
(6.1)
1
¼
K
K
$K
K
$K $K
a1
a2 a1 a3 a2 a1
1þ
þ
þ
þ
2
3
H
½
þ
þ
ꢂ
H
½
H
½
ꢂ
ꢂ
cA¼
(7.1)
(8.1)
(9.1)
K
$c
AH
þ
2.3. Determination of pKa values
a1
½
þ
H
ꢂ
In order to determine the pKa values of all pigments, titrations
were performed in water/ethanol (75:25) mixtures. The titration
was fast enough to avoid significant precipitation (which is a
relatively slow process). The UVeVis spectrum of each pigment was
obtained for the pH range between 1 and 12. However, the study at
the basic region (pH > 9) was limited by precipitation, which was
visible by the typical increasing of the absorption due to the scat-
tered light.
cAꢀ
K
$
c
a2
A
¼
þ
H
½
ꢂ
cA2ꢀ
K
$c
a3 A
ꢀ
¼
þ
H
½
ꢂ
The absorbance at a given wavelength shows a pH dependence
which is given by the following equation:
For all compounds there is no apparent hydration, which is in
agreement
with
the
available
data
for
other
Acalc ¼ εAHþ
$
cAHþ þ εA$cA þ εAꢀ
$
cAꢀ þ εA2ꢀ
$
cA2ꢀ
(10)
pyranoanthocyanins.19,25e27 For that reason, only acid-base equi-
libria were considered. On this basis, the flavylium cation can un-
dergo protonation/deprotonation reactions. The global process can
be accounted by equations (1)e(3):
where εAHþ ; εA; εAꢀ ; εA2ꢀ are the mole absorption coefficients of the
individual species or components at the considered wavelength.
The pKa values were determined by fitting the calculated
absorbance values by equation (10) to the ones obtained experi-
mentally. Solver function was then used to minimize the sum of the
difference between theoretical and experimental values squared.
Determination of the acidity constants was not obvious, because
the spectral variations in some cases are very small. In general, the
acidity constants are well separated in pH units and it is possible to
identify isosbestic points. Eventual solutions of the fitting, which
were not coherent with the pH interval given by the isosbestic
points, were discarded.
AHþ þ H2O4A þ H3Oþ Ka1 First proton transfer
A þ H2O4Aꢀ þ H3Oþ Ka2 Second proton transfer
(1)
(2)
(3)
A
ꢀ þ H2O4A2ꢀ þ H3Oþ Ka3 Third proton transfer
where

6024
J.L.C. Sousa et al. / Tetrahedron 73 (2017) 6021e6030
Fig. 2. (a) Spectral variations as a function of pH of compound 4 in the pH range 1.58 < pH < 5.86; (b) the same for 5.86 < pH < 9.05; (c) Fitting was achieved for pKa1 ¼ 5.0 0.1;
pKa2 ¼ 6.5 0.1; (d) chemical equilibria of compound 4 in water/ethanol (75:25) (v/v) between pH 1 and 9.
2.3.1. Compound 4
for a similar compound.26
Fig. 2 (a) and (b) show the evolution of the UVeVis spectra of
solutions of pigment 4 at different pH values measured immedi-
ately after a pH change. Similarly to the pyranoanthocyanins ana-
logues,19 no evidence of hydration process was observed after the
pH increase, up to slightly basic solutions. In particular, no evidence
of the formation of B hemiketal (appearance of a band with ab-
sorption at 280 nm) is given in the recorded UVeVis spectra. Such
behavior was observed for all compounds studied through this
work. As the pH increases, two sets of spectra exhibiting isosbestic
points are observed. The presence of isosbestic points in different
pH ranges are detected when the pKa's are sufficiently separated
and is a further confirmation of the goodness of the fittings shown
in Fig. 2 (c). The first set of spectra exhibiting isosbestic points, Fig. 2
(a), is attributed to the equilibrium between AHþ and A, and the
second set, Fig. 2 (b) to the equilibrium between A and Aꢀ species,
as summarized in Fig. 2 (d).
In Fig. 2 (c), the molar fractions of each species and the exper-
imental absorbance values at three chosen wavelengths are plotted
for pigment 4 as a function of pH, together with the respective
fittings. The first pKa should be assigned to the deprotonation of 7-
OH group and the second one to the deprotonation of the 40-hy-
droxyl substituent in B ring as it is described in the literature for
7,40-dihydroxyflavylium analogue28e31 [Fig. 2 (d)]. Further de-
terminations were not possible because of the precipitation of the
compound. Moreover, for pH higher than 11, a degradation
occurred with a yellow color appearance. This could be explained
by the methyl group deprotonation as it had been already reported
2.3.2. Compound 5
For compound 5, there are four potential positions where can
occur deprotonation. Once again, the precipitation prevented de-
terminations at higher pH values. However, the UVeVis spectral
variations only gave two pKa's: pKa1 and pKa2 correspond to
deprotonations of 7-OH (4.7 0.1) and 40-OH (7.9 0.1), respec-
tively. The third pKa should occur at pH higher than 10 and would
take place at the p-hydroxyl group of the catechol moiety attached
at C-10. These data are in agreement with the ones reported in the
literature for glycosylated analogues namely pyranomalvidin de-
rivatives with catechol substituents.19 (Fig. 3).
2.3.3. Compound 6
For compound 6, the UVeVis spectral variations allowed us to
determine three equilibrium constants. The first one is compatible
with the protonation of the flavylium cation AHþ in the amine
group when increasing concentrations of HCl were added (from 0.1
to 6 M) to the water/ethanol 75:25 (v/v) solution, yielding a dica-
tionic species (AH22þ). The resulting spectra indicated an acid-base
equilibria at these very acidic pH values. The fitting allowed to
determine the first pKa to be less than 1, as recently reported for an
analogous pigment32 (Fig. 4).
The other two equilibrium constants were established between
the flavylium cation, AHþ (non-protonated in the amine) and the
quinoidal base, A, and the third is related to the formation of the
ionized quinoidal base, Aꢀ (Fig. 5).

J.L.C. Sousa et al. / Tetrahedron 73 (2017) 6021e6030
6025
Fig. 3. (a) Spectral variations as a function of pH of compound 5 in the pH range 1.68 < pH < 6.07; (b) the same for 6.07 < pH < 9.56; (c) Fitting was achieved for pKa1 ¼ 4.7 0.1;
pKa2 ¼ 7.9 0.1; (d) chemical equilibria of the compound 5 in water/ethanol (75:25) (v/v) between pH 1 and 9.
deprotonation of the other carboxylic acid substituent in position
4’. Once again, the spectral variations regarding to the deprotona-
tion of the carboxylic acid in compounds 7 and 8 are very small and
thus, the respective acidity constants have a higher error. An error
of 0.3 was estimated while in the case of the zwitterion flavylium
cation and the negative charged quinoidal base, the error was 0.1.
For this system, fitting of experimental data was performed at 320
and 470 nm and thus, it was possible to determine three deproto-
nations in the window of the studied pH (Fig. 7).
The pKa values for pyrano-3-deoxyanthocyanins 4e8 are
resumed in Table 2.
Globally, for the studied dyes, the flavylium cation AHþ is
slightly more stable when a methyl group is attached to the D ring
rather than a catechol, a dimethylaminophenyl or a carboxyl sub-
stituents. In fact, oppositely to the catechol group in compound 5,
Fig. 4. UVeVis spectra of pigment 6 for different HCl solutions (from 0.1 to 6 M) with
methyl group is an electron-donating essentially by inductive ef-
fect, stabilizing the flavylium cation and increasing pKa1. In contrast
for pigment 6, the first deprotonation occurs at the R3 substituent at
very acidic conditions. For this pigment, the substituent seems to
stabilize the neutral quinoidal base (A), which could explain the
decrease of pKa1 and the increase of pKa2. Concerning the sub-
stitutions in the B ring, the presence of a p-carboxylic acid or a
catechol group does not produce significant differences in the
deprotonations of 7-OH and 40-OH. Overall, the nature of the sub-
stituent in the pyrano-3-deoxyanthocyanin core structure, mainly
at the C-10 position, greatly changes the type of equilibrium that
could be established as well as the thermodynamic constant values.
Comparing with analogous pigments already studied, namely
pyranomalvidin derivatives with glucose residue (4a-7a), it was
observed that the glycosylated molecules tend to have lower pKa1
(except the non-natural pigment 6a) which means that flavylium
25% ethanol.
2.3.4. Compound 7
In Fig. 6 (a) the spectral variations are very small. As expected,
the deprotonation of the carboxylate substituent of the flavylium
cation gives rise to a zwitterion compound, AHþ-. The estimated
error for this equilibrium is circa 0.3 pH units. A similar behavior
was reported for the absorption spectra of the chalcones of com-
pound 40-carboxy-7-hydroxyflavylium.33 The other deprotonation
is compatible with formation of the quinoidal base Aꢀ from the 7-
OH position.
2.3.5. Compound 8
This compound is similar to the previous one and at higher pH
values the spectral variations are compatible with the

6026
J.L.C. Sousa et al. / Tetrahedron 73 (2017) 6021e6030
Fig. 5. (a) Spectral variations as a function of pH of compound 6 in the pH range 1.64 < pH < 2.34; (b) the same for 2.34 < pH < 6.84; (c) Fitting was achieved for pKþa (NH)<1;
pKa1 ¼ 3.6 0.1; pKa2 ¼ 8.4 0.1; (d) chemical equilibria of the compound 6 in water/ethanol (75:25) (v/v) between pH 1 and 9.

J.L.C. Sousa et al. / Tetrahedron 73 (2017) 6021e6030
6027
Fig. 6. (a) Spectral variations as a function of pH of compound 7 in the pH range 0.77 < pH < 3.27; (b) the same for 3.27 < pH < 7.02; for higher pH values precipitation occurs; (c)
Fitting was achieved for pKa
¼ 1.6 0.3 and pKa1 ¼ 4.9 0.1; (d) Chemical equilibria of the compound 7 in water/ethanol (75:25) (v/v) between pH 1 and 7.
(COOH)
cation (AHþ) is more destabilized comparing with the respective
pyrano-3-deoxyanthocyanins.
parameters of these compounds is crucial to determine their sta-
bility for further applications, namely as pH-sensors for food
packaging applications, in dye-sensitized solar cells, food, textile,
hair dyes, among others.
4. Experimental section
4.1. Reagents
A universal buffer of Theorell and Stenhagen34 was made by
dissolving 2.25 mL of phosphoric acid (85% w/w), 7.00 g of mono-
hydrated citric acid, 3.54 g of boric acid and 343 mL of NaOH 1 M
solution in millipore water until 1 L. Solutions were made using
millipore water. All other chemicals were obtained from Sigma-
Aldrich (Madrid, Spain).
4.2. Synthesis
3. Conclusions
4.2.1. Pyranoluteolinidin-derived dyes (4e7)
The synthesis of the pyranoluteolinidin-type pigments 4e7
followed a procedure described elsewhere.35 Briefly, these pig-
ments were obtained by the reaction of individual mixtures of
luteolinidin 3a (30.7 mg, 0.10 mmol) with 1,3-acetonedicarboxylic
acid (50 equiv), caffeic acid (10 equiv), 4-(dimethylamino)cin-
namic acid (10 equiv) or oxaloacetic acid (30 equiv.) in aqueous
solution with 20% of ethanol (50 mL) at pH 3.5 (except for pigment
7, which pH was set to 2.6). The reactions were stirred at room
temperature during one week. The formation of the respective
pyranoluteolinidins 4e7 was monitored every 24 h by HPLC-DAD.
When the products reach their maximum intensity, the reaction
were stopped, ethanol was evaporated and the samples were pre-
purified on a Buchner funnel loaded with C18 silica gel. The final
The chemical equilibrium of five pyrano-3-deoxyanthocyanins
in hydroalcoholic aqueous solutions at different pH values were
studied for the first time by means of UVeVis spectroscopy. These
compounds do not undergo hydration reactions but only acid-base
reactions as it had been already described for other pyrano-
anthocyanin pigments. These results contrast with the behavior
of anthocyanins when compared with 3-deoxyanthocyanins which
have significant differences.1 It is worth to mention that the ther-
modynamic constants achieved among the several pigments in
pyrano-3-deoxyanthocyanins (as well as in pyrano-anthocyanins)
are greatly affected by the presence of different substituent
groups in
D ring. The knowledge of the physical-chemical

6028
J.L.C. Sousa et al. / Tetrahedron 73 (2017) 6021e6030
Fig. 7. (a) Spectral variations as a function of pH of compound 8 in the pH range 0.84 < pH < 2.52; (b) the same for 2.52 < pH < 6.04; (c) the same in the pH region 5.7 < pH < 9.02;
(d) Fitting was achieved for pKa (COOH) ¼ 1.2 0.3, pKa1 ¼ 4.9 0.1 and pKa2 ¼ 8.1 0.1; (e) Chemical equilibria of the compound 8 in water/ethanol (75:25) (v/v) between pH 1 and
9.
purification was achieved by column chromatography using
reversed-phase C18 silica gel or Toyopearl HW-40S as stationary
phase (250 ꢃ 16 mm i.d.), using mixtures of acidified water (with
HCl 2%) with increasing percentages of methanol as eluent.
Pyranoluteolinidin-based dyes 4e6 presented identical spectro-
scopic data to those previously reported in the literature. 10-
Carboxy-pyranoluteolinidin (7) was obtained as an orange solid
DMSO-d6/TFA 90:10), d (ppm): 169.9 (C-10), 169.9 (COOH), 168.1 (C-
7), 160.5 (C-2), 154.3 (C-40), 153.9 (C-5), 153.0 (C-8a), 146.9 (C-30),
122.9 (C-60), 120.5 (C-10), 117.1 (C-50),115.0 (C-20),109.4 (C-4a),109.3
(C-3), 103.2 (C-9), 100.8 (C-8 and C-6). LC-DAD/ESI-MS: 7 [M]þ m/z
339, lmax 492 nm.
4.2.2. 40-Carboxy-5,7-dihydroxy-flavylium (3b)
(3.6 mg, 9.6%). 1H NMR (600.13 MHz, DMSO-d6/TFA 90:10),
d (ppm):
4-Acetylbenzoic acid 2b (106.7 mg, 0.65 mmol) was added to a
solution of 2,4,6-trihydroxybenzaldehyde 1 (100.6 mg, 0.65 mmol)
in AcOH/H2SO4 (4:1) (5 mL). The reaction mixture was stirred at
room temperature for 24 h. The crude reaction was pre-purified on
7.89 (s,1H, H-9), 7.72 (dd, J ¼ 2.2 and 8.6 Hz,1H, H-60), 7.69 (s,1H, H-
3), 7.65 (d, J ¼ 2.2 Hz, 1H, H-20), 7.04 (d, J ¼ 8.6 Hz, 1H, H-50), 7.30
and 7.16 (2d, J ¼ 2.0 Hz, 2H, H-6 and H-8). 13C NMR (125.77 MHz,

J.L.C. Sousa et al. / Tetrahedron 73 (2017) 6021e6030
6029
Table 2
4.4. NMR spectroscopy
pKa values of pyrano-3-deoxyanthocyanins (4e8) and malvidin analogues with
glucoside residue (4a-7a) obtained from UVeVis spectroscopya.
1H NMR (600.13 MHz) and 13C NMR (125.77 MHz) spectra were
recorded in DMSO-d6/TFA (90:10) and MeOD-d4/TFA (90:10) on a
Bruker-Avance 600 spectrometer at 303 K and with TMS as an in-
Pigment
pKa (R3 substituent)
pKa1 (7-OH)
pKa2 (40-OH)
4
5
6
7
e
5.0 ( 0.1)
4.7 ( 0.1)
3.6 ( 0.1)
4.9 ( 0.1)
4.9 ( 0.1)
4.57 ( 0.07)
4.20 ( 0.07)
5.4 ( 0.1)
4.42 ( 0.01)
6.5 ( 0.1)
7.9 ( 0.1)
8.4 ( 0.1)
e
e
ternal standard [chemical shifts (d) in parts per million, coupling
<1
1.6 ( 0.3)
1.2 ( 0.3)
e
constants (J) in hertz]. Multiplicities are recorded as singlets (s),
doublets (d) and double doublets (dd). 1H chemical shifts were
assigned using 2D NMR (COSY, NOESY) experiments while 13C
resonances were assigned using 2D NMR techniques (gHMBC and
gHSQC). The delay for the long range C/H coupling constant was
optimized to 7 Hz.
8
8.1 ( 0.1)
8.23 ( 0.04)
7.84 ( 0.05)
9.5 ( 0.1)
7.78 ( 0.02)
4a26
5a19
6a32
7a25
e
<1
1.18 ( 0.07)
a
Values in parentheses are the standard deviations of the curve-fitting
procedure.
4.5. Absorption spectroscopy
UVeVis absorption spectra were recorded in a Varian-Cary 100
Bio spectrophotometer. Spectroscopic absorbance curves were
recorded at 25 ꢁC for all the solutions from 200 to 800 nm with a
1 nm sampling interval using a quartz cell cuvette of 3.5 mL ca-
pacity and 1 cm of optical path.
a Buchner funnel loaded with reversed-phase C18 silica gel, eluting
with acidified aqueous solutions containing increasing percentages
of methanol. Then, the desired compound was purified with acid-
ified aqueous/methanolic solutions 50e60%. After the evaporation
of methanol, the fractions were freeze-dried and the compound 3b
was obtained as an orange solid (122.1 mg, 59%). 1H NMR
4.6. Titrations
(600.13 MHz, MeOD-d4/TFA 90:10),
d
(ppm): 9.31 (d, J ¼ 8.9 Hz, 1H,
H-4), 8.43 (d, J ¼ 8.7 Hz, 2H, H-20,60), 8.29 (d, J ¼ 8.7 Hz, 2H, H-30,50),
The acidic thermodynamic constants were determined by
spectrophotometric titrations. It was subsequently added to the cell
cuvette 1 mL of NaOH solution (0.1 M), 1 mL of universal buffer
solution at pH 1 and 1 mL of stock solution of the pyrano-
deoxyanthocyanins (with 75% ethanol), where the final concen-
tration of each pigment was 0.03 mM and 25% ethanol. The first
UVeVis spectrum was immediately recorded. Increasing amounts
of NaOH solution (1 M) were added to the cell cuvette achieving a
pH range between 1 and 12. After each addition, the mixture was
rapidly shacked, a UVeVis spectrum was recorded and the pH was
measured. Final volumes were appropriately corrected. All pH
measurements were made in a Radiometer Copenhagen PHM240
pH/ion meter. The fitting for pKa determination was carried out
using Solver program from Microsoft Excel.
8.24 (d, J ¼ 8.9 Hz, 1H, H-3), 7.04 and 6.75 (2d, J ¼ 1.8 Hz, 2H, H-6
and H-8). 13C NMR (125.77 MHz, MeOD-d4/TFA 90:10),
d (ppm):
168.7 (C-2), 166.7 (40-COOH), 159.8 (C-8a and C-5), 136.0 (C-10),
132.9 (C-40), 130.3 (C-30,50), 128.2 (C-20,60), 115.8 (C-4a), 110.2 (C-3),
102.4 (C-8 or C-6), 94.8 (C-6 or C-8), C-7 not assigned. LC-DAD/ESI-
MS: 3b [M]þ m/z 283, lmax 400 nm.
4.2.3. Pyrano-40,10-dicarboxy-7-hydroxy-flavylium (8)
Oxaloacetic acid (1.6 mmol, 10 eq.) was added to a solution of 40-
carboxy-5,7-dihydroxyflavylium 3b (0.16 mmol, 50.75 mg) in a
mixture of H2O/EtOH (80:20) (v/v) (52 mL) and pH was set to 2.6.
The reaction mixture was left at room temperature for 3 days. Then,
the ethanol was evaporated and the aqueous fraction was extracted
with diethyl ether. The final purification was made by column
chromatography using Toyopearl HW-40S gel as stationary phase
(250 ꢃ 16 mm i.d.). The desired compound was recovered with 60%
of methanol acidified with HCl 2%. The pyrano-3-deoxyanthocyanin
8 was obtained as an orange solid (13.5 mg; 22%). 1H NMR
Acknowledgments
ꢀ
The authors thank Dr. Zelia Azevedo for the MS analysis and Dr.
Mariana Andrade for the NMR analysis. This research was sup-
ported by a research project grant (PTDC/QEQ-QFI/1971/2014) with
financial support from FCT/MEC through national funds and co-
financed by FEDER, under the Partnership Agreement PT2020
(UID/QUI/50006/2013 e POCI/01/0145/FEDER/007265). Luís Cruz
gratefully acknowledges FCT for his post-doc grant (SFRH/BPD/
72652/2010) and an investigator contract (NORTE-01-0145-FEDER-
000011).
(600.13 MHz, DMSO-d6/TFA 90:10),
d
(ppm): 8.34 (d, J ¼ 8.6 Hz, 2H,
H-20,60), 8.23 (s, 1H, H-3), 8.20 (d, J ¼ 8.6 Hz, 2H, H-30,50), 7.82 (s, 1H,
H-9), 7.30 and 7.44 (2d, J ¼ 1.9 Hz, 2H, H-6 and H-8). 13C NMR
(125.77 MHz, DMSO-d6/TFA 90:10),
d (ppm): 169.9 (C-7), 167.5 (C-
2),166.6 (40-COOH),159.9 (C-10),156.2 (10-COOH),154.5 (C-8a or C-
5), 153.7 (C-5 or C-8a), 136.0 (C-40), 133.2 (C-10), 130.6 (C-30,50),
128.4 (C-20,60), 111.6 (C-4a), 109.6 (C-9), 105.6 (C-3), 101.6 (C-8 or C-
6), 101.4 (C-6 or C-8). LC-DAD/ESI-MS: 8 [M]þ m/z 351, lmax 439 nm.
Appendix A. Supplementary data
4.3. LC-DAD/ESI-MS
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.tet.2017.08.051.
LC-DAD/ESI-MS analyses were performed on a Finnigan Sur-
veyor series liquid chromatograph equipped with Finnigan LCQ
(Finnigan Corp., San Jose, Calif., USA) mass detector and an API
source using an ESI interface. The samples were analyzed on a
References
reversed-phase column (150 ꢃ 4.6 mm, 5
m
m, C18) at 25 ꢁC using
1. Pina F, Melo MJ, Laia CAT, Parola AJ, Lima JC. Chem Soc Rev. 2012;41:869e908.
2. Pina F, Petrov V, Laia CAT. Dyes and Pigments. 2012;92:877e889.
3. Brouillard R, Chassaing S, Fougerousse A. Phytochemistry. 2003;64:1179e1186.
4. Cruz L, Basílio N, Mateus N, Pina F, de Freitas V. J Phys Chem B. 2015;119:
2010e2018.
5. Cruz L, Mateus N, De Freitas V. Tetrahedron Lett. 2013;54:2865e2869.
6. Brouillard R, Dubois JE. J Am Chem Soc. 1977;99:1359e1364.
7. Brouillard R, Delaporte B. J Am Chem Soc. 1977;99:8461e8468.
8. Brouillard R, Delaporte B, Dubois JE. J Am Chem Soc. 1978;100:6202e6205.
9. Fulcrand H, Benabdeljalil C, Rigaud J, Cheynier V, Moutounet M. Phytochemistry.
the same eluents, gradients and flow rates referred for HPLC anal-
ysis. The capillary voltage was 4 V and the capillary temperature
275 ꢁC. Spectra were recorded in positive ion mode between m/z
120 and 1500. The mass spectrometer was programmed to do a
series of three scans: a full mass (MS), a zoom scan of the most
intense ion in the first scan (MS2), and a MS-MS of the most intense
ion using relative collision energy of 30 and 60 (MS3).

6030
1998;47:1401e1407.
10. Cruz L, Teixeira N, Silva AMS, Mateus N, Borges J, De Freitas V. J Agric Food
Chem. 2008;56:10980e10987.
11. Hakansson AE, Pardon K, Hayasaka Y, de Sa M, Herderich M. Tetrahedron Lett.
2003;44:4887e4891.
12. Schwarz M, Jerz G, Winterhalter P. Vitis. 2003;42:105e106.
13. Mateus N, Silva AMS, Rivas-Gonzalo JC, Santos-Buelga C, De Freitas V. J. Agric.
Food Chem. 2003;51:1919e1923.
J.L.C. Sousa et al. / Tetrahedron 73 (2017) 6021e6030
22. Mazza G, Minati E. Anthocyanins in Fruits, Vegetables and Grains. Boca Raton-
Florida (USA): CRC Press; 1993.
23. Al Bittar S, Mora N, Loonis M, Dangles O. Tetrahedron. 2016;72:4294e4302.
24. Al Bittar S, Mora N, Loonis M, Dangles O. Molecules. 2014;19:20709e20730.
25. Oliveira J, Mateus N, de Freitas V. Tetrahedron Letters. 2013;54:5106e5110.
26. Oliveira J, Petrov V, Parola AJ, et al. J Phys Chem B. 2011;115:1538e1545.
27. Oliveira J, Mateus N, Silva AMS, De Freitas V.
J Phys Chem B. 2009;113:
11352e11358.
14. Oliveira J, Azevedo J, Silva AMS, et al. J Agric Food Chem. 2010;58:5154e5159.
15. Sousa A, Araujo P, Mateus N, de Freitas V. Tetrahedron Lett. 2013;54:
4785e4788.
28. McClelland RA, McGall GH. J Org Chem. 1982;47:3730e3736.
29. Pina F, Melo MJ, Ballardini R, Flamigni L, Maestri M. New J Chem. 1997;21:
969e976.
16. Chassaing S, Isorez-Mahler G, Kueny-Stotz M, Brouillard R. Tetrahedron.
2015;71:3066e3078.
30. Pina F, Melo MJ, Parola AJ, Maestri M, Balzani V. Chem A Eur J. 1998;4:
2001e2007.
17. Chassaing S, Isorez G, Kueny-Stotz M, Brouillard R. Tetrahedron Lett. 2008;49:
6999e7004.
31. Pina F, Roque A, Melo MJ, Maestri I, Belladelli L, Balzani V. Chem A Eur J. 1998;4:
1184e1191.
18. Azevedo J, Oliveira J, Cruz L, et al. J Agric Food Chem. 2014;62:7002e7009.
19. Cruz L, Petrov V, Teixeira N, Mateus N, Pina F, De Freitas V. J Phys Chem B.
2010;114:13232e13240.
32. Oliveira J, Araújo P, Fernandes A, et al. Dyes and Pigments. 2017;141:479e486.
33. Gavara R, Gago S, Jordao N, Pina F. J Phys Chem A. 2014;118:4723e4731.
34. Küster WF, Thiel A. Tabelle per le analisi chimiche e chimico-fisiche. 12 ed.
~
20. Sarni-Manchado P, Fulcrand H, Souquet JM, Cheynier V, Moutounet M. J Food
Sci. 1996;61:938e941.
21. Oliveira J, Mateus N, De Freitas V. Dyes and Pigments. 2014;100:190e200.
Milano: Hoepli; 1982.
35. Cruz L, Sousa JLC, Marinho A, Mateus N, de Freitas V. Tetrahedron Lett. 2017;58:
159e162.