On the thermal degradation of anthocyanidins: Cyanidin

Article abstract of DOI:10.1039/c3ra47809b

Cyanidin was studied by direct pH jumps (from equilibrated solutions at very low pH values to higher pH values) and reverse pH jumps (from equilibrated or not equilibrated solutions at higher pH values to very low ones). The kinetic steps of the direct and reverse pH jumps were followed by stopped flow, absorption spectroscopy and HPLC, at different timescales. The pH dependent rate constant of the slower kinetic process to reach the equilibrium follows a bell shaped curve as described for many synthetic flavylium compounds. Unlike anthocyanins, it was proved that there is no pH dependent reversibility in the system, since the chalcone suffers an irreversible degradation process. The mathematical expression to describe the bell shaped behaviour was deduced. These results contribute to explain why in plants glycosylation is crucial for the stabilization of the anthocyanins.

Full text of DOI:10.1039/c3ra47809b

RSC Advances
View Article Online
View Journal | View Issue
PAPER
On the thermal degradation of anthocyanidins:
cyanidin
Cite this: RSC Adv., 2014, 4, 18939
a
b
a
*
Luis Cabrita, Vesselin Petrov and Fernando Pina
Cyanidin was studied by direct pH jumps (from equilibrated solutions at very low pH values to higher pH
values) and reverse pH jumps (from equilibrated or not equilibrated solutions at higher pH values to very
low ones). The kinetic steps of the direct and reverse pH jumps were followed by stopped flow,
absorption spectroscopy and HPLC, at different timescales. The pH dependent rate constant of the
slower kinetic process to reach the equilibrium follows a bell shaped curve as described for many
synthetic flavylium compounds. Unlike anthocyanins, it was proved that there is no pH dependent
reversibility in the system, since the chalcone suffers an irreversible degradation process. The
mathematical expression to describe the bell shaped behaviour was deduced. These results contribute to
explain why in plants glycosylation is crucial for the stabilization of the anthocyanins.
Received 19th December 2013
Accepted 3rd April 2014
DOI: 10.1039/c3ra47809b
www.rsc.org/advances
species of the network are the avylium cation and the
quinoidal base, the last one red shied in comparison with
Introduction
the former.
Anthocyanidins are 2-phenyl-1-benzopyrylium derivatives
possessing the same structure as anthocyanins but lacking
any sugar substituents, see Scheme 1 for cyanidin. Differently
from anthocyanidins, the network of chemical reactions
involving anthocyanins (those possessing a sugar in position
3 or two sugars in position 3 and 5) is now rmly established.
The avylium cation, AH+, is the dominant species at low pH
values. Raising the pH two parallel reactions, deprotonation
to give the quinoidal base (A) and hydration leading to the
hemiketal (B), take place. These two reactions are in
competition but occur in different timescales: proton trans-
fer in the micro-second and hydration in minutes to sub-
seconds, depending on pH. By consequence upon a pH jump
from very acid solutions (where AH+ is the dominant species)
here on designated direct pH jump, A appears as a kinetic
product that usually fades with time, because it is not the
most stable species at the equilibrium. This is due to the
crucial achievement on these systems made by Brouillard
and Dubois,1 who discovered that the species A is not reactive
in acidic to moderately acid medium.
The hemiketal B (chromene) is involved in a tautomeric
process that leads to the ring opening with formation of a cis-
chalcone form. This reaction usually occurs in the sub-second
timescale. Finally the cis-chalcone isomerizes and gives the
trans-chalcone in the timescale of seconds or many days
depending on the avylium substitution pattern. The coloured
aREQUIMTE, Departamento de Quımica, Faculdade de Ciencias
e Tecnologia,
´
ˆ
Universidade Nova de Lisboa, 2829-516 Monte de Caparica, Portugal. E-mail: fp@
fct.unl.pt
bLaboratory of Organic Photochemistry, Faculty of Chemistry, University of Soa,
Scheme 1 Network of chemical reactions of anthocyanins and related
Soa, Bulgaria
compounds. Kn ¼ kn/k (n ¼ a, h, t, i).
ꢀn
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 18939–18944 | 18939

View Article Online
RSC Advances
Paper
The analysis of the thermodynamic equilibria of the a-
vylium compounds becomes easier if eqn (1) is considered.2
Eqn (1) is equivalent to a single acid–base equilibrium
between avylium cation and a conjugate base, CB, dened
as the sum of the concentrations of the other species in the
network,
Scheme 3 Network of chemical reactions when the cis–trans isom-
erization barrier is small.
½CBꢁ ¼ ½Aꢁ þ ½Bꢁ þ ½Ccꢁ þ ½Ctꢁ
AHþ þ H2O # CB þ H3Oþ
(1)
k1 ¼ ka + kꢀa[H+]
(2)
(3)
Ka0 ¼ Ka þ Kh þ KhKt þ KhKtKi
½Hþꢁ
1
Â
Ã
k2 ¼
kh þ
k
Hþ
ꢀh
½Hþꢁ þ Ka
1 þ Kt
Thermodynamics
The second kinetic process observed in anthocyanins is
controlled by the hydration, eqn (3). The tautomerization is also
much faster than hydration (unless for very low pH values) and
thus B and Cc behave as a single species in equilibrium during
this step. At the end of this process all the species except Ct are
in (pseudo)equilibrium.
A convenient way to account for the thermodynamic of the
network of chemical reactions involving avylium derivatives is
presented in Scheme 2.3
Once the equilibrium constants are calculated all the
network species are positioned from the relationship,
DG0 ¼ ꢀRT ln K where DG0 is the standard Gibbs energy, R the
gas constant, T the absolute temperature and K the equilibrium
constant for each process of the network,4 see Scheme 2. The
relative position of the several species is very dependent on the
substituents of the avylium core. For example in the case of
anthocyanins, B is the most stable species at higher pH values,
while in 40,7-dihydroxyavylium it is Ct.5
The last kinetic process corresponds to the formation of Ct
and is given by eqn (4)
KhKt
½Hþꢁ þ Ka þ Khð1 þ KtÞ
k3 ¼
ki þ k
(4)
ꢀi
Fast formation of trans-chalcone
Kinetics
This is the case of many synthetic and some natural avylium
derivatives.6 In this case the pseudo-equilibrium is not formed
and the steps 2 and 3 above described are transformed in one.
Assuming that the equilibrium between AH+ and A as well B and
Cc is reached during the kinetic process, the following equation
can be deduced,7 see Scheme 3.
This is equivalent to the following equilibrium, eqn (5)
permitting to apply the steady state hypothesis to Y (B and Cc)
leading to eqn (6)
Regarding the kinetic process taking place in the network of
chemical reactions involving avylium derivatives two situa-
tions according to the existence or not of a thermal barrier for
the cis–trans isomerization have been described: (i) the former
followed by anthocyanins, (ii) the last by some synthetic avy-
lium compounds.
Slow formation of trans-chalcone
In anthocyanins three kinetic processes can be identied upon
a pH jump from equilibrated solutions at pH ¼ 1.0 to higher pH
values (direct pH jumps).6 Immediately aer the direct pH jump
the proton transfer takes place, to form A, within micro-
seconds, faster than the mixing time of a stopped ow appa-
ratus, eqn (2). In this step AH+ and A are in fast equilibrium, the
respective mole fraction distribution depending on pH. More-
over, they behave as a single species in the subsequent kinetic
steps.
(5)
½Hþꢁ
Â
Ã
KhKtki þ k Hþ
Ktki
ꢀi
½Hþꢁ þ Ka
kobs
¼
(6)
½Hþꢁ þ
k
ꢀh
Representation of eqn (6) as a function of pH leads to a
bell shaped curve similar to the one reported in Fig. 2B, see
below.
In the case of anthocyanidins the situation is different since
these compounds are highly unstable at moderately acidic pH
values. In previous work it was reported the instability of
anthocyanidins and some suggestions about the respective
degradation mechanism were made.8 In this work strong
evidences for the degradation mechanism of cyanidin are given
together with a detailed study of the respective kinetics.
Scheme 2 Energy level diagram for the flavylium network of chemical
reactions at the equilibrium.
18940 | RSC Adv., 2014, 4, 18939–18944
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
v/v). The gradient employed started with 10% B, then increasing
linearly to 100% B at 30 min. Flow rate was 0.6 mL minꢀ1 and
oven temperature was 40 ꢃC. All samples were ltered with
0.45 mm GHP syringe lters (Acrodisc).
Experimental
Chemicals
`
Cyanidin chloride was purchased from Extrasynthese (Genay,
France), 2,4,6-trihydroxybenzaldehyde was purchased from
Aldrich, 2,4,6-trihydroxybenzoic acid and 3,4-dihydroxybenzoic
acid were purchased from Fluka.
Results and discussion
All aqueous solutions were prepared with type I water.
Methanol and ethanol were of HPLC grade. A universal buffer of
Theorell Stenhagen was made by dissolving 2.25 mL of phos-
phoric acid (85% w/w), 7.00 g of monohydrated citric acid, 3.54
g of boric acid and 343 mL of 1 M NaOH solution in water,
completed until 1 L.9
Fig. 1 shows the absorption spectra taken 20 ms aer a direct
pH jump followed by stopped ow. The spectral variations are
compatible with an acid base equilibrium between AH+ and A.
Representation of the absorbance as a function of pH (inset of
Fig. 1) gives a pKa ¼ 4.8.
The spectra shown in Fig. 1 evolve according to Fig. 2A. The
AH+/A species disappear to give other species absorbing pref-
erentially in the UV. Representation of the rate constant of this
process as a function of pH is a bell shaped curve as shown in
Fig. 2B.
Instrumentation and procedures
A stock solution of cyanidin chloride 1.50 ꢂ 10ꢀ4 M was prepared
by dissolving the appropriate amount of cyanidin in EtOH acid-
ied with 1% concentrated HCl. The use of an ethanolic instead
of an aqueous solution conferred an increased stability to the
stock solution and avoided precipitation during the experiments.
All pH measurements were made in a Crison BASIC 20+ pH-
meter tted with a Crison electrode. The calibration was made
with standard buffers at pH 4.00, 7.00 and 9.00 purchased from
CRISON.
The data shown in Fig. 2B would suggest a behaviour of
cyanidin similar to compounds lacking cis–trans isomerization
barrier, like many synthetic avylium compounds and also
some natural.6 However, a deeper analysis of the system proves
that it is not the case. The question is that in eqn (5) and (6) the
complete reversibility of the system is a requirement, which is
followed in the systems already reported.6 The data shown in
Fig. 3 clearly shows that this system is not reversible: aer a
direct pH jump and upon a certain delay, if a reverse pH jump is
applied back to pH ¼ 1 the absorption of the avylium is not
completely recovered, and increasing the delay decreases the
extent of absorption recovery.
The UV-vis measurements were made in a CARY 5000 spec-
trophotometer (VARIAN) using quartz cuvettes (1 cm path)
at 20 ꢃC, in the 220–800 nm range and at a scan rate of
1010 nm minꢀ1
.
The stopped ow experiments were conducted in an Applied
Photophysics SX20 stopped ow spectrometer provided with a
PDA.1/UV photodiode array detector with a minimum scan time
of 0.65 ms and a wavelength range of 200–700 nm.
All thermodynamic and kinetic constants of the cyanidin
chloride solution were determined by a spectrophotometric
method, as described elsewhere.10 A stock solution of cyanidin
chloride (1.50 ꢂ 10ꢀ4 M) in acidied ethanol was prepared as
described above.
Nevertheless, if the reverse pH jump is carried out within 1
min of delay aer the direct jump, the re-appearance of (some)
avylium cation can be monitored by stopped ow, Fig. 4. As
reported previously, the reverse pH jump converts all A into AH+
within a few milliseconds (mixing time of the stopped ow),
accounting for the absorbance of avylium/quinoidal base aer
the delay (75%).6 The rst observable kinetic process corresponds
to the recovery of more AH+ (16%) from B, since at low pH values
hydration step is faster than tautomerization (change of regime).
Direct pH jumps (from pH 1 to higher pH values) were carried
out by pippeting 0.250 mL of cyanidin stock solution into a
quartz cuvette, then rapidly adding 2.25 mL of a universal buffer
at the target pH value, and start collecting UV-vis spectra
immediately. This corresponds to a 10-fold dilution and the
analyzed solutions were therefore 10% in EtOH (v/v). Minor pH
adjustments were made by adding a few microliters of aqueous
HCl 10 M or NaOH 1 M solutions. The ionic strength was kept
constant. Reverse pH jumps (from higher pH values to pH 1) were
performed by adding 0.100 mL HCl (10 M) to cyanidin solutions
that previously underwent direct pH jumps. Throughout the
whole experiment the temperature was kept at 20 ꢃC.
The HPLC system was a Merck-Hitachi comprising a L-6200A
pump, a L-4500 Diode Array Detector, a L-5025 column oven, a D-
6000 interface controlled with the DSM soware, and a Rheo-
dyne 7125 manual injection valve with a 20 mL injection loop. The
analysis was performed in a Phenomenex Gemini column (C18,
150 ꢂ 4.6 mm, 3 mm). The solvents used were A (water : formic
Fig. 1 pH dependent absorption spectra obtained 20 ms after a pH
acid, 9 : 1, v/v) and B (water : formic acid : methanol, 4 : 1 : 5, jump followed by stopped flow.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 18939–18944 | 18941

View Article Online
RSC Advances
Paper
Fig. 2 (A) Spectral modifications taking place after a direct pH jump;
(B) representation of the rate constants as in (A) as a function of pH.
Fig. 3 (A) The sequence of pH jumps: initial solution of cyanidin at pH
0.8 (a), followed by a direct pH jump to pH 4.6 (b); after a delay of 1 min
at this pH, a reverse jump back to pH ¼ 0.9 (c). Only 63% of the fla-
vylium absorption was recovered; (B) the same sequence but using a
delay of 20 min at pH ¼ 4.6 before the reverse pH jump. In this case
only a small fraction of flavylium absorption was recovered.
The second and last process is the further formation AH+ (9%)
now from Cc via B. The ratio of the amplitudes of the last kinetic
step and the rst is equal to the equilibrium constant Kt. Some
kinetic information is also obtained from Fig. 4. The rst and
faster process corresponds to the hydration reaction and is given
½Hþꢁ
KhKtkdiss
½Hþꢁ þ Ka
by eqn (3), while the second one and slower is assigned to k
ꢀt
kobs
¼
(7)
Ktkdiss
½Hþꢁ þ
(there is no reversibility from B to Cc since once B is formed is
k
immediately transformed into AH+).†
ꢀh
As shown in Fig. 3 the bell shaped curve cannot be attributed
to a system as shown in eqn (5) and (6). However, using a slight
modication to take into account the experimental details of
this system a bell shaped curve is also obtained, see Scheme 4.
This modication accounts for the existence of irreversible
process(es) resulting in the formation of degradation products.
From a kinetic point of view, the irreversible step could take
place from either Cc or Ct. Comparing eqn (6) with eqn (5) the
systems are kinetically the same if ki is substituted by kdiss and
A global tting using pKa from Fig. 1, Kt, k and eqn (3) in
Fig. 4 as well as eqn (7) permits to calculate all the parameters,
see Table 1.
ꢀt
Degradation of cyanidin
The transformations occurring in solutions of cyanidin upon
different delays aer direct pH jumps were also followed by
HPLC with DAD detection, Fig. 5A. The peaks of some of the
network species and degradation products of cyanidin were
monitored and identied based on retention time, absorption
spectra and co-chromatography, and peak areas were plotted
against time, Fig. 5B.
k
is neglected.‡ On this basis, the observed rate constant
ꢀi
should be given by eqn (7).
† At this pH value the acid and basic catalysis was neglected, since it is not
expected to be relevant.
It is clear the disappearance of cyanidin (C) in accordance to
the results reported in Fig. 4. The variation of peak area as a
‡ If there is decomposition the constant ki lacks of signicance.
18942 | RSC Adv., 2014, 4, 18939–18944
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Fig. 4 Reverse pH jump to pH ¼ 1.5 after a direct pH jump from 1.0 to
4.7 (delay 1 minute) monitored by stopped flow.
Scheme
4 Network of chemical reactions in the case of
anthocyanidins.
Table 1 Equilibrium and rate constants of cyanidin at pH 4.8 in water
(10% EtOH)
kh/sꢀ1
k
ꢀh/Mꢀ1sꢀ1
400
kt/sꢀ1
k
ꢀt/sꢀ1
1.6
kdiss
Fig. 5 (A) HPLC chromatograms extracted at 280 nm showing peaks
separated from cyanidin solutions at pH ¼ 2.25 as a function of time:
3,4-dihydroxybenzoic acid (5.8 min); 2,4,6-trihydroxybenzoic acid
(8.2 min); 2,4,6-trihydroxybenzaldehyde (12.1 min); hemiketal/cis-
chalcone (13.0 min); cyanidin (22.7 min) and (B) representation of peak
areas of network and degradation products of cyanidin as a function of
time at pH 2.25. Legend: (C) cyanidin; (A) 2,4,6-trihydroxy-
benzaldehyde (brown curve); (X) chalcone (green curve); (,) 3,4-
dihydroxybenzoic acid (black curve); (B) 2,4,6-trihydroxybenzoic acid
(grey curve) and (*) impurity.
pKa
pKh
Kt
4.8
4.6
0.56
0.009
0.9
0.1
function of time suggests the initial formation of 2,4,6-trihy-
droxybenzaldehyde and 3,4-dihydroxybenzoic acid, and a chal-
cone, which we were not able to distinguish if it is cis or trans.
The chalcone(s) build up to 120 min and then decrease with
time, suggesting they are being consumed in subsequent
degradation process. The 2,4,6-trihydroxybenzoic acid seems to
be a secondary degradation product resulting from the oxida-
tion of the corresponding benzaldehyde. These results are in
good agreement with previous work.8
Inspection of the bell shaped curve suggests that there is
not reversibility from Ct (kꢀi z 0). Otherwise the curve would
not be zero at very low pH values, because the limit of eqn (6)
when the proton concentration is very high is kꢀi, as observed
previously in the case of synthetic avylium compounds.6 In
order to account for the experimental data, two possibilities
can be considered: (i) Cc leads to Ct and this last one
decomposes much faster than the backward reaction Ct to
Cc, kꢀi, and in this case kdiss should be assigned to ki, and (ii)
Cc decomposes without giving Ct, and in this case kdiss
represents the decomposition rate of the former. We were
not able to distinguish which of these two assumptions
take place.
Scheme 5 Proposed mechanism of cyanidin degradation.
RSC Adv., 2014, 4, 18939–18944 | 18943
This journal is © The Royal Society of Chemistry 2014

View Article Online
RSC Advances
Paper
A possible explanation for the anthocyanidins degradation is acidic medium. The existence of the sugar in position 3 is thus a
shown in Scheme 5. The rst step regards an addition of water crucial requirement to the expression of colour in plants.
to C2, resulting in an intermediate which could be easily
oxidized into the corresponding products.
In the case of the anthocyanins the rst step would be
impeded explaining their higher stability.
Acknowledgements
˜
ˆ
This work has been supported by Fundaçao para a Ciencia e
Tecnologia (FCT) and REQUIMTE through Grant no. PEst-C/
EQB/LA0006/2011 and Project PTDC/QUI-QUI/117996/2010: VP
thanks FP7 project “Beyond Everest” for nancial support.
Conclusions
The rate of cyanidin disappearance as a function of pH follows a
bell shaped curve apparently similar to those previously repor-
ted for other avylium derivatives. However, the latter show
reversibility, recovering avylium cation when equilibrated
solutions at higher pH values are acidied. The former is not
reversible giving rise to degradation products. The growing of
the rate at lower pH values is due to the increasing of the mole
fraction of the chalcones; once formed, these species degrade
irreversibly. Kinetically this is equivalent to a kinetic control by
the cis–trans isomerization as observed in other avylium
compounds. The descending branch of the bell shaped curve
can be attributed to the hydration control; by increasing the pH
the rate of chalcone formation decreases via decreasing of the
hemiketal formation.
The kinetic and HPLC results strongly suggest that the
chalcones are responsible for the formation of degradation
products. The degradation leads to 2,4,6-trihydroxy-
benzaldehyde and 3,4-dihydroxybenzoic acid. The compound
2,4,6-trihydroxybenzoic acid was also found in a subsequent
kinetic process most probably from the oxidation of 2,4,6-tri-
Notes and references
1 R. Brouillard and J.-E. Dubois, J. Am. Chem. Soc., 1977, 99,
1359.
2 R. Brouillard and J. Lang, Can. J. Chem., 1990, 68, 755.
3 F. Pina, M. J. Melo, M. Maestri, R. Ballardini and V. Balzani,
J. Am. Chem. Soc., 1997, 119, 5556.
4 F. Pina, J. Chem. Soc., Faraday Trans., 1998, 94, 2109;
correction 1998, 94, 3781.
5 V. Petrov, S. Stanimirov, I. K. Petrov, A. Fernandes, V. de
Freitas and F. Pina, J. Phys. Chem. A, 2013, 117, 10692.
6 F. Pina, M. J. Melo, C. A. T. Laia, A. J. Parola and J. C. Lima,
Chem. Soc. Rev., 2012, 41, 869.
7 V. Petrov and F. Pina, J. Math. Chem., 2010, 47, 1005.
8 P. Furtado, P. Figueiredo, H. Chaves das Neves and F. Pina,
J. Photochem. Photobiol., A, 1993, 75, 113.
¨
9 F. W. Kuster and A. Thiel, Tabelle per le Analisi Chimiche e
Chimico-Fisiche, Hoepli, Milano, 12th edn, 1982, p. 157.
hydroxybenzaldehyde. This result contrasts with the behavior of 10 F. Nave, V. Petrov, F. Pina, N. Teixeira, N. Mateus and V. de
cyanin and other anthocyanins, which present reversibility in
Freitas, J. Phys. Chem. B, 2010, 114, 13487.
18944 | RSC Adv., 2014, 4, 18939–18944
This journal is © The Royal Society of Chemistry 2014