A simple synthesis of 3-deoxyanthocyanidins and their O-glucosides

Article abstract of DOI:10.1016/j.tet.2016.05.076

This work deals with the chemical synthesis of simple analogs of anthocyanins, the main class of water-soluble natural pigments. Flavylium ions with hydroxyl, methoxyl and β-D-glucopyranosyloxyl substituents at positions 4′ and 7 have been prepared by straightforward chemical procedures. Moreover, the two 3-deoxyanthocyanidins of red sorghum apigeninidin (4′,5,7-trihydroxyflavylium) and luteolinidin (3′,4′,5,7-tetrahydroxyflavylium) were synthesized in a one-step protocol. Attempts to synthesize luteolinidin O-β-D-glucosides resulted in a mixture of the 5-O- and 7-O-regioisomers in low yield. A preliminary study of the 4′-β-D-glucopyranosyloxy-7-hydroxyflavylium and 7-β-D-glucopyranosyloxy-4′-hydroxy-flavylium ions shows that simply changing the glucosidation site can profoundly affect the color intensity and stability.

Full text of DOI:10.1016/j.tet.2016.05.076

Accepted Manuscript
A simple synthesis of 3-deoxyanthocyanidins and their O-glucosides
Sheiraz Al Bittar, Nathalie Mora, Michèle Loonis, Olivier Dangles
PII:
S0040-4020(16)30493-8
10.1016/j.tet.2016.05.076
TET 27803
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 19 February 2016
Revised Date: 21 May 2016
Accepted Date: 30 May 2016
Please cite this article as: Al Bittar S, Mora N, Loonis M, Dangles O, A simple synthesis of 3-
deoxyanthocyanidins and their O-glucosides, Tetrahedron (2016), doi: 10.1016/j.tet.2016.05.076.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
A simple synthesis of 3-deoxyanthocyanidins and their O-glucosides
Sheiraz Al Bittar, Nathalie Mora, Michèle Loonis, Olivier Dangles*
Graphical Abstract

ACCEPTED MANUSCRIPT
A simple synthesis of 3-deoxyanthocyanidins and their O-
glucosides
Sheiraz Al Bittar, Nathalie Mora, Michèle Loonis, Olivier Dangles*
University of Avignon, INRA, UMR408, 84000 Avignon, France
Corresponding author. Tel.: +33 490 14 44 46.
E-mail address: olivier.dangles@univ-avignon.fr (O. Dangles)

ACCEPTED MANUSCRIPT
ABSTRACT
This work deals with the chemical synthesis of simple analogs of anthocyanins, the main
class of water-soluble natural pigments. Flavylium ions with hydroxyl, methoxyl and β-D-
glucopyranosyloxyl substituents at positions 4’ and 7 have been prepared by
straightforward chemical procedures. Moreover, the two 3-deoxyanthocyanidins of red
sorghum apigeninidin (4’,5,7-trihydroxyflavylium) and luteolinidin (3’,4’,5,7-
tetrahydroxyflavylium) were synthesized in a one-step protocol. Attempts to synthesize
luteolinidin O-β-D-glucosides resulted in a mixture of the 5-O- and 7-O-regioisomers in
low yield. A preliminary study of the 4’-β-D-glucopyranosyloxy-7-hydroxyflavylium and
7-β-D-glucopyranosyloxy-4’-hydroxy-flavylium ions shows that simply changing the
glucosidation site can profoundly affect the color intensity and stability.
Keywords: Anthocyanin; 3-Deoxyanthocyanidin; Glycoside; Synthesis; Sorghum; Color
1

ACCEPTED MANUSCRIPT
1. Introduction
For centuries, food colorants have been used to attract the consumer’s eye, to reproduce
color lost during processing and to mask the heterogeneity of food formulations.1
Chemistry has provided the food industry with diverse, cheap, and stable synthetic
colorants. However, during the last decades, the safety of these synthetic colorants has
been questioned. For instance, attention deficit hyperactivity disorder (ADHD) was
observed in children having consumed red drinks supplemented with azo dyes2 and
products including these synthetic colorants have now to issue specific warnings on their
packaging. Overall, natural colorants enjoy a much more positive image to consumers than
artificial ones and the search for stable natural colorants remains a current challenge.
Anthocyanins are the most important group of water-soluble plant pigments. Stored in
cell vacuoles, they are responsible for most of the red, purple and blue colors found in the
plant kingdom.3 Anthocyanins are generally found in Nature as anthocyanidin O-
glycosides and the anthocyanidin aglycone is typically represented by its major colored
form in acidic conditions, i.e. the 3,4’,5,7-tetrahydroxyflavylium ion (pelargonidin) and
common parent ions with additional OH or OMe groups at C3’ and/or C5’. Among these 6
chromophores, anthocyanidins having a 3’,4’-dihydroxy substitution (e.g., cyanidin as the
simplest) are especially important from two viewpoints: they can bind hard metal ions
(Al3+, Fe3+, Mg2+) to form deeply colored chelates, one of the main mechanisms for natural
color variation. They are good electron donors (as, more generally, polyphenols and
flavonoids bearing catechol nuclei) and thus potential antioxidants.4,5
While each anthocyanidin OH group can be potentially glycosylated in anthocyanins,
glycosidation at C3-OH suffers no exception. Indeed, a free C3-OH is associated with a
high chemical instability (except in highly acidic conditions, pH < 2) and the
corresponding C-ring is rapidly cleaved to yield a mixture of simple colorless phenols.
Hence, hydrolysis of the glycosidic bond at C3-OH, either chemically during thermal
processing, or enzymatically by human or bacterial glycosidases, is rapidly followed by
irreversible loss of the typical chromophore.6,7 This is one of the main mechanisms
underlying the well-known low chemical stability and apparent bioavailability of
anthocyanins.
2

ACCEPTED MANUSCRIPT
Although much less common than anthocyanins, 3-deoxyanthocyanidins (3-DAs) and
their O-glycosides have been found in high concentration in some food sources, especially
red sorghum.8,9 The lack of OH group at C3 makes these pigments much less sensitive to
water addition at C2,10,11 which reversibly leads to the colorless hemiketal and chalcone
forms, and also more resistant to irreversible chemical degradation.12 Thus, in mildly
acidic to neutral conditions, 3-DAs express more intense and more stable colors than
common anthocyanins. On the other hand, the applications of 3-DAs as food colorants
could be limited by their low water solubility, which can be markedly improved by
conjugation to D-glucose. Red sorghum actually also contains 3-DA glucosides, such as
luteolinidin 5-O-β-D-glucoside.13
Extraction of anthocyanins from various natural sources typically affords complex
extracts combining anthocyanins, colorless polyphenols and other hydrophilic plant
components, such as sugars and acids. Hence, the purification of these extracts for the
preparation of anthocyanins at the gram scale remains challenging and time-consuming. In
this respect, the simple chemical synthesis of anthocyanins and simpler analogs remains
interesting, if not for direct industrial application, at least for facilitating the evaluation of
their potential as natural colorants for foods and cosmetic products, and as nonessential
micronutrients, whose bioavailability and bioactivity are extensively studied.14,15
In this present work, we report on the efficient chemical synthesis of 3-DAs and their O-
β-D-glucosides (Scheme 1), including the characteristic red sorghum pigments
apigeninidin (APN, 4’,5,7-trihydroxyflavylium) and luteolinidin (LTN, 3’,4’,5,7-
tetrahydroxyflavylium), as well as a series of simpler analogs. After careful optimization of
reaction conditions (time, temperature, solvent composition…), the chemical synthesis of
APN and LTN could be achieved in an essentially one-step procedure, i.e. via a much
simpler route than previously described in the literature.16,17 A preliminary investigation of
the coloring properties of two 3-DA glucosides is also reported, showing the large
influence of the sugar position on the resulting color and its stability.
3

ACCEPTED MANUSCRIPT
Pigment
P118
P218
P3
X3’
OH
OH
H
R4’
X5
H
R7
Yield (%) a)
H
H
56
75 b)
42 b)
23 b)
77
H
H
β-D-Glc
H
H
β-D-Glc
H
Me
H
P4
β-D-Glc
H
H
H
Me
H
P5
β-D-Glc
H
H
77
P6
β-D-Glc
H
H
H
H
45
P7
H
H
Me
H
90
P8
H
Me
H
H
85
P9
APN
LTN
H
OH
OH
H
34
OH
H
H
65
a) Yield of isolated pigment in the final (chromophore-building) condensation step
b) Yield also includes the step of sugar deprotection
Scheme 1. Pigments synthesized in this work (except P1 and P218)
2. Results and discussion
4

ACCEPTED MANUSCRIPT
The chemical synthesis 3-DAs and analogs has already been described over the last 3 -
4 decades.16-19 The simplest route typically involves the acid-catalyzed aldol condensation
of a 2-hydroxybenzaldehyde and an acetophenone bearing additional OH, OMe or O-
glycosyl substituents. Using gaseous HCl bubbled into the cooled solution of both reagents
in ethylacetate, the corresponding flavylium chloride precipitates and is recovered by
simple filtration. Alternatively, 3-DAs were prepared by reduction of the corresponding
flavones20 and from condensation between cinnamaldehyde derivatives and
phloroglucinol.21 Finally, 3-deoxyanthocyanidins can also be prepared by condensation
between phenol derivatives and arylethynylketones.22,23
As natural anthocyanins all possess a 5,7-dihydroxy substitution (A-ring), aldol
condensations involving 2,4,6-trihydroxybenzaldehyde are of particular interest.
Unfortunately, 2,4,6-trihydroxybenzaldehyde is poorly soluble in AcOEt and prone to
acid-catalyzed dimerization (possibly, oligomerization) by electrophilic aromatic
substitution, leading to yellow – orange xanthylium pigments. To circumvent these
disadvantages, less simple routes have been devised, some still based on 2,4,6-
trihydroxybenzaldehyde but requiring its preliminary partial O-acylation to improve its
solubility in weakly polar solvent (convenient for the precipitation of the flavylium salts)
and decrease the nucleophilic character of its aromatic ring.17,24 In this work, special
attention has been devoted to optimizing the direct condensation of (unprotected) 2,4,6-
trihydroxybenzaldehyde with an acetophenone to ensure a reasonable yield of flavylium
while keeping at a minimum the side-reaction leading to xanthylium pigments. Important
factors in the optimization are the solvent composition (ensuring both 2,4,6-
trihydroxybenzaldehyde solubility and flavylium precipitation), the temperature, the
reagent concentration and the proton donor. For the synthesis of the red sorghum pigments
APN and LTN, AcOEt / MeOH (2:1) mixtures were used and gaseous HCl was in situ
generated by reaction between MeOH and TMSCl added in controlled excess to a solution
of both reagents in a closed flask cooled to 0°C. In such conditions, LTN was formed in
65% yield and isolated after 30 min by simple filtration. LTN was contaminated by small
quantities (ca. 5%) of 2,4,6,8-tetrahydroxyxanthylium chloride (Scheme 2) characterized
by UPLC-MS: tR = 4.2 min, λmax = 440 nm, m/z = 244.9 (M+) and by a very simple
spectrum 1H-NMR (in CD3OD): δ (ppm ) = 6.0 (2H, s, H1, H9), 6.4 (2H, s, H3, H7), 9.5
5

ACCEPTED MANUSCRIPT
(1H, s, H4).25 Its formation from 2,4,6-trihydroxybenzaldehyde is not fully clear as it
involves a decarbonylation step (CO loss). Solubilization in EtOH and precipitation by
AcOEt addition afforded pure LTN.
Remarkably, a similar procedure yielded APN devoid of xanthylium pigment and only
contaminated by low amounts of unreacted 2,4,6-trihydroxybenzaldehyde. Repeating the
reaction over 60 min with 2 equiv. of 4-hydroxyacetophenone gave pure APN after
purification in EtOH / AcOEt.
Scheme 2. Structure of 2,4,6,8-tetrahydroxyxanthylium chloride.
Based on this improved procedure, a series of flavylium ions was efficiently prepared
with the typical OH and O-β-D-Glc substituents of natural anthocyanins at positions 4’ and
7. Although much less common at those positions, the corresponding O-methylethers were
also synthesized. The chemical synthesis of aglycones P7, P8 and P9 was achieved in
AcOEt / MeOH (1:1). As neither 2,4-dihydroxybenzaldehyde nor 2-hydroxy-4-
methoxybenzaldehyde is prone to competing dimerization, the aldol condensation was
prolonged over 3 days at 4°C to ensure total flavylium precipitation and a maximal yield
(up to 90% for P8).
As for the glucosides, a preliminary glucosidation step of 2,4-dihydroxybenzaldehyde
and 4-hydroxyacetophenone was required (Scheme 3). With the former, satisfactory yields
were obtained using phase transfer conditions in a 1:1 mixture of saturated aqueous K2CO3
(pH 12) and CH2Cl2 using 2,3,4,6-tetra-O-acetyl-α-D-glucopyranosyl bromide as the
glucosyl donor.18 Due to the strong hydrogen bond between the carbonyl group and C2-
OH, the glucosidation of 2,4-dihydroxybenzaldehyde was regioselective and took place at
C4-OH. Such strongly alkaline conditions did not allow us to convert 4-
hydroxyacetophenone into its glucoside, possibly because of competing enolate formation.
6

ACCEPTED MANUSCRIPT
A less alkaline aqueous phase (NaHCO3 / KCl (1:1), pH 8.7) was thus used19 and the
glucoside was obtained, although in low yield.
Scheme 3. Chemical synthesis of pigments: (i) TMSCl, AcOEt, then MeONa, MeOH, then
aq. HCl. (ii) tris(2-(2-methoxyethoxy)ethyl)amine in CH2Cl2 / 1 M aq. NaHCO3, 1 M aq.
KCl, then MeONa, MeOH. (iii) TMSCl, AcOEt / MeOH.
To prepare pigments P3 and P4, aldol condensation was achieved in AcOEt / MeOH
(4:1) in the presence of TMSCl (10 equiv.) and the mixture kept 3 days at -18°C to
minimize possible deglycosidation. A next step of glucose deacetylation was carried out
using MeONa in excess (8 equiv.) as part of the reagent is consumed by reversible addition
7

ACCEPTED MANUSCRIPT
to the pyrylium ring and/or phenolate formation. After acidification, P3 and P4 were
obtained in moderate over all yields (42% and 23%, respectively).
Actually, the final step (pigment deacetylation) is rather difficult to monitor and push to
completion. It also requires strongly alkaline conditions and subsequent acidification and
concentration that may be damaging to the pigment. Hence, to prepare pigments P5 and
P6, another route was tested in which deacetylation was carried out prior to condensation.
For the condensation step, solvent conditions have to be tuned so as to ensure a sufficient
solubility for the deprotected glucoside (4-O-β-D-glucopyranosyloxyacetophenone 4)
while still permitting the precipitation of the flavylium ions in the medium. Thus, 3
(protected 4) was first deacetylated in good yield, then successfully condensed with 2,4-
dihydroxybenzaldehyde or 2-hydroxy-4-methoxybenzaldehyde in MeOH / AcOEt (2:1)
(Scheme 3).
Red sorghum contains not only 3-DAs but also some of their glucosides, in particular
luteolinidin 5-O-β-glucoside (LTN-5Glc).13 To ensure glucosidation of 2,4,6-
trihydroxybenzaldehyde at C2-OH and also lower its hydrophilic character and thus favor
its transfer into the organic phase, selective protection at C4-OH was attempted.
Benzoylation with benzoylchloride / t-BuOK in THF at 0°C afforded an inseparable
mixture of 2- and 4-regioisomers along with the dibenzoylated product. Unfortunately, the
subsequent step of glucosidation failed because of insufficient stability of the benzoyl
groups in the phase transfer conditions. Benzylation with benzyl bromide / K2CO3 in DMF
turned out to take place mainly at C2-OH, yielding a 4:1 inseparable mixture of 2-
benzyloxy-4,6-dihydroxybenzaldehyde and
4-benzyloxy-2,6-dihydroxybenzaldehyde
(Scheme 4). It may be speculated that the phenolate anion resulting from deprotonation at
C4-OH is stabilized through extensive conjugation with the CHO group while its tautomer
formed by deprotonation at C2-OH could bear a larger negative charge on its O-atom,
making it more nucleophilic.
Glucosidation of the benzylether regioisomers led to pure 2-benzyloxy-6-hydroxy-4-
(2’,3’,4’,6’-tetra-O-acetyl-β-D-glucopyranosyloxy)-benzaldehyde (6) in acceptable yield,
which suggests that the minor 4-benzyloxy-2,6-dihydroxybenzaldehyde is unreactive in the
phase transfer reaction (confirmed by isolation and independent attempt of glycosidation).
Then, hydrogenolysis was carried out in soft conditions (30 min, 2 bars of H2) to avoid
aldehyde reduction. Aldol condensation with 3,4-dihydroxyacetophenone gave 7-
8

ACCEPTED MANUSCRIPT
(2’,3’,4’,6’-tetra-O-acetyl-β-D-glucopyranosyloxy)-3’,4’,5-trihydroxyflavylium
in
acceptable yield (ca. 60%). Unfortunately, the final deprotection step turned out to be slow
and incomplete even in strongly alkaline conditions (excess MeONa in MeOH or KOH in
MeOH/H2O (1:1)24). A possible explanation would be that all phenolic groups are
deprotonated in those conditions, so that the nucleophilic attack of methoxide or hydroxide
anions onto the acetate groups is impeded by electrostatic repulsion (especially with the
nearby C5-O- group). Overall, despite prolonged reaction time and repeated deprotection,
partially protected pigments remained in solution and had to be removed by
chromatography on TSK gel. Given these difficulties, the overall yield of the condensation
– deacetylation – purification sequence is very disappointing (< 10%).
9

ACCEPTED MANUSCRIPT
Scheme 4. Chemical synthesis of luteolinidin glucosides: (i) BnBr (1 equiv.), K2CO3,
DMF (ii) tris(2-(2-methoxyethoxy)ethyl)amine in CH2Cl2 / 1 M aq. NaHCO3, 1 M aq. KCl.
(iii) H2, Pd/C, THF - MeOH. (iv) TMSCl, AcOEt, then MeONa, MeOH, then aq. HCl.
The alternative route (deprotection followed by condensation) was also tested.
Deacetylation of 7 actually gave 2,6-dihydroxy-4-β-D-glucopyranosyloxybenzaldehyde.
Unfortunately, this glucoside was insoluble in most organic solvents and the final
condensation could not be carried out.
10

ACCEPTED MANUSCRIPT
Although obtained in low yield, the LTN O-glucoside could be isolated for full
structural characterization. Surprisingly, UPLC-MS analysis of the deprotected product
revealed that it actually was a ca. 3:2 mixture of LTN-7-Glc and the naturally occurring
LTN-5-Glc. The peak assignment was confirmed by co-injection with a sample of pure
LTN-7Glc obtained by reduction20 of the corresponding flavone luteolin7-O-β-D-
glucoside. Both glucosides were characterized by NMR in comparison with the literature.13
To our knowledge, this unexpected isomerization in alkaline conditions has no equivalent
in the literature.
Hydroxylated flavylium ions (AH+, typically, red) undergo two competitive pathways in
mildly acidic solution featuring the natural medium of anthocyanins,11,27 i.e. the vacuole of
plant cells: the first one is fast proton loss and concomitant formation of purple quinonoid
bases (A), the second one is the much slower reversible addition of water at the
electrophilic C2 center with concomitant proton loss and formation of a colorless
hemiketal (B) in fast equilibrium with a cis-chalcone (Cc), itself in slow equilibrium with
the corresponding trans-chalcone (Ct) (Scheme 5).
11

ACCEPTED MANUSCRIPT
Scheme 5. Structural transformations of flavylium ions in mildly acidic aqueous solution.
With 3-DAs, B and Cc are only transient non-accumulating intermediates and Ct comes
up as the only significant colorless form.18,28. A simplified scheme can thus be proposed in
which A and Ct are respectively the kinetic and thermodynamic products of the concurrent
transformations of AH+ in mildly acidic solution. The set of flavylium ions synthesized in
this work offers a good opportunity to investigate the influence on these competitive
12

ACCEPTED MANUSCRIPT
pathways of the typical substituents of anthocyanins at C4’-OH and C7-OH, in particular
the β-D-glucosyl group, and their consequences in terms of color expression and stability.
This point will be briefly exemplified with P3 and P5.
When a small volume of pigment solution in acidified MeOH (100% flavylium form) is
diluted into a pH 7.4 phosphate buffer, the corresponding quinonoid base is
instantaneously formed. P3 and P5 bearing only one OH group, a single deprotonation is
observed (no further proton loss leading to anionic bases) and a single base is formed
(instead of 2 or 3 possible tautomers as with natural anthocyanins). As the first pKa of
flavylium ions is typically in the range 4 - 5,11 the proton loss can be assumed quantitative.
Thus, the spectra recorded immediately after dilution (Fig. 1A) can be ascribed to pure A
forms resulting from proton loss at C4’-OH (A4’, P3) or at C7-OH (A7, P5). Interestingly,
while both flavylium ions display almost the same λmax in the visible range (ca. 450 nm),
the visible spectrum of the P3 base (λmax = 505 nm) is shifted by 20 nm to longer
wavelengths, compared to the P5 base (λmax = 485 nm). Moreover, the higher molar
absorption coefficient of the P3 vs. P5 flavylium ion (ca. +70%) is translated and amplified
(almost a factor 3) into the corresponding bases. Therefore, although proton loss reduces
the HOMO – LUMO gap in both pigments, this decrease and thus the bathochromic shift
are larger when deprotonation occurs at C4’-OH. Moreover, the HOMO – LUMO match,
which must be mainly responsible for the visible absorption, is better in P3 than in P5,
regardless of the acid - base form.
13

ACCEPTED MANUSCRIPT
2.0
A
A
1.5
1.0
0.5
0.0
300
350
400
450
500
550
600
λ(nm)
2.0
A
B
1.5
1.0
0.5
0.0
0
5
10
15
20
25
30
Time (min)
Fig. 1. A: UV-visible spectra (pigment concentration = 50 µM). a) Quinonoid bases (pH
7.4 phosphate buffer, 25°C): P3 (─) and P5 (─) immediately after dilution in the buffer, P3
(---) and P5 (---) after 30 min. b) Flavylium ions (0.1 M HCl, 25°C): P3 (─), P5 (---). B:
decay of the visible absorbance of the bases at λmax. P3 (■, detection at 505 nm), P5 (●,
detection at 485 nm). Solid lines result from curve-fitting assuming apparent first-order
kinetics: kobs(P3) = 15.4 (± 0.2) x10-4 s-1, kobs(P5) = 19.4 (± 0.2) x10-6 s-1, r = 0.9998.
14

ACCEPTED MANUSCRIPT
Remarkably, the advantage of P3 over P5 in terms of color variation and intensity is not
translated in terms of color stability in neutral conditions. Indeed, while the visible
spectrum of the P5 base remains essentially unchanged over 30 min (half-life of P5 base ≈
10h based on the observed first-order decay, Fig. 1B), the one of P3 is markedly weakened
(half-life of P3 base ≈ 7.5 min) with concomitant appearance of a new absorption band
around 360 nm featuring the P3 trans-chalcone (Fig. 1A). Thus, P5 appears as
exceptionally resistant to water addition and subsequent fading in neutral conditions.
The color loss features the overall conversion of the colored forms into the trans-
chalcone. Its rate is governed by the slow step of chalcone cis – trans isomerization.
Hence, the observed rate constant of fading can be written as Eqn. (1), in which the
fraction represents the percentage of cis-chalcone at pseudo-equilibrium with the colored
forms and hemiketal:27
KhKt
kobs = ki
+ k−i
(1)
[H + ]+ Ka + Kh (1+ Kt )
With the 3’,4’-dihydroxy-7-O-β-D-glucopyranosyloxyflavylium ion (P2), which is
closely related to P3, the following values were estimated for the different parameters:28 Kh
= 2.5x10-5 M, Ka = 1.3x10-5 M, Kt = 3.3, ki = 1.2x10-3 s-1, k-i = 1.6x10-5 s-1(Ki = ki / k-i =
75). At pH 7.4, one has: [H+] << Ka + Kh(1 + Kt). Moreover, Ct is much more stable than Cc
(ki << k-i). Hence, Eqn. (1) can be simplified to give Eqn. (2):
Kt
kobs = ki
Ka
1+ Kt +
Kh
(2)
From Eqn. (2), one obtains: kobs = 8.2x10-4 s-1. This value is actually close to our
estimate for P3: kobs = 15.4 (± 0.2) x10-4 s-1. More work is needed to unveil the cause for
the much higher stability of P5. As there is no reason for Kt and Kh to be dramatically
different for P3 and P5 or for the cis-trans chalcone isomerization barrier to be higher for
P5 than for P3, it may be speculated that the very slow fading of P5 stems from a much
higher Ka value due to proton loss being much more favorable from C7-OH than from C4’-
OH. Consistently, the pKa values of the 4’-hydroxyflavylium and 7-hydroxyflavylium ions
15

ACCEPTED MANUSCRIPT
were reported to be 5.5 and 3.55, respectively.11 In other words, in the competition
between proton loss and water addition, the quinonoid base of P5 is probably both the
thermodynamic and kinetic product of the flavylium ion (Ka > Kh) and the concurrent route
leading to the colorless forms is strongly inhibited.
3. Conclusion
Efficient syntheses have been devised for the red sorghum pigments luteolinidin and
apigeninidin, as well as for a series of flavylium ions substituted by the typical OH, OMe
and O-β-D-Glc groups of naturally occurring anthocyanins. Due to their high water
solubility, the glucosides are convenient analogs of natural pigments for systematic
investigations of the impact of substitution on color. Based on the preliminary study of two
regioisomers, it seems clear that the glucosidation site has a considerable influence on the
resulting color and its stability in aqueous solution. The detailed physico-chemical
investigation of these pigments is under way.
4. Experimental
4.1. Materials and instruments
All starting materials were obtained from commercial suppliers and were used without
purification.
4’-Hydroxyacetophenone,
4’-methoxyacetophenone,
2,4-dihydroxy-
benzaldehyde, 2-hydroxy-4-methoxybenzaldehyde and chlorotrimethylsilane were
purchased from Aldrich-Sigma (France). 2,4,6-Trihydroxybenzaldehyde was purchased
from Extrasynthese (France). 3’,4’-Dihydroxyacetophenone and 4-(2’,3’,4’,6’-tetra-O-
acetyl-β-D-glucopyranosyloxy)-2-hydroxybenzaldehyde (1) were prepared as already
reported.18 TLC analyses were performed on silica gel 60 F254 or C-18 silica gel F254s.
Detection was achieved by UV light (254 nm) and by charring after exposure to a 5%
16

ACCEPTED MANUSCRIPT
H2SO4 solution in EtOH. Purifications of intermediates were performed by column
chromatography on silica gel 60 (40-63 µm, from Merck). Pigments were purified by
elution on C18 silica gel cartridges (bond elut from Varian).
1H- and 13C-NMR spectra were recorded on an AscendTM400 Bruker apparatus at
400.18 MHz (1H) or 100.62 MHz (13C). Chemical shifts (
δ) are in ppm relative to
tetramethylsilane using the deuterium signal of the solvent (CDCl3, CD3OD) for
1
13
calibration. H-1H coupling constants (J) are in Hz. C-NMR signals of the glucosylated
pigments were assigned from their HSQC spectra.
UPLC-MS analyses were performed on an Acquity Ultra Performance LC™ apparatus
from Waters, equipped with an UV-visible diode array detector (DAD) and coupled with a
Bruker Daltonics HCT ultra ion trap mass spectrometer monitoring in the positive
electrospray ionization (ESI) mode. Separation was conducted on a 1.7 ꢀm (2.1 - 50 mm)
Acquity UPLC BEH C18 column thermostated at 30°C. Mass spectra were generated in
the Ultrascan mode in the m/z range 100 - 1000. The ion source parameters were: nebulizer
pressure = 50 psi, capillary voltage = 2 kV, drying gas flow = 10 L min-1, drying gas
temperature = 365°C. The mobile phase consisted of H2O / HCO2H (99:1, v/v) (eluent A)
and MeCN / H2O (60:40) + 1% HCO2H (eluent B) at a flow rate of 0.5 mL min-1. The
elution program was: 5 - 20% B (0 - 5 min), 20 - 100% B (5 - 10 min), 100 - 5% B (10 - 11
min), 5% B (11 - 14 min).
HR-MS analysis was carried out on Qstar Elite mass spectrometer (Applied Biosystems
SCIEX, Foster City, CA, USA). Mass detection was performed in the positive ESI mode.
UV-Vis absorption spectra were recorded on an Agilent 8453 diode array spectrometer
equipped with a magnetically stirred quartz cell (optical path length = 1 cm). The
temperature in the cell was controlled by means of a water-thermostated bath at 25°C.
4.2
Chemical synthesis
These syntheses prolong our previous work18 about the chemical synthesis of 3’,4’,7-
trihydroxyflavylium
chloride
(
P1)
and
3’,4’-dihydroxy-7-O-
β
-D-
glucopyranosyloxyflavylium chloride (P2).
17

ACCEPTED MANUSCRIPT
4.2.1. 4’-Hydroxy-7-O-
β
-D-glucopyranosyloxyflavylium chloride (P3
)
Equimolar amounts (1 mmol) of
1
and 4-hydroxacetophenone were dissolved in 5 mL
of AcOEt / MeOH (4/1 v/v) and cooled to 0°C. After addition of TMSCl (0.8 mL, 10
equiv.), the solution was stirred for 10 min and kept for 3 days at -18°C. The pigment was
precipitated by adding an access AcOEt. After filtration, the solid was dissolved in MeOH
(20 mL) under Ar and MeONa (8 equiv.) was added until pH 9 (wet pH paper). After
stirring for 3h at room temperature, 1 M HCl was added (until pH 1). The mixture was kept
for 12h at 4°C for complete cyclization, then concentrated under reduced pressure. The
residue was dissolved in 0.01 M HCl and loaded on a C18 cartridge. After elution with 100
mL of 0.01 M HCl to remove contaminating NaCl, the compound was eluted with 0.2 M
HCl in MeOH. After evaporation, the solid was crystallized in MeOH / AcOEt to give P3
as a dark red powder. Yield = 42%, Rf = 0.18 (BAW, i.e. n-BuOH / AcOH / H2O, 3/2/1,
1
v/v/v). H-NMR (CD3OD / TFA-d, 95/5):
δ = 9.13 (1H, d, J = 8.8, H4), 8.51 (2H, d, J =
8.7, H2’, H6’), 8.45 (1H, d, J = 8.8, H3), 8.20 (1H, d, J = 8.6, H5), 7.90 (1H, d, J = 2.2, H8),
7.60 (1H, dd, J = 8.6 and 2.2, H6), 7.12 (2H, d, J = 8.7, H3’, H5’), 5.34 (1H, d, J = 7.1, H1”),
3.98 (1H, broad d, J = 9.8, H6”B), 3.71 - 3.76 (2H, m, H5”, H6”A), 3.55-3.60 (2H, m, H2”,
13
H3”), 3.45 (1H, broad t, J = 9.4, H4”). C-NMR (CD3OD / TFA-d, 95/5):
δ = 175.1 (C2),
169.0 (C7), 167.3 (C9), 159.3 (C4’), 154.6 (C4), 134.9 (C2’, C6’), 133.1 (C5), 122.5 (C6),
121.1 (C1’), 121.0 (C10), 118.3(C3’, C5’), 115.3 (C3), 105.0 (C8), 101.9 (C1”), 78.8 (C5”), 77.8
(C3”), 74.6 (C2”), 71.3 (C4”), 62.6 (C6”). UV/VIS (0.01M HCl in MeOH): ε(465 nm) =
42800 M-1 cm-1. The purity of P3 was checked by UPLC-DAD-MS, tR = 2.4 min, λmax
455 nm, m/z (M+) = 400.9.
=
4.2.2. 4’-Methoxy-7-O-
β
-D-glucopyranosyloxyflavylium chloride (P4
)
Equimolar amounts (1 mmol) of
1
and 4-methoxyacetophenone were dissolved in 6 ml
of AcOEt/MeOH (1/1, v/v) and cooled to 0°C. After addition of TMSCl (0.8 mL, 10
equiv.), the mixture was stirred for 30 min and kept for 3 days at -18°C. The pigment was
precipitated by adding an access of AcOEt. After filtration, the same deacetylation
procedure as for P3 was carried out. P4 was obtained as a dark red powder after
1
crystallization in MeOH / AcOEt / Et2O. Yield = 23%, Rf = 0.37 (BAW). H-NMR
(CD3OD / TFA-d, 95/5):
δ
= 9.22 (1H, d, J = 8.8, H4), 8.60 (2H, d, J = 8.8, H2’, H6’), 8.45
18

ACCEPTED MANUSCRIPT
(1H, d, J = 8.8, H3), 8.53 (1H, d, J = 8.8, H5), 7.96 (1H, d, J = 2.2, H8), 7.65 (1H, dd, J =
8.8 and 2.2, H6), 7.31 (2H, d, J =8.8, H3’, H5’), 5.35 (1H, d, J = 7.3, H1”), 4.04 (3H, s,
OCH3), 3.98 (1H, broad d, J = 10.0, H6”B), 3.68 - 3.75 (2H, m, H5”, H6”A), 3.53 - 3.60 (2H,
m, H2”, H3”), 3.45 (1H, t, J = 9.0, H4”). 13C-NMR (CD3OD / TFA-d, 95/5):
δ = 175.1 (C2),
169.4 (C7), 167.7 (C9), 159.7 (C4’), 155.4 (C4), 133.3 (C2’, C6’), 133.2 (C5), 122.9 (C6),
122.5 (C1’), 121.4 (C10), 117.3, 117.1 (C3’, C5’), 115.4 (C3), 105.0 (C8), 102.0 (C1”), 78.8
(C5”), 77.9 (C3”), 74.6 (C2”), 71.2(C4”), 62.6 (C6”), 57.0 (OMe). UV/VIS (0.01 M HCl in
MeOH):
ε
(458 nm) = 35700 M-1 cm-1. The purity of P4 was checked by UPLC-DAD-MS,
max = 455 nm, m/z (M+) = 414.9.
tR = 3.8 min,
λ
4.2.3. 4-(2’,3’,4’,6’-Tetra-O-acetyl-
β-D-glucopyranosyloxy)acetophenone (3)
A solution of (1.68 g, 4.1 mmol) in distilled CH2Cl2 (5 mL) was added to a solution of
2
4-hydroxyacetophenone (1.49 g, 1.5 equiv.) and tris(2-(2-methoxyethoxy)ethyl)amine
(3.52 mL, 1.5 equiv.) in 1 M NaHCO3 / 1 M KCl (1:1) (pH 9, 5 mL). The mixture was
refluxed for 72h. After addition of H2O (10 mL) and extraction with CH2Cl2 (3x10 mL),
the combined organic phases were successively washed with 1 M HCl (2x10 mL), then
with H2O (2x10 mL), dried over Na2SO4 and concentrated. The syrupy residue was
purified on silica gel using cHex/ AcOEt (8/2 to 7.5/2.5, then 7/3, v/v). Finally,
crystallization in MeOH / H2O gave
AcOEt, 6/4, v/v). 1H-NMR (CDCl3):
3
as white crystals. Yield = 25%, Rf = 0.21 (cHex /
δ
= 7.87 (2H, d, J = 8.8, H2’, H6’), 6.96 (2H, d, J = 8.8,
H3’, H5’), 5.24 (2H, m, H3’’, H4’’), 5.12 (2H, m, H1’’, H2’’), 4.22 (1H, dd, J = 12.3 and 5.4,
H6’’B), 4.10 (1H, dd, J = 12.3 and 2.3,H6’’A), 3.85 (1H, m, H5’’), 2.50 (3H, s, OCH3), 1.99
(12H, broad s, 4Ac). 13C-NMR (CDCl3):
δ = 196.6 (C=O), 160.2 (C1’), 132.5 (C4’),
130.5(C2’, C6’), 116.3(C3’, C5’), 98.2 (C1’’), 72.6 (C3’’), 72.3 (C5”), 71.1 (C2’’), 68.2 (C4’’),
61.9 (C6’’), 20.7 (OMe), 20.6 (Me of Ac).
4.2.4. 4-O-
β
-D-Glucopyranosyloxyacetophenone (
4)
Compound
3
(320 mg, 0.69 mmol) was dissolved in dry MeOH (10 mL) and treated
with a catalytic amount of MeONa (pH 8-9 using wet pH paper). After stirring for 4 h,
total deacetylation was confirmed by TLC (cHex / AcOEt, 3/2, v/v).The mixture was then
acidified with ion exchange resin (Amberlite IRC50, H+ form) to pH 1-2 (wet pH paper),
filtered and evaporated. White crystals were obtained after crystallization in MeOH / Et2O.
1
Yield = 88%, Rf = 0.88 (cHex / AcOEt, 6/4, v/v). H NMR (CD3OD):
δ
= 8.98 (2H, d, J =
19

ACCEPTED MANUSCRIPT
8.9, H2’, H6’), 7.17 (2H, d, J = 8.9, H3’,H5’), 5.03 (1H, d, J = 7.5, H1’’), 3.90 (1H, dd, J =
12.0 and 2.2, H6’’B), 3.70 (1H, dd, J = 12.0 and 5.6, H6’’A), 3.48 - 3.50 (3H, m,H2’’,
13
H3’’,H5’’), 3.40 (1H, dd, J = 5.4 and 2.0, H4’’), 2.56 (3H, s, CH3). C-NMR (CD3OD):
δ =
199.4 (C=O), 163.1 (C1’), 132.7 (C4’), 131.6 (C2’, C6’), 117.3 (C3’, C5’), 101.6 (C1’’), 78.3
(C5’’), 77.3 (C3”), 74.8 (C2’’), 71.3 (C4’’), 62.5 (C6’’), 49.5 (OMe), 26.5 (Me of Ac).
4.2.5. 4’-O-
β
-D-Glucopyranosyloxy-7-hydroxyflavylium chloride (P5
)
Equimolar amounts (0.5 mmol) of
4
and 2,4-dihydroxybenzaldehyde were dissolved in
15 mL of AcOEt / MeOH (1/2, v/v). After addition of TMSCl (1.6 mL, 20 equiv.), the
solution was stirred for 30 min at room temperature. The pigment was precipitated by
adding AcOEt and isolated as a red powder after filtration. Yield = 77%, Rf = 0.14 (BAW).
1H-NMR (CD3OD / TFA-d, 95/5):
δ = 9.15 (1H, d, J= 8.9, H4), 8.48 (2H, d, J = 9.1, H2’,
H6’), 8.39 (1H, d, J = 8.9, H3), 8.19 (1H, d, J = 8.9, H5), 7.54 (1H, d, J = 2.2, H8), 7.42 (1H,
dd, J = 8.9 and 2.2, H6), 7.40 (2H, d, J = 9.1, H3’, H5’), 5.17 (1H, d, J = 7.2, H1’’), 3.93 (1H,
dd, J = 12.0 and 2.2, H6’’B), 3.72 (1H, dd, J = 12.0 and 5.8, H6’’A), 3.55-3.57 (3H, m, H2’’,
H3’’, H5’’), 3.45 (1H, t, J = 9.2, H4’’).13C-NMR (CD3OD / TFA-d, 95/5):
δ = 173.4 (C2),
171.0 (C7), 165.9 (C9), 160.9 (C4’), 155.4 (C4), 134.2 (C5), 133.1 (C2’, C6’), 124.1 (C1’),
123.0 (C6), 120.8 (C10), 119.1 (C3’, C5’), 113.7 (C3), 103.7 (C8), 101.5 (C1”), 78.6 (C5”), 77.9
(C3”), 74.7 (C2”), 71.2 (C4”), 62.5 (C6”). UV/VIS (0.01 M HCl in MeOH), ε(460 nm) =
37600 M-1 cm-1. The purity of P5 was checked by UPLC-DAD-MS: tR = 2.6 min, λmax
450 nm, m/z (M+) = 400.9.
=
4.2.6. 4’-O-
β
-D-Glucopyranosyloxy-7-methoxyflavylium chloride (P6
)
Equimolar amounts (1 mmol) of
4
and 2-hydroxy-4-methoxybenzaldehyde were
dissolved in 15 mL of AcOEt / MeOH (1/2, v/v). After addition of TMSCl (20 equiv.), the
mixture was stirred for 30 min at room temperature. P6 as a dark red powder was obtained
by simple filtration. Yield = 77%, Rf = 0.29 (BAW). 1H-NMR (CD3OD / TFA-d, 95/5):
δ =
9.22 (1H, d, J= 8.7, H4), 8.55 (2H, dd, J = 9.1, H2’, H6’), 8.50 (1H, d, J= 8.7, H3), 8.22 (1H,
d, J = 8.9, H5), 7.88 (1H, d, J = 2.2, H8), 7.56 (1H, dd, J = 8.9 and 2.2, H6), 7.42 (2H, d, J =
9.1, H3’, H5’), 5.18 (1H, d, J = 7.4, H1’’), 4.17 (3H, OCH3), 3.92 (1H, dd, J = 12.0 and 2.2,
H6’’B), 3.71 (1H, dd, J = 12.0 and 5.9, H6’’A), 3.56 - 3.60 (3H, m, H2’’, H3’’, H5’’), 3.53 (1H,
20

ACCEPTED MANUSCRIPT
t, J = 9.2, H4’’). 13C-NMR (CD3OD / TFA-d, 95/5):
δ
= 174.1 (C2), 171.2 (C7), 166.2 (C9),
160.9 (C4’), 155.7 (C4), 133.5 (C5), 133.2 (C2’, C6’), 124.0 (C1’), 123.1 (C6), 121.3 (C10),
119.2 (C3’, C5’), 114.8 (C3), 101.6 (C8), 101.5 (C1”), 78.6 (C5”), 77.9 (C3”), 74.7 (C2”), 71.2
(C4”), 62.5 (C6”), 58.1 (OMe). UV/VIS (0.01 M HCl in MeOH):
ε
(455 nm) = 25400 M-1
cm-1. The purity of P6 was checked by UPLC-DAD-MS: tR = 3.4 min,
λ
max = 450 nm, m/z
(M+) = 414.9.
4.2.7. 4’,7-Dihydroxyflavylium chloride (P7
)
Equimolar amounts (1 mmol) of 2,4-dihydroxybenzaldehyde and 4-
hydroxyacetophenone were dissolved in 5 mL of AcOEt / MeOH (4/1, v/v). After cooling
to 0°C, TMSCl (10 equiv.) was added. The mixture was stirred for 10 min, then kept at
4°C for 3 days. Precipitation of pigment was achieved by adding AcOEt. After filtration,
P7 was obtained as a dark red powder. Yield = 45%, Rf = 0.52 (ABFW, i.e. AcOH / n-
1
BuOH / HCO2H / H2O, 20/2/1/1, v/v/v). H-NMR (CD3OD / TFA-d, 95/5):
δ = 9.04 (1H,
d, J = 8.8, H4), 8.42 (2H, dd, J = 8.9 and 2.0, H2’, H6’), 8.30 (1H, d, J = 8.8, H3), 8.13 (1H,
d, J = 8.9, H5), 7.49 (1H, d, J = 1.9, H8), 7.40 (1H, broad d, J = 8.9, H6), 7.10 (2H, dd, J =
8.9 and 2.0, H3’, H5’). 13C-NMR (CD3OD / TFA-d, 95/5):
(C9), 160.2 (C4’), 154.4 (C4), 134 (C2’, C6’), 133.9 (C5), 122.3 (C6), 121.2 (C1’), 120.0 (C10),
118.6 (C3’, C5’), 113.4 (C3), 103.7 (C8). UV/VIS (0.01 M HCl in MeOH): (470 nm) =
δ = 173.8 (C2), 170.2 (C7), 168.2
ε
59700 M-1 cm-1.The purity of P7 was checked by UPLC-DAD-MS: tR = 4.8 min, λmax
465 nm, m/z (M+) = 238.7.
=
4.2.8. 7-Hydroxy-4’-methoxyflavylium chloride (P8
)
Equimolar amounts (1 mmol) of 2,4-dihydroxybenzaldehyde and 4-
methoxyacetophenone were reacted in the same conditions as for P7. P8 was obtained as a
1
dark red powder. Yield = 90%, Rf = 0.18 (EBFW). H-NMR (CD3OD / TFA-d, 95/5):
δ =
9.10 (1H, d, J = 8.7, H4), 8.47 (2H, dd, J = 7.6 and 2.2, H2’, H6’), 8.35 (1H, d, J = 8.7, H3),
8.15 (1H, d, J = 9.0, H5), 7.51 (1H, d, J = 2.2, H8), 7.42 (1H, dd, J = 9.0 and 2.2, H6), 7.25
13
(2H, dd, J = 7.6 and 2.0, H3’, H5’), 3.99 (3H, OCH3). C-NMR (CD3OD / TFA-d, 95/5):
δ
= 173.5 (C2), 170.7 (C7), 168.6 (C9), 160.5 (C4’), 154.9 (C4), 134.0 (C5), 133.4 (C2’, C6’),
122.7 (C6), 122.5 (C1’), 120.4 (C10), 117.1 (C3’, C5’), 113.5 (C3), 103.7 (C8), 56.8 (OCH3).
21

ACCEPTED MANUSCRIPT
UV/VIS (0.01 M HCl in MeOH):
ε
(467 nm) = 57500 M-1 cm-1. The purity of P8 was
checked by UPLC-DAD-MS: tR = 5.9 min, λmax = 465 nm, m/z (M+) = 252.7.
4.2.9. 4’-Hydroxy-7-methoxyflavylium chloride (P9
)
Equimolar amounts (1 mmol) of 2-hydroxy-4-methoxybenzaldehyde and 4-
hydroxyacetophenone were reacted in the same conditions as for P7. P9 was obtained as a
1
dark red powder. Yield = 85%, Rf = 0.18 (EBFW). H-NMR (CD3OD / TFA-d, 95/5):
δ =
9.10 (1H, d, J = 8.8, H4), 8.46 (2H, dd, J = 7.5 and 2.0, H2’, H6’), 8.39 (1H, d, J = 8.8, H3),
8.15 (1H, d, J = 9.0, H5), 7.8 (1H, d, J = 2.2, H8), 7.49 (1H, broad d, J = 9.0, H6), 7.1 (2H,
13
dd, J = 7.5 and 2.0, H3’, H5’), 4.15 (3H, s, OCH3). C-NMR (CD3OD / TFA-d, 95/5):
δ =
174.3 (C2), 170.4 (C7), 168.6 (C9), 160.2 (C4’), 154.4 (C4), 134.4 (C2’, C6’), 133 (C5), 121.1
(C1’), 122.3 (C6), 120.5 (C10), 118.8 (C3’, C5’), 114.4 (C3), 101.5 (C8), 58.0 (OMe). UV/VIS
(0.01 M HCl in MeOH):
ε
(467 nm) = 50400 M-1 cm-1. The purity of P9 was checked by
max = 465 nm, m/z (M+) = 252.7.
UPLC-DAD-MS: tR= 6.1 min,
λ
4.2.10. 4’,5,7-Trihydroxyflavylium chloride (apigeninidin)
A solution of 2,4,6-trihydroxybenzaldehyde (1 mmol) and 4-hydroxyacetophenone (2
equiv.) in 3 mL of AcOEt / MeOH (2/1, v/v) was cooled to 0°C and TMSCl (20 equiv.)
was added. After 60 min of stirring, complete precipitation of APN was achieved.
Crystallization in MeOH / AcOEt afforded APN as a red powder. Yield = 34%, Rf = 0.83
(BAW). 1H-NMR (CD3OD / TFA-d, 95/5):
8.9 and 2.0, H2’, H6’), 8.05 (1H, d, J = 8.7, H3), 7.08 (2H, dd, J = 8.9 and 2.0, H3’, H5’), 6.95
(1H, d, J = 2.1, H8), 6.66 (1H, d, J = 2.1, H6). 13C-NMR (CD3OD / TFA-d, 95/5):
= 172.9
δ = 9.10 (1H, d, J = 8.7, H4), 8.33 (2H, dd, J =
δ
(C2), 172.3 (C7), 167.3 (C9), 160.4 (C5), 160.1 (C4’), 149.6 (C4), 133.2 (C2’, C6’), 121.4
(C1’), 118.4 (C3’, C5’), 113.8 (C10), 110.4 (C6), 103.1 (C3), 96.1 (C8). UV/VIS (0.01 M HCl
in MeOH): ε
(480 nm) = 38400 M-1 cm-1.The purity of APN was checked by UPLC-DAD-
MS: tR = 5.7 min, λmax = 475 nm, m/z (M+) = 254.7. HRMS-ESI: m/z(M+) calcd =
255.0652, found = 255.0651.
4.2.11. 3’,4’,5,7-Tetrahydroxyflavylium chloride (luteolinidin)
Equimolar amounts (1 mmol) of 2,4,6-trihydroxybenzaldehyde and 3,4-
dihydroxyacetophenone were dissolved in 3 mL of AcOEt / MeOH (2/1, v/v). After
cooling to 0°C, TMSCl (20 equiv.) was added and the solution stirred for 30 min at 0°C.
22

ACCEPTED MANUSCRIPT
LTN was totally precipitated by adding AcOEt. After filtration, LTN was isolated as a red
powder (yield = 65%). To remove the yellow xanthylium contaminant (ca. 5%), LTN was
1
crystallized in EtOH / AcOEt. Rf = 0.83 (BAW). H-NMR (CD3OD / TFA-d, 95/5):
δ =
9.02 (1H, d, J = 8.7, H4), 7.97 (1H, d, J = 8.7, H3), 7.88 (1H, dd, J = 8.7 and 2.3, H6’), 7.73
(1H, d, J = 2.3, H2’), 7.04 (1H, d, J = 8.7, H5’), 6.89 (1H, d, J = 2.1, H8), 6.62 (1H, d, J =
13
2.1, H6). C-NMR (CD3OD / TFA-d, 95/5):
δ = 173.6 (C2), 173.1 (C7), 161.3 (C9), 160.8
(C5), 157.2 (C4’), 149.9 (C4), 149.1 (C3’), 126.1 (C6’), 122.6 (C1’), 118.7 (C5’), 116.8 (C2’),
114.5 (C10), 111.4 (C6), 104.0 (C3), 96.9 (C8). UV/VIS (0.01 M HCl in MeOH): (500 nm)
ε
= 24000 M-1 cm-1. The purity of LTN was checked by UPLC-DAD-MS: tR = 4.9 min, λmax
= 485 nm, m/z (M+) = 270.7. HRMS-ESI: m/z(M+) calcd=271.0601, found = 271.0603.
4.2.12. 4-Benzyloxy-2,6-dihydroxybenzaldehyde (5) and 2-benzyloxy-4,6-dihydroxy-
benzaldehyde (5’
)
A solution of 2,4,6-trihydroxybenzaldehyde (3 g, 19.5 mmol) and dry K2CO3 (2.69 g, 1
equiv.) in 20 mL of anhydrous DMF was cooled to 0°C and stirred for 15 min. Then,
benzyl bromide (2.31 mL, 1 equiv.) was added drop-wise and the mixture was kept at 0°C
under vigorous stirring during 2.5h. Glacial 0.1 M HCl (100 mL) was added and the
aqueous phase extracted with AcOEt (3×100 mL). The combined organic phases were
dried over Na2SO4, then concentrated. The crude product was purified by column
chromatography on silica gel (cHex / AcOEt, 7/3, v/v) to afford the mixture of two
regioisomers
5
and 5’ (1/4 molar ratio) as a white amorphous powder. Total yield = 45%,
1
Rf = 0.57 (cHex / AcOEt, 6/4, v/v).
5
. H-NMR(CD3OD):
δ
= 10.05 (1H, s, CHO), 7.41
= 191.1
(CHO), 167.5 (C2, C6), 163.5 (C4), 136.2 (1CPh), 128.3, 128.2, 127.2 (5CPh), 95.1 (C3, C5),
(5H, m, Ph), 5.97 (2H, s, H3, H5), 5.09 (2H, s, CH2). 13C-NMR (CD3OD):
δ
70.2 (CH2). 5’. 1H-NMR (CD3OD):
δ
= 10.08 (1H, s, CHO), 7.41 (5H, m, Ph), 6.01 (1H, d,
13
J = 2.0, H5), 5.86 (1H, d, J = 2.0, H3), 5.07 (2H, s, CH2). C-NMR (CD3OD):
δ = 191.9
(CHO), 165.9 (C6), 164.2 (C4), 163.5 (C2), 136.3 (1CPh), 128.2, 127.9, 127.8 (5CPh), 105.6
(C1), 93.3 (C5), 91.8 (C3), 69.8 (CH2).
4.2.13.
benzaldehyde (
A solution of tetra-O-acetyl-α-D-glucopyranosylbromide (
distilled CH2Cl2 (10 mL) was added to a solution of and 5’ (1 g, 3.28 mmol) and tris[2-
2-Benzyloxy-6-hydroxy-4-(2’,3’,4’,6’-tetra-O-acetyl-β-D-glucopyranosyloxy)-
6
)
2
, 2.03 g, 1.5 equiv.) in
5
(2-methoxyethoxy)ethyl]amine (TMEA) (1.58 mL, 1.5 equiv.) in a saturated aqueous
K2CO3 solution (10 mL). The mixture was refluxed for 18 h. After the same treatment as
23

ACCEPTED MANUSCRIPT
for
3
, the syrupy residue was purified on silica gel (cHex / AcOEt, 65/35, v/v), then
crystallized in CH2Cl2 / Et2O to afford
6 as white crystals. Yield = 49%, Rf = 0.52 (cHex /
AcOEt, 6/4, v/v). 1H-NMR(CDCl3):
δ
= 12.36 (1H, s, OH), 10.19 (2H, s, CHO), 7.39 (5H,
m, Ph), 6.59 (1H, d, J = 2.2, H5), 6.54 (1H, d, J = 2.2, H3), 5.10 (2H, s, CH2), 5.06-5.28
(4H, m, H1’, H2’, H3’, H4’), 4.15 - 4.22 (2H, 2dd, J = 12.1 and 6.1, J = 12.1 and 2.1,H6’A,
H6’B), 3.87 (1H, m, H5’), 2.02 - 2.07 (12H, m, 4Ac). 13C-NMR (CDCl3):
δ = 192.5 (CHO),
169.6-170.2 (4 C=O), 166.1 (C2), 165.9 (C6), 164.3 (C4), 135.5 (1CPh), 128.8, 128.5, 128.1,
127.9, 127.1 (5CPh), 107.3 (C1), 95.7 (C1’), 97.5 (C5), 93.3 (C3), 72.6 (C3’), 72.4 (C5’), 70.9
(C2’), 70.7 (CH2), 68.2 (C4’), 61.9 (C6’), 20.6 (4Me).
4.2.14. 2,6-Dihydroxy-4-(2’,3’,4’,6’-tetra-O-acetyl-β-D-glucopyranosyloxy)benzaldehyde
(7)
Compound
6 (1 g, 1.74 mmol) was dissolved in 10 mL of THF / MeOH (4/1, v/v) and
10% Pd on activated charcoal (0.124 g, 0.12 mmol) was added. Debenzylation was carried
out in a hydrogenation reactor under 2 bars of H2 for 30 min. The mixture was diluted in
AcOEt (30 mL), filtered over celite, concentrated and purified on silica gel (AcOEt / cHex,
45/55, v/v). After crystallization in AcOEt / n-hexane,
Yield = 90%, Rf = 0.27 (cHex / AcOEt, 6/4, v/v). 1H-NMR (CDCl3):
7
was obtained as white crystals.
= 12.36 (2H, s, OH),
δ
10.19 (1H, s, CHO), 6.01 (2H, s, H3, H5), 5.28 - 5.33 (2H, m, H3’, H4’), 5.13 - 5.20 (2H, m,
H1’, H2’), 4.18 - 4.33 (2H, 2dd, J = 11.0 and 5.7, 11.0 and 2.0, H6’A, H6’B), 3.92 (1H, m,
H5’), 2.07 - 2.13 (12H, m, 4Ac). 13C-NMR (CDCl3):
δ = 192.5 (CHO), 169.8 - 171.3
(4C=O), 164.2 (C2, C4, C6), 106.8 (C1), 97.4 (C1’), 95.6 (C3, C5), 72.6 (C3’), 72.3 (C5’), 71.0
(C2’), 68.3 (C4’), 62.0 (C6’), 20.6 (4Me).
2.2.15.
7-(2’’,3’’,4’’,6’’-Tetra-O-acetyl-β-D-glucopyranosyloxy)-3’,4’,5-trihydroxy-
flavylium chloride (
8
)
Condensation between equimolar quantities (1 mmol) of
7
and 3,4-
dihydroxyacetophenone in the presence of TMSCl (20 equiv.) was achieved in 5 mL of
AcOEt / MeOH (4/1, v/v). After 30 min at 0°C,
8 was isolated by filtration as a red
1
powder. Yield = 60%. H-NMR (CD3OD):
δ = 9.13 (1H, d, J = 8.6, H4), 8.20 (1H, d, J =
8.6, H3), 8.00 (1H, d, J = 8.9 and 1.9, H6’), 7.84 (1H, d, J = 1.9, H2’), 7.31 (1H, broad s,
H8), 7.06 (1H, d, J = 8.9, H5’), 6.73 (1H, broad s, H6), 5.73 (1H, d, J = 7.8, H1’’), 5.46 (1H,
t, J = 9.3, H3’’), 5.27 (1H, dd, J = 9.3 and 7.8, H2’’), 5.18 (1H, t, J = 7.8, H4’’), 4.22 - 4.36
(3H, m, H5’’, H6’’A, H6’’B), 2.07 - 2.13 (12H, m, 4 Me).
24

ACCEPTED MANUSCRIPT
2.2.16. 5-(
β
-D-glucopyranosyloxy)-3’,4’,7-trihydroxyflavylium chloride (LTN-5Glc) and 7-
(β
-D-glucopyranosyloxy)-3’,4’,5-trihydroxyflavylium chloride (LTN-7Glc).
The same deacetylation procedure as for P3 was carried out. UPLC-MS analysis of the
red powder obtained showed a mixture of two deprotected isomers and several less polar
partially acetylated pigments. Purification over TSK gel with elution by 1% aqueous HCl
afforded the two inseparable deprotected isomers. Overall yield (deprotection +
purification) = 10%. The purity of the LTN-Glc mixture was checked by UPLC-DAD-MS.
LTN-7Glc (60%): tR = 1.85 min,
1.90 min,
max = 485 nm, m/z (M+) = 433.1. LTN-7Glc (major). 1H-NMR (CD3OD / TFA-
d, 95/5): = 9.15 (1H, d, J = 8.8, H4), 8.10 (1H, d, J = 8.8, H3), 7.91 (1H, dd, J = 9.2 and
λ
max = 485 nm, m/z (M+) = 433.1. LTN-5Glc (40%): tR =
λ
δ
2.2, H6’), 7.74 (1H, d, J = 2.2, H2’), 7.06 (1H, broad s, H8), 7.01 (1H, d, J = 9.2, H5’), 7.00
(1H, d, J = 1.5, H6), 5.18 (1H, d, J = 7.2, H1”), 3.98 (1H, broad d, J = 11.9, H6”B), 3.76 (1H,
dd, J = 11.9 and 5.4, H6”A), 3.64 (1H, m, H5”), 3.51 - 3.61 (2H, m, H2”,H3”), 3.43 (1H, t, J =
9.4, H4”). 13C-NMR (CD3OD / TFA-d, 95/5):
δ = 171.4 (C2), 169.1 (C7), 158.1 (C9), 157.5
(C5), 157.1 (C4’), 149.1 (C4), 148.4 (C3’), 126.0 (C6’), 121.5 (C1’), 117.8 (C5’), 116.2 (C2’),
113.5 (C10), 111.9 (C3), 105.0 (C6), 101.8 (C1’’), 98.1 (C8), 78.7 (C3’’), 77.8 (C5’’), 74.6
1
(C2’’), 71.2 (C4’’), 62.5 (C6’’). LTN-5Glc (minor). H-NMR (CD3OD / TFA-d, 95/5):
δ =
9.09 (1H, d, J = 8.8, H4), 8.14 (1H, d, J = 8.8, H3), 7.96 (1H, dd, J = 9.2 and 2.2, H6’), 7.79
(1H, d, J = 2.2, H2’), 7.28 (1H, broad s, H8), 7.03 (1H, d, J = 9.2, H5’), 6.90 (1H, d, J = 1.5,
H6), 5.24 (1H, d, J = 7.2, H1”), 3.98 (1H, broad d, J = 11.9, H6”B), 3.76 (1H, dd, J = 11.9
and 5.4, H6”A), 3.64 (1H, m, H5”), 3.51 - 3.61 (2H, m, H2”,H3”), 3.43 (1H, t, J = 9.4, H4”).
13C-NMR (CD3OD / TFA-d, 95/5):
δ = 173.3 (C7), 171.4 (C2), 159.6 (C5), 158.1 (C9),
157.1 (C4’), 149.5 (C4), 149.0 (C3’), 126.5 (C6’), 121.5 (C1’), 118.0 (C5’), 116.5 (C2’), 114.0
(C10), 112.7 (C3), 103.8 (C6), 101.8 (C1’’), 97.7 (C8), 78.7 (C3’’), 77.7 (C5’’), 74.6 (C2’’),
71.1 (C4’’), 62.4 (C6’’).
5. Supplementary Information
1H-NMR and UPLC-DAD-MS analyses of all pigments.
25

ACCEPTED MANUSCRIPT
6. References
1.
2.
Cavalcanti, R. N.; Santos, D. T.; Meireles, M. A. A. Food Res. Int. 2011, 44, 499-
509.
McCann, D.; Barrett, A.; Cooper, A.; Crumpler, D.; Dalen, L.; Grimshaw, K.;
Kitchin, E.; Lok, K.; Porteous, L.; Prince, E.; Sonuga-Barke, E.; Warner, J. O.;
Stevenson, J. The Lancet 2007, 370, 1560-1567.
3.
4.
5.
Yoshida, K.; Mori, M.; Kondo, T. Nat. Prod. Rep. 2009, 26, 857-964.
Kähkönen, M. P.; Heinonen, M. J. J. Agric. Food Chem. 2003, 51, 628-633.
Goupy, P.; Bautista-Ortin, A. B.; Fulcrand, H.; Dangles, O. J. Agric. Food Chem.
2009, 57, 5762-5770.
6.
Fleschhut, J.; Kratzer, F.; Rechkemmer, G.; Kulling, S. E. Eur. J. Nutr. 2006, 45, 7-
18.
7.
8.
Sadilova, E.; Carle, R.; Stintzing, F. C. Mol. Nutr. Food Res. 2007, 51, 1461-1471.
Taylor, J. R. N.; Belton, P. S.; Beta, T.; Duodu, K. G. J. Cereal Sci. 2014, 59, 257-
275.
9.
Petti, C.; Kushwaha, R.; Tateno, M.; Harman-Ware, A. E.; Crocker, M.; Awika, J.;
DeBolt, S. J. Agric. Food Chem. 2014, 62, 1227-1232.
10. Mazza, G.; Brouillard, R. J. Agric. Food Chem. 1987, 35, 422–426.
11. Pina, F.; Melo, M. J.; Laia, C. A. T.; Parola, A. J. Lima, J. C. Chem. Soc. Rev. 2012
41, 869-908.
,
12. Yang, L.; Dykes, L.; Awika, J. M. Food Chem. 2014, 160, 246-254.
13. Iwashina, T.; Kitajima, J.; Matsumoto, S. Bull. Natl. Mus. Sci. 2010, 36, 61-64.
14. Tsuda, T. Mol. Nutr. Food Res. 2012, 56, 159-170.
15. Del Rio, D.; Rodriguez-Mateos, A.; Spencer, J. P. E.; Tognolini, M.; Borges, G.;
Crozier, A. Antioxid Redox Signal. 2013, 18, 1818-1892.
16. Sweeny, J.G.; Iacobucci, G.A. Tetrahedron 1981, 37, 1481-1483.
26

ACCEPTED MANUSCRIPT
17. Carbonneau, M. A.; Cisse, M.; Mora-Soumille, N.; Dairi, S.; Rosa, M.; Michel, F.;
Lauret, C.; Cristol, J. P.; Dangles, O. Food Chem. 2014, 145, 701-709.
18 Mora-Soumille, N.; Al Bittar, S.; Rosa, M.; Dangles, O. Dyes Pigments 2013, 96, 7-15.
19. El Hajji, H.; Dangles, O.; Figueiredo, P.; Brouillard, R. Helv. Chim. Acta 1997, 80,
398-413.
20. Bjoroy, O.; Rayyan, S.; Fossen, T.; Kalberg, K.; Andersen, O. M. Phytochemistry
2009, 70, 278-287.
21. Sousa, A.; Mateus, N.; De Freitas, V. Tetrahedron Lett. 2012, 53, 1300-1303.
22.
Kueny-Stotz, M.; Isorez, G.; Chassaing, S.; Brouillard, R. Synlett 2007, 1223-1226.
23. Chassaing, S.; Kueny-Stotz, M.; Isorez, G.; Brouillard, R. Eur. J. Org. Chem. 2007
,
2438-2448.
24. Dangles, O.; El Hajji, H. Helv. Chim. Acta 1994, 77, 1595-1610.
25 Dangles, O.; Brouillard, R. New J. Chem. 1994, 18, 287-296.
26. Cruz, L.; Mateus, N.; De Freitas, V. Tetrahedron Lett. 2013, 54, 2865-2869.
27. Pina, F. J. Agric. Food Chem. 2014, 62, 6885–6897.
28. Petrov, V.; Gavara, R.; Dangles, O.; Al Bittar, S.; Mora-Soumille, N.; Pina, F.
Photochem. Photobiol. Sci. 2013, 12, 576–581.
27