Protecting-group-free solid-phase anchoring of polyphenolic C-glucosidic ellagitannins and synthesis of 1-alkylamino-vescalagin derivatives

Article abstract of DOI:10.1002/ejoc.201402444

The C-glucosidic ellagitannins vescalagin and vescalin are among the plant polyphenols that have been shown to interact with protein targets in a specific manner. This work describes a protecting-group-free method to immobilize these protein-binding natural products on a solid support, and a functionalizing cleavage method, which is adaptable to the solution-phase synthesis of vescal(ag)in derivatives, and which advantageously permitted the preparation of 1-alkylamino derivatives of vescalagin.

Full text of DOI:10.1002/ejoc.201402444

FULL PAPER
DOI: 10.1002/ejoc.201402444
Protecting-Group-Free Solid-Phase Anchoring of Polyphenolic C-Glucosidic
Ellagitannins and Synthesis of 1-Alkylamino-Vescalagin Derivatives
Céline Douat,[a] Emanuela Berni,[a] Rémi Jacquet,[a] Laurent Pouységu,[a] Denis Deffieux,[a]
and Stéphane Quideau*[a]
Keywords: Natural products / Polyphenols / Ellagitannins / Solid-phase synthesis / Amines
The C-glucosidic ellagitannins vescalagin and vescalin are
among the plant polyphenols that have been shown to inter-
act with protein targets in a specific manner. This work de-
scribes a protecting-group-free method to immobilize these
protein-binding natural products on a solid support, and a
functionalizing cleavage method, which is adaptable to the
solution-phase synthesis of vescal(ag)in derivatives, and
which advantageously permitted the preparation of 1-alk-
ylamino derivatives of vescalagin.
annin subclass such as vescalagin (1β) and its congener ves-
calin (2β; Figure 1)[5] strongly bind to and inhibit human
DNA topoisomerase IIα, a target of anticancer drugs such
Introduction
In recent years, plant polyphenols[1] have received a great
deal of attention from the scientific community because of
their antioxidant and presumed preventive actions against
cancer, cardiovascular and neurodegenerative diseases, and
their occurrence in plant-derived foods and beverages, as
well as in traditional herbal medicines. Plant polyphenols
are well-known for their capacity to interact with proteins
but have long been relegated to molecules only capable of
engaging proteins in nonspecific complexation processes
that lead to precipitation phenomena.[1] For example, this
so-called “tanning action” underlies the cross-linking of
collagen molecules during the conversion of animal skins
into leather,[1] as well as the precipitation of salivary pro-
teins at the origin of the mouth feeling of astringency.[1,2]
However, there is compelling evidence from numerous re-
searches that plant polyphenolic molecules can also affect
cellular biochemical processes by interacting in more spe-
cific manners with various types of proteinaceous targets
(e.g., enzymes, cell receptors, transcription factors, struc-
tural proteins, or peptides).[1b] Ellagitannins are among
those plant polyphenols that can express remarkable bio-
logical activities, including antiviral and antitumor ef-
fects.[3,4] In particular, members of the C-glucosidic ellagit-
[a] Université de Bordeaux, Institut des Sciences Moléculaires
(UMR-CNRS 5255) and Institut Européen de Chimie et
Biologie (IECB),
2 rue Robert Escarpit, 33607 Pessac cedex, France
E-mail: s.quideau@iecb.u-bordeaux.fr
Figure 1. Structures of the two protein-binding C-glucosidic ellagit-
annins vescalagin (1β) and vescalin (2β) and their C-1 epimers cas-
talagin (1α) and castalin (2α).
http://www.u-bordeaux.fr/
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201402444.
Eur. J. Org. Chem. 2014, 4963–4972
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4963

S. Quideau et al.
FULL PAPER
as doxorubicin and etoposide.[6] We have also recently re-
ported that 1β can bind to cellular actin filaments (F-actin)
and dismantle the supramolecular association of this struc-
tural protein into fibrillar aggregates.[7]
Results and Discussion
The starting (–)-vescalagin (1β) was isolated in several
hundred milligrams from oak heartwood, and (–)-vescalin
From a structural point of view, 1β adopts a globular (2β) was generated by acidic hydrolysis of 1β, as described
shape and features an open-chain glucose core triply con- previously.[6c,9] Our earlier work on these C-glucosidic ellag-
nected at positions O-2, O-3, and O-5 to an atropisomeric itannins revealed that their benzylic secondary alcohol
terarylic nonahydroxyterphenoyl (NHTP) unit, which is ad- function at C-1 can be readily and cleanly substituted by
ditionally C–C linked to C-1 of the glucose core (Figure 1). various carbon-, oxygen-, and sulfur-based nucleophiles un-
Vescalagin (1β) also bears an atropisomeric biarylic hexa- der mild acidic conditions without any prior protection of
hydroxydiphenoyl (HHDP) unit doubly connected at the their phenolic and aliphatic hydroxy groups.[6b–6d] Thus, we
glucose positions O-4 and O-6, whereas this HHDP unit is decided to exploit this windfall chemoselectivity to attach
lacking in the structure of 2β. These unique structural fea- such polyphenolic natural products to a resin support. As
tures prompted us to contemplate the use of these protein- thiols appeared to be particularly efficient for the formation
binding C-glucosidic ellagitannins as scaffolds for the prep- of thioethers from either 1β or 2β by such a nucleophilic
aration of new bioactive derivatives along the lines of the substitution reaction,[6b,6c,7] we chose to prepare a resin
biology-oriented synthesis (BIOS) concept.[8] In this vein of equipped with thiol groups. Moreover, the selection of the
investigation, we first wanted to identify appropriate chemi- type of polymer composing the resin core was also of cru-
cal means to immobilize these polyphenolic natural prod- cial importance for convenient monitoring of the reaction.
ucts on a support for solid-phase (parallel) synthesis of de- In this regard, we planned to rely on the high-resolution
rivatives by taking advantage of the chemoselective reactiv- magic-angle spinning (HRMAS) NMR spectroscopic tech-
ity expressed at their benzylic secondary alcoholic C-1 posi- nique, which has emerged as a powerful tool for the in situ
tion.[6b–6d,7]
characterization of resin-bound compounds.[10] Thus, the
We wish to report herein a convenient protecting-group- resin had to be made of a polymeric matrix material with
free immobilization method to anchor both 1β and 2β on a 1H NMR resonances that would not interfere with the attri-
(biocompatible) polyethylene glycol polyamide (PEGA) bution of the ellagitannin signals and at the same time pres-
resin, together with a functionalizing cleavage method, ent good swelling and mobility properties to improve the
1
which offers a practical solution that is adaptable to both overall quality of the NMR spectra.[11] The PEG H NMR
solid- and solution-phase syntheses for the preparation of signals of the amine-terminated polyethylene glycol poly-
1-amino derivatives of vescalagin.
amide (PEGA) resin[12] do not interfere with those of the
Scheme 1. Preparation of the thiol resin 5 and its use in the synthesis of the resin-bound ellagitannins 6 and 7. Reactions conditions:
(i) 6-bromohexanoyl chloride, DIPEA, CH2Cl2, r.t., 1 h; (ii) KSCOCH3, DMF, r.t., overnight; (iii) LiBH4, THF, r.t., 2ϫ 3 h; (iv) 1β, THF/
TFA (1.5 vol.-%), 50 °C, 4 h; (v) 2β, THF/TFA (1.5 vol.-%), 50 °C, 4 h.
4964
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 4963–4972

Protecting-Group-Free Solid-Phase Anchoring of Ellagitannins
glucose core of 1β and 2β, and PEGA also possesses good was present anymore on the resin. The resin-bound vescala-
swelling and solvation properties in a wide range of sol- gin derivative 6 was then characterized by 1H HRMAS
vents, including aqueous media; hence, it was selected as analysis, and the resulting spectrum was compared with
the solid support.
that of the starting resin 5 (Figure 2). An inspection of the
The installation of thiol groups on this PEGA resin was 1H HRMAS spectrum of 6 swollen in (CD3)2SO ([D6]-
inspired from related reports on the preparation of thiol DMSO) revealed the presence of three sharp singlets be-
resins[13] and was performed as depicted in Scheme 1. We tween δ = 6.4 and 6.8 ppm. These signals are assigned to
chose to append a six-carbon aliphatic linker between the the aromatic protons of the HHDP (2 H) and NHTP (1 H)
resin terminal amino groups and the thiol functions to en- units of 1β and, thus, confirm its attachment onto the resin.
hance the accessibility of the ellagitannin molecules to these We have no direct evidence of the β-orientation of the car-
reactive sites. This was achieved by an acylation reaction bon–sulfur bond at C-1, but we assume this configuration
with 6-bromohexanoyl chloride. The resulting bromide to be correct on the basis of the retention of configuration
resin 3 was then treated with potassium thioacetate to give systematically observed in related nucleophilic substitution
the thioacetate resin 4, which was finally reduced with reactions with 1β (or 2β) and different carbon-, oxygen-,
LiBH4 to furnish the desired thiol resin 5.
This resin 5 was then allowed to react with either 1β or tions.[6b–6d,7]
2β (1 equiv. each relative to the resin loading) under the The same procedure was applied to immobilize 2β onto
and sulfur-based nucleophiles under similar reaction condi-
1
same acidic conditions as those used for analogous conden- the thiol resin 5. Analysis of the resulting material by H
sation reactions in solution (Scheme 1).[6b,6c,7] The progress HRMAS again confirmed the successful loading of 2β onto
of these immobilization reactions was initially monitored this solid support; this time, only one aromatic singlet at δ
by reversed-phase high-performance liquid chromatography = 6.55 ppm was observed for the resonance of the NHTP
(RP-HPLC) analysis of the supernatant, which indicated a aromatic proton of 2β (see Supporting Information). Ini-
complete consumption of vescalagin (1β) after 4 h. The tially, we wanted to prepare this resin-bound vescalin deriv-
completion of the vescalagin immobilization was also veri- ative 7 by hydrolyzing the HHDP unit of the resin-bound
fied by an Ellman’s test, which was negative as no free thiol vescalagin derivative 6 (Scheme 1). This attractive option
1
Figure 2. H HRMAS NMR spectra (with presaturation of the PEG resonances at 3.5 ppm) of (a) the thioacetate resin 4, (b) the free
thiol resin 5, and (c) the resin-bound vescalagin 6 in [D6]DMSO. The signals of the thioacetate resin intermediate 4 are depicted in red
for the acetate moiety, in green for the S-methylene group, and in dark violet for the methylene group α to the amide function of the
resin. The signals of the three aromatic protons of the resin-bound vescalagin 6 are marked with red stars.
Eur. J. Org. Chem. 2014, 4963–4972
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4965

S. Quideau et al.
FULL PAPER
could have enabled us to bypass the solution-phase hemi- of 1β or 2β from the resin but also allows the simultaneous
synthesis of 2β from 1β and the chromatographic separation introduction of a new substituent at the C-1 point of attach-
of the ellagic acid derived from the released HHDP unit.[6c] ment. As pointed out by Oki and Kobayashi,[15] the cleav-
Unfortunately, our attempts to apply solution-phase acidic age of a thioether C–S bond by a sulfenyl chloride first
hydrolysis conditions [e.g., 0.1 m HCl (5 vol.-%) in trifluoro- generates a reactive thiosulfonium ion intermediate that can
acetic acid/tetrahydrofuran (TFA/THF 1:1 v/v), 40 °C] to lead to the formation of a carbenium ion, which can then
the resin-bound vescalagin derivative 6 failed to cleanly de- either be trapped by a nucleophile (Nu, e.g., the sulfenyl-
liver 7. The rate of hydrolysis was much slower than that in derived chloride anion) in the case of an SN1-type reaction
solution, and a significant portion of vescalagin was con- or evolve into an olefin product through an E1-type mecha-
comitantly released from the resin because of the cleavage nism. The cleavage of the C–S bond can also occur in a
of the thioether linkage under these acidic conditions (see concerted fashion with the formation of a C–Nu bond in
Supporting Information, Figure S11). Notwithstanding this an SN2-type reaction, which should be accompanied by an
failure, we were pleased to have successfully managed to inversion of configuration. A preference for one or the
independently attach these polyphenolic and glycolic natu- other mechanistic path expresses itself according to the
ral products on a PEGA resin in a protecting-group-free structural criteria of the starting thioether and the stability
manner. As this resin exhibits good permeability toward of the carbenium ion possibly derived therefrom, which also
macromolecules in aqueous media,[12c,14] these two PEGA- determines the site of cleavage at the C–S–C thioether mo-
bound ellagitannins 6 and 7 could also serve in affinity- tif. A well-known application of this reaction is the removal
based screening and identification of protein targets.
of S-protecting groups (e.g., tert-butyl, trityl, and acet-
However, the use of these systems in BIOS-related proto- amidomethyl) on the cysteine thiol function in peptide syn-
cols would still necessitate an efficient means to release the thesis, for which 3-nitro-2-pyridinesulfenyl chloride
compounds from the resin at the end of the exercise. This (NpysCl) has emerged as an alternative C–S cleaving rea-
cleavage step should occur in a stereoselective manner, gent.[16]
ideally with retention of configuration to afford exclusively
Thus, we treated the PEGA-bound vescalagin derivative
analogues of 1β and 2β, not their C-1 epimeric forms, 6 in suspension in dry CH3CN with NpysCl (ca. 1 equiv.).
(–)-castalagin (1α) and (+)-castalin (2α; Figure 1).[5] In- The resin changed from brown to yellow after 2 h at room
spired by the seminal work of Oki and Kobayashi on the temperature. The supernatant was then filtered, and the
reactivity of thioethers with sulfenyl chlorides,[15] we devel- resin was washed with dry CH3CN to remove any remain-
oped a methodology that not only allows the direct release ing NpysCl. RP-HPLC analysis of the combined CH3CN
1
1
Figure 3. (a) H NMR spectrum of 13 in [D6]DMSO; H HRMAS NMR spectra of (b) the resin-bound thiosulfonium ion 8 and (c) the
after-cleavage 3-nitro-2-pyridine disulfide resin 9. The signals of the three nitropyridine protons of resins 8 and 9 are marked with red
stars.
4966
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 4963–4972

Protecting-Group-Free Solid-Phase Anchoring of Ellagitannins
fractions did not show any sign of ellagitannin derivatives. butylvescalagin (13). The resin-bound species 8 does indeed
1
However, H HRMAS analysis of a resin aliquot in [D6]- possess the 1-H proton, the signal of which is slightly down-
DMSO clearly indicated that the vescalagin-derived moiety field from that of 13 (see Figure 3a and b). Thus, we con-
was still linked to the resin and revealed the presence of firmed that the resin-bound vescalagin-bearing thiosulfon-
three additional signals belonging to a 3-nitro-2-pyridine ium ion 8 is a species stable enough to be observed by
sulfur unit. Thus, we tentatively attributed these signals to HRMAS analysis.
the resin-bound thiosulfonium ion 8 (Figure 3b).
This resin material 8 was then suspended in water at
These observations indicate that the chloride anion of room temperature for 20 min, after which time the superna-
the NpysCl reagent does not participate in the release of tant was filtered and then analyzed by RP-HPLC coupled
the ellagitannin moiety from the resin by engaging itself in to electrospray ionization mass spectrometry (ESIMS). A
any SN-type reaction under these conditions, despite the major compound was observed with the same retention
possibility of passing through a stable benzylic C-1 carbo- time and molecular mass as those of vescalagin (1β). To
cation intermediate 10 (Scheme 2); no chloroellagitannin of optimize the recovery of 1β by extraction from the inside of
type 11 was observed. Alternatively, as can also be inferred the resin beads, the aqueous supernatant was removed, and
from the original work of Oki and Kobayashi,[15b] the chlor- the resin was washed three times with anhydrous N,N-di-
ide anion could deprotonate the C-1 position of the thiosul- methylformamide (DMF), in which the swelling of the
fonium ion 8 to furnish the sulfonium ylide 12. Although PEGA resin is enhanced. The amount of 1β recovered after
analogous ylides usually rapidly evolve in solution through evaporation and lyophilization of the combined water and
Pummerer-like processes,[15b] we nevertheless wondered if a DMF washes indicated that its release from the resin was
resin-bound version such as 12 could also be stable enough quite efficient (84% yield, Scheme 2). The stereochemistry
1
to be observed by HRMAS analysis. We discarded this al- at C-1 was checked by H NMR analysis of the recovered
ternative by verifying the presence of the 1-H proton by molecule in D2O. The observation of a narrow doublet at δ
1
comparing the H HRMAS NMR spectrum of 8 with the = 4.82 ppm for H-1 with a small coupling constant JH1,H2
1
solution H NMR spectrum of the thioether model β-1-S- of ca. 2 Hz confirmed the β-orientation of the OH group at
Scheme 2. Mechanistic options for the treatment of the resin-bound vescalagin derivative 6 with NpysCl.
Eur. J. Org. Chem. 2014, 4963–4972
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4967

S. Quideau et al.
FULL PAPER
C-1.[5d,6c,17] Hence, as we assumed that the initial instal- presence of the 3-nitro-2-pyridine disulfide moiety bound
lation of 1β onto the thiol resin 5 proceeded with retention to this resin 9 (Figure 3c).
of configuration at C-1 on the basis of related experimental
We then focused our efforts on the exploitation of the
precedents from our own laboratory (vide supra),[6b,6c,7] the resin-bound vescalagin derivative 6 to introduce amino
recovery of 1β and not that of its epimer 1α does imply groups at the C-1 position of 1β. In contrast to the intro-
again that the configuration is retained during the water- duction of carbon-, oxygen-, and sulfur-based nucleophiles
mediated release from the resin-bound thiosulfonium ion 8. at this position, which is conveniently achieved by acid-
From a mechanistic point of view, this retention of configu- catalyzed SN1-type reactions in solution, amines are recalci-
ration means that the use of water in place of dry CH3CN trant to this chemistry because their protonation to ammo-
as solvent facilitated the cleavage of the C-1–S bond of 8 nium ions shuts down their nucleophilicity. Thus, the ex-
with departure of the 3-nitro-2-pyridine disulfide resin 9 ploitation of a sulfenyl-mediated cleavage of the C-1–S
(Scheme 2). The resulting carbenium intermediate 10 would bond of 6 in the presence of an alkylamine could offer a
then be quenched by water in this SN1-type scenario, which solution to access 1-alkylamino-vescalagin derivatives at the
is known to proceed in a diastereoselective fashion in favor time of the release of the vescalagin unit from the resin. We
of an attack from the β-exo face of 10 (Scheme 2).[6c] More- first validated the feasibility of this chemistry in solution
over, this sequence of NpysCl addition and water-mediated before implementing it in the solid phase concomitantly
cleavage was also performed with the thioether 13 in solu- with the release of the ellagitannin from the resin. Thus, the
1
tion in THF. The comparison of the H NMR spectrum of thioether 13 was treated with NpysCl (1.0 equiv.) in [D8]-
the released ellagitannin with that of 1α again confirmed THF with the aim of monitoring in situ the formation of
that the released ellagitannin was vescalagin 1β (see the the expected thiosulfonium ion intermediate 14 by 1H
Supporting Information for details, including 1H NMR NMR analysis (see Scheme 3 and Figure S5). This reaction
spectra in D2O). Finally, we also verified the nature of the reached completion after 15 min at room temperature. The
after-cleavage resin by HRMAS analysis and confirmed the nitropyridine moiety was clearly detected by 1H NMR spec-
Scheme 3. Solution-phase synthesis of 15a: (i) NpysCl, [D8]THF, r.t.; (ii) benzylamine, [D8]THF, r.t.
4968
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 4963–4972

Protecting-Group-Free Solid-Phase Anchoring of Ellagitannins
Scheme 4. Preparation of 1-alkylamino-vescalagin derivatives by (a) the functionalizing resin-cleavage method [(i) NpysCl (1 equiv.),
CH3CN, r.t., 2 h; (ii) RNH2 (6 equiv.), CH3CN, r.t., 30 min; estimated HPLC yields] and (b) its adaptation to solution-phase conditions
[(i) NpysCl (3 equiv.), THF, r.t., 30 min; (ii) RNH2 (5 equiv.), THF, r.t., 30 min; isolated yields].
troscopic analysis, which also showed the expected down- with anhydrous DMF. Unfortunately, we could not avoid
field shift of the 1-H proton of the vescalagin unit by ca. the formation of 1β, which was evidently caused by some
1.1 ppm relative to that of 13 and the conservation of a residual water either trapped within the PEGA resin or still
weak coupling constant JH1-H2 of ca. 2 Hz as an indication present in the reaction (CH3CN) or wash (DMF) solvents.
of the β-orientation of the thiosulfonium unit at C-1 (Fig- This water-mediated competing production of 1β inevitably
ure S5b). Benzylamine (3.5 equiv.) was next added, and affected the yields of the desired alkylamino-vescalagin de-
analysis of the reaction mixture by RP-HPLC indicated rivatives. Nevertheless, five of the six expected derivatives
complete consumption of 14 (and 13) after 15 min and the 15a–15f were identified by RP-HPLC-ESIMS analysis, and
formation of two products, the desired benzylamino-vescal- their yields were estimated by HPLC analysis (Scheme 4).
agin derivative 15a and the mixed disulfide 16 (Scheme 3). These low-to-moderate yields discouraged us from scaling-
This mixture was then quickly separated by semipreparative up these solid-phase reactions for proper isolation and puri-
RP-HPLC to afford a partially purified derivative 15a, fication of these derivatives. We instead moved to solution-
which was analyzed in [D8]THF by 1H NMR spectroscopy. phase synthesis conditions starting from the thioether 13
The observation of a broad singlet at δ = 4.97 ppm for 1-H and using NpysCl (3.0 equiv.) and the same amines
is indicative of a β-orientation of the amino group at (5.0 equiv.) in anhydrous THF at room temperature. Thus,
C-1 (Figure S5c). This result confirmed the feasibility of the benzylamino- (15a), “clickable” propargylamino- (15b),
the intended chemistry and, hence, encouraged us to pursue n-butylamino- (15c), furanomethylamino- (15d), thiofura-
its implementation in both solid- and solution-phase syn- nomethylamino- (15e), and fluorophenylmethylaminovesca-
theses with the aim of developing a chemo- and stereoselec- lagin (15f) derivatives were all conveniently produced, puri-
tive amination of vescalagin.
fied by semipreparative RP-HPLC [eluted with MeOH/H2O
Thus, a series of five additional primary amines were se- containing HCOOH (0.1 vol.-%)], and hence isolated as
lected for this purpose (Scheme 4). The resin-bound vescal- ammonium formate salts in generally much higher yields
agin derivative 6 was first treated with NpysCl (1.0 equiv.) (Scheme 4), despite the usual difficulties in the chromatog-
in freshly distilled CH3CN at room temperature for 2 h, af- raphy of such polyphenolic compounds without significant
ter which time the resulting thiosulfonium ion 8 was re- loss of materials.
moved by filtration and washed once with CH3CN prior to
the addition of the amines (6.0 equiv.). The release of the 1-
alkylamino-vescalagin products 15a–15f from the resin was
monitored by RP-HPLC analysis, and their full extraction
Conclusions
from the resin material after a reaction time of ca. 30 min
A protecting-group-free strategy was developed to immo-
was achieved by washing the leftover resin several times bilize the polyphenolic C-glucosidic ellagitannins vescalagin
Eur. J. Org. Chem. 2014, 4963–4972
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4969

S. Quideau et al.
FULL PAPER
rotor and swollen in [D6]DMSO. The spectra were acquired with a
Bruker Avance DRX 500 MHz spectrometer equipped with a 4 mm
HRMAS probe at 25 °C. The samples were spun at 5 kHz. The
chemical shifts were referenced to the residual solvent signals at δ
= 2.50 ppm. The spectra were acquired with 32K data points, which
were zero-filled to 64K data points prior to Fourier transform.
(1β) and vescalin (2β) on a biocompatible PEGA resin.
These resin-bound polyphenols should suitably serve as nat-
ural protein-binding scaffolds for the affinity-based screen-
ing of protein targets and for biology-oriented synthesis of
diverse analogues. Furthermore, the resin can be cleaved in
a functionalizing manner that notably permits the chemo-
and stereoselective introduction of amino groups to the ves-
calagin core. The adaptation of this protecting-group-free
method to solution-phase conditions enabled a convenient
and generally higher-yielding preparation of 1-alkylamino-
vescalagin derivatives, which are not otherwise accessible by
the usual acid-catalyzed SN1-type reactions used to intro-
duce carbon-, oxygen-, and sulfur-based nucleophiles at the
same position of the vescalagin core. Further exploitation
of these methodologies will offer access to other amino-
vescalagin derivatives for biological screening.
Thiol Resin 5: Amino PEGA resin (100 mg, 0.4 mmol/g,
0.04 mmol) was suspended in dry CH2Cl2 and swollen for 20 min.
After collection by filtration, the resin was again suspended in dry
CH2Cl2 (5 mL), and N,N-diisopropylethylamine (DIPEA, 0.02 mL,
0.12 mmol) and 6-bromohexanoyl chloride (0.012 mL, 0.08 mmol)
were added. The resin was then gently shaken for 1 h. After fil-
tration, the bromo resin intermediate 3 was successively washed
twice with CH2Cl2 and MeOH, followed by CH2Cl2 and DMF.
The efficiency of the loading was confirmed by a negative 2,4,6-
trinitrobenzenesulfonic acid (TNBS) test. The resin 3 was next sus-
pended in DMF (5 mL), and KSCOCH3 (14 mg, 0.12 mmol) was
added; shaking was maintained overnight. After collection by fil-
tration, the thioacetate resin 4 was successively washed twice with
DMF, MeOH, CH2Cl2, and finally with CH3CN. The resin 4 was
next solvated in dry THF (5 mL) and LiBH4 (0.1 mL, 2 m solution
in THF) was added. The resin was gently shaken for 3 h, then col-
lected by filtration and washed several times with MeOH and THF.
This deprotection step was repeated once to furnish the thiol resin
5 after final CH2Cl2 washings. The presence of the free thiol func-
tion on the resin was confirmed by Ellman’s test. Modified PEGA
resins 4 and 5 were characterized by 1H HRMAS NMR spec-
troscopy (see Supporting Information).
Experimental Section
General: All reagents were purchased from Sigma–Aldrich. The
PEGA resin was purchased from Novabiochem. Most reactions
were performed under N2 or Ar with anhydrous solvents, obtained
either by distillation from the appropriate drying agent (CH3CN
from CaH2, DMF from 4 Å molecular sieves, and MeOH from
CaCl2) or by passage through a column of neutral alumina
(CH2Cl2 and THF) under N2. The PEGA resin modification was
realized manually in a glass reactor to obtain resin 5 as described
below. The immobilization, derivatization, and release from the
resin of vescalagin and vescalin were performed with a Chemspeed
ASW100 automated synthesizer under N2. HPLC-quality meth-
anol and Milli-Q (Millipore) water were used for reversed-phase
(RP) HPLC analyses and separations. The mobile phases were
composed of solvent A [H2O/HCOOH (999:1)] and solvent B
[MeOH/HCOOH (999:1)]. HPLC analyses were performed with a
Thermo system equipped with P1000 XR pumps, an AS3000 au-
tosampler, a UV 6000 LP diode array detector, and a Macherey–
Nagel Nucleodur C-18 Pyramid column (250ϫ4.6 mm, 5 μm).
Gradient elution (0–20 min: 0 to 20% solvent B; 20–35 min: 20 to
100% solvent B, 35–40 min: 100% solvent B) was applied at a flow
rate of 1 mL/min, and the column effluent was monitored by UV
detection at 280 nm. Semipreparative HPLC separations were per-
formed with a Varian ProStar system equipped with PrepStar SD1
pumps, a 2 mL injection loop, a ProStar 320 UV/Vis detector, and a
Macherey–Nagel Nucleodur C-18 Pyramid column (250ϫ21 mm,
5 μm). The same gradient elution as for HPLC analyses was applied
at a flow rate of 16 mL/min, and the column effluent was moni-
tored by UV detection at 280 nm. RP-HPLC-ESIMS analyses were
performed with a ThermoElectron LCQ Advantage spectrometer
equipped with an ion-trap mass analyzer and coupled to a Thermo-
Electron Surveyor HPLC system by a flow-split system (1/10). The
column and the conditions of elution and effluent monitoring were
Resin-Bound Vescal(ag)in 6 and 7: Resin 5 (42 mg, 0.38 mmol/g,
0.0158 mmol) was suspended in TFA/THF (1.5% v/v, 2 mL) under
N2 and allowed to swell for 10 min. After filtration, vescalagin (1β,
14.8 mg, 0.0158 mmol) or vescalin (2β, 10.0 mg, 0.0158 mmol) in
TFA/THF (1.5% v/v, 1.5 mL) was added, and the reaction mixture
was heated at 50 °C for 4 h. The disappearance of 1β or 2β from
the supernatant was monitored by analysis of an aliquot by RP-
HPLC. The resulting resin-bound vescalagin or vescalin 6 or 7 was
filtered, washed several times with MeOH and CH2Cl2, dried by
lyophilization, and analyzed by 1H HRMAS NMR spectroscopy
(see Figure 2 and the Supporting Information).
General Procedure for the Release of Vescal(ag)in (1β and 2β) from
the Resin: The resin-bound vescalagin 6 (7 mg, 0.283 mmol/g,
1.9 μmol) was suspended in freshly distilled CH3CN (0.350 mL)
and swollen for 15 min. A 14.0 mm solution of NpysCl (0.160 mL,
2.2 μmol) in distilled CH3CN was next added to the resin 6. The
supernatant solution became yellow. The resin was then gently
shaken for 2 h at room temperature; it progressively changed from
brown to yellow, and the supernatant lost its yellow color. The
supernatant was then removed, and the resin was washed several
times with distilled CH3CN to remove any remaining NpysCl. H2O
(1 mL) was added, and after 20 min at room temperature, the aque-
ous supernatant was removed, and the resin was washed three times
with DMF. The aqueous and DMF supernatant solutions were
gathered, evaporated under reduced pressure, and lyophilized to
furnish 1β in 84% yield. The same procedure was followed to re-
lease 2β from resin 7 in a comparable yield.
1
the same as those for HPLC analyses. The H and 13C NMR spec-
tra of samples in solution were recorded at 300 MHz with a Bruker
Avance II spectrometer or at 400 MHz with a Bruker Avance DPX
spectrometer. The electrospray ionization (ESI) high-resolution
mass spectrometric (HRMS) data were obtained from the Centre
d’Etudes Structurales et d’Analyses des Molécules Organiques
(CESAMO), Université de Bordeaux, Talence and from the Mass
Spectrometry Laboratory at the European Institute of Chemistry
and Biology (IECB), Pessac, France. For 1H HRMAS NMR ex-
periments, the resin samples (5 mg) were introduced into a 4 mm
General Aminating Resin-Cleavage Procedure: The resin-bound ves-
calagin 6 (10 mg, 0.308 mmol/g,[18] 3.1 μmol) was suspended in
freshly distilled CH3CN (0.35 mL) and swollen for 15 min. Then, a
14.0 mm solution of NpysCl (0.22 mL, 3.1 μmol) in freshly distilled
CH3CN was added, and this reaction mixture was gently shaken
for 2 h at room temperature. The supernatant was then removed,
and the resulting thiosulfonium resin 8 was washed several times
4970
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 4963–4972

Protecting-Group-Free Solid-Phase Anchoring of Ellagitannins
with freshly distilled CH3CN. The resin was next suspended in
freshly distilled CH3CN (0.35 mL), and each primary amine
(18.6 μmol, 6 equiv.) in freshly distilled CH3CN was added. Each
reaction was run at room temperature for 30 min, after which the
supernatant was removed, and the resin was washed three times
with freshly distilled CH3CN (3ϫ 250 μL) and three times with
distilled DMF (3ϫ 250 μL). The CH3CN and DMF supernatant
solutions were collected, evaporated under reduced pressure, redis-
solved in water, and lyophilized. The release of the amino-vescala-
gin derivatives 15a–15f was monitored by RP-HPLC, and their
identification was confirmed by ESIMS.
edged. E. B. wishes to thank the Conseil Régional d’Aquitaine for
her postdoctoral fellowship. The authors also wish to thank Axelle
Grelard (UMR-CNRS 5248) for her help in performing the
HRMAS NMR experiments.
[1] a) S. Quideau, Plant Polyphenols, eLS, 2013 [http://
www.els.net]; DOI: 10.1002/9780470015902.a0001913.pub2; b)
S. Quideau, D. Deffieux, C. Douat-Casassus, L. Pouységu, An-
gew. Chem. Int. Ed. 2011, 50, 586–621; Angew. Chem. 2011,
123, 610; c) E. Haslam, Practical Polyphenolics: From Structure
to Molecular Recognition and Physiological Action, Cambridge
University Press, Cambridge, 1998; d) E. Haslam, J. Nat. Prod.
1996, 59, 205–215.
[2] a) V. de Freitas, N. Mateus, Curr. Org. Chem. 2012, 16, 724–
746; b) P. Bandyopadhyay, A. K. Ghosh, C. Ghosh, Food
Funct. 2012, 3, 592–605; A. Nayak, G. H. Carpenter, Physiol.
Behav. 2008, 95, 290–294.
β-1-S-Butylvescalagin (13): To a solution of 1β (30 mg, 0.032 mmol)
in anhydrous THF (15 mL) under Ar was added TFA (225 μL) and
butanethiol (40 μL, 0.32 mmol). The reaction mixture was heated
to 60 °C and stirred for 7 h, after which the solvent was evaporated
under reduced pressure. The resulting residue was directly purified
by semipreparative HPLC to furnish 13 (27 mg, 86%) as a white
foam after lyophilization. 1H NMR (300 MHz, [D6]DMSO): δ =
9.61–7.70 (br s, 15 H), 6.50 (s, 1 H), 6.44 (s, 1 H), 6.36 (s, 1 H),
5.42 (d, J = 7.41 Hz, 1 H), 5.00 (s, 1 H), 4.92 (t, J = 7.1 Hz, 1 H),
4.86 (d, J = 11.05 Hz, 1 H), 4.47 (d, J = 7.41 Hz, 1 H), 4.42 (d, J
= 1.44 Hz, 1 H), 3.97 (d, J = 12.46 Hz, 1 H), 2.81 (td, J = 6.96,
1.5 Hz, 2 H), 1.62 (m, 2 H), 1.42 (m, 2 H) 0.88 (t, J = 7.3 Hz, 3 H)
ppm. 13C NMR (100 MHz, CD3OD): δ = 170.49, 168.18, 167.50,
166.67, 166.06, 147.96, 146.16, 146.11, 146.01, 145.23, 145.01,
144.98, 144.60, 144.41, 139.17, 138.40, 137.64, 136.99, 136.07,
127.82, 126.84, 125.06, 124.76, 124.29, 117.82, 117.05, 116.03,
115.44, 115.20, 115.08, 113.88, 109.39, 109.19, 107.90, 77.49, 71.75,
71.35, 70.36, 66.08, 43.88, 32.77, 32.18, 22.90, 14.02 ppm. HRMS
(ESI): calcd. for C45H33SO25 [M – H]– 1005.1037; found 1005.1034.
[3] a) S. Quideau (Ed.), Chemistry and Biology of Ellagitannins:
An Underestimated Class of Bioactive Plant Polyphenols, World
Scientific, Singapore, 2009.
[4] a) N. Vilhelmova, R. Jacquet, S. Quideau, A. Stoyanova, A. S.
Galabov, Antiviral Res. 2011, 89, 174–181; b) S. Quideau, M.
Jourdes, D. Lefeuvre, P. Pardon, C. Saucier, P.-L. Teissedre, Y.
Glories, in: Recent Advances, in: Polyphenol Research, vol. 2
(Eds.: C. Santos-Buelga, M. T. Escribano-Bailon, V. Lattan-
zio), Wiley-Blackwell, Oxford, UK, 2010, p. 81–137; c) T.
Okuda, Phytochemistry 2005, 66, 2012–2031; d) M. N. Clifford,
A. Scalbert, J. Sci. Food Agric. 2000, 80, 1118–1125; e) T.
Okuda, T. Yoshida, T. Hatano, Phytochemistry 2000, 55, 513–
529; f) K.-i. Miyamoto, T. Murayama, T. Hatano, T. Yoshida,
T. Okuda, in: Plant Polyphenols 2 – Chemistry, Biology, Phar-
macology, Ecology (Eds.: G. G. Gross, R. W. Hemingway, T.
Yoshida), Kluwer Academic/Plenum Publishers, New York,
1999, p. 643–663; g) S. Quideau, K. S. Feldman, Chem. Rev.
1996, 96, 475–503; h) T. Okuda, T. Yoshida, T. Hatano,
Fortschr. Chem. Org. Naturst. 1995, 66, 1–117; i) T. Okuda, T.
Yoshida, T. Hatano, Phytochemistry 1993, 32, 507–521; j) K.-
H. Lee, Y. Kashiwada, G.-i. Nonaka, I. Nishioka, M. Nish-
zawa, T. Yamagishi, A. J. Bodner, R. E. Kilkuskie, Y.-C. Cheng,
in: Natural Products as Antiviral Agents (Eds.: C. K. Chu, H. G.
Cutler), Plenum Press, New York, 1992, p. 69–90; k) T. Okuda,
T. Yoshida, T. Hatano, in: Phenolic Compounds in Food and
their Effects on Health, vol. II, ACS Symposium Series 507
(Eds.: M.-T. Huang, C.-T. Ho, C. Y. Lee), American Chemical
Society, Washington DC, 1992, p. 160–183.
[5] a) W. Mayer, W. Gabler, A. Riester, H. Korger, Justus Liebigs
Ann. Chem. 1967, 707, 177–181; b) W. Mayer, F. Kuhlmann,
G. Schilling, Justus Liebigs Ann. Chem. 1971, 747, 51–59; c) W.
Mayer, H. Seitz, J. C. Jochims, K. Schauerte, G. Schilling,
Justus Liebigs Ann. Chem. 1971, 751, 60–68; d) G.-i. Nonaka,
T. Sakai, T. Tanaka, K. Mihashi, I. Nishioka, Chem. Pharm.
Bull. 1990, 38, 2151–2156.
[6] a) C. Auzanneau, D. Montaudon, R. Jacquet, S. Puyo, L. Pou-
ységu, D. Deffieux, A. Elkaoukabi-Chaibi, F. De Giorgi, F.
Ichas, S. Quideau, P. Pourquier, Mol. Pharmacol. 2012, 82,
134–141; b) C. Douat-Casassus, S. Chassaing, C. Di Primo, S.
Quideau, ChemBioChem 2009, 10, 2321–2324; c) S. Quideau,
M. Jourdes, D. Lefeuvre, D. Montaudon, C. Saucier, Y. Glo-
ries, P. Pardon, P. Pourquier, Chem. Eur. J. 2005, 11, 6503–6513;
d) S. Quideau, M. Jourdes, C. Saucier, Y. Glories, P. Pardon, C.
Baudry, Angew. Chem. Int. Ed. 2003, 42, 6012–6014; Angew.
Chem. 2003, 115, 6194; e) Y. Kashiwada, G.-I. Nonaka, I. Ni-
shioka, K. J.-H. Lee, I. Bori, Y. Fukushima, K. F. Bastow, K.-
H. Lee, J. Pharm. Sci. 1993, 82, 487–492.
General Solution-Phase Aminating Procedure: The 1-butylthioves-
calagin derivative 13 (ca. 48 mg, 48 μmol, 1 equiv.) was placed in a
round-bottomed flask in the presence of NpysCl (ca. 28 mg,
147 μmol, 3 equiv.) under Ar, and the mixture was then dissolved
in anhydrous THF (25 mL, ca. 2 mmol/L). After 30 min of stirring
at room temperature, an aliquot (ca. two drops) of the resulting
yellow reaction mixture was taken, diluted with water (1 mL), and
then analyzed by RP-HPLC to verify the complete formation of
the thiosulfonium ion intermediate 14. Note: Intermediate 14
rapidly reacts with water to give back vescalagin (1β), so the obser-
vation of 1β and not starting 13 by HPLC detection is a good
indication of the full conversion of 13 into 14. Each of the six
freshly distilled amines (ca. 257 μmol, 5.4 equiv.) was then added,
and the reaction mixture immediately changed from yellow to red-
dish orange, and a white precipitate started to form. Stirring was
continued at room temperature for 30 min, after which the solvent
was evaporated, and the residue was directly submitted to semipre-
parative HPLC to furnish the resulting alkylamino-vescalagin de-
rivatives 15a–15f as ammonium formate salts after lyophilization
(white foams, see Supporting Information for details on the charac-
terization of these compounds).
Supporting Information (see footnote on the first page of this arti-
cle): Additional experimental details, characterization data and
1
copies of the H and 13C NMR spectra for key intermediates and
final compounds.
Acknowledgments
[7] S. Quideau, C. Douat-Casassus, D. M. Delannoy López, C.
Di Primo, S. Chassaing, R. Jacquet, F. Saltel, E. Génot, Angew.
Chem. Int. Ed. 2011, 50, 5099–5104; Angew. Chem. 2011, 123,
5205.
[8] a) A. Nören-Müller, I. Reis-Corrêa Jr., H. Prinz, C. Rosen-
baum, K. Saxena, H. J. Schwalbe, D. Vestweber, G. Cagna, S.
Financial supports from the Agence Nationale de la Recherche
(ANR) (ANR-06-BLAN-0139), the Conseil Régional d’Aquitaine
(grant numbers 20071301020 and 20091102004) and La Ligue Con-
tre le Cancer (Comité Dordogne 2006) are gratefully acknowl-
Eur. J. Org. Chem. 2014, 4963–4972
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4971

S. Quideau et al.
FULL PAPER
Schunk, O. Schwarz, H. Schiewe, H. Waldmann, Proc. Natl.
Acad. Sci. USA 2006, 103, 10606–10611; b) S. Wetzel, R. S.
Bon, K. Kumar, H. Waldmann, Angew. Chem. Int. Ed. 2011,
50, 10800–10826; Angew. Chem. 2011, 123, 10990.
[13] a) L. A. McAllister, S. Brand, R. de Gentile, D. J. Procter,
Chem. Commun. 2003, 2380–2381; b) J.-M. Becht, A. Wagner,
C. Mioskowski, Tetrahedron Lett. 2004, 45, 7031–7033.
[14] a) J. F. Tolborg, L. Petersen, K. J. Jensen, C. Mayer, D. L. Jake-
man, R. A. J. Warren, S. G. Withers, J. Org. Chem. 2002, 67,
4143–4149; b) J. Kress, R. Zanaletti, A. Amour, M. Ladlow,
J. G. Frey, M. Bradley, Chem. Eur. J. 2002, 8, 3769–3772; c)
R. M. Kohli, C. T. Walsh, M. D. Burkart, Nature 2002, 418,
658–661.
[9] a) S. Quideau, T. Varadinova, D. Karagiozova, M. Jourdes, P.
Pardon, C. Baudry, P. Genova, T. Diakov, R. Petrova, Chem.
Biodiversity 2004, 1, 247–258; b) A. Scalbert, L. Duval, S. Peng,
B. Monties, C. Du Penhoat, J. Chromatogr. A 1990, 502, 107–
119.
[10] a) J. F. Espinosa, Curr. Top. Med. Chem. 2011, 11, 74–92; b)
D. D. Laws, H.-M. L. Bitter, A. Jerschow, Angew. Chem. Int.
Ed. 2002, 41, 3096–3129; Angew. Chem. 2002, 114, 3224; c)
W. L. Fitch, G. Detre, C. P. Holmes, J. N. Shoolery, P. A. Ke-
ifer, J. Org. Chem. 1994, 59, 7955–7956.
[11] a) P. A. Keifer, J. Org. Chem. 1996, 61, 1558–1559; b) J. Furrer,
K. Elbayed, M. Bourdonneau, J. Raya, D. Limal, A. Bianco,
M. Piotto, Magn. Reson. Chem. 2002, 40, 123–132.
[12] a) M. Meldal, Tetrahedron Lett. 1992, 33, 3077–3080; b) M.
Meldal, F.-I. Auzanneau, O. Hindsgaul, M. Palcic, J. Chem.
Soc., Chem. Commun. 1994, 1849–1850; c) D. Deffieux, S. Gau-
drel-Grosay, A. Grelard, C. Chalumeau, S. Quideau, Tetrahe-
dron Lett. 2009, 50, 6567–6571.
[15] a) M. Oki, K. Kobayashi, Bull. Chem. Soc. Jpn. 1970, 43, 1223–
1229; b) M. Oki, K. Kobayashi, Bull. Chem. Soc. Jpn. 1970,
43, 1229–1234.
[16] a) J. J. Pastuszak, A. Chimiak, J. Org. Chem. 1981, 46, 1868–
1873; b) R. Matsueda, S. Higashida, R. J. Ridge, G. R. Mats-
ueda, Chem. Lett. 1982, 921–924.
[17] T. Yoshida, H. Ohbayashi, K. Ishihara, W. Ohwashi, K. Haba,
Y. Okano, T. Shingu, T. Okuda, Chem. Pharm. Bull. 1991, 39,
2233–2240.
[18] The loading of the resin 6 varies because different batches of
the PEGA resin were used during the course of this work.
Received: April 18, 2014
Published Online: June 26, 2014
4972
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 4963–4972