Enzymatic synthesis of α-glucosides of resveratrol with surfactant activity

Article abstract of DOI:10.1002/adsc.201000968

We report the synthesis of a series of α-glucosyl derivatives of resveratrol (3,5,4′-trihydroxystilbene) by a transglycosylation reaction catalyzed by the enzyme cyclodextrin glucanotransferase (CGTase) using starch as glucosyl donor. Several reaction parameters (temperature, solvent composition, enzyme concentration and starch/resveratrol ratio) were optimized. The yield of α-glucosylated products reached 50% in 24h. The structures of the derivatives were determined by a combination of amyloglucosidase-hydrolysis tests, MS and 2D-NMR. Three families of products were obtained: glucosylated at 3-OH, at 4′-OH and at both 3-OH and 4′-OH. The bonds between glucoses were basically α(1→4). Interestingly, the water solubilities of the α-glucosylated derivatives were at least 65- and 5-fold higher than those of resveratrol and the natural β-glucosylated derivative (piceid), respectively. In contrast with piceid, the synthesized α-glucosylated compounds exhibited surfactant activity, with critical micelle concentration (CMC) values in the range 0.5-3.6mM. Although the incorporation of a glucosyl moiety caused a loss of antioxidant activity (more pronounced in the position 3-OH compared with 4′-OH), the fact that the glycosides need to be converted into the aglycones before they are absorbed minimizes such an effect. In contrast, the modification of physicochemical properties such as solubility and partition coefficient by glycosylation could exert a positive influence on the bioavailability of resveratrol. Copyright

Full text of DOI:10.1002/adsc.201000968

FULL PAPERS
DOI: 10.1002/adsc.201000968
Enzymatic Synthesis of a-Glucosides of Resveratrol with
Surfactant Activity
Pamela Torres,a Ana Poveda,b Jesus Jimenez-Barbero,c Jose Luis Parra,d
Francesc Comelles,d Antonio O. Ballesteros,a and Francisco J. Ploua,*
a
Instituto de Catꢀlisis y Petroleoquꢁmica, CSIC, Marie Curie 2, 28049 Madrid, Spain
Fax : (+34)-91-585-4760; phone: (+34)-91-585-4869; e-mail: fplou@icp.csic.es
Servicio Interdepartamental de Investigaciꢂn, UAM, Francisco Tomꢀs y Valiente 7, 28049 Madrid, Spain
Centro de Investigaciones Biolꢂgicas, CSIC, Ramiro de Maetzu, 28040 Madrid, Spain
Instituto de Quꢁmica Avanzada de CataluÇa, CSIC, Jordi Girona 18, 08034 Barcelona, Spain
b
c
d
Received: December 23, 2010; Revised: February 10, 2011; Published online: May 3, 2011
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.201000968.
Abstract: We report the synthesis of a series of a- with piceid, the synthesized a-glucosylated com-
glucosyl derivatives of resveratrol (3,5,4’-trihydroxy- pounds exhibited surfactant activity, with critical mi-
ACHTUNGTRENNUNGstilbene) by a transglycosylation reaction catalyzed celle concentration (CMC) values in the range 0.5–
by the enzyme cyclodextrin glucanotransferase 3.6 mM. Although the incorporation of a glucosyl
(CGTase) using starch as glucosyl donor. Several re- moiety caused a loss of antioxidant activity (more
action parameters (temperature, solvent composi- pronounced in the position 3-OH compared with 4’-
tion, enzyme concentration and starch/resveratrol OH), the fact that the glycosides need to be convert-
ratio) were optimized. The yield of a-glucosylated ed into the aglycones before they are absorbed mini-
products reached 50% in 24 h. The structures of the mizes such an effect. In contrast, the modification of
derivatives were determined by a combination of physicochemical properties such as solubility and
amyloglucosidase-hydrolysis tests, MS and 2D-NMR. partition coefficient by glycosylation could exert a
Three families of products were obtained: glucosylat- positive influence on the bioavailability of resvera-
ed at 3-OH, at 4’-OH and at both 3-OH and 4’-OH. trol.
The bonds between glucoses were basically a
Interestingly, the water solubilities of the a-glucosy-
G
lated derivatives were at least 65- and 5-fold higher Keywords: antioxidants; cyclodextrin glucanotrans-
than those of resveratrol and the natural b-glucosy- ferases (CGTases); glycosylation; piceid; resveratrol;
lated derivative (piceid), respectively. In contrast surfactants
Introduction
found in the root of P. cuspidatum[3] and the major re-
sveratrol derivative in grapes.[4]
Resveratrol (1) is a phytoalexin synthesized by plants
Resveratrol possesses a variety of antioxidant,[5]
as a defence mechanism in certain cases (fungal at- anti-inflammatory,[6] estrogenic,[7] anticancer,[8] cardio-
tacks, injury, UV exposure, etc.).[1] It is composed of a protective,[9] neuroprotective[10] and immunomodula-
stilbene core with three phenolic groups in positions tory[11] bioactivities. Its ability to activate various de-
3, 5 and 4’. As occurs with flavonoids,[2] resveratrol is acetylase enzymes (sirtuins) has been recently discov-
commonly present in plants bound to sugars as b-gly- ered and could be responsible for the mentioned
cosides. Thus, several derivatives of resveratrol properties[12], and in particular of a delaying effect on
(Figure 1) have been identified in the roots of Poligo- aging.[13]
num cuspidatum such as piceid 2 (3-O-b-d-glucosyl-
Glycosylation of resveratrol may cause different ef-
resveratrol), resveratroloside 3 (4’-O-b-d-glucosyl- fects on its bioavailability. On the one hand, the sugar
resveratrol), piceatannol 4 (3,5,3’,4’-tetra-O-hydroxy- moiety of polyphenol glycosides plays a major role in
estilbene) and its glucosylated derivative 5 (4’-O-b-d- their absorption.[14] Polyphenols are hydrophobic scaf-
glucosyl-piceatannol). Piceid is the major polyphenol folds exhibiting poor absorption, resulting in a very
Adv. Synth. Catal. 2011, 353, 1077 – 1086
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1077

Pamela Torres et al.
FULL PAPERS
Figure 1. Polyphenols found in the root of Polygonum cuspidatum.
low concentration in the circulatory streams.[15] The Results and Discussion
modification of physicochemical properties such as
solubility and partition coefficient by glycosylation Screening and Optimization of Transglycosylation
seems to exert a positive influence on the entry of
polyphenols into enterocytes.[16] For example, 3-O-gly- We screened different enzymes with transglycosyla-
cosides of the flavonoid quercetin exhibited unexpect- tion activity (Table 1) in order to conjugate resvera-
edly improved bioavailability.[17] On the other hand, it trol with a glycosyl moiety. Due to the low solubility
is well reported that glucosylated polyphenols are of resveratrol in water (approx. 30 mg per litre), it
first deglycosylated at the intestinal wall before diffus- was necessary to add a cosolvent. Dimethyl sulfoxide
ing into the cells. After absorption, polyphenols agly- (DMSO) showed the best properties in terms of sub-
cones are conjugated with glucuronic acid or sulphate strates solubility and reaction yield (data not shown),
in the intestinal enterocytes and in the liver.[2] Glyco- in accordance with previous studies on the enzymatic
sylation may also exert other benefits for the applica- glycosylation of flavonoids.[22] Compared with other
tion of bioactive compounds, including formulation, cosolvents, mixtures buffer/DMSO were very efficient
protection from oxidation by masking phenolic to solubilize resveratrol and the glycosyl donor to a
groups, and improvement of bioavailability of drugs significant extent. The initial composition of the reac-
that need to pass through the blood-brain barrier.[18]
tion mixture was resveratrol (10 mg), glycosyl donor
The stereo- and regioselectivity of enzymes is con- (50 mg), enzyme (65 mL) in 1 mL final volume. The
sidered a valuable alternative to chemical synthesis reaction medium was a 50/50 v/v mixture of DMSO
for the preparation of structurally well-defined glyco-
conjugates.[19] For the in vitro glycosylation of natural
products, glycosyltransferases and glycosidases have
Table 1. Screened enzymes for the glycosylation of resvera-
been successfully employed. The only previous re-
ports on the biocatalytic glycosylation of resveratrol
were performed using whole cells of Bacillus cereus
yielding the piceid 2[20] and Streptococcus mutans that
formed the alpha-derivative 3-O-a-d-glucosyl-resvera-
trol.[21]
In the present work, we report the enzymatic syn-
thesis of various a-glucosyl-resveratrol derivatives by
a transglycosylation reaction catalyzed by cyclodex-
trin glucanotransferases (CGTases, EC 2.4.1.19). The
antioxidant and surfactant properties of the novel
compounds were evaluated.
trol.
Enzyme
Donor Transferred
group
a-glucosidase (Aspergillus niger)
dextransucrase (Leuconostoc mes- sucrose glucosyl-
enteroides)
CGTase (Thermoanaerobacter)
CGTase (Bacillus macerans)
b-fructofuranosidase (Aspergillus
aculeatus)
b-galactosidase (Bacillus circulans) lactose galactosyl-
b-galactosidase (Aspergillus
oryzae)
maltose glucosyl-
starch glucosyl-
starch glucosyl-
sucrose fructosyl-
lactose galactosyl-
1078
asc.wiley-vch.de
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 1077 – 1086

Enzymatic Synthesis of a-Glucosides of Resveratrol with Surfactant Activity
and 0.2M sodium acetate (pH 5.6). Reactions were in-
cubated at 408C and analyzed by HPLC.
Under the above conditions, most enzymes did not
catalyze the formation of any product, except for
CGTases with starch as glucosyl donor. A similar con-
clusion was recently reported by Markosyan et al. ex-
ploring the glycosylation of benzoquinazolines with
different glycosidic enzymes.[23] Glycosylation of phe-
nolic compounds using simple carbohydrates as
donors has been described only in very few cases.
However, glucosylation of several flavonoids such as
catechin, hesperidin, naringin, rutin and luteolin has
been achieved with glucosyltransferases from Strepto-
coccus strains[24]
,
a-glucosidases[25] and dextran-
ACHTUNGTRENNUNG
choice to glucosylate phenolic compounds due to its
broad acceptor specificity.[26]
The transglucosylation activity of CGTase is known
to be very dependent on the enzyme source. In our
work, the CGTase from Thermoanaerobacter sp. (Tor-
uzyme 3.0L) gave a higher yield than that from Bacil-
lus macerans (CGTase Amano). In contrast with the
CGTase from Bacillus macerans (the HPLC chroma-
togram showed four different compounds with in-
creasing concentration, indicating the formation of
the so-called analogous series, Figure 2), the CGTase
from Thermoanaerobacter sp. did not present the typi-
cal chromatographic profile of a mixture of deriva-
tives with increasing glucosylation degree,[27] suggest-
ing a more complex reaction scheme. However, the
yield with Thermoanaerobacter CGTase was extreme-
ly small, not surpassing 2% in 24 h.
Figure 3. Effect of temperature (A) and DMSO concentra-
tion (B) on resveratrol glucosylation. The total concentra-
tion of glucosylated products is depicted in y axis.
In order to improve the yield of glucosylated prod-
ucts, several parameters were optimized (temperature,
percentage of cosolvent, enzyme concentration and
w/w ratio starch/resveratrol). An increase of the reac- Then, the activity decayed drastically, with negligible
tion temperature was favourable on yield until 608C. conversion at 808C (Figure 3A).
We tested several concentrations of cosolvent (Fig-
ure 3B) concluding that 20% (v/v) DMSO was the
optimum in terms of total yield of glucosylated prod-
ucts. This solvent composition represented a compro-
mise between substrate solubility and enzyme effi-
ciency. Blackwood and Bucke reported that CGTases
exhibit good performance in the presence of small
amounts of polar organic solvents.[28]
After the optimization of the different parameters,
the best conditions were as follows: 10 mgmLꢀ1 re-
sveratrol (44 mM), 60 mgmLꢀ1 starch, 0.2M sodium
acetate (pH 5.6)/DMSO 80/20 (v/v), 560 units CGTase
per gram resveratrol (measured in the formation of b-
cyclodextrin, see Experimental Section). Under the
above conditions, 50% of the initial resveratrol was
converted into glucosylated products in 24 h. The
yield was significantly higher than that described by
Figure 2. HPLC chromatograms (96 h of reaction) showing
Shim et al. with Streptococcus mutans cells; they re-
the formation of resveratrol glucosyl derivatives catalyzed
by CGTase from: (I) Thermoanaerobacter sp. and (II) Bacil-
lus macerans.
ported 18% of 3-O-a-d-glucosyl-resveratrol in 48 h[21]
,
but using only 1 mgmLꢀ1 (4.4 mM) resveratrol.
Adv. Synth. Catal. 2011, 353, 1077 – 1086
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1079

Pamela Torres et al.
FULL PAPERS
Scaling-Up and Enzymatic Hydrolysis of the
Transglycosylation Products
oglucoside derivative and the appearance of free re-
sveratrol. These results indicated that most bonds be-
tween glucoses were a
A
We scaled-up the reaction with the CGTase from
Considering that the glycosides need to be convert-
Thermoanaerobacter sp. in order to isolate and char- ed into the aglycones before they are absorbed into
acterize the different glucosides. The reaction was the blood circulation,[14] the above results suggest that
scaled to 20 mL as described in the Experimental Sec- amylolytic enzymes in the intestinal tract will be able
tion, and the products purified by semipreparative to hydrolyze a
HPLC. Figure 4 shows the typical chromatogram ob-
tained. Monoglucosides (6a, 6b), diglucosides (7a–7d),
triglucosides (8a–8d) and tetraglucosides (9a, 9*)
were further identified using mass spectrometry (ESI-
TOF, see Supporting Information). The isolated yields
are given in Table 2.
N
In previous reports on CGTase-catalyzed transgly-
cosylations, the bonds between glucosyl groups were
in the majorly aACHTUNGTRENNUNG
presence of such bonds in our polyglucosylated prod-
ucts, an enzymatic hydrolysis method based on amylo-
glucosidase specificity was performed.[31,32] We incu-
bated a mixture of glucosylated resveratrol deriva-
tives (6b, 7c, 8d) with a pure amyloglucosidase [previ-
ously we checked that the enzyme was able to hydro-
lyze only a
G
maltotriose, and was not contaminated with other ac-
tivities]. As illustrated in Figure 5, the HPLC analysis
showed that, after 22.5 h incubation with amylogluco-
sidase, di- and triglucoside peaks (7c, 8b) disappeared
Figure 4. Semipreparative HPLC chromatogram showing
the different products obtained at 15 h with the CGTase
substantially, with a concomitant increase of the mon- from Thermoanaerobacter sp.
Table 2. Structure assignation and yield of the main products obtained in the CGTase-catalyzed glucosylation of
resveratrol.[a]
Compound Mass
Structure
Type Name
Isolated yield [mg][b]
6a
I
3-O-a-d-glucosyl-resveratrol
28.4
20.5
12.0
MonoGlc
6b
7a
7b
7c
II
I
4ꢄ-O-a-d-glucosyl-resveratrol
3-O-a-d-maltosyl-resveratrol
I
3-O-[a-d-glucosyl-(1?2)-a-d-glucosyl]-resveratrol n.d.
DiGlc
II
4ꢄ-O-a-d-maltosyl-resveratrol
10.5
4.1
7d
IV
3,4ꢄ-di-O-a-d-glucosyl-resveratrol
8a
8b
I
3-O-a-d-maltotriosyl-resveratrol
n.d
6.1
n.d
II
IV
4ꢄ-O-a-d-maltotriosyl-resveratrol
TriGlc
3-O-a-d-maltosyl-,4’-O-a-d-glucosyl-resveratrol
8c/8d
IV
II
3-O-a-d-glucosyl,4’-O-a-d-maltosyl-resveratrol
4’-O-a-d-maltotetraosyl-resveratrol
structure not determined
n.d
n.d
n.d
9a
9*
[a]
TetraGlc
IV
Conditions: 200 mg resveratrol, 1.4 g starch, 2.4 mL Toruzyme, 4 mL DMSO, 13.6 mL 0.2M sodium acetate (pH 5.6),
608C, 15 h.
[b]
n.d.=not determined: recovered yield was less than 1 mg.
1080
asc.wiley-vch.de
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 1077 – 1086

Enzymatic Synthesis of a-Glucosides of Resveratrol with Surfactant Activity
Figure 5. HPLC analysis showing the enzymatic hydrolysis
of glucosylated resveratrol derivatives by the action of amy-
loglucosidase after 22.5 h.
more slowly, the bond between resveratrol and glu-
cose, yielding the free resveratrol and thus presuma-
bly favouring its absorption.
Figure 6. Possible substitution patterns in resveratrol gluco-
sylation.
The chromatogram in Figure 5 also indicated the
presence of several diglucosides not hydrolyzed by The triglucosides 8c/8d could not be univocally identi-
amyloglucosidase. These compounds could contain fied. The tetraglucosides 9* (see Figure 4) showed an
two glucosyl moieties in different resveratrol phenolic important heterogeneity, so it was not possible to de-
positions, or other bonds between glucoses could be termine the position of the glucosyl groups, except in
involved. In this context, we assayed the amylogluco- the case of 9a, which was characterized as a malto-
sidase Spirizyme from Novozymes, which is also able
ACHTUNGTRENtNGNU etraosyl derivative at the 4’-OH position (see Sup-
to hydrolyze kojibiose (2-O-a-d-glucopiranosyl-d-glu- porting Information for NMR analysis).
cose) apart from maltose and maltodextrins. When
The diglucoside 7b was difficult to isolate in pure
the mixture of glucosylated resveratrol derivatives form as its concentration in the reaction was very low,
was treated with Spirizyme (data not shown), we ob- so the assignation of an a
served the disappearance of other diglucoside peaks, two glucoses was based on amyloglucosidase-cata-
G
suggesting that the minor derivatives could contain a- lyzed hydrolysis assays.
ACHTUNGTRENNUNG
Kinetics of Resveratrol Glucosylation
The kinetics of transglucosylation was studied with
Five different patterns of glucosylated derivatives the two CGTases. Reactions were followed during
could be initially expected (Figure 6): Type I, 3-O-glu- 50 h, and the 8 main glucosylated derivatives were
cosyl-resveratrol; Type II, 4’-O-glucosyl-resveratrol; quantified by HPLC (Table 3). The maximum yield
Type III, 3,5-O-glucosyl-resveratrol; Type IV, 3,4’-O- was obtained at 5 h with the CGTase from Thermo-
glucosyl-resveratrol and Type V, 3,5,4’-O-glucosyl-
ACHTUNGTRENaNGNU naerobacter sp., whereas in the case of CGTase from
resveratrol. The complete elucidation of the structure B. macerans the highest production was achieved at
of the isolated compounds was carried out by 2D- 26 h.
NMR.
A
combination of homonuclear (COSY,
The presence of a mixture of glucosylated com-
TOCSY, NOESY, ROESY) and heteronuclear se- pounds may have interest in terms of absorption of
quences (HSQC, HMBC) was used, assisted by 1D- resveratrol at the intestinal wall when they are orally
selective NOE or TOCSY-type experiments.
administered. Recently Biasutto et al. observed that
None of the elucidated structures presented either when glucosyl groups were incorporated to resvera-
Type III or Type V patterns. This was due, probably, trol via a succinate linker, the absorption was delayed
to the steric hindrance between two glucosyl moieties and the curve showing the concentration of resvera-
in the spatially close 3- and 5-positions. The structures trol derivatives in blood vs. time was shifted to longer
of the isolated products are summarized in Table 2. times in comparison with resveratrol.[15] They suggest-
Adv. Synth. Catal. 2011, 353, 1077 – 1086
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1081

Pamela Torres et al.
FULL PAPERS
Table 3. Concentration of the 8 main products obtained in
the CGTase-catalyzed glucosylation of resveratrol, deter-
mined by HPLC.[a]
OH. This fact is in concordance with our previous re-
sults studying a series of resveratrol esters.[33] Further-
more, the antioxidant activity decreased more sub-
stantially when increasing the number of glucosyl
moieties. The TEAC values, calculated from the
slopes of linear regressions of Figure 7, are summar-
ized in Table 4. The derivative with the higher TEAC
value (6b) has an antioxidant activity of only 69%
compared with resveratrol.
Compound Concentration (mM)[b] Concentration (mM)[c]
6a
6b
7a
7c
7d
8a
8b
9a
6.6
4.0
3.8
2.4
0.5
0.7
1.3
1.1
1.4
0.6
0.9
0.4
0.2
0.3
0.4
0.1
Surface Tension and and Critical Micelle
Concentration (CMC)
[a]
Conditions: 10 mgmLꢀ1 resveratrol (44 mM), 60 g/L
starch, 0.2M sodium acetate (pH 5.6)/DMSO 80/20 (v/v),
5.6 units CGTase per ml.
Using CGTase from Thermoanaerobacter sp. in 5 h.
Using CGTase from Bacillus macerans in 26 h.
It was expected that the incorporation of one or vari-
ous glucosyl moieties to resveratrol 1 could impart
surfactant properties to such a hydrophobic molecule.
We measured the surfactant properties of several a-
glucosylated derivatives (the monoglucosides 6a and
[b]
[c]
ed the administration of a mix of the aglycone 6b, and the diglucosides 7a, 7c and 7d) as well as the
(resveratrol) and its glycosylated derivative to pro- natural b-glucosylated piceid 2. Surprisingly, piceid
duce a long-lasting increase in circulating levels of the did not exhibit surfactant properties (Figure 8), which
polyphenols and its metabolites.
could be attributed to the spatial configuration of the
In our case, the presence of derivatives with differ- b-anomer.
ent degrees of glycosylation may even extend the
half-life of resveratrol and its metabolites in the circu-
latory streams. This has been attributed to the fact
that the time needed for complete hydrolysis of the
different glycosylated compounds in the intestinal
tract, which is a previous requirement before polyphe-
nol absorption, may vary substantially depending on
the glycosylation pattern.
Trolox Equivalent Antioxidant Capability (TEAC)
Although it is expected that the synthetic compounds
are completely deglycosylated before absorption,[15]
we studied their antioxidant activity (TEAC assay) in
order to analyze the role of different phenolic groups
on the properties of resveratrol.
Because resveratrol is a lipophilic compound that
has a very low solubility in water, the TEAC assay
was developed in ethanolic solutions. The antioxidant
activities of resveratrol and seven of its glucosylated
derivatives were measured. The antioxidant mecha-
nism of these compounds is associated with the pres-
ence of hydroxy groups that are able to donate a hy-
drogen to another compound. The results of the assay
are represented in Figure 7A for those compounds
glucosylated totally or partly at 3-OH, and Figure 7B
for those glucosylated exclusively at 4’-OH.
The incorporation of a glucosyl moiety to the posi-
tion 3 (Figure 7A) caused a higher loss of antioxidant
activity than in position 4’ (Figure 7B), which seems
to imply that phenolic 3-OH is more important for
the antioxidant activity than the corresponding 4’- at 3-OH and 4’-OH; (B) glucosylated exclusively at 4’-OH.
Figure 7. Effect of different resveratrol derivatives on
ABTSC+ reduction: (A) glucosylated only at 3-OH, or both
1082
asc.wiley-vch.de
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 1077 – 1086

Enzymatic Synthesis of a-Glucosides of Resveratrol with Surfactant Activity
Table 4. TEAC values of resveratrol and its a-glucosides.
corresponds to its CMC, which was quite similar to
the value of compound 6a. The diglucosides 7a and 7c
also showed surfactant activity (plots not shown). In
the case of 7d, the CMC was calculated (0.49 mM)
but the surface tension (g) decreased only to values
close to 60 mN/m. This molecule can be related to the
so-called bolamphiphilic surfactants[35,36] due to the
presence of two glucoses in opposite sides of the stil-
bene ring (structural Type IV). Comparing the three
diglucosides of resveratrol, 7a caused a higher de-
crease of the surface tension (ꢁ48.8 mN/m) than 7c
and 7d (ꢁ66.7 and 59.6 mN/m, respectively). The
main results of this study are summarized in Table 5.
It is noteworthy that the solubilities in water of the
Compound
TEAC
1
1.70
0.75
1.17
0.84
1.00
0.36
0.98
0.26
6a
6b
7a
7c
7d
8b
9a
The 3-OH monoglucoside 6a showed a typical sur- synthesized derivatives (a-glucosylated) and the natu-
factant behaviour, characterized by a linear decrease ral compounds (b-glucosylated) differ significantly.
of the surface tension vs. logarithm of concentration The solubility of resveratrol in water was approx.
until reaching the CMC. In the case of 6b, we ob- 0.03 g/L (30 ppm), whereas that of piceid was 0.37 g/
served two inflexion points. The first one is commonly L. However, the solubilities of the novel a-glucosides
due to the micelle preaggregation (denominated criti- were higher than 2 g/L in all cases, which represents
cal aggregation concentration or CAC[34]); the second at least a 65-fold increase with respect to resveratrol.
Figure 8. Variation of surface tension vs. concentration for different glucosylated resveratrol derivatives.
Adv. Synth. Catal. 2011, 353, 1077 – 1086
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1083

Pamela Torres et al.
FULL PAPERS
Table 5. Surfactant properties of some glucosyl derivatives
of resveratrol.
chased form Sigma. All other reagents were of the highest
available purity and used as purchased.
6a
6b
7d
Determination of Enzymatic Activity
CAC (mM)
CMC (mM)
gCMC (mN/m)[b]
C20 (mM)[c]
pC20
n.d.[a]
3.33
50.0
4.72
3.59
50.5
2.18
n.d.[a]
0.49
59.6
–
Among the compounds that can be synthesized by CGTases,
b-cyclodextrin (a cyclic oligosaccharide composed of seven
d-glucopyranoside units) can enclose a molecule of phe-
nophthalein inside the structure, changing a pink solution
(at basic pH) to a colourless one. An ethanolic solution of
phenophthalein (3.75 mM) was diluted with 0.2M sodium
carbonate buffer (pH 9.7) to obtain a 55 mM concentration.
The reaction mixture was composed of 0.2M sodium car-
bonate pH 9.7 (0.45 mL), 50 g/L starch (0.5 mL) and 50 mL
enzyme (or buffer). This mixture was incubated at 508C for
15 min. After that, an aliquot of 20 mL was added to 180 mL
of phenolphthalein solution (55 mM) and the absorbance
was measured at 554 nm in a microplate reader (Versamax,
Molecular Devices). One enzyme unit (U) was defined as
that catalyzing the formation of 1 mmol of b-cyclodextrin
per min.
2.56
2.59
2.66
–
G (mol/cm2)[d]
3.5·10ꢀ10
46.9
4.6·10ꢀ10
35.8
3.0·10ꢀ10
56.0
>0.82
A (ꢇ2)[e]
Solubility (g/L)
>2.5
>2.0
[a]
Not detected.
[b]
Surface tension at CMC.
[c]
[d]
[e]
Concentration needed to decrease the surface tension of
water by 20 mNmꢀ1.
G: surface excess concentration according to the Gibbs
equation.
A: area occupied per molecule adsorbed at the saturated
interface.
General Procedure for Enzymatic Reactions on an
Analytical Scale
Conclusions
A stock resveratrol solution was prepared in DMSO, and
starch was dissolved (10–100 mgmLꢀ1) in 0.2M sodium ace-
tate buffer (pH 5.6). Both solutions were mixed in a sealed
15-mL dark vial to a final resveratrol concentration of
44 mM (10 mgmLꢀ1), using different DMSO/buffer ratios.
The mixture was kept under nitrogen at 40–808C with
150 rpm orbital stirring (SI50, Stuart Scientific). The enzyme
was added to a final concentration of 1–10 UmLꢀ1. Aliquots
(200 mL) were withdrawn at intervals, filtered using an Ep-
pendorf tube containing a Duraporeꢅ 0.45 mm filter and the
progress of the reaction analyzed by HPLC.
A series of a-glucosides of resveratrol have been en-
zymatically synthesized, purified and characterized
using MS and NMR. Most of them are novel com-
pounds. The glucosides showed a remarkable surfac-
tant activity which make it useful to include them in
different delivery systems and administration forms.[16]
Considering that the glycosides need to be converted
into the aglycones before they are absorbed into the
blood circulation, the presence of a mixture of gluco-
sylated compounds may exert a long-lasting increase
in circulating levels of the polyphenols and its metab-
olites, as the time needed for complete hydrolysis of
the different glycosylated compounds in the intestinal
tract, may vary substantially depending on the glyco-
Preparative-Scale Enzymatic Reactions
The reaction mixture contained resveratrol (200 mg,
44 mM), starch (1.4 g), Toruzyme 3.0L (2.4 mL, 112 U),
sylation pattern. To demonstrate that the bioavailabil- DMSO (4 mL), 0.2M sodium acetate buffer 13.6 mL
(pH 5.6). The mixture was incubated at 608C under similar
conditions to those described above and monitored by
HPLC. After 15 h, the mixture was cooled, filtered, the sol-
vent evaporated and the crude product chromatographied,
yielding different pure compounds.
ity of resveratrol can be modified upon glucosylation,
pharmacokinetics studies of the synthesized deriva-
tives are required.
Experimental Section
HPLC Analysis
HPLC analysis was performed using a ternary pump (model
9012, Varian) coupled to a thermostatted (258C) autosam-
pler (model l-2200, VWR International). The temperature
of the column was kept constant at 408C (MEF-01 oven,
Analisis Vinicos, Spain). Detection was performed using a
photodiode array detector (ProStar, Varian) in series with
an evaporative light scattering detector (ELSD, model
2000ES, Alltech), and integration was carried out using the
Varian Star LC workstation 6.41. For the analytical reaction,
the column was a Nucleosil C18 (4.6ꢆ250 mm, 5 mm, Anali-
sis Vinicos) and mobile phase was H2O (containing 0.1% of
acetic acid) at 0.7 mLminꢀ1 for 4 min. Then, a gradient from
this mobile phase to 50:50 (v/v) H2O/methanol was per-
Materials
Resveratrol from Polygonum cuspidatum was purchased
from Shangai Seebio Biotechnology. Amyloglucosidase from
Aspergillus niger (Spirizyme Fuel) and CGTase from Ther-
moanaerobacter sp. (Toruzyme 3.0L) were kindly donated
by Novozymes A/S. CGTase from Bacillus macerans
(CGTase Amano) was kindly supplied by Amano Enzyme
Inc. Pure amyloglucosidase from Aspergillus niger (ref.
10115) was purchased from Sigma. Partially hydrolyzed
starch from potato (Passelli SA2) was from Avebe (Foxhol,
The Netherlands). 2,2’-Azino-bis(3-ethylbenzthiazoline)-6-
sulfonic acid (ABTS), kojibiose and maltodextrins were pur-
1084
asc.wiley-vch.de
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 1077 – 1086

Enzymatic Synthesis of a-Glucosides of Resveratrol with Surfactant Activity
formed in 8 min, and this eluent was maintained during
9 min. Finally, a gradient from this composition to the initial
H2O was performed in 2 min. The resveratrol derivatives
were quantified at 308 nm. The elution of carbohydrates was
followed by the ELSD detector. For the preparative-scale
tion was applied in both dimensions. Prior to Fourier trans-
form the data matrixes were zero filled up to 1024 points in
the 13C-dimension. The spectral widths for the HMBC spec-
1
tra were 5000 Hz and 25000 Hz for the H- and 13C-dimen-
sions, respectively. The number of collected complex points
1
reactions, the column was
a
Mediterranea-C18 (21.2ꢆ
was 1028 for the H-dimension with a recycle delay of 2 s,
250 mm, 5 mm, Teknokroma, Spain). The initial mobile
phase was 70:30 (v/v) methanol/H2O (H2O contained 0.1%
of formic acid) at 9 mLminꢀ1 for 10 min. Then, a gradient
from this mobile phase to 45:55 (v/v) H2O/methanol was
performed in 5 min, and this eluent was maintained during
15 min. Finally, a gradient from this composition to the ini-
tial mobile phase 70:30 (v/v) metanol/H2O was performed in
5 min. For the amyloglucosidase-hydrolysis test the column
was a Luna-NH2 (4.6ꢆ250 mm, 5 mm, Phenomenex) and
mobile phase was 80:20 (v/v) acetonitrile:water at
1 mLminꢀ1 during 30 min. The resveratrol derivatives were
detected at 308 nm. The elution of carbohydrates was fol-
lowed by the ELSD detector.
and 256 time increments were always recorded in the 13C-di-
mension. The J-coupling evolution delay was set to 66 ms.
Prior to Fourier transform the data matrixes were zero filled
up to 1024 points in the 13C-dimension. The number of tran-
sients in every experiment was set according to the concen-
tration of the sample.
TEAC assay
The trolox equivalent antioxidant capability (TEAC) assay
described by Re et al.[37] was employed to measure the anti-
oxidant activity of the resveratrol derivatives, with some
modifications to adapt to 96-well plates. This assay is based
on the ability of antioxidants in reducing ABTS radical.
Briefly, ABTS (7 mM final concentration) was added to an
aqueous solution of 2.45 mM potassium persulphate and
kept in the dark at room temperature for 15 h to obtain the
ABTS radical, which was stable for 2 days. The ABTSC+ so-
lution was diluted with ethanol to get an absorbance of 0.70
(ꢂ0.02) at 734 nm, and equilibrated at room temperature.
In each well, 20 mL of a Trolox solution (standard) or of the
antioxidants (0.5–10 mM) in ethanol were added to 230 mL
of adjusted ABTSC+ solution. The decrease of absorbance of
ABTSC+ solution was monitored at 734 nm during 6 min
using a microplate reader (model Versamax, Molecular De-
vices) and the decrease of absorbance (DA734nm) for each
concentration was determined using the area under the
curve. The concentration vs. DA734nm curve was plotted for
the different compounds and used to calculate the equiva-
lent Trolox concentration. The TEAC value was determined
as the ratio between the slopes of concentration-DA734nm
curves for the corresponding antioxidant and Trolox.
ESI-MS
Samples were analyzed on a mass spectrometer (model
LCQ Deca XP plus, Thermo Instruments) that contained an
ionic tramp. Samples were introduced using a needle pump
at 10 mL/min.
Hydrolysis Test
In order to elucidate the structure of the new compounds an
enzymatic hydrolysis method was used. Two amyloglucosi-
dases from Aspergillus niger were assayed: 10115 (Sigma)
and Spirizyme (Novozymes A/S) The reaction mixture con-
tained the glucosylated derivatives (40 mL of a 5 g/L solu-
tion) and the enzyme (40 mL of a solution 2.0 UmLꢀ1 in
50 mM sodium acetate buffer pH 4.5). The mixture was in-
cubated at 408C with 200 rpm orbital stirring (Vortemp 56,
Labnet). After 24 h, the mixture was filtered using an ep-
pendorf tube containing a Duraporeꢅ 0.45 mm filter and the
reaction analyzed by HPLC.
Surface Tension and CMC
The surface tension measurements of the aqueous solutions
were performed in a Krꢈss K-12 tensiometer using the Wil-
helmy plate method. Stock solutions were prepared for each
compound, which were conveniently diluted to measure the
surface tension. The CMC values were taken from the inter-
section of the linear sections obtained in the plot of surface
tension vs. logarithm of the concentration. Besides the
CMC, we calculated the surface excess concentration (G) ac-
cording to the Gibbs equation [G=ꢀ(dg/dlog c)/2.303 nRT]
and the area occupied per molecule adsorbed at the saturat-
NMR
NMR spectra of the different compounds were recorded on
a Bruker DRX 500 spectrometer using deuterated water
(D2O) or DMSO-d6 as solvent. A temperature of 298 K was
employed with concentrations around 10 mM. Chemical
shifts were reported in ppm, and referenced versus the DSS
signal. Vicinal proton-proton coupling constants were esti-
mated from first order analysis of the spectra. The 2D-
TOCSY experiment (60 ms mixing time) was performed
using a data matrix of 256ꢆ1 K to digitalize a spectral width
of 5000 Hz. 2D-NOESY (600 ms) and 2D-T-ROESY experi-
ments (500 ms) used the standard sequences provided by
the manufacturer, and the data matrixes described above. In
all cases, squared cosine-bell apodization functions were ap-
plied in both dimensions. The spectral widths for the HSQC
ed interface (A), expressed in ꢇ2: A=1016/
ACTHNUTRGNENUG(NAꢆG).
Acknowledgements
spectra were 5000 Hz and 18000 Hz for the H- and 13C-di-
mensions, respectively. The number of collected complex
points was 1028 for the H-dimension with a recycle delay
of 2 s, and 256 time increments minimum were always re-
corded in the 13C-dimension. The J-coupling evolution delay
was set to 3.2 ms. A squared cosine-bell apodization func-
We thank Drs. Juan Carlos Espꢀn (CEBAS, CSIC, Murcia,
Spain), Juan Carlos Morales (IIQ, CSIC, Sevilla, Spain) and
Isabel Medina (IIM, CSIC, Vigo, Spain) for technical infor-
mation and suggestions. We thank Ramiro Martꢀnez (Novo-
zymes A/S, Madrid, Spain) for supplying Toruzyme 3.0L
and for useful suggestions. Comunidad de Madrid is ac-
1
1
Adv. Synth. Catal. 2011, 353, 1077 – 1086
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1085

Pamela Torres et al.
FULL PAPERS
knowledged for a research contract to P. Torres. This research
was supported by the CSIC (Project “Intramural”
200680F0132) and Spanish Ministry of Science and Innova-
tion (Project BIO2007-67708-C04-01).
[19] F. J. Plou, A. Gꢂmez de Segura, A. Ballesteros, in: In-
dustrial Enzymes: Structure, Function and Applications,
(Eds.: J. Polaina, A. P. MacCabe), Springer, New York,
2007, pp. 141–157.
[20] R. H. Cichewicz, S. A. Kouzi, J. Nat. Prod. 1998, 61,
1313–1314.
[21] H. Shim, W. Hong, Y. Ahn, Bull. Korean Chem. Soc.
2003, 24, 1680–1682.
[22] A. Bertrand, S. Morel, F. Lefoulon, Y. Rolland, P.
Monsan, M. Remaud-Simeon, Carbohydr. Res. 2006,
341, 855–863.
[23] A. A. Markosyan, L. A. Abelyan, A. I. Markosyan,
V. A. Abelyan, Appl. Biochem. Microbiol. 2009, 45,
130–136.
[24] a) G. H. Meulenbeld, S. Hartmans, Biotechnol. Bioeng.
2000, 70, 363–369; b) K. Nakahara, M. Kontani, H.
Ono, T. Kodama, T. Tanaka, T. Ooshima, S. Hamada.
Appl. Environ. Microbiol. 1995, 61, 2768–2770.
[25] a) C. Gao, P. Mayon, D. A. MacManus, E. N. Vulfson,
Biotechnol. Bioeng. 2001, 71, 235–243; b) T. Sato, H.
Nakagawa, J. Kurosu, K. Yoshida, T. Tsugane, S. Shi-
mura, K. Kirimura, K. Kino, S. Usami, J. Biosci.
Bioeng. 2000, 90, 625–630.
[26] a) F. J. Plou, M. T. Martin, A. Gomez de Segura, M. Al-
calde, A. Ballesteros, Can. J. Chem. 2002, 80, 743–752;
b) H. Nakano, S. Kitahata, in: Handbook of Industrial
Biocatalysis, (Ed.: C. T. Hou), CRC Press, New York,
2005, pp. 22-1–22-17; c) T. Kometani, T. Nishimura, T.
Nakae, H. Takii, S. Okada, Biosci. Biotechnol. Bio-
chem. 1996, 60, 645–649; d) Y. Suzuki, K. Suzuki,
Agric. Biol. Chem. 1991, 55, 181–187.
References
[1] a) K. H. Fritzemeier, H. Kindl, Planta 1981, 151, 48–
52; b) R. Gonzalez-Barrio, M. L. Vidal-Guevara, F. A.
Tomas-Barberan, J. C. Espin, Inn. Food Sci. Emerg.
Technol. 2009, 10, 374–382.
[2] a) P. C. H. Hollman. Pharm. Biol. 2004, 42, 74–83;
b) R. Lucas, D. Alcantara, J. C. Morales, Carbohydr.
Res. 2009, 344, 1340–1346.
[3] R. M. Lamuela-Raventos, A. I. Romero-Perez, A. L.
Waterhouse, M. C. Torre-Boronat, J. Agric. Food
Chem. 1995, 43, 281–283.
[4] P. Romero, G. Ibern, R. Lamuela, M. C. Torre-Boronat,
J. Agric. Food Chem. 1999, 47, 1533–1536.
[5] Z. Kerem, D. Chetrit, O. Shoseyov, G. Regev-Shoshani,
J. Agric. Food Chem. 2006, 54, 10288–10293.
[6] a) M. Murias, N. Handler, T. Erker, K. Pleban, G.
Ecker, P. Saiko, T. Szekeres, W. Jager, Bioorg. Med.
Chem. 2004, 12, 5571–5578; b) L. M. Szewczuk, L.
Forti, L. A. Stivala, T. M. Penning, J. Biol. Chem. 2004,
279, 22727–22737.
[7] B. D. Gehm, J. M. McAndrews, P. Y. Chien, J. L. James-
on, Proc. Natl. Acad. Sci. USA 1997, 94, 14138–14143.
[8] T. E. Kraft, D. Parisotto, C. Schempp, T. Efferth. Crit.
Rev. Food Sci. 2009, 49, 782–799.
[9] J. M. Wu, Z. R. Wang, T. C. Hsieh, J. L. Bruder, J. G.
Zou, Y. Z. Huang, Int. J. Mol. Med. 2001, 8, 3–17.
[10] J. C. Jung, E. Lim, Y. Lee, J. M. Kang, H. Kim, S. Jang,
S. Oh, M. Jung, Eur. J. Med. Chem. 2009, 44, 3166–
3174.
[27] M. T. Martin, M. A. Cruces, M. Alcalde, F. J. Plou, M.
Bernabe, A. Ballesteros, Tetrahedron 2004, 60, 529–
534.
[28] A. D. Blackwood and C. Bucke, Enzyme Microb. Tech-
nol. 2000, 27, 704–708.
[29] T. K. Kim, D. C. Park, Y. H. Lee, J. Microbiol. Biotech-
[11] R. E. King, J. A. Bomser, D. B. Min, Compr. Rev. Food
Sci. Food Saf. 2006, 5, 65–70.
[12] M. Porcu and A. Chiarugi, Trends Pharmacol. Sci.
2005, 26, 94–103.
[13] a) M. Shakibaei, K. B. Harikumar, B. B. Aggarwal,
Mol. Nutr. Food Res. 2009, 53, 115–128; b) M. H.
Karol, Biomol. Ther. 2009, 17, 113–124.
nol. 1997, 7, 310–317.
[30] C. Monthieu, A. Guibert, F. R. Taravel, R. Nardin, D.
Combes, Biocatal. Biotransform. 2003, 21, 7–15.
[31] M. Sato, T. Matsuo, N. Orita, Y. Yagi, Biotechnol. Lett.
1991, 13, 69–74.
[32] H. Aga, M. Yoneyama, S. Sakai, I. Yamamoto, Agric.
Biol. Chem. 1991, 55, 1751–1756.
[14] K. D. R. Setchell, N. M. Brown, L. Zimmer-Nechemias,
W. T. Brashear, B. E. Wolfe, A. S. Kirschner, J. E.
Heubi, Am. J. Clin. Nutr. 2002, 76, 447–453.
[15] L. Biasutto, E. Marotta, A. Bradaschia, M. Fallica, A.
Mattarei, S. Garbisa, M. Zoratti, C. Paradisi, Bioorg.
Med. Chem. Lett. 2009, 19, 6721–6724.
[16] D. V. Ratnam, D. D. Ankola, V. Bhardwaj, D. K.
Sahana, M. N. V. R. Kumar, J. Controlled Release 2006,
113, 189–207.
[17] I. C. W. Arts, A. L. A. Sesink, M. Faassen-Peters,
P. C. H. Hollman, Br. J. Nutr. 2004, 91, 841–847.
[18] E. Lim and D. Bowles, BIOforum Eur. 2008, 9, 32–34.
[33] P. Torres, A. Poveda, J. Jimenez-Barbero, A. Balles-
teros, F. J. Plou, J. Agric. Food Chem. 2010, 58, 807–
813.
[34] T. Sakai, Y. Kaneko, K. Tsujii, Langmuir 2006, 22,
2039–2044.
[35] V. M. Garamus, G. Milkereit, S. Gerber, V. Vill, Chem.
Phys. Lett. 2004, 392, 105–109.
[36] C. M. Kausch, Y. Kim, V. M. Russell, R. E. Medsker,
R. R. Thomas, Langmuir 2003, 19, 7182–7187.
[37] R. Re, N. Pellegrini, A. Proteggente, A. Pannala, M.
Yang, C. Rice-Evans, Free Radical Biol. Med. 1999, 26,
1231–1237.
1086
asc.wiley-vch.de
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 1077 – 1086