Enzymatic fructosylation of aromatic and aliphatic alcohols by Bacillus subtilis levansucrase: Reactivity of acceptors

Article abstract of DOI:10.1016/j.molcatb.2011.02.002

Levansucrases from Bacillus subtilis (BS-LVS) and Leuconostoc mesenteroides ssp. mesenteroides ATCC 8293 (LevC), inulosucrase from Leuconostoc citreum (IslA) and an invertase from Saccharomyces cerevisiae (Inv) were evaluated in acceptor reactions with non-sugar acceptors. Among them, BS-LVS was selected for the fructosylation of aromatic or aliphatic alcohols due to its high activity and stability. The effects of acceptor concentration, enzyme concentration and the presence of a co-solvent in the fructosylation efficiency of hydroquinone were evaluated. It was demonstrated that this reaction is kinetically controlled, producing the best yields of phenolic fructosides when 500 mM of acceptor and 5 U mL^-1 of enzyme were employed. Higher enzyme loads resulted in the rapid hydrolysis of the products. Increased amounts of organic co-solvent up to 50% (v/v) reduced fructoside yield due to a concomitant decrease in the thermodynamic activity of the acceptor, as confirmed by theoretical calculations using COSMO-RS; moreover, increased fructose transfer to levan and reduced hydrolysis were observed. It was found that BS-LVS preferentially fructosylates aromatic over aliphatic alcohols. A maximum fructoside production (19-35 mM) was obtained with dihydroxybenzene acceptors such as hydroquinone, whereas reactions with primary alcohols, such as benzyl alcohol resulted in lower fructosylation efficiency. This selectivity was also demonstrated by the fact that 4-hydroxybenzylalcohol, a bifunctional acceptor, was fructosylated at a rate ten times faster on its aromatic hydroxyl group. BS-LVS selectivity over phenol fructosylation was inversely correlated with the acceptor pK_a value.

Full text of DOI:10.1016/j.molcatb.2011.02.002

Journal of Molecular Catalysis B: Enzymatic 70 (2011) 41–48
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Enzymatic fructosylation of aromatic and aliphatic alcohols by Bacillus subtilis
levansucrase: Reactivity of acceptors
Arlette Mena-Arizmendia , Joel Aldereteb , Sergio Águilaa , Alain Martyc,d,e , Alfonso Miranda-Molinaa ,
Agustín López-Munguíaa, Edmundo Castilloa,∗
a Departamento Ingeniería Celular y Biocatálisis, Instituto de Biotecnología, Universidad Nacional Autónoma de México, Apartado Postal 510-3, Av. Universidad 2001, Cuernavaca,
Morelos 62250, Mexico
b Grupo de Química Teórica y Computacional, Facultad de Ciencias Químicas, Universidad de Concepción, Casilla 160-C, Concepción, Chile
c Université de Toulouse, INSA, UPS, INP LISBP, 135 Avenue de Rangueil, F-31077 Toulouse, France
d INRA, UMR792 Ingénierie des Systèmes Biologiques et des Procédés, F-31400 Toulouse, France
e CNRS, UMR5504, F-31400 Toulouse, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Levansucrases from Bacillus subtilis (BS-LVS) and Leuconostoc mesenteroides ssp. mesenteroides ATCC
8293 (LevC), inulosucrase from Leuconostoc citreum (IslA) and an invertase from Saccharomyces cere-
visiae (Inv) were evaluated in acceptor reactions with non-sugar acceptors. Among them, BS-LVS was
selected for the fructosylation of aromatic or aliphatic alcohols due to its high activity and stability. The
effects of acceptor concentration, enzyme concentration and the presence of a co-solvent in the fructo-
sylation efficiency of hydroquinone were evaluated. It was demonstrated that this reaction is kinetically
controlled, producing the best yields of phenolic fructosides when 500 mM of acceptor and 5 U mL−1 of
enzyme were employed. Higher enzyme loads resulted in the rapid hydrolysis of the products. Increased
amounts of organic co-solvent up to 50% (v/v) reduced fructoside yield due to a concomitant decrease
in the thermodynamic activity of the acceptor, as confirmed by theoretical calculations using COSMO-
RS; moreover, increased fructose transfer to levan and reduced hydrolysis were observed. It was found
that BS-LVS preferentially fructosylates aromatic over aliphatic alcohols. A maximum fructoside pro-
duction (19–35 mM) was obtained with dihydroxybenzene acceptors such as hydroquinone, whereas
reactions with primary alcohols, such as benzyl alcohol resulted in lower fructosylation efficiency. This
selectivity was also demonstrated by the fact that 4-hydroxybenzylalcohol, a bifunctional acceptor, was
fructosylated at a rate ten times faster on its aromatic hydroxyl group. BS-LVS selectivity over phenol
fructosylation was inversely correlated with the acceptor pKa value.
Received 7 December 2010
Received in revised form 25 January 2011
Accepted 1 February 2011
Available online 26 February 2011
Keywords:
Levansucrase
Activity coefficients
COSMO-RS
Organic solvents
Acceptor reaction
Phenols
Phenolic fructosides
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
in nature and engaged in diverse functions, mostly as secondary
metabolites [1,2]. In recent years, natural phenolic glycosides have
Glycosides are compounds formed by a carbohydrate covalently
linked to a non-sugar moiety (aglycon). They are widely distributed
generated industrial interest due to their diverse biological activ-
ities in humans. For instance, glycosides of flavonoids are widely
used as radical-scavenging molecules [3–6], whereas ␤-arbutin is
used as a skin-whitening agent [7]. Glycosylation strategies can also
be useful to modify the physicochemical properties of the agly-
con (i.e., to increase its water solubility or stability), modulating its
pharmacological properties [8,9]. Despite the great interest in using
phenolic glycosides as therapeutic agents, antioxidants or free rad-
ical scavengers, the most important limitation in their industrial
application is their low content in plants and the complexity of their
synthesis [10–12]. Indeed, chemical strategies involve multi-step
reactions for the protection and deprotection of hydroxyl groups
as well as the use of heavy metal salts [13,14], which result in a
complex mixture of products and toxic by-products [1].
Abbreviations:
BS-LVS, B. subtilis levansucrase; Inv, S. cerevisiae ␤-
d-fructofuranosidase; IslA, L. citreum inulosucrase; LevC, L. mesenteroides
ssp. mesenteroides ATCC 8293 levansucrase; GlyTFs, glycosyltransferases;
FTFs, fructosyltransferases; GTFs, glucosyltransferases; 2M2P, 2-methyl-
2-propanol; Hq, hydroquinone; Rsr, resorcinol; Cat, catechol; 4mPh,
4-methoxyphenol; Ph, phenol; Bnz, benzyl alcohol; Bu, butanol; 4HB, 4-
hydroxybenzyl
alcohol;
Fru-Hq,
4-hydroxyphenyl-␤-d-fructofuranoside;
Fru-Rsr, 3-hydroxyphenyl-␤-d-fructofuranoside; Fru-Cat, 2-hydroxyphenyl-␤-
d-fructofuranoside; Fru-4mPh, 4-methoxyphenyl-␤-d-fructofuranoside; Fru-Ph,
phenyl-␤-d-fructofuranoside; Fru-Bnz, benzyl-␤-d-fructofuranoside; Fru-Bu, n-
butyl-␤-d-fructofuranoside; Fru1-4HB, 4-(hydroxy)-benzyl-␤-d-fructofuranoside;
Fru2-4HB, 4-(hydroxymethyl)phenyl-␤-d-fructofuranoside.
To overcome these limitations, enzymatic glycosylation has
emerged as an interesting alternative, taking advantage of the
∗
Corresponding author. Tel.: +52 55 56 22 76 09; fax: +52 777 17 23 88.
E-mail address: edmundo@ibt.unam.mx (E. Castillo).
1381-1177/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2011.02.002

42
A. Mena-Arizmendi et al. / Journal of Molecular Catalysis B: Enzymatic 70 (2011) 41–48
intrinsic selectivity of enzymes. Throughout the past decades, gly-
cosidases have been the most frequently reported enzymes for the
glycosylation of primary and secondary alcohols, whereas the gly-
cosylation of phenols has been seldom addressed [8,15]. In general,
low yields of phenol glycosides have been reported [16,17], a result
that has been attributed to the low nucleophilicity of these acceptor
molecules [8,15].
USA). Silica gel 0.04–0.063 mm was obtained from Macherey–Nagel
(Düren, Germany); SephadexTM G10 and Lipophilic SephadexTM
LH20 were purchased from Sigma–Aldrich, Inc. (MO, USA). ␤-
d-Fructofuranosidase (EC 3.2.1.26) from S. cerevisiae (Inv) was
purchased from Sigma Chemical Co. (MO, USA).
2.2. Enzymes
Alternatively, non-Leloir glycosyltransferases (GlyTFs) have
been successfully applied for glycosylation reactions using sucrose
as a donor for the transfer of the glycosyl moiety to the accep-
tor molecule [18]. GlyTFs include fructosyltransferases (FTFs) and
glucosyltransferases (GTFs), which are enzymes capable of trans-
ferring fructose or glucose, respectively, to an acceptor such as
a glycan growing chain (polymerisation), water (hydrolysis) or
another suitable compound added to the reaction medium (accep-
tor reaction), resulting in a glycoside
Although there are some studies reporting the glycosylation of
several phenolic molecules such as flavonoids and dihydroxyben-
zenes, this has only been achieved at low acceptor concentrations
(from 3 to 10 mM). In consequence, productivity remains low and
limited by the acceptor solubility. This is the case for flavonoids
such as catechin, luteolin, quercetin, myricetin and epigallocat-
echin, which were glucosylated in the phenolic group using
glucansucrases from Streptococcus sp. and Leuconostoc sp. [19–22].
To cope with this limitation, different water-miscible organic sol-
vents have been used to enhance acceptor solubility, but low
amounts of glucosides were observed [20].
For small phenolic molecules, such as catechol, 4-
methylcatechol and 3-methoxycatechol, which have higher
solubility in water, glucosylation was successful using the
glucosyltransferase-d from S. mutans GS-5 in an aqueous medium,
with yields of 65%, 50% and 75%, respectively, using 40 mM of
acceptor [23]. However, high concentrations of these compounds
(200 mM) led to a dramatic loss in enzyme activity. More recently,
the fructosylation of hydroquinone was reported using a levansu-
crase from Leuconostoc mesenteroides at a high initial concentration
of the acceptor (350 mM) but with a rather low fructoside final
concentration (4 mM) [24].
In this context, the availability of biocatalysts with high trans-
ferase activity and specificity as well as high stability in the
presence of high concentrations of phenolic compounds is still
required to enhance the glycosylation yields of hydroxylated aro-
matic compounds. Levansucrase from Bacillus subtilis (BS-LVS) has
been proved to be a robust biocatalyst, showing high activity and
stability in non-conventional media, such as organic solvents [25],
and high efficiency in acceptor reactions [26]. In this work, the enzy-
matic fructosylation of phenolic compounds such as hydroquinone,
catechol and resorcinol catalysed by FTFs is reported. The selectiv-
ity of BS-LVS for the fructosylation of different aromatic or aliphatic
alcohols was studied and explained in terms of the chemical and
thermodynamic properties of the acceptors.
Levansucrase (E.C. 2.4.1.10) from B. subtilis (BS-LVS) was pro-
duced from a strain derived from B. subtilis Marburg 168 as
previously reported [25]. This strain, designated with the genotype
ꢀnpr, ꢀapr, CmR, degU32 (Hy), overexpresses the BS-LVS gene
(sacB). After fermentation, the extracellular enzyme was precip-
itated using 25% (w/v) PEG (Mw: 5000), centrifuged (22,100 × g,
25 min) and recovered in phosphate buffer 50 mM, pH 6.0 [26,27].
Inulosucrase (E.C. 2.4.1.9) from L. citreum (IslA) in cell-associated
form was used. The microorganism was grown as previously
reported [28], the culture was centrifuged (15,300 × g, 10 min),
and cells containing IslA activity were washed and recovered in
phosphate buffer 50 mM, pH 6.0. The gene coding for LevC, a lev-
ansucrase from L. mesenteroides ssp. mesenteroides ATCC 8293, was
expressed in Escherichia coli TOP 10 under the conditions previ-
ously reported [29]. The culture was centrifuged (8630 × g, 10 min),
and cells were recovered in 50 mM phosphate buffer, pH 6.0. After
cell rupture (French press, 1125 psi) and centrifugation (22,100 × g,
10 min), the supernatant containing levansucrase activity was
directly used in the reaction. A commercial ␤-d-fructofuranosidase
(EC 3.2.1.26) from S. cerevisiae (Inv) was employed.
2.3. Analytical methods
2.3.1. Enzyme activity
The activity of the enzymes studied was determined from the
initial rate of reducing sugar release (fructose and glucose) from
a 292 mM sucrose solution in phosphate buffer 50 mM, pH 6.0 for
FTFs, or acetate buffer 50 mM pH 4.5 for ␤-d-fructofuranosidase
using the dinitrosalicylic acid (DNS) method with glucose as a
standard [30]. One activity unit (U) was defined as the amount
of enzyme releasing the equivalent of 1 ␮mol of reducing sugar
per minute (␮mol min−1). The global enzyme activity in reactions
containing phenolic acceptors was evaluated as the initial rate of
sucrose consumption by high-performance liquid chromatography
(HPLC). Samples from these reactions were inactivated in boiling
water for 10 min and analysed. Fructoside yields were calculated
from sucrose conversion as determined by HPLC. All of these assays
and further experiments were performed in triplicate and error
bars are included in graphs.
2.3.2. TLC analysis of fructosylation reactions
TLC was performed using HPTLC silica gel 60 glass plates (Merck,
Darmstadt, Germany) and 85:15 (v/v) acetonitrile/water as the
mobile phase. Plates were eluted twice, and the aromatic substrates
and product spots were visualised with iodine vapour. Subse-
quently, sugars and products were revealed by spraying ␣-naphtol
solution (2.4% ␣-naphtol in an ethanol/water/H2SO4 83:7:11 mix-
ture) and heating at 100 ◦C for 1 min.
2. Experimental
2.1. Materials
Hydroquinone, benzyl alcohol, 4-hydroxybenzyl alcohol, 4-
methoxyphenol, sucrose, fructose, glucose and arbutin were
supplied by Sigma–Aldrich Inc. (MO, USA). Phenol was obtained
from Merck (Edo. de México, México). Catechol and resorcinol were
purchased from Fluka (Buchs, Switzerland). Butanol, 2-methyl-
2-propanol, diisopropyl ether and methyl tert-butyl ether were
acquired from J.T. Baker (Edo. de México, México). All substances
were of high purity grade (≥98%). Acetonitrile and acetone were
of HPLC grade and were supplied by Burdick & Jackson (MI,
2.3.3. HPLC analysis of fructosylation reactions
HPLC analyses were performed using a Waters HPLC system
(Waters Corp., MA, USA) equipped with a 600E system con-
troller, using an Alltech Prevail Carbohydrate ES column (5 ␮m,
250 mm × 4.6 mm) (Alltech Associates, Inc., IL, USA) or a Waters
Spherisorb ODS-2 column (5 ␮m, 250 mm × 4.6 mm, Waters Corp.,
MA, USA). System was coupled to a 410 refractive index (RI) detec-
tor and a 996 photodiode array detector (Waters Corp., MA, USA) at
280 nm (or 223 nm). Fructosylation products were quantified using

A. Mena-Arizmendi et al. / Journal of Molecular Catalysis B: Enzymatic 70 (2011) 41–48
43
the corresponding purified fructosides as standards. See Supporting
2.9. Structural analysis (RMN)
Information (SI) file for details.
NMR spectra were acquired either on a Varian Unity NMR Spec-
trometer operating at 400 MHz for 1H and 100 MHz for 13C nuclei or
on an Eclipse JEOL (E) NMR spectrometer operating at 300 MHz for
1H and 75 MHz for 13C nuclei. Chemical shifts are reported in parts
per million (ppm) relative to H2O and were made on the basis of 1H-
1H COSY, HMBC and HSQC spectral analysis as required. HRFABMS
spectra in a matrix of m-nitrobenzyl alcohol were recorded on a
JEOL JMX-AX 505 HA mass spectrometer. Purified fructosides struc-
ture was elucidated on the basis of the 1D (1H, 13C) and 2D (COSY,
HSQC and HMBC) NMR experiments. All products corresponded
to monofructosides. The 4-(hydroxy)-benzyl-␤-d-fructofuranoside
(Fru1-4HB) was used as a model for structural description. Struc-
tural details of all synthesised compounds are included in SI file.
2.4. Enzyme selection
To select the most active and efficient enzyme for these acceptor
reactions, the already described enzymes (three FTFs and one fruc-
tosidase) were evaluated in reactions containing different amounts
of hydroquinone, benzyl alcohol or butanol as acceptors. Reac-
tions containing 0–400 mM acceptor, sucrose 292 mM and enzyme
1 U mL−1 in the appropriate buffer were incubated at 30 ◦C using
an Eppendorf Thermomixer comfort (Hamburg, Germany). In all
cases, reactions without an acceptor were carried out as controls.
Enzyme performance was quantitatively evaluated in terms of the
enzyme activity described in Section 2.3.1; the substrates and prod-
ucts were identified by TLC.
2.10. Nucleophilicity index calculation
Calculations were carried out using the GAUSSIAN 09 program.
The geometry of the neutral and anionic species were optimised at
the B3LYP/6-311+G (d, p) level; the inclusion of diffuse functions
has been shown to be important for anionic species [31]. The nucle-
ophilicity index (ω−) was evaluated from the frontier molecular
orbital values, according to Jaramillo et al. [32,33]:
2.5. Effect of substrate concentration on acceptor reactions
Hydroquinone was used as a model substrate to select the reac-
tion conditions for BS-LVS. Reactions containing 400 mM sucrose
and 100–500 mM hydroquinone as an acceptor with 1 U mL−1 of BS-
LVS in phosphate buffer 50 mM and pH 6.0 were incubated for 20 h
at 30 ◦C. Samples were withdrawn, heat-inactivated and diluted for
product and sugar analysis by HPLC.
Stability of BS-LVS during storage in hydroquinone was evalu-
ated by incubation of the enzyme in a 500 mM hydroquinone in
phosphate buffer 50 mM, pH 6.0. Aliquots were withdrawn from
0 to 24 h, and the global enzyme activity measured as already
described. The size of the samples from the incubation solution
was calculated to dilute the hydroquinone content in the reaction
medium to 25 mM. This concentration was considered low enough
to avoid any effect on the enzyme activity. A control reaction was
carried out using the enzyme without previous contact with hydro-
quinone. Residual enzyme activity was measured from the initial
rate of sucrose conversion by HPLC during 45 min of reaction.
1(ꢁa − ꢁb)2
Áa
ω−
=
2(Áa + Áb)2
where ꢁ ≈ (εH + εL)/2 is the electronic chemical potential;
Á ≈ (εL − εH) is the hardness; the subscript a is the nucleophile,
either neutral or anionic; b the electrophile (fructose). εH and εL
correspond to the HOMO and LUMO energies, respectively.
2.11. COSMO-RS calculations
A full description of COSMO-RS (COnductor-like Screening
MOdel for Realistic Solvents) theory developed by Klamt et al.
is provided elsewhere [34–36]. In this work, Turbomol 5.9.1 and
COSMOtherm version C2.1 release 01.07 software were employed
(COSMOlogic GmbH & Co. KG, Leverkusen, Germany, 2007). Three-
dimensional molecular structures of the compounds and solvents,
as well as a minimisation of their molecular conformations,
were implemented with Marvin Sketch 5.3.8 software (v.2010
ChemAxon, http://www.chemaxon.com/). Structural conformers
were not calculated due to the structural simplicity of the eval-
uated compounds. The generation of molecular COSMO files was
performed with the Turbomole 5.9.1 program package at the den-
sity functional theory level using the BP (B88-VWN-P86) functional
with a triple-z valence polarised basis set (TZVP). The infinite
dilution activity coefficients estimation was computed with a non-
iterative mode carried out on COSMOtherm version C2.1 release
01.07.
2.6. Effect of enzyme concentration on acceptor reactions
The effect of enzyme concentration (1–20 U mL−1 BS-LVS)
on fructoside production was evaluated in reactions containing
400 mM sucrose and 500 mM hydroquinone in phosphate buffer
50 mM, pH 6.0. Reactions were incubated 20 h at 30 ◦C. The fructo-
side concentration in the samples was determined by HPLC.
2.7. Acceptor reaction in aqueous–organic mixtures
The effect of 2-methyl-2-propanol as an organic co-solvent in
the acceptor reactions was evaluated in solutions containing up to
50% (v/v) of the solvent, 400 mM sucrose, 500 mM hydroquinone
and 5 U mL−1 BS-LVS in phosphate buffer 50 mM pH 6.0 incubated
at 30 ◦C.
3. Results and discussion
3.1. Enzyme selection
2.8. Isolation and purification of fructosides
The enzyme activities of three FTFs (BS-LVS, IslA and LevC)
and one ␤-fructofuranosidase (Inv) were evaluated in the pres-
ence of sucrose and different concentrations of hydroquinone (Hq),
butanol (Bu) and benzyl alcohol (Bnz) as fructosyl acceptors. These
were compared to the activity of the control reaction without an
acceptor (Fig. 1). The higher water solubility of Bu allowed the use
of higher concentrations of this acceptor.
Fructosides were produced by BS-LVS (5 U mL−1) in reactions
with 400 mM sucrose and 500 mM acceptor (850 mM for butanol
and 200 mM for 4-hydroxybenzyl alcohol and phenol). Purification
techniques consisted in gel chromatography, flash chromatography
or preparative HPLC using a Waters Spherisorb®S5 ODS2 semi-prep
column (5 ␮m, 250 mm × 20 mm, Waters Corp., MA, USA). See SI file
for details.
Under these conditions, BS-LVS was the enzyme retaining the
highest activity at high acceptor concentration. Indeed, BS-LVS

44
A. Mena-Arizmendi et al. / Journal of Molecular Catalysis B: Enzymatic 70 (2011) 41–48
Fig. 1. Relative enzyme activity of FTFs and Inv in acceptor reactions with (a) hydroquinone, (b) benzyl alcohol and (c) butanol. (ꢀ) BS-LVS; (᭹) Inv; (ꢁ) IslA; (♦) LevC.
retained more than 95% of its activity in the presence of 400 mM Hq
(Fig. 1a) and more than 80% in 400 mM Bnz (Fig. 1b) and 850 mM
Bu (Fig. 1c).
␤-fructofuranosidase. As a consequence, BS-LVS was selected for
further analysis in acceptor reactions.
Conversely, a drastic decrease in activity was observed for Inv,
IslA and LevC when concentrations of Bnz and Hq were increased.
Previous reports have dealt with the low stability of some gly-
cosyltransferases and glycosidases in the presence of phenolic
compounds as acceptors [16,23]. The high activity of BS-LVS in the
presence of these phenolic compounds is in agreement with its high
activity and stability in the presence of organic solvents [25].
To evaluate the fructosylation efficiency, all four enzymes were
evaluated in the presence of 200 mM Hq and analysed by TLC, using
arbutin, a monoglucosylated hydroquinone derivative, as a stan-
dard. These results are shown in Fig. 2, where it is observed in the
TLC plates that all four enzymes retained a fraction of their activity;
different extents of glucose, fructose and polymer formation can be
observed.
3.2. Effect of substrate concentration on acceptor reactions
It has been previously reported that the efficiency of GTFs on
acceptor reactions is highly influenced by the nature and con-
centration of the acceptor [37,38]. Considering that Fru-Hq is an
efficient tyrosinase inhibitor [24], the enzymatic fructosylation of
hydroquinone mediated by BS-LVS was studied as a reference reac-
tion to establish suitable conditions for acceptor reactions. For
this purpose, the sucrose concentration was set at 400 mM: high
concentrations are reported to favour transfer over hydrolysis reac-
tions [26,28]. Hq concentration was varied from 100 to 500 mM,
which is close to the limit of Hq solubility. After 20 h of reaction,
sucrose was converted, and the fructosylated products were quan-
tified by HPLC.
All enzymes were capable of forming a fructosylated prod-
uct that elutes at the same Rf as arbutin, which corresponds to
monofructosylated Hq (Fru-Hq). However, a qualitative compari-
son with fructoside spots showed that the largest amount of Fru-Hq
was obtained with BS-LVS. On the other hand, very low amounts
of fructoside were observed for FTFs IslA and LevC, in agreement
with low amounts of fructose and/or glucose released. It is notable
that, although Inv is rather stable in 200 mM Hq, the acceptor reac-
tion leading to the formation of Fru-Hq takes place at the lowest
efficiency among the assayed enzymes. It may therefore be con-
cluded that among these enzymes, BS-LVS presents the highest
stability in the presence of phenolic compounds and the highest
Hq fructosylation efficiency. In addition, these results demonstrate
that, in general, FTFs are better catalysts for Hq fructosylation than
In all of the reactions studied, only one fructosylated product
was identified as Fru-Hq, with the highest concentration (22 mM)
obtained from 500 mM Hq (black bars in Fig. 3).
It is worth mentioning that at this acceptor concentration
sucrose conversion was the lowest (gray bars in Fig. 3), suggest-
ing low stability of BS-LVS at high concentrations of Hq. Therefore,
BS-LVS storage stability in 500 mM Hq at 30 ◦C was evaluated over
24 h. Samples were withdrawn at different time intervals, and the
initial rate of sucrose conversion in the absence of an acceptor was
assayed as described (see Section 2.5). During the first two hours
of incubation in Hq, more than half of the BS-LVS activity was lost
(Fig. S1 in Supporting Information file), and all of the activity disap-
peared after 24 h. This time-dependent inactivation of BS-LVS was
Fig. 3. Fructoside synthesised and sucrose converted by BS-LVS after 20 h of reac-
tion as a function of the initial concentration of Hq. The concentrations of Fru-Hq
produced are shown as black bars; the percentage of sucrose converted is shown as
gray bars. Reactions were carried out in 400 mM sucrose and 1 U mL−1 of enzyme at
30 ◦C.
Fig. 2. TLC analysis of the fructosylation reaction catalysed by FTFs and Inv using
Hq as acceptor. (a) Layer was visualized with iodine (only two lines are shown); (b)
the same layer was visualized with ␣-naphtol. (Std) Sucrose and arbutin standards.
(1) BS-LVS, (2) Inv, (3) IslA, and (4) LevC.

A. Mena-Arizmendi et al. / Journal of Molecular Catalysis B: Enzymatic 70 (2011) 41–48
45
Fig. 4. Effect of enzyme concentration on the kinetics of fructoside production.
Reactions were carried out in 400 mM sucrose and 500 mM Hq at 30 ◦C.
Fig. 5. Evolution of hydroquinone fructoside (Fru-Hq) synthesis and sucrose con-
sumption by BS-LVS in reactions containing 400 mM sucrose, 500 mM Hq and
5 U mL−1 of enzyme at 30 ◦C.
not reversible because after incubation, the enzyme was diluted to
very low Hq concentrations to measure the activity. Although these
results show an unfavourable effect of Hq on BS-LVS, it should be
noted that the efficiency of the enzyme under operational condi-
tions is usually higher due to the sucrose stabilisation effect on
the enzyme [39]. This can be observed in Fig. 3, where in opera-
tional conditions a relatively high conversion of sucrose with the
concomitant production of Fru-Hq is observed.
2M2P concentrations enhance the initial rate and lead to a higher
transfer/hydrolysis ratio due to the depletion of thermodynamic
water activity [25]. Therefore, the reactions with phenolic accep-
tors were carried out in 0–50% 2M2P in the reaction medium and
evaluated in terms of Fru-Hq yield and sucrose conversion.
Unexpectedly, while sucrose conversion was practically equiva-
lent for all of the 2M2P concentrations evaluated (data not shown),
Fru-Hq synthesis was drastically reduced as 2M2P concentration
increased (black bars in Fig. 6). It is worth mentioning that the
fructose released decreased with increasing 2M2P content (results
not shown), indicating that the co-solvent favours the transfer of
fructosyl groups to the growing levan polymer chain, reducing the
transfer to water (hydrolysis).
The effect of solvent mixtures on enzymatic processes has been
evaluated in terms of reaction rates and these, in turn, have been
evaluated as a function of thermodynamic activity coefficients (ꢂ)
of the substrates [41]. Actually, this coefficient may be consid-
ered as a measure of how much the substrate interacts with the
medium and, consequently, of how it is available for the reaction.
A robust and reliable way to predict thermophysical data for liq-
uid systems, such as ꢂ coefficients, has recently been reported
[42]. The COnductor-like Screening MOdel for Realistic Solva-
tion (COSMO-RS) method is based on a combination of quantum
chemical calculations for solutes and solvents and statistical ther-
modynamics procedures. This method allows for the estimation
3.3. Effect of enzyme concentration
The effect of enzyme concentration on sucrose conversion
and fructoside production was studied in reactions containing
400 mM sucrose and 500 mM Hq. As shown in Fig. 4, an impor-
tant improvement on Fru-Hq initial production rate is observed
withincreasingBS-LVS concentration, withthehighest Fru-Hq pro-
duction obtained at 5 U mL−1. Although at higher concentrations of
enzyme (10 and 20 U mL−1) a high initial Fru-Hq synthesis rate is
observed, it is followed by a decrease in rate proportional to enzyme
activity, suggesting that Fru-Hq synthesis is subjected to kinetic
control.
From these results, it is clear that the selection of enzyme con-
centration in the reaction medium plays a major role in defining
Fru-Hq yield and productivity. In fact, the higher the enzyme activ-
ity in the reaction, the higher the initial fructosylation rate and
sucrose conversion, but the faster Fru-Hq is lost by hydrolysis. The
effect of enzyme concentration was further studied following the
evolution of the acceptor reaction with a BS-LVS activity of 5 U mL−1
(Fig. 5). In this case, Fru-Hq concentration reaches 30 mM with an
almost total conversion of sucrose after 10 h of reaction, as com-
pared to 22 mM Fru-Hq and 65% sucrose conversion obtained with
1 U mL−1 (Fig. 3). In all cases, when sucrose is exhausted, the fruc-
toside concentration decreases. When purified Fru-Hq was treated
with BS-LVS in the absence of sucrose, it was readily hydrolysed
(results not shown). However, because the affinity of the enzyme
for sucrose is higher, to maximise the Fru-Hq synthesis and avoid
hydrolysis, further experiments were performed at 500 mM accep-
tor, 400 mM sucrose and 5 U mL−1 BS-LVS.
3.4. Effect of organic co-solvent on the acceptor reaction
One of the most frequently used strategies to enhance yields
in enzymatic processes involving hydrophobic substrates in aque-
ous media is the use of organic co-solvents. The general idea when
using this approach is that these organic solvents may increase
the solubility of the substrates and, therefore, their availability
for the enzymatic reaction [16,20,40]. An additional effect of co-
solvents has been reported in reactions with BS-LVS, where high
Fig. 6. Fructosylation of Hq with BS-LVS in aqueous–organic mixtures: the predicted
thermodynamic activity coefficients (ꢂ) of Hq are shown as white bars. Reactions
were carried out in 400 mM sucrose, 500 mM Hq and 5 U mL−1 of enzyme in the
appropriate aqueous–organic mixture. In all reactions, conversion was higher than
85%.

46
A. Mena-Arizmendi et al. / Journal of Molecular Catalysis B: Enzymatic 70 (2011) 41–48
Fig. 7. Acceptors used in fructosylation reactions with BS-LVS: Hq, hydroquinone; Rsr, resorcinol; Cat, catechol; 4mPh, 4-methoxyphenol; Ph, phenol; Bnz, benzyl alcohol;
Bu, butanol; 4HB, bifunctional acceptor 4-hydroxybenzyl alcohol.
Fig. 8. Fructosylation of aromatic and primary alcohol acceptors with BS-LVS. (a) Initial rate of sucrose conversion with 500 mM acceptor. (b) Total fructoside produced at
400 mM (white bars), 500 mM (gray bars) and 850 mM (black bars) initial acceptor concentration. All reactions were carried out with 400 mM sucrose and 5 U mL−1 of enzyme
at 30 ◦C.
of solvent–solute molecular surface interactions and, from these
estimations, the prediction of thermodynamic properties for liquid
systems. Thus, to explain why the presence of 2M2P results in a
negative effect on fructosylation, ꢂ values of Hq in water/2M2P
mixtures were predicted using the COSMO-RS method and are
reported in Fig. 6, where it may be observed that ꢂ values decrease
almost linearly with increasing 2M2P concentration (white bars in
Fig. 6).
Indeed, the hydrophobic character of Hq leads to a stronger
molecular interaction with 2M2P than with water, decreasing the
ꢂ value as the amount of 2M2P increases. Hence, it may be con-
cluded that the reduction in Hq fructosylation is due to the lower
availability of the acceptor in this co-solvent. Conversely, the pres-
ence of 2M2P decreases the water activity, favouring the transfer
of fructose to levan with the concomitant decrease of hydrolysis
of sucrose. Therefore, the addition of a hydrophobic organic co-
solvent in acceptor reactions with Hq is not an adequate strategy
to improve fructosylation, water being a better solvent.
benzenes or 850 mM Bu). Surprisingly, high initial rates of sucrose
conversion (Fig. 8a) and high fructoside concentration (Fig. 8b)
were obtained in reactions with Hq, Cat and Rsr, dihydroxyben-
zenes of reported low nucleophilicity, whereas reactions with Bu
and Bnz, which are reportedly better nucleophiles [15,43], resulted
in lower fructoside concentrations. In these conditions, negligible
initial rates were observed when phenol was used as the acceptor.
It is important to note that, although the initial rates of sucrose
conversion for Bu and Bnz are similar to those observed for dihy-
droxybenzenes (Fig. 8a), the final fructoside concentrations were
considerably lower (Fig. 8b). This suggests that even if Bu and Bnz
do not seem to affect BS-LVS activity, these primary alcohols are
not good acceptors: all sucrose converted is either hydrolysed or
transferred to levan.
Interestingly, even when the activity of BS-LVS in the presence of
4mPh was greatly reduced, the fructoside yield relative to sucrose
converted was higher than that observed for Bnz, leading to the
conclusion that 4mPh is a better acceptor than primary alcohols, as
shown in Table 1 for 500 mM of 4mPh. In this context, BS-LVS may
be considered an interesting catalyst for fructosylation and other
synthetic applications, as fructoside yields (i.e., 30 mM Fru-Hq)
3.5. Acceptor reactions with hydrophobic substrates
To extend the fructosylation strategy to other aromatic
or aliphatic alcohols, resorcinol (Rsr), catechol (Cat), 4-
methoxyphenol (4mPh), phenol (Ph), benzyl alcohol (Bnz)
and butanol (Bu) were also evaluated in acceptor reactions with
BS-LVS (Fig. 7). These reactions were carried out under the previ-
ously selected conditions, i.e., 400 mM sucrose, 5 U mL−1 of enzyme
and an adequate concentration of the acceptor. All substrates are
soluble at 500 mM, except 4mPh and Bnz, whose maximum
solubility in water is 320 mM and 370 mM (at 25 ◦C), respectively;
the high water solubility of Bu allowed for the evaluation of the
effect of higher concentrations of this acceptor (850 mM).
Table 1
Nucleophilicity (ω−) and pKa values for the different acceptors evaluated.
a
Acceptor
pKa
Nucleophilicity (ω−
)
Fructosylation
yield (%)b
Butanol
16.95
15.20
9.94
9.68
9.34
8.91
0.0194
0.0102
0.0065
0.0104
0.0080
0.0066
1.91
0.81
6.25
9.51
6.86
11.70
Benzyl alcohol
4-Methoxyphenol
Hydroquinone
Catechol
Resorcinol
a
Nucleophilicity calculated for the deprotonated species.
In all cases, the highest fructoside concentrations were obtained
at the highest concentration of acceptor (500 mM for dihydroxy-
b
Yield relative to sucrose conversion, using 500 mM acceptor and 400 mM
sucrose.

A. Mena-Arizmendi et al. / Journal of Molecular Catalysis B: Enzymatic 70 (2011) 41–48
47
relationship between fructosylation yield and pKa values. Indeed,
the fructosylation of 4HB on the aromatic hydroxyl (pKa = 9.48;
ω
− = 0.0041) proceeds 10 times faster than the fructosylation of
its primary hydroxyl (pKa = 15.21; ω− = 0.009). This indicates that
the higher reactivity of phenolic groups may be explained by their
ability to be deprotonated and become activated nucleophiles.
In terms of BS-LVS catalytic behaviour, these facts suggest that
the higher acidity of phenolic compounds facilitates the loss of their
acidic proton, especially when they are confronted with the basic
environment of the catalytic site in BS-LVS, that is, when Glu342 is
deprotonated. Therefore, phenolic molecules become better accep-
tors, as higher efficiencies in enzymatic fructosylation reactions
with BS-LVS were observed.
4. Conclusions
Among several enzymes that are able to fructosylate hydropho-
bic compounds, BS-LVS was selected due to its activity in the
presence of aromatic or aliphatic alcohol acceptors. In particu-
lar, BS-LVS was very efficient in fructosylation reactions due to
its higher stability in the presence of hydrophobic compounds,
allowing concentrations of up to 500 mM in the reaction medium.
Three different activities are commonly present in BS-LVS with
sucrose: transfer of fructose to levan, transfer of fructose to an
acceptor and sucrose hydrolysis. Surprisingly, the common strategy
to improve glycosylation of hydrophobic molecules by the addition
of organic co-solvents resulted in an important reduction of sucrose
hydrolysis and also a decrease in the fructosylation yield, favour-
ing the fructose transfer to levan. This behaviour was theoretically
explained in terms of the effect of the solvent on the acceptor ther-
modynamic activity, and this concept may become a useful tool in
predicting and designing acceptor reactions in organic media.
Several fructosides were produced and purified through accep-
tor reactions in aqueous media; the synthesis of Fru-Cat, Fru-Rsr,
Fru-4mPh, Fru2-4HB and Fru-Ph is reported for the first time.
Indeed, to our knowledge, we report the highest concentration
of Fru-Hq (30 mM) using a fructosyltransferase as catalyst. It was
found that phenolic compounds are fructosylated with higher
efficiencies than primary alcohols, and this reaction selectivity cor-
relates with the pKa value of the acceptor, with higher yields at low
pKa values, where the amount of deprotonated acceptor is higher.
The present enzymatic procedures and design tools represent a
good reference for the synthesis of novel glycosylated products,
with potential use in cosmetics and pharmaceuticals.
Fig. 9. Evolution of the fructosylation products in acceptor reactions of BS-LSV with
4HB. Reaction conditions: 400 mM sucrose, 200 mM 4HB, 5 U mL−1 at 30 ◦C.
are higher than those previously reported for phenolic compounds
(4 mM) [24].
3.6. Influence of acceptor nucleophilicity on the fructosylation
reaction
To explain the better acceptor selectivity shown by BS-LVS for
phenols over primary alcohols, a bifunctional compound such as 4-
hydroxybenzyl alcohol (4HB) was used as an acceptor (Fig. 7). This
substrate allows a direct selectivity comparison because it holds
a phenol and a primary hydroxyl within the same molecule. The
reaction was carried out in a medium saturated with 200 mM 4HB
(water solubility of 4HB is around 54 mM at 20 ◦C).
Two different fructosylated products Fru1-4HB and Fru2-4HB
were identified, corresponding to the fructosylation of 4HB on the
primary and aromatic –OH, respectively, as confirmed by NMR and
HPLC.
When evaluating the reaction kinetics, it was observed that
the initial rate of Fru1-4HB formation was 10 times lower
(5.03 ␮mol L−1 min−1) than the initial rate of Fru2-4HB formation
(50.64 ␮mol L−1 min−1), demonstrating the higher specificity of BS-
LVS for aromatic hydroxyls (Fig. 9).
According to the acid/base reaction mechanism proposed for BS-
LVS [44,45], in the first step, the Glu342 residue donates a proton to
the glycosyl oxygen atom of the leaving glucosyl residue, whereas
the nucleophile Asp86 residue forms a typical glycosyl-enzyme
intermediate. In the second step, the deprotonated Glu342 acts
as a general base to activate (deprotonate) the acceptor molecule,
making it able to attack the fructose C2 to form the new glycosidic
linkage (Fig. S2 in SI file).
Acknowledgments
E. Castillo thanks DGAPA-UNAM for a sabbatical fellowship
and INSA-Toulouse for an invited researcher fellowship. A. Mena-
Arizmendi thanks CONACyT for Ph.D. fellowship 169648. This work
has been supported by the grants IN226706-3 PAPIIT-UNAM and
CONACyT 40609-Z. Authors also thank M.E. Rodríguez and F. Gon-
zalez from IBt-UNAM for technical assistance, as well as Victoria
Labastida and Gabriela Vargas from CIQ-UAEM, for mass spec-
troscopy and NMR analysis, and J.M. Hurtado from IBt-UNAM for
assistance with computing resources.
Deprotonated hydroxyls in the acceptor molecules are regarded
as better nucleophiles than their non-deprotonated species. There-
fore, a critical step in the acceptor reaction is the ability of a
hydroxylated acceptor molecule to donate a proton and become
a better nucleophile. This property may be associated with the pKa
of the acceptor.
To explain the different reactivity of primary and aromatic alco-
hols in acceptor reactions with BS-LVS, the nucleophilicity (ω−)
values of the deprotonated species of some of the acceptors studied
were calculated according to Jaramillo et al. [32,33] and compared
to pKa values reported for different aromatic and primary alcohols.
These values are given in Table 1, where it is clear that the values
of ω− are in the same order of magnitude for both aromatic and
primary deprotonated alcohols, indicating that once the aromatic
and primary alcohols are deprotonated, both species may behave as
good nucleophiles. Nevertheless, the same data revealed an inverse
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcatb.2011.02.002.
References
[1] B.M. de Roode, A.C.R. Franssen, A. van der Padt, R.M. Boom, Biotechnol. Prog.
19 (2003) 1391–1402.

48
A. Mena-Arizmendi et al. / Journal of Molecular Catalysis B: Enzymatic 70 (2011) 41–48
[2] P.R. Cheeke, Encyclopedia of Life Sciences, John Wiley & Sons Ltd., 2001.
[23] G.H. Meulenbeld, S. Hartmans, Biotechnol. Bioeng. 70 (2000) 363–369.
[24] J. Kang, Y.-M. Kim, N. Kim, D.-W. Kim, S.-H. Nam, D. Kim, Appl. Microbiol.
Biotechnol. 83 (2009) 1009–1016.
[25] E. Castillo, A. Lopez-Munguía, J. Biotechnol. 114 (2004) 209–217.
[26] M. Canedo, M. Jimenez-Estrada, J. Cassani, A. López-Munguía, Biocatal. Bio-
transform. 16 (1999) 475–485.
[27] M.A. Perez-Oseguera, L. Guereca, A. López-Munguía, Appl. Microbiol. Biotech-
nol. 45 (1996) 465–471.
[28] M.E. Ortiz-Soto, V. Olivares-Illana, A. Lopez-Munguia, Biocatal. Biotransform.
22 (2004) 275–281.
[3] R.W. Hofmann, E.E. Swinny, S.J. Bloor, K.R. Markham, K.G. Ryan, B.D. Campbell,
B.R. Jordan, D.W. Fountain, Ann. Bot. 86 (2000) 527–537.
[4] H. Matsuda, T. Wang, H. Managi, M. Yoshikawa, Bioorg. Med. Chem. 11 (2003)
5317–5323.
[5] B. Mishra, K.I. Priyadarsini, M.S. Kumar, M.K. Unnikrishnan, H. Mohan, Bioorg.
Med. Chem. 11 (2003) 2677–2685.
[6] C. Rice-Evans, Curr. Med. Chem. 8 (2001) 797–807.
[7] P.H. Wu, R. Giridhar, W.T. Wu, Biotechnol. Bioeng. 95 (2006) 1138–
1147.
[8] F. van Rantwijk, M.W.V. Oosterom, R.A. Sheldon, J. Mol. Catal. B: Enzym. 6 (1999)
511–532.
[29] C. Olvera, S. Centeno-Leija, A. Lopez-Munguia, Antonie van Leeuwenhoek 92
(2007) 11–20.
[9] M. Schottelius, H.J. Wester, J.C. Reubi, R. Senekowitsch-Schmidtke, M.
Schwaiger, Bioconjug. Chem. 13 (2002) 1021–1030.
[10] A.V. Demchenko, Curr. Org. Chem. 7 (2003) 35–79.
[11] F.A. Shaikh, S.G. Withers, Biochem. Cell Biol. 86 (2008) 169–177.
[12] M. Skerget, P. Kotnik, M. Hadolin, H.R. Hras, M. Simonic, Z. Knez, Food Chem.
89 (2005) 191–198.
[30] J.B. Summer, S.F. Howell, J. Biol. Chem. 108 (1935) 51–54.
[31] J. Chandrasekhar, J.G. Andrade, P.V.R. Schleyer, J. Am. Chem. Soc. 103 (1981)
5612–5614.
[32] P. Jaramillo, P. Pérez, P. Fuentealba, J. Phys. Chem. A 113 (2009) 6812–6817.
[33] P. Jaramillo, P. Pérez, R. Contreras, W. Tiznado, P. Fuentealba, J. Phys. Chem. A
110 (2006) 8181–8187.
[13] Y.G. Du, G.H. Wei, R.J. Linhardt, J. Org. Chem. 69 (2004) 2206–2209.
[14] K. Toshima, K. Tatsuta, Chem. Rev. 93 (1993) 1503–1531.
[15] B.M. de Roode, H. Zuilhof, M.C.R. Franssen, A. van der Padt, A. de Groot, J. Chem.
Soc. Perkin Trans. 2 (2000) 2217–2224.
[34] A. Klamt, J. Phys. Chem. 99 (1995) 2224–2235.
[35] A. Klamt, F. Eckert, Fluid Phase Equilibr. 172 (2000) 43–72.
[36] A. Klamt, COSMO-RS: From Quantum Chemistry to Fluid Phase Thermodynam-
ics and Drug Design, Elsevier, Amsterdam, 2005.
[16] R. Vijayakumar, S. Divakar, Biotechnol. Lett. 27 (2005) 1411–1415.
[17] R. Prodanovic, N. Milosavic, D. Sladic, M. Zlatovic, B. Bozic, T.C. Velickovic, Z.
Vujcic, J. Mol. Catal. B: Enzym. 35 (2005) 142–146.
[37] A. Gómez de Segura, M. Alcalde, M. Bernabé, A. Ballesteros, F.J. Plou, J. Biotech-
nol. 124 (2006) 439–445.
[38] J. Seibel, R. Beine, R. Moraru, C. Behringer, K. Buchholz, Biocatal. Biotransform.
[18] F.J. Plou, A.G. de Segura, A. Ballesteros, in: J. Polaina, A.P. MacCabe (Eds.), Indus-
trial Enzymes, Springer, Valencia, Spain, 2007, pp. 141–157.
[19] K. Nakahara, M. Kontani, H. Ono, T. Kodama, T. Tanaka, T. Ooshima, S. Hamada,
Appl. Environ. Microbiol. 61 (1995) 2768–2770.
24 (2006) 157–165.
[39] A. Madlová, M. Antosˇová, M. Polakovic, V. Bálesˇ, Chem. Pap. 54 (2000) 339–344.
[40] J. Dudíková, M. Mastihubová, V. Mastihuba, N. Kolarova, J. Mol. Catal. B: Enzym.
45 (2007) 27–33.
[20] A. Bertrand, S. Morel, F. Lefoulon, Y. Rolland, P. Monsan, M. Remaud-Simeon,
Carbohydr. Res. 341 (2006) 855–863.
[21] G.H. Meulenbeld, H. Zuilhof, A. van Veldhuizen, R.H.H. van den Heuvel, S. Hart-
mans, Appl. Environ. Microbiol. 65 (1999) 4141–4147.
[41] G.C. Sandoval, J.-S. Condoret, A. Marty, AIChE J. 47 (2001) 718–726.
[42] A. Klamt, F. Eckert, W. Arlt, Annu. Rev. Chem. Biomol. Eng. 1 (2010) 101–122.
[43] R. Macomber, Organic Chemistry, vol. II, University Science Books, 1996, pp.
160–183.
[22] Y.H. Moon, G. Kim, J.H. Lee, X.J. Jin, D.W. Kim, D. Kim, J. Mol. Catal. B: Enzym.
40 (2006) 1–7.
[44] G.Y. Meng, K. Fütterer, Nat. Struct. Biol. 10 (2003) 935–941.
[45] R. Chambert, G. Treboul, R. Dedonder, Eur. J. Biochem. 41 (1974) 285–300.