Highly efficient and regioselective enzymatic synthesis of β-(1→3) galactosides in biosolvents

Article abstract of DOI:10.1039/c3ra40860d

A green synthesis for β-(1→3) galactosyldisaccharides that combines the use of a biodegradable biocatalyst, aqueous solutions, and solvent recycling (renewable and derived from biomass) has been developed. The use of biomass-derived solvents allows good catalytic activity in the synthesis of Gal-β-d-(1→3)GlcNAc and Gal-β-d-(1→)3GalNAc (99% and 95% yields respectively) with β-Gal-3-NTag β-galactosidase, preventing hydrolytic activity and with full regioselectivity. This represents a considerable improvement over the use of an aqueous buffer or conventional organic solvents. Furthermore, reaction scaling up and biosolvent recycling are feasible without losing catalytic action. In order to understand the role of these green solvents in the enzyme's synthetic behaviour, different structural studies were performed (fluorescence and molecular modelling) in the presence of some selected biosolvents to conclude that the presence of green biosolvents in the reaction media modifies the enzyme's tertiary structure allowing better substrate disposition in the active site, most probably due to solvation effects, explaining the behaviour observed. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c3ra40860d

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Highly efficient and regioselective enzymatic synthesis
of b-(1A3) galactosides in biosolvents3
Cite this: RSC Advances, 2013, 3,
12155
a
b
a
c
´
´
´
´
´
´
Carlos Bayon, Alvaro Cortes, Antonio Aires-Trapote, Concepcion Civera and Marıa
a
´
Jose Hernaiz*
A green synthesis for b-(1A3) galactosyldisaccharides that combines the use of a biodegradable
biocatalyst, aqueous solutions, and solvent recycling (renewable and derived from biomass) has been
developed. The use of biomass-derived solvents allows good catalytic activity in the synthesis of Gal-b-D-
(1A3)GlcNAc and Gal-b-D-(1A)3GalNAc (99% and 95% yields respectively) with b-Gal-3-NTag b-galacto-
sidase, preventing hydrolytic activity and with full regioselectivity. This represents a considerable
improvement over the use of an aqueous buffer or conventional organic solvents. Furthermore, reaction
scaling up and biosolvent recycling are feasible without losing catalytic action. In order to understand the
role of these green solvents in the enzyme’s synthetic behaviour, different structural studies were
performed (fluorescence and molecular modelling) in the presence of some selected biosolvents to
conclude that the presence of green biosolvents in the reaction media modifies the enzyme’s tertiary
structure allowing better substrate disposition in the active site, most probably due to solvation effects,
explaining the behaviour observed.
Received 20th February 2013,
Accepted 3rd May 2013
DOI: 10.1039/c3ra40860d
www.rsc.org/advances
the Leloir pathway and are difficult to obtain and have limited
stability. Moreover, they require expensive cofactors as glycosyl
Introduction
There is a current need in the chemical and biochemical
industry to develop green synthetic routes, involving renew-
able raw materials, to replace unfriendly ones that usually
generate toxic residues and/or involve harsh reaction condi-
tions.
donors. Glycosyl hydrolases (glycosidases) can also be used to
synthesize oligosaccharides in a kinetically controlled reac-
tion, in which a glycosyl donor is used to transfer its glycosyl
residue to a sugar acceptor, present in the reaction med-
ium.16,17 In spite of the increasing amount of work carried out
with glycosyl hydrolases, their main drawbacks are low yields
and lack of regioselectivity,18 which limits their use for
synthetic purposes. In an important contribution to glycosi-
dase-catalyzed oligosaccharides synthesis, Ito et al.19 reported
for the first time the use of a b-galactosidase from Bacillus
circulans to synthesise some b-D-(1A3) galactosyldisaccharides
bearing a GlcNAc or a GalNAc residue at the reducing end.
Some b-D-(1A6) linkages were formed as well, but to the best
of our knowledge, it was the first example of a preparative
scale b-D-galactosyl transfer that occurs preferentially at the
O-3 position using a galactosyl hydrolase. There is consider-
able synthetic interest in the use of this galactosidase to
perform this type of galactosyl transfer.
Nowadays Gal-b-D-(1A3)GalNAc and Gal-b-D-(1A3)GlcNAc
are very well known epitopes associated with carcinoma cells
and other crucial intercellular recognition events.1–6 Both are
also important constituents of mucin type or complex type
glycoproteins7. Therefore, there is a great interest in the
synthesis of these compounds in order to deepen our
knowledge of their biological functions. Organic chemical
methods for obtaining them have been developed8–13 but they
involve several elaborate protection and deprotection proce-
dures. On the other hand, enzymatic synthesis using glycosyl
transferases has also been reported.14–16 Glycosyl transferases
are widely used to perform regiospecific galactosylation and
silylation on a preparative scale, but these enzymes belong to
Recently, it has been shown that the use of green
solvents,20,21 such as biomass derived solvents or ionic liquids,
as co-solvents in the enzymatic reaction lead to a variation of
the enzyme activity.22–29
aDepartment of Organic and Pharmaceutical Chemistry, Faculty of Pharmacy,
Complutense University of Madrid, Campus de Moncloa, 28040 Madrid, Spain
b
´
´
Unidad de Bioinformatica. Centro de Biologıa Molecular ‘‘Severo Ochoa’’ (CBMSO),
´
´
CSIC, Autonoma University of Madrid (UAM), C/Nicolas Cabrera 1, 28049, Madrid,
Spain
Biomass derived solvents show some new and attractive
advantages, as there are made from a renewable source, they
are (presumably) less toxic, exhibit low volatility, and possess
tuneable physicochemical properties.20,21,30–32 We have pre-
viously described the use of some of these biomass derived
cDepartment of Physics and Chemistry II, Faculty of Pharmacy, Complutense
University of Madrid, Campus de Moncloa, 28040 Madrid, Spain.
E-mail: mjhernai@ucm.es; Fax: (+) 34 9139 41822; Tel: (+) 34 91139 41821
Electronic supplementary information (ESI) available. See DOI: 10.1039/
3
c3ra40860d
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 12155–12163 | 12155

View Article Online
Paper
RSC Advances
solvents in the transglycosylation reaction catalyzed by
Biolacta b-galactosidase, using p-nitrophenyl-b-D-galactopyra-
noside (p-NP-b-Gal) as donor and GlcNAc as acceptor,
changing the regioselectivity of the enzyme from a mixture
of b(1A4) and b(1A6) towards the synthesis of only b(1A6)
oligosaccharides.26
Thus, the aim of the present work was to produce a easily
purified b-Gal-3 from Bacillus circulans ATCC 31382 and
explore the possibility of combining the advantages of using
of this enzyme, a biodegradable biocatalyst, and solvents from
renewable sources to the efficient and regioselective synthesis
of b-D-(1A3) galactooligosaccharides.
Results
Scheme 1 Synthesis of Gal-b-(1A3)-GlcNAc and Gal-b-(1A3)-GalNAc catalyzed
by b-Gal-3-NTag.
Production of recombinant enzyme b-Gal-3 from Bacillus
circulans ATCC 31382
Two recombinant b-Gal-3 enzymes were obtained, one cloned
in pET28b+ carrying an N-terminal 6-histidine tag (b-Gal-3-
NTag), and the other, cloned in pET22b+ carrying a C-terminal
one (b-Gal-3-CTag). After purification (Fig. 1), the two enzymes
were assayed to determine the influence of the HistTag on the
hydrolytic activity. b-Gal-3-NTag showed an activity of 12.2 U
mg21 and b-Gal-3-CTag of 12.5 U mg21. Fujimoto et al.
characterized the wild type enzyme, obtaining an activity of
Enzymatic synthesis in presence of green solvents with b-Gal-3
from Bacillus circulans ATCC 31382
The transglycosylation reaction of p-nitrophenyl-b-D-galacto-
pyranoside (p-NP-b-Gal) as donor and N-acetyl glucosamine
(GlcNAc) or N-acetyl galactosamine (GalNAc) as acceptors,
catalyzed by b-GaL-3-NTag from Bacillus circulans ATCC 31382
will be used throughout this work (Scheme 1).
5.13 U mg .
21 18 This lower activity indicates that the technology
The regioselectivity showed by the enzyme was investigated
when the reaction was carried out in 50 mM sodium
phosphate buffer at pH 6 (Fig. 2 and 3). This reaction can
afford Gal-b-(1A3)-GlcNAc as a major product (50.7% conver-
sion) and 48.9% of hydrolysis product (Gal) when GlcNAc was
used as acceptor, while with GalNac the main product formed
was Gal-b-(1A3)-GalNAc with similar yields (49.3%) and 48.4%
of hydrolysis product.
Three different groups of solvents (Scheme 2) were screened
in order to evaluate their effect over b-Gal-3-NTag activity
during transglycosylation reactions: glycerol based solvents
with cyclic structures (S1–S3), glycerol based solvents with
open chain (S4–S12) and 3-N,N- dimethylamide based solvents
(S13–S15). Main features of these solvents (density, Log P and
solubility 2M) are summarized in Table 1.
based on histidine tags achieves better enzyme purification
than ion exchange and gel permeation techniques previously
used. Although our recombinant enzymes showed equivalent
activity, the different antibiotics added to the growing medium
have an effect on the recombinant enzyme production. Larger
amounts of enzyme are obtained in kanamycin medium than
in ampicillin, 21 mg and 9 mg per litter of grown medium
respectively. Because of that, the enzyme used in subsequent
assays is b-Gal-3-NTag, cloned in pET28b+.
Transglycosylation reactions catalyzed by the b-GaL-3-NTag
were carried out following the general procedure described in
the experimental section and monitored by HPLC. The
concentration of green solvents was fixed to 2 M in the
sodium phosphate buffered mixture. The results obtained are
summarized in Fig. 2 (GlcNAc as acceptor) and Fig. 3 (GalNAc
as acceptor).
The use of only buffer as reaction medium leads to a good
conversion of p-NP-b-Gal but there is a considerable amount of
hydrolytic product, galactose (48.9 and 48.4% respectively).
In presence of green solvents, the Gal-b-(1A3)-GlcNAc was
formed as main product (Fig. 2), except for S3, S5, S6 and S14.
In fact, S6 and S14 give rise to a total loss of enzymatic activity
when used as co-solvents in this reaction. However, solvents
S10 and S13 lead to the best yields for the synthesis of Gal-b-
Fig. 1 DS-electrophoresis of the purification of b-Gal-3. Lane A, non-induced cell
extract. Lane B, IPTG induced cell extract. Lane C, purified b-Gal-3-NTag. Lane D,
b-Gal-3-CTag. Lane E, molecular weight standard.
12156 | RSC Adv., 2013, 3, 12155–12163
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
Fig. 2 Transglycosylation yields obtained with b-Gal-3-NTag using buffer 2 M solvents derived from biomass with p-NP-Gal as donor and GlcNAc as acceptor.
Fig. 3 Transglycosylation yields obtained with b-Gal-3-NTag using buffer-2 M solvents derived from biomass with p-NP-Gal as donor and GalNAc as acceptor.
(1A3)-GlcNAc (93% and 99%, respectively), with no appear-
ance of hydrolysis product. This means not only an important
increase of enzymatic activity in comparison to the enzyme’s
natural behaviour in 50 mM phosphate buffer at pH 6, but also
an improvement in the selectivity of the reaction (no
hydrolysis product), with full regioselectivity (b-(1A3) isomer).
The synthetic activity of b-galactosidase was examined when
GalNAc was used as acceptor. Results are shown in Fig. 3, and
are similar of that obtained with GlcNAc. In the presence of
solvents S4, S9 and S11, Gal-b-(1A3)-GalNAc was formed as
main product, but for S3, S5, S6, S7 and S14 no enzymatic
activity was observed. Again, solvents S10 and S13 lead to the
best yields for the synthesis of Gal-b-(1A3)-GlcNAc (86% and
95%, respectively), with no appearance of hydrolysis product.
Table 1 Density, Log P and solubility 2 M of the different solvents
Solvent
Solvent density (g ml21
)
Log P
System composition
S1
S2
S3
S4
S5
S6
S7
S8
1.24
1.41
1.06
1.07
0.94
0.91
1.12
1.36
1.27
0.90
0.89
0.89
1.06
1.15
0.91
20.057
20.024
0.030
20.60
0.14
0.27
1.14
1.42
1.71
1.93
2.07
Monophasic
Monophasic
Monophasic
Monophasic
Monophasic
Monophasic
Biphasic
Biphasic
Biphasic
Biphasic
Biphasic
S9
S10
S11
S12
S13
S14
S15
2.48
20.69
1.41
Biphasic
Monophasic
Monophasic
Biphasic
Scheme 2 Structure of the different green solvents employed in transglycosyla-
tion reaction with b-Gal-3-NTag.
1.42
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 12155–12163 | 12157

View Article Online
Paper
RSC Advances
Fig. 5 HPLC chromatogram of the reaction catalyzed by b-Gal-3-NTag using
buffer-2 M S13 with p-NP-Gal as donor and GlcNAc as acceptor. (a) Reaction
debris; (b) GlcNAc; (c) Gal; (d) Gal-b-(1A3)-GlcNAc.
solvent elimination, based on direct hexane precipitation at
2196 uC without prior extraction. The pellets obtained
(composed by the mixture of saccharides) were purified by
filtration. S13 was separated by rotary evaporation of the
hexane, recovering the total amount of both solvents. After the
purification process, S13 can be reused in further reactions.
On the other hand, an advantage of solvents S9 and S10,
compared with other solvents used in this study, is that under
these experimental conditions an biphase mixture is created
between these solvents and aqueous buffer (Table 1), then
after reaction these solvents can be separated from reaction
media by centrifugation. Moreover, carbohydrate compounds
in the reaction media are not soluble in these solvents phase
and remains in the aqueous phase. Centrifugation becomes a
very useful tool for the isolation of these solvents from the
reaction media, allowing its reuse in further reactions.
With these optimized separation procedures in hand, we
proceeded to the systematic semi-preparative experiments in
80 ml of reaction media in presence of S9, S10 and S13. In all
cases the yield of pure disaccharide after purification through
a carbon-celite column, compared with initial p-NP-Gal, was
85% (2,17 g of Gal-b-(1A3)-GlcNAc).
Fig. 4 Transglycosylation yields (%) obtained with b-Gal-3-NTag at different S13
concentrations (1M, 2M and 3M) with p-NP-Gal as donor and GlcNAc as
acceptor in (a) and GalNAc in (b).
The use of S13 showed yields over 95% in the synthesis of both
disaccharides, therefore this solvent was selected for further
assays.
There is not a clear relationship between solvent structures
and the available physical parameters (log P and density) with
the noticeable increase of synthetic activity (Table 1).
Effect of the co-solvent concentration on the
transglycosylation reactions
In order to study the influence of co-solvent concentration on
the yield boost, the reaction was carried out with different
amounts of solvent S13 (1, 2 and 3 M). Results obtained are
shown in Fig. 4. As it can be seen, in a S13 2 M concentration,
the best yields for (Gal-b-(1A3)-GlcNAc and Gal-b-(1A3)-
GalNAc) were obtained (99% and 95% respectively). The same
effect was previously observed in b-galactosidase from Thermus
thermophilus and b-galactosidase from Escherichia coli when
these biosolvents were used as co-solvents. It was attributed to
a conformational change in the enzyme secondary and tertiary
structures caused by the co-solvent.28,29
Solvents effects in tertiary structure by fluorescence study
To explore the effect of S13 on the enzyme structure and its
potential conformational changes, we measured the fluores-
cence variations at different S13 concentrations. The enzyme
was first dissolved in 10 mM sodium phosphate buffer (pH
6.00) to record a native emission spectrum, then some aliquots
of solvent were added sequentially. As a result, a red shift was
observed when the enzyme is placed in 2 M S13 (concentration
for optimal transglycosylation yield), the maximum fluores-
cence emission is shifted to 345.87 nm from 340.63 nm
(Fig. 6). The spectra obtained at different solvent concentra-
tions tested for transglycosylation reactions (Fig. 6) show a
maximum peak at 344.76 nm for 1 M solvent concentration
and a maximum peak on 346.59 nm for 3 M. These results
suggest a modification of the tryptophan chemical environ-
ment which is the main fluorescence residue, which most
likely involves a conformational modification. However, it is
well known that red-shift of the emission wavelength is due to
Purification stage
The feasibility of the target molecule isolation from the
reaction medium must be evaluated after the synthesis of an
oligosaccharide. Fig. 5 shows a typical HPLC chromatogram of
a sample that was withdrawn during an enzymatic synthesis of
Gal-b-(1A3)-GlcNAc using S13 as a cosolvent. After the
enzymatic synthesis of Gal-b-(1A3)-GlcNAc in the presence
of 2 M S13, a mixture of Gal-b-(1A3)-GlcNAc and GlcNAc
(which is used overconcentrated) is obtained as observed in
Fig. 5.
Solvent S13 is water soluble and cannot be eliminated by
lyophilisation, percolation, rotary evaporation or organic
extraction. An alternative protocol was developed for green
12158 | RSC Adv., 2013, 3, 12155–12163
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
Fig. 6 Normalized emission fluorescence spectra for b-Gal-3-NTag at different
concentrations of S13.
Fig. 7 Modelling of the b-Gal-3-NTag active site.
the catalytic glutamic acid residues pair (Fig. 7). The sugar core
is located on top of the aromatic residue while the N-acetyl
chain is involved in a hydrogen bond. These interactions allow
the hydroxyl groups to be located within a distance of the
glycosylated glutamic acid and the free one, and could allow
the reaction to occurs. Accommodation of the substrate in the
active site may evolve very differently depending on the solvent
and the flexibility conditions of the enzyme, being the reason
to perform molecular dynamic simulations to obtain some
insight into these.
Molecular dynamics. In previous studies, we have found
some correlation between reactant groups distance and
regioselectivity and/or reaction yields.29 Measuring the dis-
tances between the glycosylated glutamic residue (Glu 233-
galactose intermediate) and the hydroxyl group in the O3
position of GlcNAc in the presence of water as solvent and a
mixture of S13-water, we found that shorter distances are
translated in enhanced conversion (Fig. 8).
an increased exposure of tryptophan residues to a more polar
environment, which is usually related with protein denatura-
tion/enzyme deactivation.33 Nevertheless, since CD spectrum
of S13 in the far UV range (200–250 nm) was too noisy to reveal
any change of secondary structure of protein, we have studied
the thermostability of the b-gal 1–3 in buffer and with S13 2 M
by fluorescence spectroscopy (see Fig. S1 in ESI ). As it is shows
3
in this Figure the presence of S13 at higher temperatures of 60
uC shows a sharper slope indicating that S13 destabilizes the
protein, however, in fact this change occurs at high tempera-
ture not at 37 uC, our running temperature. According to this
result, the effect of S13 on the enzyme’s structure could be
attributed to a more open conformation of the enzyme that
would destabilize the protein.
Molecular modelling studies
In previous work, our group found that the change of
regioselectivity cannot be easily related to the biomass-derived
solvent’s physical parameters (density, system composition,
log P, etc.).29 To get some insights on the origin of the activity
boost observed in some of the solvents tested, a molecular
dynamics simulation along with a fluorescence study were
carried out, comparing pure water with the S13-water mixture,
as solvating medium of the b-gal-3 from Bacillus circulans
ATCC 31382.
The solvent could play a key role in these changes modifying
the interaction pattern between substrate and enzyme.
Therefore, we measured the calculated non-bonded energy
terms of substrate and protein to try to estimate the different
contributions that could improve the binding energy and
therefore, the overall conversion.
Active site. Since no ligand or any other indication of the
active site was available, we have performed multiple sequence
alignment and structure superimposition with the b-galacto-
sidase (lacZ) of Escherichia coli. Two glutamic residues
appeared to be conserved across the compared sequences
and corresponded roughly in the structure superimposition
with the catalytic pair of glutamic acids in the lacZ. These two
residues, Glu-157 and Glu-233 were located in the center of a
well-defined cavity in the homology model and were selected
as the catalytic pair. Glu-233 corresponds with the nucleophilic
residue in lacZ while Glu-157 is aligned with the acid–base
catalyst (Fig. 7).
Docking results. Most populated docking results for the
second substrate in the glycosylated enzyme may confirm the
role of the previously mentioned residues, Trp-235 and Tyr-
450, in the active site substrate recognition and orientation to
Fig. 8 Distances between the glycosylated glutamic residue and the hydroxyl
group in the O3 position of GlcNAc in water (black) and a mixture of S13-water
(blue).
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 12155–12163 | 12159

View Article Online
Paper
RSC Advances
Table 2 Average force field energy terms in molecular dynamics simulation
Solvent
Electrostatic term (kJ mol21
)
Van der Waals term (kJ mol21
)
Total MM energy (kJ mol21
)
Water
S13-water
213.27
214.64
276.54
274.62
289.81
289.27
As it can be seen in Table 2, the total calculated energy is
almost equal in the mixture of S13-water. However, in
electrostatic terms it is larger in mixed solvents than in water
alone and by the Circe effect principle, which states that an
increase of reaction speed in enzymatic environments can be
mostly explained by a increase in the electrostatic interaction
between enzyme and substrate, can provide a reasonable
explanation of the enhanced conversion rate under mixed
solvent conditions. Also, in the mixture case, the cost of
desolvation is predicted to be less unfavourable since the
substrate and the active site are both polar in nature and
therefore interact strongly with the water molecules.
After this analysis we can conclude that the great regios-
electivity changes during transglycosylation reactions observed
in this study are caused by the interaction of the co-solvents
with the enzyme, causing the modification of its secondary
and tertiary structures (Fig. 6), also explaining the molecular
modelling and docking results (Fig. 8).
tion was done by Bradford method,36 using bovine serum
albumin as standard.
Hydrolytic reactions
Hydrolytic activity was determined by spectrophotometrical
quantification of p-NP liberated from the hydrolysis of p-NP-
b-galactopyranoside 5 mM in sodium phosphate buffer 50
mM, pH 7 in a 300 mL cell by measuring the increase in
absorbance at 410 nm during
3 min at 37 uC. Each
experimental assay was run at least three times with standard
deviation under 5% of the samples average. One enzyme unit
(U) was defined as the amount (mg) of protein that hydrolyzes
1.0 mmol of substrate per minute.
Transglycosylation reactions
p-NP-b-Gal (85 mM) and GlcNAc or GalNAc (425 mM) were
dissolved in 1.00 mL of buffer sodium phosphate 50 mM pH 6
and 2M of solvents from biomass as final concentration, and
pre-warmed at reaction temperature (37 uC). Reaction started
by addition of biocatalyst to the mixture: 5 U of recombinant
enzyme. In order to check the effect of different concentra-
tions of S13, reactions were run in 1.00 mL of buffer sodium
phosphate 50 mM pH 6 and 1M, 2M and 3M of solvent as final
concentration.
Aliquots (50 mL) were withdrawn from reaction media at
different times. The reaction was stopped after 3 h by heating
to 100 uC for 5 min and conserved immediately at 220 uC.
Analytical determination of products were performed by HPLC
using a NH2P50-4E amino column (Asahipak, Japan) using
three detectors: ELSD (Evaporative Light Scattering), UV-Vis at
317 nm and CD (Circular Dichroism).
Materials and methods
Materials and reagents
Solvents S1, S2, S3, S4, S5, S6, S7, S8, S9, S10, S11 and S12 were
synthesized using the same procedures previously described.34
S13, S14 and S15 were a gift from COGNIS IP Management
GmbH (Germany).
Bovine serum albumin (BSA), p-nitrophenol (p-NP), p-nitro-
phenyl-b-D-galactopyranosyde (p-NP-b-Gal), N-acetyl-D-glucosa-
mine (GlcNAc), N-acetyl-D-galactosamine (GalNAc) and
analytical standards of monosaccharides for HPLC were
purchased from Sigma Aldrich. All other chemicals were of
analytical grade.
Purification of the reaction products in semi-preparative
conditions
The scaled-up reaction mixture was composed by p-NP-b-Gal
85 mM and GlcNAc 425 mM dissolved in 80 mL of buffer
sodium phosphate 50 mM pH 6 with S13 2M. The reaction was
stopped after 3 h by heating to 100 uC for 5 min. After that, it
was lyophilized with the aim of eliminating the water. The
crude obtained (30 ml) was mixed with 30 mL of hexane and
freeze at 2196 uC using liquid nitrogen. The resulting pellet
was recovered by filtration and dried. Solvent from biomass
were separated from hexane by rotary evaporation, recovering
the total amount of both solvents.
The pellet was directly loaded onto a carbon–celite column
(50% m/m), eluted with a linear gradient from 0% to 15% (v/v)
of ethanol in water. Solvents were eliminated and disaccharide
stored in freeze. 30 mg were dissolved in D2O to be
characterized by 1H-NMR and 13C-NMR spectroscopy (D2O,
Production of the enzyme and purification
Recombinant b-galactosidase from Bacillus circulans 31382 was
cloned in Escherichia coli BL21 using pET28b+ and pET22b+
vectors (Novagen). E. coli cultures were grown aerobically at 37
uC in LB broth with kanamycin (30 mg L21) and induced with
IPTG (isopropyl b-D-thiogalactopyranoside, 1 mM) at 37 uC for
5 h. Cells were disrupted by sonic disruption, unbroken cells
and insoluble debris were eliminated by centrifugation (14.000
g for 15 min at 4 uC). The solution obtained was passed
through a Ni2+-agarose column (3 mL) according to manufac-
turer’s protocol (BioRad). Fractions were monitored by
absorbance at 280 nm, pooled, and concentrated and desalted
in an Amicon ultra centrifuge filter (Millipore). The purifica-
tion process was followed by SDS-PAGE.35 Protein quantifica-
12160 | RSC Adv., 2013, 3, 12155–12163
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
700 MHz-500 MHz). Spectra were consistent with previous
references.15,17,26,37
Center for Biotechnology Information (NCBI) web interface.
The results (e-value: 1e-146) yielded a 42% of sequence identity
with the b-galactosidase of Bacterioides thethaiotaomicron (PDB
ID: 3D3A, resolution 2.15 A) which was selected for performing
the model. We have used two automatic modelling servers:
SWISS-Model and CPHmodels-3.0 to build the model based on
the alignment with the b-galactosidase of Bacterioides thethaio-
taomicron. These two generated models were refined using the
GROMOS 96 43a1 force field38 to energetically minimize the
structure using the 1000 steps of the Steepest descent
algorithm. For validation and selection of the best model, we
have employed an analysis of the Ramachandran plot, the
ERRAT score and the Verify3D residue plots. After comparison,
the model generated with CPH models was selected since it
was the most compliant with the parameters found in
naturally folded proteins for the validation criteria.
Gal-b(1A3)GalNAc (Galacto-N-Biose, GNB)15 1H-NMR (700
.
MHz, D2O): 1.92 (s, Ac), 4.58 (d, J1b, 2 = 8.47 Hz, H-1b), 5.11 (d,
J1a 2 = 3.71 Hz, H-1a). 13C-NMR (700 MHz, D2O): 21.96 (Me of
Ac, a), 22.19 (Me of Ac, b), 48.92 (C-2a), 52.39 (C-2b), 60.89 (C-
6a), 60.93 (C-6b), 61.11 (C-69), 68.01 (C-49), 68.51 (C-4a), 68.68
(C-4b), 70.13 (C-5a), 70.57 (C-29), 72.48 (C-39), 74.78 (C-5b),
74.90 (C-59), 76.99 (C-3a), 80.01 (C-3b), 91.13 (C-1, a), 95.12 (C-
1b), 104.64 (C-19b), 104.81 (C-19a), 174.61 (CLO of Ac, a), 174.91
(CLO of Ac, b).
Gal-b(1A3)GlcNAc (Lacto-N-Biose, LNB)15,17 1H-NMR (500
.
MHz, D2O): 1.96 (s, 3H, Ac), 5.11 (d, J1a, 2 = 3.45 Hz, H-1a). 13C-
NMR (500 MHz, D2O): 22.39 (Me of Ac, a), 22.64 (Me of Ac, b),
53.28 (C-2a), 56.02 (C-2b), 60.98 (C-6), 61.39 (C-69), 68.94 (C-49),
69.10 (C-4), 71.12 (C-29), 71.62 (C-5a), 72.95 (C-39), 75.63 (C-59),
75.85 (C-5b), 80.57 (C-3a), 83.01 (C-3b), 91.42 (C-1a), 95.11 (C-
1b), 103.83 (C-19b), 103.96 (C-19 a), 174.93 (CLO of Ac, a),
175.19 (CLO of Ac, b).
Active site search and docking
The active site of the enzyme was obtained from structure
superimposition with E. coli b-galactosidase (PDB code 1DP0),
sequence conservation analysis and blind docking, yielding a
match of the catalytic pair Glu-157 and Glu-233 (nucleophilic
residue). Bearing in mind the catalytic mechanism of the
enzyme, and as we have done in other previous works,28,29 we
followed a two steps modelling on the enzyme. First, we
performed a docking simulation of p-nitro-phenol-b-galacto-
pyranose in the active site using Autodock 4.2,39 the default
Lamarckian genetic algorithm parameters and a grid of 60 6
60 6 60 points (spacing 0.375 Å). The best pose according to
both energetic (favourable docking energy) and geometrical
criteria (close enough to Glu-233 to allow any reaction) was
Gal-b(1A6)GlcNAc (N-Acetyl-allolactosamine)26,37 1H-NMR
.
(700 MHz, D2O): 1.94 (s, Ac), 4.33 (d, H19, J19, 29 = 7,91 Hz),
4.61 (d, H1b, J 1b,2 = 8,47 Hz) 5.09 (d, J1a, 2 = 3.58 Hz, H-1a).
13C-NMR: 21.81 (Me of Ac, a), 22.08 (Me of Ac, b), 53.95 (C-2 a),
56.52 (C-2b), 60.94 (C-69), 68.49 (C-49), 68.57 (C-6b) 68.59 (C-
6a), 69.62 (C-3a), 69.77 (C-5a), 70.52 (C-4b), 70.56 (C-4a), 70.56
(C-2), 72.58 (C-3b), 72.61 (C-39), 74.85 (C-5b), 75.11 (C-59), 90.81
(C-1a), 94.93 (C-1b), 103.26 (C-19b), 103.27 (C-19a),174.45 (CLO
of Ac a), 174.70 (CLO of Ac b).
Solvents effects in tertiary structure by fluoresence study
Fluorescence emission spectra were recorded on a Perkin
Elmer LS50B 295 nm was set as excitation wavelength because
its high selectivity for tryptophan. Emission was acquired from
310 to 400 nm. The excitation and emission bandwidth were
adjusted at 3 nm, measurements were done in 1 cm path
length cell at 25 uC. First, the fluorescence spectrum of S13 was
recorded using a 2M solution of this solvent, then for b-Gal-3-
NTag (0.1 mg mL21) in 10 mM sodium phosphate buffer at pH
6.00. The S13 emission spectrum was negligible compared to
the one of the enzyme in buffer solution. Then, several
additions of S13 were performed in the cuvette in order to
evaluate the effect of this solvent in the tryptophan exposition
of the enzyme. Thermal ramps of enzyme in buffer and at 2M
S13 were performed from 20 to 90 uC at a rate of and 60 uC
selected to build
a covalent complex model using the
galactopyranose moiety and the Glu-233 residue. Then, a
second docking simulation, using the glycosylated enzyme,
was carried out with the GlcNAc, selecting the most favourable
pose according to its predicted docking energy. This complex
was used as the starting point for molecular dynamics studies.
Solvent parameters
The solvent used in the simulations was already parameterized
for the GROMOS 96 43a1 force field,40 briefly: we first
optimized the solvent molecules using the Density
Functional Theory (DFT) at the B3LYP 6-31G** level, then
the bond lengths and angles obtained from this calculation
were used and adapted to the force field to match both
experimental density and the enthalpy of vaporization.
h
21.Spectra were acquired in a PTI modular spectrometer with
thermoelectric temperature regulator TLC 50, Spectra analysis
were done using Felix 32 software. (supplementary material)
Molecular dynamics
The final complex obtained by docking the two molecules
implicated in the catalytic mechanism of the enzyme was used
as the initial model for molecular dynamics simulation.
The protein was parameterized using the GROMOS 96 43a1
parameters, including galactose-glutamic parameters already
used in other works.28 GlcNac parameters were generated by
the Dundee PRODRG 2.5 Server, carefully checking the
generated topology as pointed out by Lemkul et al.41
Computational methods
Homology modelling
Due to the lack of a public structure with atomic resolution for
the b-gal-3 b-galactosidase of Bacillus circulans ATCC 31382, we
have generated a three dimensional homology model. With
this aim, we have located the closest homologue structure
deposited in the Protein Data Bank performing a Basic Local
Alignment Search Tool (BLAST) search with the National
For both, water and mixed solvent systems, we first solvated
the complex with the selected solvent and then we performed
the following protocol: a energy minimization with 2000 steps
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 12155–12163 | 12161

View Article Online
Paper
RSC Advances
7 A. Takada, K. Ohmori, T. Yoneda, K. Tsuyuoka,
A. Hasegawa, M. Kiso and R. Kannagi, Cancer Res, 1993,
53, 354–361.
8 R. U. Lemieux and H. Driguez, J. Am. Chem. Soc., 1975, 97,
4063–4069.
9 D. P. Khare, O. Hindsgaul and R. U. Lemieux, Carbohydr.
Res., 1985, 136, 285–308.
10 M. Wilstermann and G. Magnusson, Carbohydr. Res., 1995,
272, 1–7.
of SD algorithm and 7000 steps of Polak–Ribiere Conjugate
gradients (CG) was performed; a equilibration of 300 ps at 300
K (NVT), using the Particle-Mesh-Ewald (PME) to deal with
long range electrostatics, the v-rescale thermostat and
restrained positions in protein and ligand heavy atoms,
followed by 300 ps at 300 K and 1 atm (NPT) using the
Parrinello–Rahman barostat with the same conditions; and
finally, a 2 ns production simulation recording energy and
coordinates each 1 ps and a step of 2 fs.40 All simulations were
performed using the GROMACS v4.0.7 suite.38
11 R. R. Schmidt and E. Ru¨cker, Tetrahedron Lett., 1980, 21,
1421–1424.
12 R. R. Schmidt, Angew. Chem., Int. Ed. Engl., 1986, 25,
212–235.
13 H. Paulsen, Angew. Chem., Int. Ed. Engl., 1982, 21, 155–173.
14 G. Baisch, R. Ohrlein, M. Streiff and F. Kolbinger, Bioorg.
Med. Chem. Lett., 1998, 8, 751–754.
15 H. Yu, V. Thon, K. Lau, L. Cai, Y. Chen, S. Mu, Y. Li, P.
G. Wang and X. Chen, Chem. Commun., 2010, 46,
7507–7509.
16 A. D’Almeida, M. Ionata, V. Tran, C. Tellier, M. Dion and
C. Rabiller, Tetrahedron: Asymmetry, 2009, 20, 1243–1246.
17 A. Vetere, M. Miletich, M. Bosco and S. Paoletti, Eur. J.
Biochem., 2000, 267, 942–949.
Conclusions
An efficient strategy for the synthesis of Gal-b-D-(1A3)GlcNAc
and Gal-b-D-(1A)3GalNAc, combining renewable raw materials
and recycled solvents (derived from biomass), buffer solutions
and a biodegradable biocatalyst has been explored. The use of
biosolvents allows good yields in the enzymatic synthesis of b-
D-(1A3) linkages, preventing hydrolysis and with full regios-
electivity. Yields of up to 99% and 95% of Gal-b-D-
(1A3)GlcNAc and Gal-b-D-(1A)3GalNAc were obtained with
total substrate conversion. These reactions represent a con-
siderable improvement over the use of aqueous buffer or
conventional organic solvents. Optimization of the reaction
conditions allows the reaction’s scale up to 80 ml with further
solvent recycling and reuse. This recovery strategy could, in
principle, be extrapolated to other enzymatic reactions carried
out in these kind of green solvents.
18 H. Fujimoto, M. Miyasato, Y. Ito, T. Sasaki and K. Ajisaka,
Glycoconjugate J., 1998, 15, 155–160.
19 Y. Ito and T. Sasaki, Biosci., Biotechnol., Biochem., 1997, 61,
1270–1276.
´
´
20 M. J. Hernaiz, A. R. Alcantara, J. I. Garcıa and J.
V. Sinisterra, Chem.–Eur. J., 2010, 16, 9422–9437.
21 M. Perez, J. V. Sinisterra and M. J. Hernaiz, Curr. Org.
Chem., 2010, 14, 2366–2383.
In order to understand the possible effect of these green
solvents on the synthetic behaviour of the enzyme, we
performed different structural studies (fluorescence and
molecular modelling) in the presence of selected biosolvents.
We are able to conclude that the presence of green co-solvents
in the reaction media modifies the enzyme tertiary structure,
permitting a better disposition of the substrates in the active
site, due to solvation effects, that could explain the behaviour
observed.
22 S. V. Kamat, E. J. Beckman and A. J. Russell, J. Am. Chem.
Soc., 1993, 115, 8845–8846.
23 K. Ryu and J. S. Dordick, Biochem., 1992, 31.
24 S. Tawaki and A. M. Klibanov, J. Am. Chem. Soc., 1992, 114,
1882–1884.
25 C. R. Wescott and A. M. Klibanov, J. Am. Chem. Soc., 1993,
115, 1629–1631.
´
´
´
26 M. Perez-Sanchez, M. Sandoval, A. Cortes-Cabrera,
´
´
´
H. Garcıa-Marın, J. V. Sinisterra, J. I. Garcıa and M.
J. Hernaiz, Green Chem., 2011, 13, 2810–2817.
27 M. Perez-Sanchez, M. Sandoval and M. J. Hernaiz,
Tetrahedron, 2012, 68, 2141–2145.
Notes and references
28 M. Sandoval, A. Cortes, C. Civera, J. Trevino, E. Ferreras,
M. Vaultier, J. Berenguer, P. Lozano and M. J. Hernaiz, RSC
Adv., 2012, 2, 6306–6314.
1 P. K. Mandal and W. B. Turnbull, Carbohydr. Res., 2010,
346, 2113–2120.
2 G. F. Springer, Science, 1984, 224, 1198–1206.
3 A. Holeman and P. H. Seeberger, Curr. Opin. Biotechnol.,
2004, 15, 615–622.
4 E. Rapoport, S. Khaidukov, O. Baidina, V. Bojenko,
E. Moiseeva, G. Pasynina, U. Karsten, N. Nifant’ev,
J. LePendu and N. Bovin, Eur. J. Cell Biol., 2003, 82,
295–302.
5 H. Shirato, S. Ogawa, H. Ito, T. Sato, A. Kameyama,
H. Narimatsu, Z. Xiaofan, T. Miyamura, T. Wakita, K. Ishii
and N. Takeda, J. Virol., 2008, 82, 10756–10767.
6 M. E. C. Caines, H. Zhu, M. Vuckovic, L. M. Willis, S.
G. Withers, W. W. Wakarchuk and N. C. J. Strynadka, J.
Biol. Chem., 2008, 283, 31279–31283.
´
´
´
´
´
´
29 M. Perez-Sanchez, A. Cortes-Cabrera, H. Garcıa-Martın, J.
´
´
V. Sinisterra, J. I. Garcıa and M. J. Hernaız, Tetrahedron,
2011, 67, 7708–7712.
30 A. Wolfson, C. Dlugy and Y. Shotland, Environ. Chem. Lett.,
2007, 5, 67–71.
31 A. Wolfson, C. Dlugy, D. Tavor, J. Blumenfeld and
Y. Shotland, Tetrahedron: Asymmetry, 2006, 17, 2043–2045.
32 L. H. Andrade, L. Piovan and M. D. Pasquini, Tetrahedron:
Asymmetry, 2009, 20, 1521–1525.
33 J. T. Vivian and P. R. Callis, Biophys. J., 2001, 80, 2093–2109.
´
´
´
´
34 J. I. Garcıa, H. Garcıa-Marın, J. A. Mayoral and P. Perez,
Green Chem., 2010, 12, 426–434.
35 U. K. Laemmli, Nature, 1970, 227, 680–685.
36 M. M. Bradford, Anal. Biochem., 1976, 72, 248–254.
12162 | RSC Adv., 2013, 3, 12155–12163
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
37 N. Bridiau and T. Maugard, Biotechnol. Prog., 2011, 27,
386–394.
38 B. Hess, C. Kutzner, D. van der Spoel and E. Lindahl, J.
Chem. Theory Comput., 2008, 4, 435–447.
40 A. Amadei, A. B. M. Linssen and H. J. C. Berendsen,
Proteins: Struct., Funct., Genet., 1993, 17, 412–425.
41 J. A. Lemkul, W. J. Allen and D. R. Bevan, J. Chem. Inf.
Model., 2010, 50, 2221–2235.
39 G. M. Morris, R. Huey, W. Lindstrom, M. F. Sanner, R.
K. Belew, D. S. Goodsell and A. J. Olson, J. Comput. Chem.,
2009, 30, 2785–2791.
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 12155–12163 | 12163