A transitional hydrolase to glycosynthase mutant by Glu to Asp substitution at the catalytic nucleophile in a retaining glycosidase

Article abstract of DOI:10.1016/j.carres.2014.02.003

Glycosynthases from more than 16 glycosidase families have been developed for the efficient synthesis of oligosaccharides and glycoconjugates. β-1,3-1,4-Glucan oligo- and polysaccharides with defined sequences can be quantitatively achieved with the glycosynthases derived from Bacillus licheniformis β-1,3-1,4-glucanase. The screening of a nucleophile saturation library of this enzyme yielded the unexpected E134D mutant which has high glycosynthase efficiency (25% higher k_cat than the best glycosynthase to date, E134S) but also retains some hydrolase activity (2% relative to the wild-type enzyme). Here, we report the biochemical and structural analyses of this mutant compared to E134S and wild-type enzymes. E134D shows a pH profile of general base catalysis for the glycosynthase activity, with a kinetic pK_a (on k_cat/K_M) assigned to Glu138 of 5.8, whereas the same residue acts as a general acid in the hydrolase activity with the same pK_a value. The pK_a of Glu138 in the wt enzyme was 7.0, a high value due to the presence of the catalytic nucleophile Glu134 which destabilizes the conjugate base of Glu138. Thus, the pK_a of Glu138 drops 1.1 pH units in the mutant relative to the wild-type enzyme meaning that the larger distance between carboxylates in positions 138 and 134 (5.6 A? for wt, 7.0 A? for E134D) and/or a new hydrogen bonding interaction with a third Asp residue (Asp136) in the mutant reduces the effect of the negatively charged Asp134. In consequence, the pK _a of Glu138 has a similar pK_a value in the E134D mutant than in the other glycosynthase mutants having a neutral residue in position 134. The behavior of the E134D mutant shows that shortening the side chain of the nucleophile, despite maintaining a carboxylate group, confers glycosynthase activity. Therefore E134D is a transitional hydrolase to glycosynthase mutation.

Full text of DOI:10.1016/j.carres.2014.02.003

Carbohydrate Research 389 (2014) 85–92
Contents lists available at ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
A transitional hydrolase to glycosynthase mutant by Glu to Asp
substitution at the catalytic nucleophile in a retaining glycosidase
⇑
Hugo Aragunde, Estela Castilla, Xevi Biarnés, Magda Faijes, Antoni Planas
Laboratory of Biochemistry, Institut Químic de Sarrià, Universitat Ramon Llull, Via Augusta 390, 08017 Barcelona, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Glycosynthases from more than 16 glycosidase families have been developed for the efficient synthesis of
oligosaccharides and glycoconjugates. b-1,3-1,4-Glucan oligo- and polysaccharides with defined
sequences can be quantitatively achieved with the glycosynthases derived from Bacillus licheniformis
b-1,3-1,4-glucanase. The screening of a nucleophile saturation library of this enzyme yielded the
unexpected E134D mutant which has high glycosynthase efficiency (25% higher kcat than the best glyco-
synthase to date, E134S) but also retains some hydrolase activity (2% relative to the wild-type enzyme).
Here, we report the biochemical and structural analyses of this mutant compared to E134S and wild-type
enzymes. E134D shows a pH profile of general base catalysis for the glycosynthase activity, with a kinetic
pKa (on kcat/KM) assigned to Glu138 of 5.8, whereas the same residue acts as a general acid in the hydro-
lase activity with the same pKa value. The pKa of Glu138 in the wt enzyme was 7.0, a high value due to the
presence of the catalytic nucleophile Glu134 which destabilizes the conjugate base of Glu138. Thus, the
pKa of Glu138 drops 1.1 pH units in the mutant relative to the wild-type enzyme meaning that the larger
distance between carboxylates in positions 138 and 134 (5.6 Å for wt, 7.0 Å for E134D) and/or a new
hydrogen bonding interaction with a third Asp residue (Asp136) in the mutant reduces the effect of
the negatively charged Asp134. In consequence, the pKa of Glu138 has a similar pKa value in the
E134D mutant than in the other glycosynthase mutants having a neutral residue in position 134. The
behavior of the E134D mutant shows that shortening the side chain of the nucleophile, despite maintain-
ing a carboxylate group, confers glycosynthase activity. Therefore E134D is a transitional hydrolase to
glycosynthase mutation.
Received 18 November 2013
Received in revised form 1 February 2014
Accepted 2 February 2014
Available online 12 February 2014
Keywords:
Glycosynthase
b-Glucanase
Bacillus licheniformis
Hydrolase
b-1,3-1,4-Glucans
pKa
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
transglycosylation is severely hampered by the predominant
hydrolase activity, where the product formed is necessarily a
Retaining glycosidases (or glycoside hydrolases, GH) catalyze
the hydrolysis of glycosidic bonds of oligosaccharides, polysaccha-
rides, and glycoconjugates by general acid/base catalysis in a
double displacement reaction with net retention of the anomeric
configuration. The catalytic mechanism involves two essential res-
idues, a general acid/base and a catalytic nucleophile, the reaction
proceeding through the formation of a covalent glycosyl-enzyme
intermediate1 (Fig. 1A). A variation of this mechanism is found in
some hexosaminidases which operate by substrate-assisted
catalysis; they lack the catalytic nucleophile and the 2-acetamido
group of the substrate acts as an internal nucleophile.2,3 Retaining
GH have been extensively used for enzymatic synthesis of oligosac-
charides and glycoconjugates through their ability to catalyze
transglycosylation reactions.4 However, kinetically-controlled
hydrolysable substrate. The introduction of the glycosynthases
concept in 19985,6 overcomes this limitation for the efficient enzy-
matic synthesis of oligosaccharides and glycoconjugates.
Glycosynthases are engineered retaining glycoside hydrolases
in which the catalytic nucleophile has been replaced by a non-
nucleophilic residue. They are inactive hydrolases but efficiently
catalyze glycosyl transfer to an acceptor when using activated gly-
cosyl fluoride donors with the opposite anomeric configuration
than the original substrate of the parental wt hydrolase reaction
(Fig. 1B). Enzymes from more than 16 different GH families
(according to the CAZY database classification7) have been con-
verted into glycosynthases. They cover different specificities and
relevant applications have been developed (for reviews see8–11).
Bacillus licheniformis 1,3-1,4-b-glucanase is a retaining endo-
glycosidase belonging to CAZY family GH16, which hydrolases
mixed-linked 1,3-1,4-b-glucans.12 The catalytic machinery in-
volves a catalytic triad formed by Glu138 and Glu134 as general
⇑
Corresponding author. Tel.: +34 932672000; fax: +34 932056266.
E-mail address: antoni.planas@iqs.edu (A. Planas).
http://dx.doi.org/10.1016/j.carres.2014.02.003
0008-6215/Ó 2014 Elsevier Ltd. All rights reserved.

86
H. Aragunde et al. / Carbohydrate Research 389 (2014) 85–92
general acid/base
A
B
O
O
O
H
O
OH
OH
OH
O
O
O
H
O
HO
O
HO
O
HO
O
O
OH
R
H
OH
OH
OH
O
O
O
O
nucleophile
general base
O
O
OH
OH
O
OH
acceptor
OH
OH
H
O
O
HO
O
HO
OR
O
HO
O
OR
HO
O
O
OH
OH
OH
F
X
donor
mutated nucleophile
Figure 1. (A) Mechanism of a retaining glycosidase involving the participation of a general acid/base residue and a catalytic nucleophile residue. (B) Mechanism of a
glycosynthase mutant, where X is a non-nucleophilic substitution for the original nucleophile in the wt enzyme.
acid/base and catalytic nucleophile, respectively, and Asp136 as
assisting residue that participates in the modulation of the pKas
of the catalytic residues.13,14 Nucleophile mutants E134A and
E134S are efficient glycosynthases catalyzing condensation reac-
tions between glycosyl fluoride donors of general structure
buffer 50 mM, pH 7.0, 0.1 mM CaCl2. Proteins were purified essen-
tially as reported for the WT enzyme by means of metal affinity
chromatography of the His-tagged proteins.24 Cells were lysed in
a
cell-disrupter (Constant Systems, UK) and centrifuged at
12,000 rpm. The supernatant was loaded onto a HiTrap 1 mL col-
umn (GE Healthcare) previously equilibrated with loading buffer
(50 mM phosphate pH 7.0, 0.1 mM CaCl2). The column was rinsed
to remove unbound proteins and then eluted with a gradient of
0.5 M imidazole in 50 mM phosphate pH 7.0, 0.1 mM CaCl2. Protein
fractions were dialyzed twice against 50 mM phosphate pH 7.0,
0.1 mM CaCl2, followed by a last dialysis against water. Proteins
were lyophilized for storage, and redissolved prior to use. Enzymes
were >95% homogeneous as judged by SDS–PAGE. Concentration
was determined by absorbance at 280 nm using an extinction coef-
(Glcb4)nGlcb3GlcaF and gluco-oligosaccharide acceptors with reg-
ioselective formation of b-1,4 glycosidic bonds.6,15–18 More
recently, they have shown to be efficient catalysts for the prepara-
tion of artificial polysaccharides with regular structures through
glycosynthase-catalyzed donor self-condensation.19
During the development of a screening assay for the analysis of
mutant glycosynthase libraries (HTS based on a fluoride chemo-
sensor assay),20 we analyzed a nucleophile saturation library
(E134X) of the Bacillus licheniformis 1,3-1,4-b-glucanase. In addi-
tion to the expected and well known glycosynthase variants (Ala,
Ser, and Gly mutations), a surprising result was the finding that
E134D showed high glycosynthase activity while keeping residual
hydrolase activity. Shortening the side chain of the nucleophile res-
idue (Glu to Asp) was already known to decrease significantly the
hydrolase activity of b-glycosidases (e.g. lacZ b-galactosidase,21
Agrobacterium b-glucosidase,22 B. circulans xylanase23), but this
conservative substitution has not been previously reported as a
glycosynthase. This surprising result may be interpreted as a tran-
sitional form between the original hydrolase and its glycosynthase
variants. Here we analyze the properties of the E134D mutant 1,3-
1,4-b-glucanase with the aim of understanding how a conservative
mutation (Asp for Glu substitution) still maintaining a carboxylate
group behaves as a glycosynthase. pH studies reveal significant
aspects about the modulation of the pKas of the catalytic residues.
ficient of 3.53ꢀ105 Mꢁ1 cmꢁ125
.
2.2. Kinetics of glycosynthase reactions
The glycosyl donor and acceptor, b-D-glucopyranosyl-
(1?4)-b-
(Glcb4Glcb3Glc
D
-glucopyranosyl-(1?3)-
a
-
D
-glucopyranosyl
fluoride
a
F)25 and p-nitrophenyl b-
D-glucopyranoside
(GlcbpNP) respectively, were dissolved in phosphate–citrate buffer
(50 mM citric acid and 50 mM Na2HPO4), CaCl2 (0.1 mM) at the
indicated pH, and with a constant ionic strength of I = 0.5 M
adjusted with KCl. The solution was preincubated at 35 °C for
5 min. Reactions were initiated by addition of the correspondent
enzyme and the mixtures kept at 35 °C. Aliquots were taken at dif-
ferent time intervals and diluted 1:10 in 2% (v/v) formic acid to
stop the enzymatic reaction. HPLC analysis was done in an Agilent
HPLC 1100 chromatograph using a reverse phase column (NovaPak
2. Material and methods
C18, 4
l
m, 3.9 mm ꢂ 150 mm, Waters), eluted with 12–14% MeOH
in H2O at a flow of 1 mL/min, UV detector at 310 nm. Initial rates
2.1. Protein expression and purification
(v0) were calculated from the linear portion of the progress curves
(normalized area vs time) and expressed as
v
0/[E]ꢀ(sꢁ1).
All the proteins (wt, E134A, E134S, and E134D) were cloned in a
pET16b vector as reported.20,24 Escherichia coli BL21 DE3 Star cells
harboring the corresponding expression plasmid were grown in
500 mL LB medium (100 mg/mL ampicillin, 2% glucose) for 8 h at
37 °C and 250 rpm until late exponential phase. The medium was
changed to 500 mL LB containing 1 mM IPTG and 100 mg/mL
ampicillin for protein expression (16 h at 25 °C). Cells were
harvested by centrifugation and resuspended in 100 mL phosphate
For the reaction progress curve, mutants E134D and E134S were
assayed at 2 lM enzyme concentration, 2 mM donor and 20 mM
acceptor in phosphate buffer (50 mM pH 7 and CaCl2 0.1 mM) at
35 °C. The formation/disappearance of glycosynthase-derived
products was followed by HPLC as described above.
The pH dependence of the glycosynthase activity was deter-
mined on kcat and kcat/KM in the pH range 5 to 9. At each pH, donor
was varied (0.1 to 8 mM, 11 different concentrations) and acceptor

H. Aragunde et al. / Carbohydrate Research 389 (2014) 85–92
87
kept constant at 20 mM. Initial rates were fitted to the Michaelis–
2.5. Enzymatic polymerization
Menten equation to calculate the kinetic parameters by non-linear
regression. For E134D, kcat/KM versus pH data were fitted to an up-
ward single ionization curve (Eq. 1).
Reactions were carried out in phosphate buffer (50 mM pH 7,
0.1 mM CaCl2). The donor substrate Glcb4Glcb3Glc F (150 mM)
a
was incubated with different enzyme concentrations at 35 °C with
orbital shaking at 250 rpm for 24 h. Enzyme concentrations for
E134S and E134D were chosen as to have the same activity accord-
kcat
KM
A
¼
:
ð1Þ
aꢁpHÞ
1 þ 10ðpK
ing to the higher kcat value for E134D: 19
(E134S) for low enzyme concentration, and 47
63 M (E134S) for high enzyme concentration. A precipitate was
l
M (E134D), 25
lM
For E134S, kcat/KM versus pH data were fitted to a double ionization
curve with residual activity at high pH, as reported for the E134A
glycosynthase mutant25 (Eq. 2).
lM
(E134D),
l
formed during the reactions. It was recovered by centrifugation
at 13,000 rpm for 5 min, washed several times with cold water,
and freeze-dried. Product yields are expressed in % (weight poly-
mer/weight initial donor) for the insoluble material. Supernatants
were also lyophilized to recover soluble oligomers. Products were
analyzed by HPSEC to determine the molecular mass profiles as de-
scribed in Perez et al.19 Analyses were performed on an Agilent
1100 HPLC system equipped with a refractive index detector
a2Þ
kcat
KM
A þ B ꢀ 10ðpHꢁpK
¼
:
ð2Þ
1 þ 10ðpK ꢁpHÞ þ 10ðpHꢁpK
a2Þ
a1
2.3. Kinetics of hydrolase reactions
The hydrolase activity was determined with 4-methylumbel-
liferyl 3-O-b-cellobiosyl-b- -glucopyranoside (Glcb4Glcb3GlcbMU)
D
(RID) using a PSS Gram column (8.0 ꢂ 300 mm, 100 Å, 10
lm)
as substrate and following changes in UV absorbance due to the
release of 4-methylumbelliferone (MU) using matched 1 cm path
length cells at 35 °C in a Evolution 300 spectrophotometer (Thermo
Scientific).13 Rates of the enzyme-catalyzed reactions were deter-
mined by incubating the enzyme with the substrate in citrate/
phosphate buffer (50 mM citric acid and 50 mM Na2HPO4) and
0.1 mM CaCl2 in the pH range of 5–7 in the thermostated cell
holder. Reactions were initiated by addition of enzyme to the pre-
incubated mixture of substrate and buffer. The absorbance change
was monitored at a wavelength of k = 365 nm. Extinction coeffi-
cients for MU at 365 nm were determined at each pH under the
and a PSS Gram precolumn (8.0 ꢂ 50 mm, 100 Å, 10 m) thermo-
l
stated at 50 °C, and DMSO with lithium bromide (5 g/L) as eluent
at a flow rate of 0.5 mL/min. The calibration curve was obtained
with dextrans as standards in the range 1–55 kDa (American Poly-
mer Standards Corporation DXT1–DXT55 kDa). Freeze-dried poly-
mers from the enzymatic reactions and standards were dissolved
in DMSO and filtered. From the chromatograms, Mp (molecular
mass of the peak maximum), Mw (weight average molecular
mass), Mn (number average molecular mass), DP (degree of poly-
merization), and PDI (polydispersity index) were calculated.28
same experimental conditions. Initial rates (
from the linear portion of the progress curves. For wt, the slope
v0) were calculated
2.6. Modeling of the enzymeꢀsubstrate complexes
of v0 versus substrate concentration (0.1–0.3 mM) gave the kcat
/
The three dimensional structures of different enzymeꢀsubstrate
complexes of Bacillus licheniformis 1,3-1,4-b-glucanase were mod-
eled in silico by means of homology modeling.29 The structure of
the wtꢀsubstrate complex was previously modeled.30 E134S and
E134D variants were modeled based on the 3D structure of the free
enzyme (PDB accession code: 1GBG). The substrate was extracted
from the structure of the orthologous Bacillus macerans 1,3-1,4-b-
glucanase in complex with the tetrasaccharide product
(Glcb4Glcb4Glcb3Glc) (PDB accession code: 1U0A), which was
KM values at each pH. kcat/KM versus pH data were fitted to a
bell-shaped double ionization curve (Eq. 3)
kcat
KM
A
¼
:
ð3Þ
a2Þ
1 þ 10ðpK ꢁpHÞ þ 10ðpHꢁpK
a1
For E134D, initial rates were determined at 1 mM substrate
concentration at each pH. These rates at low substrate concentra-
tion were assumed to reflect kcat/KM. Hydrolysis rates at lower sub-
strate concentrations (to determine true kcat/KM values from the
linear dependence v0 vs [S]) were difficult to measure due to the
manually converted to an a-fluoride glycosyl donor (Glcb3GlcaF),
and its position in the structure was kept fixed during the modeling.
100 replicas of the models were generated for each variant, in
which different orientations of the side-chains were sampled dur-
ing a brief simulated annealing and geometry optimization process.
low hydrolase activity of the mutant.
data were fitted to a downward single ionization curve (Eq. 4)
v0/[E] (ꢃkcat/KM) versus pH
v0
½Eꢄ
A
¼
:
ð4Þ
aÞ
1 þ 10ðpHꢁpK
3. Results
All fittings were done with the FigP software. Estimated stan-
dard errors in pKa determinations were 0.1 pH units.
The ability of the E134D 1,3-1,4-b-glucanase to behave as a gly-
cosynthases poses a number of questions about the mechanism:
how large is the cavity created by shortening the Glu to Asp side
2.4. Chemical rescue
chain to accommodate the a-fluoride of the donor substrate?; does
This assay was performed following a modification of the proce-
dure reported in Viladot et al.26 A solution of enzyme (E134A,
E134S, or E134D) (1 lM) and sodium azide (1 M) was preincubated
in a thermostated cell holder at 30 °C in phosphate–citrate buffer
(87 mM phosphate, 6.5 mM citrate) pH = 7.2. Reactions were initi-
the mutant enzyme operate by general base catalysis thus
implying a large downward shift of the pKa of E138 (general
acid in the wt hydrolase mechanism) to behave as a base in
the glycosynthase mechanism?; which is the extent and pH
dependence of the residual hydrolase activity?; is it a practical
mutant for the enzymatic synthesis of polysaccharides by donor
self-condensation?.
ated by adding 2,4-dintrophenyl 3-O-b-cellobiosyl-b-D-glucopy-
ranoside (Glcb4Glcb3GlcbDNP) substrate (1 mM), and the release
of 2,4-dinitrophenol (DNP) was monitored by absorbance changes
at 425 nm, using an extinction coefficient of
De -
6134 Mꢁ1
cmꢁ1 26,27
.
Blank reactions without sodium azide were run in paral-
3.1. Glycosynthase activity of mutant E134D
lel to substract the background enzymatic hydrolysis rate of this
reactive substrate. Initial rate was calculated as the slope of DNP
concentration (reaction-blank) versus t.
The time course of the glycosynthase reaction of E134D at pH 7,
35 °C, is compared to that of the well characterized E134S

88
H. Aragunde et al. / Carbohydrate Research 389 (2014) 85–92
glycosynthase using Glcb4Glcb3Glc
a
F
(2 mM) as donor and
the pKa (5.8) determined for the glycosynthase activity and as-
GlcbpNP (20 mM) as acceptor substrates, where the acceptor
is in large excess to minimize donor self-condensation. As shown
in Figure 2, E134S rapidly forms the condensation product
(Glcb4Glcb3Glcb4GlcbpNP) up to 90% yield, and the product re-
mains stable upon long incubation time. E134D has the same initial
reaction rate, reaching a maximum of 90% yield, but the product is
then slowly hydrolyzed due to the still significant hydrolase activ-
ity. Kinetic parameters had been previously determined20 and are
summarized in Table 1. Under initial rate conditions (<10% conver-
sion), E134D has 25% higher kcat than E134S and E134A, but KM
(donor) is in between both Ser and Ala glycosynthase mutants.
The pH dependence of the glycosynthase activity for E134D,
compared to that of E134S, is presented in Figure 3. The E134D
mutant shows a pH dependence corresponding to general base
catalysis, with a kinetic pKa (on kcat/KM) of 5.8 and maximum activ-
ity at pH 7. At pH values higher than 7.5 the protein is less stable
and precipitates, precluding kinetic measurements. The profile up
to pH 7 is similar to that of E134S (and the previously reported
E134A mutant,17 with the same kinetic pKa than E134S (5.8) and
slightly higher than E134A (5.2). It corresponds to Glu138 acting
as general base as it was previously assigned in the E134A mutant
by titration with a water-soluble carbodiimide.25
signed to Glu138, but acting as general base (see Section 4).
3.3. Chemical rescue with an exogenous nucleophile
Alanine mutants at the catalytic nucleophile can be reactivated
using an exogenous nucleophile that can bind to the cavity left by
removal of the catalytic carboxylate (Ala for Glu mutation). The
E134A 1,3-1,4-b-glucanase is reactivated by addition of azide in a
concentration dependent manner using an activated 2,4-dinitro-
phenyl b-glycoside substrate, to give the
a-glycosyl azide
product.26,31,32 The experiment was applied to the E134S and
E134D mutants by incubating the enzymes with 1 mM
Glcb4Glcb3GlcbDNP (DNP: 2,4-dinitrophenyl) and 1 M sodium
azide at pH 7.2 and 30 °C. The release of DNP (after subtracting
the blank rate in the absence of sodium azide) is plotted in Figure 5.
E134S shows reactivation (chemical rescue) but to a lesser extent
than the E134A mutant, whereas the E134D mutant is not reacti-
vated but instead, azide inhibits the residual hydrolase activity of
the mutant.
3.4. Glycosynthase-catalyzed polymerization
The glycosynthase mutants E134A and E134S of 1,3-1,4-b-glu-
canase are able to produce artificial polysaccharides by donor
self-condensation in the absence of an acceptor.19,33 The
E134S-catalyzed polymerization of the tetrasaccharide donor
3.2. Hydrolase activity of E134D
The hydrolase activity of E134D was determined with
Glcb4Glcb3GlcbMU (MU: 4-methylumbelliferyl), a common b-aryl
glycoside substrate of the wt enzyme, by monitoring MU release.
At pH 7, 35 °C, the specific activity of E134D is about 2% of the
wt enzyme. To evaluate the kinetic pKa of the general acid residue
E138 in the mutant, the pH dependence of the hydrolytic reaction
was determined and compared to that of the wt enzyme. For the
wt enzyme, the pH profile of kcat/KM (Fig. 4A) follows a double ion-
ization curve, with a high pKa of 7.0 corresponding to the general
acid E138, and a low pKa of 5.4 for the nucleophile E134. These val-
ues are close to those previously reported at 53 °C (pKas of 7.0 and
5.5, respectively), temperature of maximum activity for the wt
enzyme.13 For E134D, rates at 1 mM substrate were assumed to re-
flect kcat/KM values (hydrolysis rates at lower substrate concentra-
tion were difficult to measure due to the low activity of the
mutant). As shown in Figure 4B, the hydrolase activity decreases
from pH 5 to 7, with a kinetic pKa of 5.9 assignable to the general
acid E138 in the hydrolase mechanism. Thus the E134D mutation
results in a pKa decrease of one pH unit. It is also consistent with
Glcb4Glcb4Glcb3GlcaF led to high molecular mass polysaccharides
(up to Mw of 30 kDa), where the degree of polymerization (DP)
was linearly dependent on enzyme activity. Since the polymers
are water-insoluble at rather low Mw, it was hypothesized that
the balance between enzymatic polymerization rate and polymer
precipitation rate dictates the molecular mass of the products.19
In keeping with that, a more active glycosynthase would in princi-
ple achieve polysaccharides with higher DP. Although the new
E134D mutant is not a practical glycosynthase for soluble oligosac-
charide synthesis due to the residual hydrolase activity, it might be
efficient for polysaccharide synthesis: insoluble polysaccharides
might be formed fast (E134D has a 25% higher kcat than E134S
(Table 1)) and their precipitation protects them from hydrolysis.
Polymerization reactions catalyzed by E134S and E134D were
performed with the trisaccharide donor Glcb4Glcb3GlcaF
(150 mM) in phosphate buffer pH 7 at 35 °C. Precipitated polysac-
charides and the soluble fractions were analyzed by HPSEC (Fig. 6).
Results are summarized in Table 2. For E134S (at low and high con-
centrations) the insoluble polymers show a monomodal distribu-
tion of molecular masses with Mw of 13 kDa, (DP 78),
corresponding to 26 condensations, whereas the soluble fraction
only contains traces of oligosaccharides. For E134D the size distri-
bution of the precipitated polymers is multimodal and dependent
100% conversion
2
of enzyme concentration. At low (19 lM) concentration, an insolu-
E134S
ble polymer with Mw 5.5 kDa is obtained, with a main peak (Mp)
at 5.3 kDa (DP 33), and a minor peak at Mp 0.6 kDa for a trisaccha-
ride, either unreacted donor or hydrolysis product. At higher en-
1.5
zyme concentration (47 lM) Mw increases up to 14 kDa, and the
1.0
E134D
molecular mass distribution becomes trimodal with Mp 32.7, 4.7,
and 0.6 kDa, corresponding to DP 201, 30, and 3, respectively. A
significant fraction of a larger polymer is obtained, but some trisac-
charide (Mp 0.6 kDa) is still present. As opposed to the case of
E134S, the soluble fractions in the E134D reactions contain a sig-
nificant amount of trisaccharide (Mp 0.6 kDa) and hexasaccharide
(Mp 1.2 kDa) at low enzyme concentration and mainly trisaccaride
at high enzyme concentration. The nature of these small molecules
may be explained by two non-excluding hypotheses: unreacted
donor or hydrolysis of the longer polymers formed. In agreement
with the progress curve of the glycosynthase reaction shown in
0.5
0
0
50
100
time (min)
150
200
Figure 2. Time-course monitoring of the glycosynthase reactions catalyzed by the
E134D and E134S mutants. Conditions: 2 mM Glcb4Glcb3Glc F donor, 20 mM
GlcbpNP acceptor, 2 M enzyme, pH 7.0, 35 °C.
a
l

H. Aragunde et al. / Carbohydrate Research 389 (2014) 85–92
89
Table 1
Kinetic parameters for the glycosynthase activity of E134D, S, and A mutants
kcat (sꢁ1
)
KM,donor (mM)
(kcat/KM
)
(Mꢁ1 sꢁ1
)
% kcat
% kcat/KM
donor
E134D
E134S
E134A
1.44 0.05
1.08 0.04
1.04 0.04
0.49 0.05
0.10 0.02
1.12 0.12
2.95ꢀ103
1.08ꢀ104
9.27ꢀ102
100
75
72
27
100
9
Conditions: 0–8 mM donor (Glcb4Glcb3Glc
a
F), 20 mM acceptor (Glc-pNP), 0.5
l
M enzyme, 50 mM phosphate pH 7.0, 0.1 mM CaCl2, 35 °C.
100
100
A
B
7.9
75
50
25
0
75
pKa 5.8
50
pKa 5.8
25
E134S
9
E134D
0
5
6
7
pH
8
5
6
7
8
pH
Figure 3. pH dependence of kcat/KM for the glycosynthase activity of E134S (A) and E134D (B) mutant enzymes. At pH >7.5, E134D activity decreases due to protein
precipitation (⁄in the graph).
WT
E134D
100
75
50
25
0
100
75
50
25
0
A
B
5.4
pKa 7.0
pKa 5.9
5
6
7
pH
8
9
5
6
7
pH
8
9
Figure 4. pH dependence of kcat/KM for the hydrolytic activity of the WT (A) and E134D mutant (B) enzymes.
observed in the soluble fraction at high enzyme concentration, all
supportive of hydrolysis of the products formed.
0.02
0.01
rescue
-
E134A+N3
3.5. Modeling the enzymeꢀsubstrate complex in the
-
glycosynthase mutants
E134S+N3
Mutants E134D and E134S with bound a-glycosyl fluoride sub-
0.00
strate were modeled in silico and compared to the modeled
Michaelis complex of the wt 1,3-1,4-b- glucanase previously
reported30,34 (Fig. 7). A b-glycoside substrate binds with a 1S3 dis-
torted conformation in the Michaelis complex of the wt enzyme
(preactivated conformation for catalysis in the hydrolase mecha-
-0.01
-0.02
-
E134D+N3
1000
inhibition
nism30,35), whereas the
a-glycosyl fluoride donor in the glyco-
synthase mutants remains in a chair (4C1) conformation (as
recently observed in the X-ray 3D structures of glycosynthase mu-
tants of rice BGlu136). The shorter side chain of the mutants at po-
0
500
1500
2000
time (s)
sition 134 relative to the wild-type creates
a cavity that
Figure 5. Chemical rescue by exogenous azide of mutants E134A, E134S, and
E134D. Conditions: 1 mM Glcb4Glcb3GlcbDNP, 1 M NaN3, 1
phosphate buffer pH 7.2, 0.1 mM CaCl2, 30 °C.
lM enzyme, citrate–
accommodates the fluoride aglycon with -configuration. The dis-
a
tance between the carboxylates in position 138 and 134 is 5.6 Å in
the wt, and 7.0 in the E134D mutant. For the Ser mutant (Fig. 7C),
the hydroxyl oxygen is at approximately 4.3 Å of the fluoride atom,
with an orientation that may allow for a weak hydrogen bonding
interaction to assist the departure of the fluoride aglycon in the
SN2 displacement reaction by the acceptor. For the Asp mutant
Figure 2 (high synthase activity but then product hydrolysis), it is
more likely that they are hydrolysis products. Moreover, the yield
of insoluble polymers was low (10–20%) for the E134D reactions
(about 90% for E134S), and only a trisaccharide product was

90
H. Aragunde et al. / Carbohydrate Research 389 (2014) 85–92
E134D
19 μM
47 μM
E134S
25 μM
63 μM
A
B
6
4
2
0
4
2
0
12
14
16
18
20
22
12
14
16
18
20
22
time (min)
time (min)
E134D
19 μM
47 μM
E134S
25 μM
63 μM
C
D
10
5
10
5
0
0
12
14
16
18
20
22
12
14
16
18
20
22
time (min)
time (min)
Figure 6. Size exclusion chromatograms (HPSEC) of the insoluble polysaccharides by glycosynthase-catalyzed polymerization, (A) with E134D, (B) with E134S, and of the
soluble fractions, (C) with E134D, (D) with E134S, at two different enzyme concentrations. Mp values (maximum of the peaks) are indicated in kDa.
Table 2
Enzymatic polymerization by E134D and E134S glycosynthase mutants
b
Mutant
E134D
[E] (
l
M)
Sample
Yield (%)
Mw (kDa)
DPa
Mn (kDa)
PDI
Mp (kDa)
19
Precipitate
Supernatant
Precipitate
Supernatant
10
90
20
80
5.5
1.0
14.0
1.7
34
6
86
10
2.8
0.7
3.5
0.8
1.9
1.4
4.0
2.3
4.8, 0.6
0.6, 1.2
37.6, 4.4, 0.6
0.6
47
E134S
25
63
Precipitate
Supernatant
Precipitate
Supernatant
80
20
90
10
12.3
7.6
12.4
4.9
76
47
76
30
4.2
1.8
4.4
1.3
2.9
4.2
2.8
3.7
15.8
—
15.4
—
Reaction conditions were 50 mM phosphate buffer pH 7, 0.1 mM CaCl2, 150 mM Glcb4Glcb3GlcaF donor, at 35 °C for 24 h.
a
Degree of polymerization (DP) expressed as the number of glucosyl units for the weight average molecular mass polysaccharide (Mw).
Mp, molecular mass of the peak maximum.
b
Figure 7. Modeled enzymeꢀsubstrate complexes, WT with Glcb4Glcb3Glcb4Glc, E134D and E134S with Glcb3Glc
aF. Dashed lines are distances between heteroatoms
expressed in Å.
(Fig. 7B), the distance between the fluoride and the carboxylate is
shorter, and may involve a halogen bonding interaction (distance
Fꢀ ꢀ ꢀO of 3.3 Å and a C1–Fꢀ ꢀ ꢀO angle of ꢃ160°, a geometry that
perfectly fits that of halogen bonding interactions found in
proteinꢀligand complexes).37 Preliminary computational results
by ab initio methods indicate that this type of electrostatic interac-
tion, although present in
a-halogen-glycosyl systems, is very weak
for fluoride (data not shown).

H. Aragunde et al. / Carbohydrate Research 389 (2014) 85–92
91
It is relevant, as it will be discussed below, that Asp136, a third
carboxylate of the catalytic triad in the wt enzyme, is hydrogen
bonding to the nucleophile Glu134 (Oꢀ ꢀ ꢀO distance of 2.7 Å,
Fig. 7A) but the same residue is predicted by the modeling to
hydrogen bond with Glu138 in the E134D and E134S mutants
(Fig. 7B and C).
equivalent to that in the covalent glycosyl-enzyme intermediate
of the wt enzyme, thus Glu138 having a low pKa. The same con-
cept applies to the other glycosynthase variants having a neutral
residue in the position of the original nucleophile, as it is the case
of the E134S mutant (with a pKa of 5.8, Fig. 3A). But what about
the E134D mutation?
Glu138 should be able to act as a base to explain that the E134D
mutant behaves as a glycosynthase. Certainly, the E134D mutant
shows a pH dependence corresponding to general base catalysis,
with a kinetic pKa of 5.8 and maximum activity at pH 7 (Fig. 3B).
This pKa value is the same as that for E134S (5.8) and slightly high-
er than that for E134A (5.2). It is also consistent with the pH profile
of the residual hydrolase activity of the mutant, with a calculated
pKa of 5.9 in the downward profile for general acid catalysis. There-
fore, E134D can be seen as a transitional hydrolase to glycosynth-
ase mutant depending on pH, with hydrolase activity at low pH
and glycosynthase activity at high pH.
4. Discussion
Typical glycosynthases arise from replacement of the catalytic
nucleophile of a retaining glycosidase by a non-nucleophilic resi-
due. Surprisingly, the conservative Glu to Asp substitution
(E134D mutant) in 1,3-1,4-b-glucanase results in an active glyco-
synthase. It was an unexpected result because the carboxylate
group is still present although in a different position due to the
shorter side chain. Two essential aspects of the glycosynthase
mechanism are that the nucleophile mutation has to create room
Possible reasons for the low pKa of Glu138 in the E134D mu-
tant are as follows: (a) the distance between the carboxylates in
positions 138 and 134 is longer in the mutant than in the wt
(Fig. 7), thus reducing the electrostatic effect of the carboxylate
of Asp134 and resulting in a lower destabilization of the conju-
gate base of Glu138. As a consequence its pKa becomes similar
to that in the E134S and E134A neutral nucleophile mutations.
(b) The carboxylate of Glu138 is stabilized by a H-bonding inter-
action with Asp136, the third auxiliary residue of the catalytic
triad in family GH16 which participates in modulating the pKas
of the catalytic residues.14 As observed in the X-ray structures
of the wt enzyme, Asp136 is hydrogen bonding with Glu134 in
the free enzyme,39,40 but it is rearranged in a covalent enzyme–
ligand complex to hydrogen bond with Glu138, and occupies an
intermediate position in the enzyme–product complex.41,42 Given
the flexibility of the side chain of Asp136 (as observed in the
crystal structures of the wt enzyme), the modeled structure of
the E134Dꢀsubstrate complex (Fig. 7B) proposes an orientation
of the Asp136 side chain that can establish a H-bond with
Glu138. Therefore, it is reasonable that, even with Asp134 being
deprotonated, the carboxylate of Glu138 might be stabilized by
Asp136, thus lowering its pKa and being able to act as a base in
the glycosynthase reaction.
to allow binding of the donor with the fluoride aglycon in
that general base catalysis is required to activate the acceptor
substrate.
Replacement of the catalytic nucleophile by Ala, Gly, or Ser in
typical glycosynthases, results in a shorter side chain that cre-
a, and
ates a cavity for the accommodation of the
a-fluoride of the do-
nor. It can be tested by chemical rescue of the hydrolytically
inactive mutant by addition of an exogenous nucleophile as so-
dium azide or formate. For the E134A mutant, addition of azide
restores the activity with an activated b-glycoside substrate
(such as 2,4-dinitrophenyl b-glycoside). Azide is able to bind into
the cavity left by removal of the nucleophile side chain and dis-
place the b-aglycone by an SN2 mechanism to give the
a-glyco-
syl azide product.26 Whereas activity is restored in the E134A
and E134S mutants, azide inhibits the hydrolase activity of the
E134D mutant (Fig. 5). This behavior is similar to that observed
with the wt enzyme.26 It indicates that the Glu to Asp mutation
does not leave enough room for azide to bind and act as a nucle-
ophile, or that the negative charge of the Asp residue prevents
proper binding of azide by electrostatic repulsion, resulting in
any case in the absence of chemical rescue. Although this exper-
iment does not prove or disprove that a cavity has been created,
the fact that the mutant has glycosynthase activity confirms that
Because E134D has a higher kcat than E134S in the glycosynth-
ase reaction (Table 1), it was tested in polymerization reactions.
The question was whether the fast donor condensation leads to
polymeric products that precipitate and become inaccessible to
the enzyme before they are hydrolyzed by the residual hydrolase
activity of the mutant. It is shown not to be the case since low
yields in polymeric products were obtained as compared to the
E134S mutant (Table 2). The polymers obtained (although in low
amount) had a different distribution but were not significantly lar-
ger than those produced by the E134S glycosynthase. Although this
new mutant is a more active glycosynthase in terms of initial rates,
it is not practical as biocatalyst.
the
a-fluoride is bound. As shown in the modeled structure
(Fig. 7B), the
for catalysis.
a-glycosyl fluoride is properly bound and oriented
According to the general mechanisms depicted in Figure 1, the
residue acting as general acid in the first step of the hydrolase
mechanism behaves as a base to activate the acceptor in the gly-
cosynthase mechanism. This change of the pKa of the same resi-
due is the result of different environments in the wt and
glycosynthase mutant due to the presence or absence of the cat-
alytic nucleophile. In the wt 1,3-1,4-b-glucanase, Glu138 is the
general acid with a pKa of 7.0 (Fig. 4A) in the first step leading
to the covalent glycosyl-enzyme intermediate in the hydrolase
mechanism.6 This high pKa is in part due to the presence of the
negatively charged catalytic nucleophile Glu134 which destabi-
lizes the conjugate base of Glu138. It drops in the glycosyl-
enzyme intermediate due to neutralization of the negative charge
in the covalent complex allowing Glu138 to act as general base in
the deglycosylation step leading to the hydrolysis product. This
pKa cycling of the general acid/base residue in retaining glycosi-
dases has been nicely demonstrated in a B. circulans xylanase
by 13C-NMR titration.38 In the glycosynthase mechanism cata-
lyzed by the original E134A mutant, Glu138 has a lower pKa of
5.225 because of the Ala for Glu134 replacement, where removal
of the carboxylate group deletes the negative charge that destabi-
lized the conjugate base of Glu138. The active site charge is
5. Conclusions
The screening of a nucleophile saturation library revealed that
the E134D mutant is a novel glycosynthase. Characterization of
the glycosynthase and residual hydrolase activities of this mutant
allow to conclude that shortening the side chain of the residue at
position 134 creates enough room for binding the a-fluoride donor,
and that Glu138 has a low pKa to act as a general base in the gly-
cosynthase mechanism similar to other neutral substitutions in po-
sition 134. E134D represents a transition between a glycosidase
and a glycosynthase, with hydrolase activity at low pH and syn-
thase activity at high pH.

92
H. Aragunde et al. / Carbohydrate Research 389 (2014) 85–92
19. Perez, X.; Faijes, M.; Planas, A. Biomacromolecules 2011, 12, 494–501.
Acknowledgments
20. Aragunde, H.; Castilla, E.; Faijes, M.; Planas, A. Biochem. J. 2014, 458, 355–363.
21. Cupples, C. G.; Miller, J. H.; Huber, R. E. J. Biol. Chem. 1990, 265, 5512–5518.
22. Withers, S. G.; Rupitz, K.; Trimbur, D. E.; Warren, R. A. J. Biochemistry 1992, 31,
9979–9985.
23. Wakarchuk, W. W.; Campbell, R. L. Protein Sci. 1994, 3, 467–475.
24. Addington, T.; Calisto, B.; Alfonso-Prieto, M.; Rovira, C.; Fita, I.; Planas, A.
Proteins 2011, 79, 365–375.
This work was supported in part by Grant BFU2010-22209-C02-
02 from Ministry of Economy and Competitiveness, Spain. H.A.
acknowledges a pre-doctoral contract (FI-DGR 2013) from the Gen-
eralitat de Catalunya.
25. Faijes, M.; Perez, X.; Perez, O.; Planas, A. Biochemistry 2003, 42, 13304–13318.
26. Viladot, J. L.; de Ramon, E.; Durany, O.; Planas, A. Biochemistry 1998, 37, 11332–
11342.
References
27. Planas, A.; Millet, O.; Palasí, J.; Pallarés, C.; Abel, M.; Viladot, J. L. Carbohydr. Res.
1998, 310, 53–64.
1. Zechel, D. L.; Withers, S. G. Acc. Chem. Res. 2000, 33, 11–18.
2. Wang, Q.; Withers, S. G. J. Am. Chem. Soc. 1995, 117, 10137–10138.
3. Tews, I.; Vanscheltinga, A. C.; Perrakis, A.; Wilson, K. S.; Dijkstra, B. W. J. Am.
Chem. Soc. 1997, 119, 7954–7959.
4. Wang, L.; Huang, W. Curr. Opin. Chem. Biol. 2009, 13, 592–600.
5. Mackenzie, L. F.; Wang, Q.; Warren, R. A. J.; Withers, S. G. J. Am. Chem. Soc 1998,
120, 5583–5584.
6. Malet, C.; Planas, A. FEBS Lett. 1998, 440, 208–212.
7. Coutinho, P. M.; Henrissat, B. In Recent Advances in Carbohydrate Bioengineering;
Gilbert, H. J., Davies, G. J., Henrissat, B., Svensson, B., Eds.; The Royal Society of
Chemistry: Cambridge, 1999; pp 3–12.
8. Perugino, G.; Trincone, A.; Rossi, M.; Moracci, M. Trends Biotechnol. 2004, 22,
31–37.
28. Carraher, C. E. Seymour/Carraher’s Polymer Chemistry; CRC Press, 2007.
29. Sali, A.; Blundell, T. L. J. Mol. Biol. 1993, 234, 779–815.
30. Biarnes, X.; Nieto, J.; Planas, A.; Rovira, C. J. Biol. Chem. 2006, 281, 1432–1441.
31. Planas, A.; Viladot, J. L.; Faijes, M. In Recent Advances in Carbohydrate
Bioengineering; Gilbert, H. J., Davies, G. J., Henrissat, B., Svensson, B., Eds.; The
Royal Society of Chemistry: Cambridge, 1999; pp 292–301.
32. Viladot, J. L.; Canals, F.; Batllori, X.; Planas, A. Biochem. J. 2001, 355, 79–86.
33. Faijes, M.; Imai, T.; Bulone, V.; Planas, A. Biochem J. 2004, 380, 635–641.
34. Biarnes, X.; Ardevol, A.; Iglesias-Fernandez, J.; Planas, A.; Rovira, C. J. Am. Chem.
Soc. 2011, 133, 20301–20309.
35. Davies, G. J.; Planas, A.; Rovira, C. Acc. Chem. Res. 2012, 45, 308–316.
36. Wang, J.; Pengthaisong, S.; Cairns, J. R.; Liu, Y. Biochim. Biophys. Acta 2013, 1834,
536–545.
37. Wilcken, R.; Zimmermann, M. O.; Lange, A.; Joerger, A. C.; Boeckler, F. M. J. Med.
Chem. 2013, 56, 1363–1388.
9. Faijes, M.; Planas, A. Carbohydr. Res. 2007, 342, 1581–1594.
10. Cobucci-Ponzano, B.; Perugino, G.; Strazzulli, A.; Rossi, M.; Moracci, M. Methods
Enzymol. 2012, 510, 273–300.
11. Armstrong, Z.; Withers, S. G. Biopolymers 2013, 99, 666–674.
12. Planas, A. Biochim. Biophys. Acta 2000, 1543, 361–382.
13. Malet, C.; Planas, A. Biochemistry 1997, 36, 13838–13848.
14. Planas, A.; Faijes, M.; Abel, M. J. Appl. Glycosci. 2003, 50, 245–251.
15. Faijes, M.; Fairweather, J. K.; Driguez, H.; Planas, A. Chem. Eur. J. 2001, 7, 4651–
4655.
16. Fairweather, J. K.; Faijes, M.; Driguez, H.; Planas, A. Chem. Biochem. 2002, 3,
866–873.
17. Faijes, M.; Pérez, X.; Pérez, O.; Planas, A. Biochemistry 2003, 42, 13304–13318.
18. Faure, R.; Saura-Valls, M.; Brumer, H., III; Planas, A.; Cottaz, S.; Driguez, H. J.
Org. Chem. 2006, 71, 5151–5161.
38. McIntosh, L. P.; Hand, G.; Johnson, P. E.; Joshi, M. D.; Korner, M.; Plesniak, L. A.;
Ziser, L.; Wakarchuk, W. W.; Withers, S. G. Biochemistry 1996, 35, 9958–
9966.
39. Keitel, T.; Simon, O.; Borriss, R.; Heinemann, U. Proc. Natl. Acad. Sci. U.S.A. 1993,
90, 5287–5291.
40. Hahn, M.; Pons, J.; Planas, A.; Querol, E.; Heinemann, U. FEBS Lett. 1995, 374,
221–224.
41. Gaiser, O.; Piotukh, K.; Ponnuswamy, M. N.; Planas, A.; Borriss, R.; Heinemann,
U. J. Mol. Biol. 2006, 357, 1211–1225.
42. Heinemann, U.; Aÿ, J.; Gaiser, O.; Müller, J. J.; Ponnuswamy, M. N. Biol. Chem.
1996, 377, 447–454.