Mechanism-Oriented Redesign of an Isomaltulose Synthase to an Isomelezitose Synthase by Site-Directed Mutagenesis

Article abstract of DOI:10.1002/cbic.201100576

An isomelezitose synthase was redesigned out of the sucrose isomerase from Protaminobacter rubrum for the synthesis of isomelezitose (6-O^F-glucosylsucrose), a potential nutraceutical. The variants F297A, F297P, R333K, F321A-F319A and E428D catalyze the formation of isomelezitose in up to 70% yield.

Full text of DOI:10.1002/cbic.201100576

DOI: 10.1002/cbic.201100576
Mechanism-Oriented Redesign of an Isomaltulose
Synthase to an Isomelezitose Synthase by Site-Directed
Mutagenesis
[
a]
Julian Gçrl, Malte Timm, and Jꢀrgen Seibel*
Dedicated to Professor Gerd Bringmann on the occasion of his 60th birthday.
An isomelezitose synthase was redesigned out of the sucrose
The variants F297A, F297P, R333K, F321A_F319A and E428D
isomerase from Protaminobacter rubrum for the synthesis of
catalyze the formation of isomelezitose in up to 70% yield.
F
isomelezitose (6-O -glucosylsucrose), a potential nutraceutical.
Introduction
F
F
The trisaccharide isomelezitose (6-O -glucosylsucrose, 1,
as well as the synthesis of 4-O -glucosylsucrose and theander-
[
1]
G
Scheme 1), recently found in honey, has been identified as a
ose (6-O -glucosylsucrose) was observed by Inohara-Ochiai
[
2]
[5]
potent nutraceutical candidate.
et al. when incubating an a-glucosidase from Bacillus sp.
SAM1606 with 60% (w/v) sucrose at 608C. Unfavorably, the
[7]
enzyme preferentially catalyzes hydrolysis (71%). Mutagene-
[5,7]
sis experiments performed by Inohara-Ochiai et al.
succeed-
ed in changing the ratio of the obtained trisaccharides in favor
of 1, but the overall yield was reduced from 8 to 2%. Further-
more, hydrolysis was still the main reaction (90%).
Results and Discussion
Enzymes are highly efficient catalysts for stereoselective syn-
theses. However, the reactions are limited due to high sub-
strate and product specificity. Our goal is to expand the syn-
thetic repertoire of enzymes. Here we provide an example
where an enzyme has been redesigned for efficient synthesis
of isomelezitose (1). With the a-glucosidase from B. sp.
SAM1606 which forms 1 as a starting point, the following
steps were performed to obtain an isomelezitose synthase.
Scheme 1. The structure of the trisaccharide isomelezitose (1).
Isomelezitose is a substrate for bifidobacteria in the colon
[
2]
and thus has probiotic properties. It is not cleaved either by
salivary enzymes or bacteria of the pharynx or in the small
intestine. Isomelezitose is noncariogenic, has a smaller calorific
value in comparison to sucrose and is suitable for inclusion in
[
2]
foods for diabetics. Due to the interest in this compound, an
industrial process has been developed which uses immobilized
cells of Protaminobacter rubrum (CBS 574.77) commonly used
for the synthesis of isomaltulose from sucrose on an industrial
* Identification of glucansucrases which exhibit high se-
quence identity to the a-glucosidase from B. sp. SAM1606
[8,9]
[10]
utilizing alignment tools (BLAST, ClustalW2 ).
* Narrowing down the query to enzymes within the PDB
which show high transfer activity and
* perform the desired a-(1,6)-glycosidic coupling reaction
(mechanism-based analysis).
* Selection of promising amino acids for site-directed muta-
genesis by sequence and structural alignments in combina-
tion with docking studies (estimation of space requirement
and possible interactions for the catalytic process).
[
3]
scale (100000 tons per annum), forming 1 in 11% yield at
[
2]
5
08C. However, drawbacks of the process are high reaction
temperature, which causes a decreased stability of the enzyme
at most 43 d compared to 200 d for isomaltulose production)
and low yield, which results in the need for further concentra-
(
[
2,3]
tion and purification steps.
As complete cells were used the
enzyme forming the trisaccharide remained unknown.
Fujii et al. described a similar method and obtained the tri-
saccharide in 8% yield using immobilized cells from S. plymuth-
[
4]
[4,5]
[a] J. Gçrl, M. Timm, Prof. Dr. J. Seibel
ica. Some a-glucosidases
of the family GH13 are responsi-
Institute of Organic Chemistry, Julius-Maximilians University Wꢀrzburg
Am Hubland, 97074 Wꢀrzburg (Germany)
E-mail: seibel@chemie.uni-wuerzburg.de
ble for the synthesis of 1 according to the carbohydrate-active
[
6]
enzymes (CAZy) database. As an example, the formation of 1
ChemBioChem 2012, 13, 149 – 156
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
149

J. Seibel et al.
*
Site-directed mutagenesis to shift the product specificity to
isomelezitose production.
Sequence and structure alignments and query refinements
Since the a-glucosidase from B. sp. SAM1606 synthesizes 1 as a
[
5]
side product its sequence was applied in several alignments
[
8,9]
[10]
(BLAST,
ClustalW2 ). The alignments revealed a large
number of a-(1,6)-glucosidases from the GH13 family (a-amy-
[
6]
lase family, CAZy database ) to be most similar (>90% query
[
8,9]
coverage according to BLAST ).
For the generation of new enzyme specificities, directed
[11–13]
evolution has been demonstrated to be a successful tool.
An alternative approach for some enzymes could involve site-
[
14]
directed mutagenesis. For this, incorporating the knowledge
of structure and function of a given enzyme involving detailed
data on the reaction mechanism is a basic requirement for the
[15]
selection of target sites for site-directed mutations.
Thus,
only enzymes within the Protein Data Bank were further exam-
[
16]
ined using 3D-Jury. The a-glucosidase from B. cereus (80%
sequence identity, 3D-Jury score 372.00) was predicted to have
the highest similarity with the glucosidase from B. sp.
SAM1606. Since the synthesis of 1 is a transfer reaction, we
then focused on enzymes with a high transfer activity. The
query revealed the sucrose isomerases (SIs, EC. 5.4.99.11, up to
[17–20]
Scheme 2. Product specificity of the different sucrose isomerases.
Reaction mechanism, linkage and structure
[
17–20]
99% transfer activity
) among which are SmuA from
P. rubrum and MutB from P. mesoacidophila MX-45 (3D-Jury
Therefore, primary, secondary and tertiary structure of the SIs
and related a-glucosidases were examined. Although the struc-
tures of the SIs are very similar (up to 80% sequence identity),
the ratio of isomaltulose (3) to trehalulose (4) varies enormous-
score: 365.00 and 361.33, respectively) to be most similar. The
[
10]
obtained sequences were aligned using ClustalW2 (Figure 1,
[
21]
visualization: Jalview ).
[19,22–24]
Next, we took a closer look on the product specificities and
the reaction mechanism. The SIs catalyze the isomerisation of
sucrose to isomaltulose (d-fructofuranosyl-(6,1)-a-d-glucopyra-
noside, 3) and trehalulose (d-fructofuranosyl-(1,1)-a-d-gluco-
pyranoside, 4, Scheme 2). The formation of 1 was not yet ob-
served and the structural determinants that enable the synthe-
sis of 1 by a-glucosidases remain unknown. By site-directed
mutagenesis we anticipated to generate a variant of the SI
which performs the synthesis of 1.
ly depending on the originating organism (Scheme 2).
The isomaltulose synthase from P. rubrum (SmuA, 45% se-
quence identity to the a-glucosidase from B. sp. SAM1606,
95% transfer activity) catalyzes the desired a-(1,6)-coupling re-
action, whereas the main product of MutB (trehalulose syn-
[18]
thase from P. mesoacidophila) is trehalulose (4) (Scheme 2).
Zhang et al. and Ravaud et al. recently published the struc-
[18,22,24]
tures of SmuA, MutB and PalI.
The enzymes consist of
three domains: domain A containing the (b/a) -barrel (TIM
8
Figure 1. Part of the ClustalW2-Alignment of the a-glucosidase from B. sp. SAM1606 to enzymes within the PDB; blue: density represents identity, magenta:
catalytic residues, orange: fructose binding site, green: entrance to active pocket; SmuA: SI from P. rubrum, MX-45: MutB, SI from P. mesoacidophila MX-45.
1
50
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2012, 13, 149 – 156

Redesign of an Isomaltulose Synthase
barrel), a loop-rich N-terminal
region (domain B) and domain C
harboring several antiperiplanar
b-sheets.
The residues D241, E295,
H368 and D369 (magenta,
SmuA numbering, Figure 1)
were identified to catalyze the
[
22,25,26]
isomerization of 2 to 3.
The reaction occurs by a so
called “ping-pong” mechanism.
D241 is the nucleophile which
attacks the glycosidic bond and
forms a covalent b-linked gluco-
Figure 2. Comparison of the surface of the a-glucosidase from B. cereus (A) and the SI SmuA (B); the entrance is
wide open in case of the a-glucosidase compared to the narrow entry of the SI preventing the hydrolysis (visuali-
zation: PyMOL 1.3).
syl–enzyme
intermediate
(“ping”). This is facilitated by
E295 (acid/base catalyst) and
D369 (transition-state stabilizer).
2
57
265
[25]
Hence, the fructosyl residue is cleaved but remains inside the
active site. The product specificity (ratio of 3:4) is proposed to
be determined by a unique fructosyl binding site (FBS,
QQQLKNFA (Na5, Lee et al. ), but excludes trisaccharide
formation (Figure 2B).
325
329
RXDRX , +1 sub-site; orange, Figure 1) which is responsible
Docking studies
for the correct orientation of the fructosyl moiety during the
isomerization step (nucleophilic attack of the 6-OH of the fruc-
In order to estimate the space requirement of sucrose as an
acceptor and to identify interactions between the enzyme and
sucrose, docking studies were performed. For this, an enzyme–
substrate (ES) complex was generated by superposition of the
crystal structure from SmuA (3GDB) with MutB (2PWE) harbor-
ing a sucrose molecule in the active pocket due to the muta-
tion E295Q. The fructosyl moiety was deleted (PyMOL) and the
resulting structure was subsequently used as template in Auto-
[
23]
tosyl moiety on the ES-complex; “pong”). An aromatic clamp
(F297, F321; green, Figure 1) at the entrance of the active
[
18,24]
pocket protects the ES-complex from attack by water.
In
addition, only low yields (<5.6%) were obtained in acceptor/
[
27]
substrate studies previously performed.
However, transfer
reactions in higher yields are possible if the SI is mutated. Lee
et al. demonstrated that isomaltose can be obtained in 19%
yield by the SmuA variant R325K (part of FBS) with glucose as
[28]
Dock 4.2. In two calculations (100 runs each) fructose and
sucrose respectively were applied as ligands in the docking
process. We suggest that the results can help us to predict
[
25]
the acceptor.
1
2
) whether sucrose fits into the narrow active pocket,
) whether the fructosyl moiety of sucrose is able to adopt
Selection of amino acids for site-directed mutagenesis
similar conformations as fructose alone and 3) which amino
acids interact with sucrose as an acceptor.
With the knowledge of these previous studies, we aimed to re-
design the isomaltulose synthase from P. rubrum into an isome-
lezitose synthase by site-directed mutagenesis. In addition to
sequence alignments (Figure 1) it was helpful to compare the
structure of SmuA with the a-glucosidase from B. sp. SAM1606.
Due to the lack of a crystal structure of the a-glucosidase from
B. sp. SAM1606, the a-glucosidase from B. cereus (80% se-
quence identity with SAM1606, 3D-Jury score 372.00) was used
as a template for further alignments. The a-glucosidase from
B. cereus (1UOK) was aligned to the SmuA (3GDB) using PyMol
Docking of fructose as ligand
Results differing by less than 1.0 ꢂ root mean square deviation
(rmsd) were clustered (top, Table 1). The numbering of the
clusters depends on the highest binding energy of a confor-
mation within a cluster.
Docking of sucrose as ligand
(Figure 2). The position of the substrate sucrose was estimated
by an additional superposition of 2PWE (MutB, 80% identity to
SmuA) to 3GDB/1UOK harboring a sucrose molecule in the
To simulate the formation of 1, sucrose was used as a ligand in
the docking-process. The docking results are listed in Table 1.
This time, conformations with a clRMS value less than 1.5 ꢂ
were clustered since deviations were mostly caused by the ori-
entation of the glucosyl moiety.
295
active site (mutant E Q, SmuA-numbering).
Although the structures of both enzymes are similar, the a-
glucosidase lacks the FBS of SmuA (Figure 2A). As a conse-
quence, the a-glucosidase exhibits a wide-open entrance to
the active pocket, favoring the hydrolysis but enabling synthe-
sis of 1. In contrast, the SmuA prevents hydrolysis due to its
According to the docking results, sucrose fits in the active
site of the glucopyranosyl–enzyme complex (Figure 3A). The
fructosyl residue of sucrose (+1 sub-site, Figure 3A) is able to
occupy similar conformations compared to fructose alone (not
aromatic clamp, the FBS and
a different loop/a-helix
ChemBioChem 2012, 13, 149 – 156
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
151

J. Seibel et al.
enlarge the +2 subsite (Figure 3A) while retaining the polarity
of the +2 subsite.
Table 1. The docking results are clustered und sorted by binding energy.
[
a]
Docking (rmds)
Cluster
Number of
Highest binding
With respect to the previous alignments, Phe297 (green, Fig-
ures 1, 3; Table 2) was mutated to Ala and Pro to enlarge the
À1 [b]
conformations energy [kcalmol
]
1
64
3
À4.77
À4.57
À4.09
À3.94
À6.52
À4.77
À1.79
À1.90
À7.06
À7.01
À6.82
À6.76
ES-complex+fructose 2–3
(
1.0 ꢂ)
4
5
1
2
3
4
1
13
20
57
13
6
24
17
12
6
Table 2. Performed mutations, their location within the protein and ob-
–13
À1
served isomelezitose production (enzyme conc. 0.1 gL , Sorensen buffer
(
50 mm, pH 7.0, 308C).
ES-complex+sucrose
1.5 ꢂ)
(
Mutation
Location/
proximity
Isomelezitose
synthesis
Remaining
transfer
–15
activity
ES-complex F297A+
sucrose (1.5 ꢂ)
2–3
4
[
a]
F297A
F297P
F321A
R333K
aromatic clamp
aromatic clamp
aromatic clamp
+2 subsite, opposite
of F297 and F321
opposite to FBS,
yes
yes
hydrolysis
yes
4%
[
a]
5
–43
65
<1%
–
[a] The numbering of the clusters depends on the highest binding
[a]
13%
energy of a conformation within a cluster; [b] The binding energy was
calculated by AutoDock 4.2.
[b]
E428D
R456K
yes
no
+
1 subsite
in between À1 and
1 subsite
[c]
inactive
+
[
[
b]
b]
shown) with C-6’ in position for the nucleophilic attack of the
F321A_F319A
F297A_R333K
aromatic clamp
aromatic clamp,
yes
yes
glucopyranosyl–enzyme complex.
+
2 subsite
[
a] No Michaelis–Menten kinetic; only valid at a concentration of 200 mm
sucrose, 100% corresponds to the activity of isomaltulose formation by
the wild-type enzyme. [b] Not determined due to low activity, but forma-
tion is observed on TLC. [c] No activity detected.
Selection of targets for site-directed mutagenesis
On one hand, for the synthesis of 1, binding of sucrose as an
acceptor must be favored compared to fructose. On the other,
we aimed to prevent hydrolysis. The isomerization of sucrose
to isomaltulose is a highly efficient reaction. We think that the
formation of the b-glucopyranosyl–enzyme complex followed
by the nucleophilic attack of the 6’-OH group of fructose is
concerted. This would explain why no hydrolysis takes place.
We conclude that the FBS is necessary for this concerted pro-
cess. Previous mutations of the FBS led to more hydrolysis,
entrance of the active pocket. Our docking studies reveal that
rotamers (not shown) of Arg333 (cyan, Figures 1, 3; +2 subsite;
Table 2), which was supposed to provide H-bonds to the fruc-
[18]
tosyl moiety, are able to interact with the fructosyl (1’-OH,
ca. 2.5 ꢂ) and the 6-OH group of the glucosyl residue of
sucrose (ca. 2.7 ꢂ) during the transfer reaction. In addition,
whereas Arg333 is conserved in the SIs it is substituted by Lys
in case of the a-glucosidase from B. sp. SAM1606. Thus, we
mutated Arg333 to Lys. In addition, Glu428 (red, Figures 1, 3;
Table 2) was substituted by Asp since it may interact with su-
[
25]
supporting our suggestions. Thus, the FBS itself should not
be mutated. Furthermore, we presumed that the bulky gluco-
syl moiety of sucrose as an acceptor might prohibit reactive
conformations of the fructosyl moiety. Therefore, we aimed to
[
29,30]
Figure 3. A) Docking of sucrose to the generated enzyme–substrate complex (visualization: VMD 1.9
). B) Within cluster 4 the fructosyl residue of sucrose
is able to obtain new, more flexible conformations compared to the wild-type enzyme.
1
52
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2012, 13, 149 – 156

Redesign of an Isomaltulose Synthase
[
33]
crose (possible H-bond with 2’-OH and 3’-OH of fructose, +1
sub-site) as well as Arg456 (orange, Figures 1, 3; Table 2) which
is positioned between the +1 and À1 subsites.
strongly from the facilitated Michaelis–Menten plot. For this
reason, KM could not be calculated and the kinetic parameters
given in Tables 2 and 3 are only valid for the described condi-
tions (200 mm sucrose).
The desired variants were created in silico (PyMOL Mutagen-
esis Tool) and similar AutoDock studies were carried out. As an
example, Figure 3B shows the docking of sucrose to the simu-
lated ES-complex of variant F297A. Cluster 1 matches the con-
formation of the sucrose in the previous docking experiments,
whereas cluster 4 adopts new conformations for the fructosyl
residue. These conformations are probably caused by a differ-
ent positioning of the glucosyl moiety. Its calculated binding
Similarly with some variants of the a-glucosidase from B. sp.
SAM1606, the formation of 1 was fast at the beginning with
no detectable production of 3, 4 or monosaccharides (hydroly-
sis). 1 can be obtained on a preparative scale in 70% yield
based on sucrose consumed (200 mm sucrose, 50 mm Soren-
sen buffer pH 6.0, 258C, 17 h). After long time the reaction pro-
ceeds with the formation of 3. To optimize the process, 1 has
to be continuously isolated.
À1
energy (À6.82 kcalmol , Table 1) is higher than the binding
À1
energy of cluster 1 (À7.06 kcalmol , Table 1), but still exceeds
The structure of 1 was confirmed by NMR. The linkages of
the three glycosyl moieties were determined by two-dimen-
sional correlation spectroscopies (HSQC, HMBC, TOCSY,
NOESY). The correlation signal between H-1 (5.42 ppm, d, 1H)
of glucose and C-2’ (104.1 ppm) of fructose (Figure 4, top) and
the highest binding energy of the wild-type enzyme
À1
(
À6.52 kcalmol , Table 1). Besides the influence of the muta-
tion on the fructose binding and hydrolysis, this could reflect
an increased possibility for the fructosyl residue to occupy re-
active conformations for isomelezitose synthesis.
Site-directed mutagenesis, determination of product
specificity and kinetic data
By site-directed mutagenesis we successfully generated multi-
ple variants, which are able to synthesize 1 and have minor hy-
drolysis activity (<5%, determined using TLC). All performed
mutations are listed in Table 2. In particular, F297A and R333K
still preserved 4 and 13% activity (calculated by the method of
[
31]
Vilozny et al., Table 3) compared to isomaltulose formation
by the wild-type enzyme. The mutations F297P and E428D as
well as the combinations F297A_R333K and F321A_F319A lead
to a significant decrease in activity (Tables 2, 3). For these var-
iants the formation of the trisaccharide was observed using
TLC and 1 was isolated in similar yields (for example, 70% for
the variant F321A_F319A), but we were unable to measure reli-
able kinetic data. R456K showed nearly no activity, whereas
only hydrolysis was observed for F321A (Tables 2, 3).
Interestingly, the Michaelis–Menten plots of the variants
show linear regressions instead of hyperbolic slopes. Even with
1
.2 m sucrose no saturation of initial velocity (v ) was observed.
i
[33]
This kind of progression has been previously observed and
can either be explained by allosteric effects or a switch in the
reaction mechanism. Since two molecules of sucrose are
needed for synthesis of 1, the kinetic reaction order can differ
Figure 4. Close-up views on the HMBC spectrum of isomelezitose.
3
À1
the J coupling constant of H-1
Table 3. Kinetic data of the wild-type and the variants (0.1 gL ) in Sorensen buffer (50 mm, pH 7.0), 308C.
(
3.90 Hz) show the a-(1,2)-link-
À1
À3
À1
À1 À1
À1
Variant
V
max [mms 10
]
K
M
[mm]
k
cat [s
]
k
cat/K
M
[mm
s
]
Specific activity [Umg ]
age between fructose and glu-
cose (=sucrose). In addition,
the correlation signal between
[
32]
3
wt (pH 5.5, 258C)
wt (pH 7.0, 308C)
–
32
22Æ1
42.1ꢃ10
1301.0
500
35.8Æ0.4
1.54Æ0.08
0.24Æ0.02
4.6Æ0.6
30Æ5
1.4Æ0.2
1.1Æ0.1
0.046Æ0.005
[
a]
F297A
F297P
R333K
–
–
–
1.3Æ0.2
–
–
3
[
a]
À3
À3
H-1’’ (4.96 ppm, d, 1H; J=
3.73 Hz) and C-6’ of fructose
verifies
0.20Æ0.03
7.2ꢃ10 Æ0.9ꢃ10
[
a]
3.8Æ0.6
–
0.14Æ0.02
[
[
[
b]
b]
b]
[b]
[b]
[b]
[b]
F321A_F319A
F297A_R333K
E428D
the
a-(1,6)-linkage
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
(
Figure 4, top). Furthermore,
the correlation signals between
[
a] No Michaelis–Menten kinetics; only valid at a concentration of 200 mm sucrose; [b] Not determined due to
the diastereotopic protons H-6’
low activity but formation is observed on TLC.
and
C-1’’
are
observed
ChemBioChem 2012, 13, 149 – 156
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
153

J. Seibel et al.
(
Figure 4, bottom). All other signals are in agreement with the
Table 4. Primers for the desired mutations; mutated bases are under-
lined.
structure.
Mutation
R333K
Primer
Conclusion
For 5’-CGA GAC TCT GAT CAA AAG TGG CGT CGA AAA GAT
TGG-3’
The redesign of enzymes for new substrates and/or functions
is a key endeavor of biocatalysis and the enzymatic synthesis
of carbohydrates has been proven to be nontrivial in this
regard. We succeeded with one single amino acid exchange,
R333K, to yield up to 70% isomelezitose (1), keeping in mind
that no activity was observed with the wild-type enzyme. In
conclusion, the scope of an enzyme was enhanced by site-
directed mutagenesis. Sequence and structure alignments as
well as mechanism-based computational docking studies were
utilized to create an isomelezitose synthase out of the sucrose
isomerase from Protaminobacter rubrum. An industrial process
by using this enzyme as an immobilized catalyst could be en-
visaged.
Rev 5’-CCA ATC TTT TCG ACG CCA CTT TTG ATC AGA GTC
TCG-3’
F297A
F297P
For 5’-CC GGT GAA ATC GCT GGC GTA CCC-3’
Rev 5’-GGG TAC GCC AGC GAT TTC ACC GG-3’
For 5’-CC GGT GAA ATC CCT GGC GTA CCC-3’
Rev 5’-GGG TAC GCC AGG GAT TTC ACC GG-3’
For 5’-CTG AAC ATT GCA GCT ACC GCT GAC TTA ATC AGA
CTC G-3’
Rev 5’-C GAG TCT GAT TAA GTC AGC GGT AGC TGC AAT
GTT CAG-3’
For 5’-CTG AAC ATT GCA TTT ACC GCT GAC TTA ATC AGA
CTC G-3’
Rev 5’-C GAG TCT GAT TAA GTC AGC GGT AAA TGC AAT
GTT CAG-3’
For 5’-C GAT GAT ATT GAC GTG AAA GGT TTT TGG C-3’
Rev 5’-G CCA AAA ACC TTT CAC GTC AAT ATC ATC G-3’
For 5’-CGC CTG ACG AGC AAG GAT AAC AGC C-3’
Rev 5’-G GCT GTT ATC CTT GCT CGT CAG GCG-3’
F319A_
F321A
F321A
E428D
R456K
Experimental Section
AutoDock studies: The docking experiments were carried out with
[
28]
AutoDockTools 1.5.4 with AutoDock version 4.2. An enzyme–sub-
strate complex was generated by superposition of the crystal struc-
ture from SmuA (3GDB) with MutB (2PWE) harboring sucrose in
the active pocket due to the mutation E295Q. The fructosyl moiety
Technologies Inc., Santa Clara, USA) following the manufacturer’s
manual. The digested (DpnI) PCR products were transformed into
electro-competent cells of E. coli TOP10 and the transformants
were selected with LB-Amp. The plasmids were isolated using the
E.Z.N.A. Plasmid Mini Kit I (OMEGA Bio-Tek, Norcross, USA). The de-
sired mutations were confirmed by DNA sequencing (GATC biotech
AG, Konstanz, Germany).
[34]
was deleted (PyMOL, version 1.3, Schrçdinger, LLC, ) and the re-
sulting complex was applied in AutoDock as a macromolecule file.
Mutated amino acids (introduced with the PyMOL mutagenesis
tool) were set to flexible as their conformation may vary. The con-
formations of sucrose and fructose (ligand) were adopted from the
crystal structure 1GI (amylosucrase from Neisseria polysaccharea,
.0 ꢂ resolution) harboring sucrose in the active site. In the case of
fructose, the glucosyl moiety was deleted from sucrose (PyMOL).
00 runs were performed for each docking experiment. The results
Enzyme production by fermentation: Precultures were grown in
test tubes containing LB-Amp (5 mL) at 378C with shaking at
2
(
50 rpm for 17 h. The precultures were transferred into LB-Amp
250 mL) and incubated at 378C with shaking at 160 rpm. The
cultures were induced with isopropyl-b-d-thiogalactopyranoside
IPTG) at an OD600 value of approximately 0.6 to a final concentra-
tion of 0.50 mm and incubated for 17 h at 208C with shaking at
2
1
(
were clustered by root mean square deviation (rmsd, 1.0 ꢂ for fruc-
tose and 1.5 ꢂ for sucrose as ligand) and ranked by binding energy
calculated with AutoDock 4.2.
1
4
60 rpm. The cells were collected by centrifugation at 6400g and
8C, washed with Sorensen buffer (10 mL, 50 mm, pH 8.0), resus-
Oligonucleotides: Oligonucleotides were obtained from biomers.
net (Ulm, Germany). Primers (Table 4) for subcloning of the sucrose
isomerase from Protaminobacter rubrum CBS 547.77 (GenBank ac-
cession number CQ765969): for_SalI 5’-TAT AGT CGA CCC CGT CAA
GGA TTG AAA ACT GC-3’ and rev_NotI 5’-ATA TGC GGC CGC TTA
TTG ATT TAG TTT ATA AAC CCC-3’. PCRs were performed with a
PCR MasterCycler personal instrument (Eppendorf).
pended in the same buffer (6.25 mL) supplemented with b-mercap-
toethanol (5 mm) and imidazole (10 mm). The cells were disrupted
by sonification (4 min pulsed operation at 50 W).
Enzyme purification: The pET-32a vector introduced an N-terminal
His tag into the expressed proteins. After sonification, cells were
6
sedimented (10 min, 13300g, 48C), and the proteins were isolated
from the supernatants using HIS-Select Nickel Affinity Gel (Sigma–
Aldrich) following the manufacturer’s manual. The eluted proteins
were rebuffered in 50 mm Sorensen buffer (pH 7.0) utilizing HiTrap
desalting columns (Amersham Biosciences Europe GmbH, Freiburg,
Germany) according to the manufacturer’s manual. The proteins
were concentrated with Vivaspin 500 (10000 MWCO PES) concen-
trators (Sartorius Stedim Biotech GmbH). Protein concentration was
determined by the Bradford assay with bovine serum albumin as
standard. Purification was checked by SDS-PAGE using pre-cast
gels (VarioGel, Anamed Elektrophorese, Grob-Bieberau, Germany)
with PageRuler unstained protein Ladder (Fermentas) as standard.
Bacterial strains, plasmids and culture conditions: Escherichia coli
TOP10 and Escherichia coli BL21 Star (both Invitrogen) were used
for cloning procedures and protein expression, respectively. Prota-
minobacter rubrum CBS 547.77 was obtained from the Centraalbur-
eau voor Schimmelcultures (Utrecht, Netherlands). The amplified
gene was digested with SalI and NotI (both Fermentas) and sub-
cloned into a pET-32a vector (Merck). The bacteria, harboring the
recombinant plasmid (wild-type or mutant) were routinely grown
at 378C in lysogeny broth medium (250 mL), supplemented with
À1
ampicillin (200 mgL , LB-Amp).
General techniques: PCRs, agarose gel electrophoresis and trans-
formations were carried out according to standard protocols.
[35]
Enzyme assays: Enzyme reactions were carried out in Sorensen
À1
buffer (50 mm, pH 6.6) containing the purified protein (0.1 gL )
À1
Site-directed mutagenesis of the SI: Mutations were introduced
and sucrose (100 mgmL ) in a total volume of 20 mL. The reaction
using the QuikChange II Site-Directed Mutagenesis Kit (Agilent
mixture was incubated at 308C with shaking at 750 rpm. Samples
1
54
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2012, 13, 149 – 156

Redesign of an Isomaltulose Synthase
(
1 mL) were collected, diluted with H O (99 mL, final sugar concen-
ics-Microflex MALDI-TOF instrument with 2,5-dihydroxybenzoic
2
À1
1
tration of 1 mgmL ) and stored at À208C.
acid (DHB) as matrix. H NMR (600 MHz, D O, acetic acid, 258C):
2
d=5.42 (d, J=3.90 Hz, 1H; H-1), 4.96 (d, J=3.73 Hz, 1H; H-1’’),
Enzyme kinetics: Kinetic data were collected utilizing an enzyme
4
4
.21 (d, J=8.60 Hz, 1H; H-3’), 4.10 (t, J=8.60 Hz, 1H; H-4’), 4.06–
.03 (m, 2H; H-5’, H-6’), 3.86 (dd, J=12.3, 2.30 Hz, 1H; H-6a), 3.86
[31]
assay developed by Vilozny et al. 2009. The boronic acid N,N’-bis-
benzyl-4-boronic acid)-4,4’-bipyridinium dibromide (4,4’-m-BBV),
(
(m, 2H; H-6’’a, H-5), 3.81–3.74 (m, 5H; H-6b, H-3, H-6’’b, H-3’’, H-6’),
[36]
which was synthesized as described elsewhere, was used as a
quencher and 8-hydroxypyrene-1,3,6-trisulfonic acid trisodium salt
(
taining 4,4’-m-BBV (2.05 mm), HPTS (16.7 mm) and enzyme
(
3
.72 (ddd, J=10.0, 4.67, 2.30 Hz, 1H; H-5’’), 3.67 (d, J=5.28, 2H; H-
’), 3.57–3.54 (m, 2H; H-2, H-2’’), 3.46–3.43 (m, 2H; H-4, H-4’’);
1
HPTS, Sigma–Aldrich) as a fluorescent dye. 20 mL reactions con-
13
C NMR (150 MHz, D O, acetic acid, 258C): 101.5 (C-2’), 96.02 (C-
2
1
7
6
’’), 89.61 (C-1), 77.17 (C-5’), 73.84 (C-3’), 72.05 (C-4’), 70.70 (C-3’’),
0.42 (C-3), 70.10 (C-5), 69.64 (C-5’’), 69.07 (C-2’’), 68.87 (C-2), 67.21,
À1
0.1 gL ) in Sorensen buffer (50 mm, pH 7.0) were performed in
3
84-well plates. For the Michaelis–Menten plot the initial concen-
7.14 (C-4, C-4’), 66.48 (C-6’), 59.28 (C-1’), 58.23 (C-6’’), 58.05 (C-6);
À1
tration of sucrose was varied from 0.0625–600 gL . The change of
fluorescence F (t=time, l =360 nm, l =535 nm) was continu-
+
MS: m/z: 527 [M+Na] .
t
ex
em
ously monitored (interval of 30 s for 15 min) in a GENios well-plate
reader (TECAN Austria GmbH) at 308C. Assays without enzyme
were used as blank. The blank-corrected ratio F /F was converted
Acknowledgements
t
0
into concentrations by using a linear calibration of fructose/isomal-
This work was supported by the DFG (SFB578).
tulose at different concentrations. The initial velocity (v ) was ob-
i
tained by linear regression of the concentration of the detected
sugar over time. The plots of the variants showed no hyperbolic
slope, instead linear regressions (R>99%) were obtained. No satu-
ration was observed even at 1.2m sucrose. For this reason, KM
could not be calculated and the kinetic parameters are only valid
Keywords: biocatalysis · enzyme engineering · isomelezitose ·
oligosaccharides · sucrose isomerase
À1
[1] J. Gꢄmez Bꢅrez, R. Garcia-Villanova, S. Elvira Garcia, A. Gonzꢅlez Par-
at 200 mm sucrose [enzyme (0.1 gL ) in Sorensen buffer (50 mm,
pH 7.0)].
amꢅs, Chromatographia 1999, 50, 461–469.
[
2] V. M. Munir, Vol. WO9922013 (Ed.: M. M. Suedzucker AG, Mannheim/Och-
senfurt, M. Vogel), Europe, 1999.
Thin-layer chromatography: To monitor the progress of product
formation diluted samples (1:100 with water) of the enzyme assays
were applied onto silica thin-layer chromatography plates (20 cm x
[
[
3] K. Buchholz, J. Seibel, Carbohydr. Res. 2008, 343, 1966–1979.
4] S. Fujii, S. Kishihara, M. Komoto, J. Shimizu, J. Sci. Food Agric. 2000, 30,
339–344.
2
0 cm TLC aluminum sheets, silica gel 60 F254 with concentrating
[
5] M. Inohara-Ochiai, M. Okada, T. Nakayama, H. Hemmi, T. Ueda, T. Iwashi-
zone with dimensions of 20 cm ꢃ 2.5 cm; Merck). Standard solu-
tions were prepared containing the monosaccharides d-glucose,
sucrose, isomaltulose and isomelezitose in different concentrations
ta, Y. Kan, Y. Shibano, T. Ashikari, T. Nishino, J. Biosci. Bioeng. 2000, 89,
431–437.
[6] B. Henrissat, G. Davies, Curr. Opin. Struct. Biol. 1997, 7, 637–644.
[7] M. Okada, T. Nakayama, A. Noguchi, M. Yano, H. Hemmi, T. Nishino, T.
Ueda, J. Mol. Catal. B 2002, 16, 265–274.
8] S. F. Altschul, T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller,
D. J. Lipman, Nucleic Acids Res. 1997, 25, 3389–3402.
À1-
À1
(
0.0625 gL -1.00 gL ). The carbohydrates were separated using
two ascents (90 min) with water/2-propanol/ethyl acetate (1:3:6 v/
v/v) as solvent system. Spots were detected by dipping the plates
into the detecting reagent 0.3% N-(1-naphthyl)ethylenediamine
[
[
9] S. F. Altschul, J. C. Wootton, E. M. Gertz, R. Agarwala, A. Morgulis, A. A.
Schaffer, Y. K. Yu, FEBS J. 2005, 272, 5101–5109.
(
Fluka) with 5% concentrated sulfuric acid in methanol and heat-
ing in an oven (1208C, 5 min). The consumption of sucrose and
the yields of the products (isomelezitose and isomaltulose) were
calculated by densitometric measurement of the spots (Bio-Imag-
ing system, biostep GmbH, Jahnsdorf, Germany) using greyscale
calibration with fructose, glucose, sucrose, isomaltulose and isome-
lezitose as standards.
[
10] R. Chenna, H. Sugawara, T. Koike, R. Lopez, T. J. Gibson, D. G. Higgins,
J. D. Thompson, Nucleic Acids Res. 2003, 31, 3497–3500.
11] K. E. Jaeger, T. Eggert, A. Eipper, M. T. Reetz, Appl. Microbiol. Biotechnol.
2001, 55, 519–530.
[
[12] M. T. Reetz, Angew. Chem. 2001, 113, 292–320; Angew. Chem. Int. Ed.
2001, 40, 284–310.
[
13] F. H. Arnold, Acc. Chem. Res. 1998, 31, 125–131.
Isolation and characterization of isomelezitose (b-d-fructofura-
nosyl-(2,1)-a-d-glucopyranosyl-(6,1)-a-d-glucopyranoside): The
enzyme activity assay was performed in Sorensen buffer (50 mm,
[14] C. P. Strube, A. Homann, M. Gamer, D. Jahn, J. Seibel, D. W. Heinz, J. Biol.
Chem. 2011, 286, 17593–17600.
[
15] M. Lehmann, L. Pasamontes, S. F. Lassen, M. Wyss, Biochim. Biophys.
Acta Protein Struct. Mol. Enzymol. 2000, 1543, 408–415.
À1
pH 7.0) with enzyme (0.1 gL ), sucrose (526 mmol, final concentra-
À1
[16] K. Ginalski, A. Elofsson, D. Fischer, L. Rychlewski, Bioinformatics 2003, 19,
015–1018.
17] Y. Nagai, T. Sugitani, K. Tsuyuki, Biosci. Biotechnol. Biochem. 1994, 58,
789–1793.
tion: 100 gL ), total volume 1.8 mL. After incubation for 17 h
1
(
258C with shaking at 750 rpm), the enzyme was inactivated by
[
[
heating to 958C for 5 min. The resulting precipitate was separated
by centrifugation and the supernatant was evaporated under
reduced pressure. The residual carbohydrates were isolated by
column chromatography (silica gel 0.04–0.063 mm, Macherey–
Nagel) using the same solvent system water/2-propanol/ethyl ace-
tate (1:3:6 v/v/v). The fractions containing the desired carbohydrate
were pooled and the solvent was evaporated under reduced pres-
sure yielding 30.0 mg (59.4 mmol, 70% based on the amount of su-
crose consumed (58 mg, 169 mmol), variant: F321A_F319A) of the
1
18] S. Ravaud, X. Robert, H. Watzlawick, R. Haser, R. Mattes, N. Aghajari,
FEBS Lett. 2009, 583, 1964–1968.
[19] L. Wu, R. G. Birch, Appl. Environ. Microbiol. 2005, 71, 1581–1590.
[20] D. Zhang, X. Li, L. H. Zhang, Appl. Environ. Microbiol. 2002, 68, 2676–
2682.
[
[
[
21] A. M. Waterhouse, J. B. Procter, D. M. A. Martin, M. Clamp, G. J. Barton,
Bioinformatics 2009, 25, 1189–1191.
22] D. Zhang, N. Li, S. M. Lok, L. H. Zhang, K. Swaminathan, J. Biol. Chem.
2
003, 278, 35428–35434.
1
13
trisaccharide. H NMR (600 MHz) and C NMR (150 MHz,) were re-
corded on Bruker AMX 600 instruments using D O (d=4.79) as the
23] D. Zhang, N. Li, K. Swaminathan, L. H. Zhang, FEBS Lett. 2003, 534, 151–
2
155.
1
13
solvent and the internal H and acetic acid (21.03 ppm) as C stan-
dard. MALDI-TOF mass spectra were obtained on a Bruker Dalton-
[24] S. Ravaud, X. Robert, H. Watzlawick, R. Haser, R. Mattes, N. Aghajari, J.
Biol. Chem. 2007, 282, 28126–28136.
ChemBioChem 2012, 13, 149 – 156
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
155

J. Seibel et al.
[
25] H. C. Lee, J. H. Kim, S. Y. Kim, J. K. Lee, Appl. Environ. Microbiol. 2008, 74,
183–5194.
26] A. Aroonnual, T. Nihira, T. Seki, W. Panbangred, Enzyme Microb. Technol.
007, 40, 1221–1227.
[33] E. J. Hehre, C. F. Brewer, D. S. Genghof, J. Biol. Chem. 1979, 254, 5942–
5950.
[34] The PyMOL Molecular Graphics System, Version 1.3, Schrodinger, LLC,
2010.
5
[
2
[
[
[
[
[
27] T. Veronese, P. Perlot, FEBS Lett. 1998, 441, 348–352.
28] M. F. Sanner, J. Mol. Graphics 1999, 17, 57–61.
29] W. Humphrey, A. Dalke, K. Schulten, J. Mol. Graph. 1996, 14, 33–38.
30] J. E. Stone, Masters Thesis, University of Missouri–Rolla (USA), 1998.
31] B. Vilozny, A. Schiller, R. A. Wessling, B. Singaram, Anal. Chim. Acta 2009,
[35] J. Sambrook, D. W. Russell, Molecular Cloning: A Laboratory Manual,
Vols. 1–3, 3rd ed., Cold Spring Harbor Laboratory Press, New York,
2
001.
[
36] D. B. Cordes, S. Gamsey, Z. Sharrett, A. Miller, P. Thoniyot, R. A. Wessling,
B. Singaram, Langmuir 2005, 21, 6540–6547.
649, 246–251.
[32] S. Ravaud, H. Watzlawick, R. Haser, R. Mattes, N. Aghajari, Acta Crystal-
Received: September 7, 2011
logr. Sect. F Struct. Biol. Cryst. Commun. 2006, 62, 74–76.
Published online on November 29, 2011
1
56
www.chembiochem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2012, 13, 149 – 156