Báo cáo khoa học: Enhancing thermostability of maltogenic amylase from Bacillus thermoalkalophilus ET2 by DNA shuffling pdf

Enhancing thermostability of maltogenic amylase from
Bacillus thermoalkalophilus ET2 by DNA shuffling
Shuang-Yan Tang
1
, Quang-Tri Le
1
, Jae-Hoon Shim
1
, Sung-Jae Yang
1
, Joong-Huck Auh
1
,
Cheonseok Park
2
and Kwan-Hwa Park
1
1 Center for Agricultural Biomaterials, and Department of Food Science and Biotechnology, School of Agricultural Biotechnology, Seoul
National University, South Korea
2 Department of Food Science and Biotechnology and Institute of Life Sciences and Resources, Kyunghee University, Yongin, South Korea
In recent years, many thermostable amylases have been
cloned from thermophilic archaea and bacteria [1–4].
The physical basis for the protein stability at high tem-
peratures has also been widely investigated [5,6]. Ther-
mostability appears to be conferred by a variety of
strategies. Studies of thermostability have been carried
out by comparing the atomic structure of a thermophi-
lic protein with its mesophilic homologues [7–9] or
by comparing between thermophilic and mesophilic
genomic sequences in a large scale [10–13]. It is
believed that surface electrostatics, amino acid compo-

sition, shorter loops, increased charged dipole on heli-
ces, and interactions between cations and aromatic
rings (cation–p interaction) are important overall
factors for increased thermostability of proteins. Many
different forces and interactions contribute to
thermostability. The challenge is to pinpoint which of
the different factors and interactions are the most crit-
ical for the specific proteins [14]. Studies of Bacillus
licheniformis a-amylase (BLA) revealed that the ther-
mostability of this highly thermostable enzyme could
be further enhanced through better hydrophobic pack-
ing and cavity-filling effects, introduction of favorable
aromatic–aromatic interactions on the surface,
improved formation of secondary structures and
release of conformational strain, stabilization of an
intrinsic metal binding site, and removal of possible
deamidating residues [15].
Site-directed mutagenesis has been consistently used
to enhance thermostability of various enzymes based
on sequence comparison between thermophilic and
mesophilic counterparts or on tertiary structural
Keywords
DNA shuffling; maltogenic amylase
(MAase); neopullulanase; site-directed
mutagenesis; thermostability
Correspondence
K H. Park, Department of Food Science and
Biotechnology, Seoul National University,
Shillim-dong, Kwanak-gu, Seoul 151–742,
Korea

Fax: +82 2 8735095
Tel: +82 2 8804854 ⁄ 8804852
E-mail:
(Received 28 March 2006, revised 22 May
2006, accepted 24 May 2006)
doi:10.1111/j.1742-4658.2006.05337.x
DNA shuffling was used to improve the thermostability of maltogenic
amylase from Bacillus thermoalkalophilus ET2. Two highly thermostable
mutants, III-1 and III-2, were generated after three rounds of shuffling and
recombination of mutations. Their optimal reaction temperatures were all
80 °C, which was 10 °C higher than that of the wild-type. The mutant
enzyme III-1 carried seven mutations: N147D, F195L, N263S, D311G,
A344V, F397S, and N508D. The half-life of III-1 was about 20 times
greater than that of the wild-type at 78 °C. The mutant enzyme III-2 car-
ried M375T in addition to the mutations in III-1, which was responsible
for the decrease in specific activity. The half-life of III-2 was 568 min while
that of the wild-type was <1 min at 80 °C. The melting temperatures of
III-1 and III-2, as determined by differential scanning calorimetry,
increased by 6.1 °C and 11.4 °C, respectively. Hydrogen bonding, hydro-
phobic interaction, electrostatic interaction, proper packing, and deamida-
tion were predicted as the mechanisms for the enhancement of
thermostability in the enzymes with the mutations.
Abbreviations
BLA, Bacillus licheniformis a-amylase; BTMA, MAase from Bacillus thermoalkalophilus ET2; b-CD, b-cyclodextrin; DSC, differential scanning
calorimetry; MAase, maltogenic amylase.
FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS 3335
information [16,17]. Compared with site-directed
mutagenesis, which is a more conventional approach,
directed evolution allows us to explore enzyme thermo-
stability for which the molecular basis is poorly under-

stood. DNA shuffling is an evolutionary protocol
wherein cycles of selection, recombination, mutation,
and amplification are employed to evolve DNA
sequences and corresponding protein structures. It is a
powerful molecular evolution technology which
enables in vitro generation of large libraries of chimeric
and mutated hybrids genes. In addition to recombina-
tion, DNA shuffling also introduces point mutations at
a controlled rate, which broadens the possibilities for
evolving improved genes [18–21]. DNA shuffling has
been used to improve protein properties such as stabil-
ity, catalytic activity, and substrate specificity [6,22,23].
Maltogenic amylase (EC 3.2.1.133, MAase) is a
multifunctional enzyme able to catalyze both hydro-
lysis and transglycosylation activities. MAase shares
similar catalytic properties with neopullulanases (EC
3.2.1.135) and cyclomaltodextrinases (EC 3.2.1.54),
and together these enzymes constitute a subfamily
in the glycoside hydrolase family 13. The specific
sequence in the fifth conserved region distinguishes this
subfamily from the oligo-1,6-glucosidase subfamily
[24]. It has been suggested that these three enzymes are
nearly the same enzymes in terms of their structures
and catalytic properties [25]. They are able to hydro-
lyze multiple carbohydrate substrates including starch,
cyclodextrin, and pullulan, with the main hydrolysis
product of starch and cyclodextrin being maltose while
that of pullulan is panose. They are also capable of
simultaneously transferring the hydrolyzed sugar moi-
ety to another sugar acceptor molecule [26].

The most widely used thermostable enzymes are the
amylases in the starch industry [27,28]. Thermostable
amylases have such an extremely large application in
the starch industry, as solubilized starch is a better
substrate for amylases and starch gelatinization only
occurs at high temperatures. Amylases from fungi and
bacteria have dominated the applications in industry
[29]. In addition to starch processing, for some high-
value compounds with poor solubility, maltogenic
amylase can transfer sugar residues to them to increase
their solubilities [30]. In these cases, thermostable mal-
togenic amylase can work at high temperatures under
which the solubility of substrates is better.
In this study, we generated a thermostable mutant
of MAase from Bacillus thermoalkalophilus ET2
(BTMA), which had an optimal reaction temperature
of 80 °C, 10 °C higher than that of the wild-type, and
had 85% activity left at 85 °C. The half-life of the
mutant was about threefold greater than that of the
mutant of maltogenic amylase from Thermus strain
(ThMA) at 85 °C. The hypothetical mechanism of
each mutation contributing to the thermostability was
predicted based on the modeled tertiary structures of
the mutant enzymes.
Results
Screening of mutants with improved optimal
reaction temperature and thermostability
Under each round at the screening temperature, the
parental enzyme showed only a very faint halo on the
starch-containing plate when stained with iodine solu-

tion. Mutants that showed a larger and stronger halo
were selected, and the optimal reaction temperatures
and half-lives of thermal inactivation of their purified
enzymes were examined. Mutants with higher optimal
reaction temperatures or longer half-lives than the par-
ental enzyme were selected and changes in their DNA
sequences were investigated.
Approximately 12 000 colonies were screened after
heat treatment at 90 °C for 60 min in the first round
of shuffling reaction. Two mutant enzymes, I-1–79
and I-5–49, whose optimal reaction temperatures were
75 °C, corresponding to 5 °C higher than the wild-type
BTMA, were selected. Their half-lives were 1.7- and
three-fold longer than that of the wild-type at 75 °C
(Table 1). I-1–79 contained N147D and I-5–49 con-
tained F397S according to DNA sequencing analysis
(Fig. 1). These mutant enzymes were used as parental
enzymes for the second round of shuffling.
Approximately 8000 colonies were screened after heat
treatment at 93 °C for 45 min and II-1–56, which has
an optimal reaction temperature of 80 °C and a 1.3-fold
longer half-life compared with I-5–49 at 75 °C, was
selected (Table 1). I95M, S550P, and A569V were found
to be introduced in addition to N147D and F397S
(Fig. 1). Through recombination using restriction
enzymes, II (N147D, F397S), II-1–56A (N147D, F397S,
I95M), and II-1–56B (N147D, F397S, S550P, A569V)
were constructed from II-1–56. The optimal reaction
temperatures of II, II-1–56A, and II-1–56B were all
80 °C, and no significant differences in their thermosta-

bilities were detected (Table 1). From these results, it
was therefore thought that the combination of N147D
and F397S mutations mainly contributed to the increase
in thermal resistance of II-1–56.
The mutant II was used as the parental enzyme for
the third round of shuffling. In this round, about 8000
colonies were screened after heat treatment at 96 °C
for 30 min, and four mutants with longer half-
lives than that of II (III-1–18, III-1–19, III-4–85, and
Enhancing MAase thermostability by DNA shuffling S Y. Tang et al.
3336 FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS
III-3–13) were selected. Their half-lives at 75 °C were
11.7-, 4.6-, 4.0-, and 1.8-fold longer than that of the
parental enzyme, respectively (Table 1). Two point
mutations (Q343H and M375T) in III-1–18, four point
mutations (Y18C, V51I, F195L, and N263S) in III-1–
19, two mutations (D311G and A344V) in III-4–85,
and two mutations (N508D and L549S) in III-3–13
were newly introduced mutations in addition to the
existing mutations in the parental gene (Fig. 1).
To investigate the contribution of the mutations
generated in the third round shuffling mutants, various
additional mutants were constructed by general recom-
bination using restriction enzymes or by site-directed
mutation. Two mutants were constructed from III-4–
85: mutants III-4–85A and III-4–85B carried D311G
and A344V, respectively, in addition to N147D and
F397S. Both mutants had shorter half-lives than III-4–
85, but had 2.9- and 2.0-fold longer half-lives than the
parental enzyme II, suggesting both D311G and

A344V contributed to the thermostability to some
extent (Table 1).
Mutations in III-1–18A and III-1–18B were Q343H
and M375T, respectively, in addition to N147D and
F397S. It was found that the half-life of III-1–18B at
75 °C was 56.02 ± 4.49 min, roughly two times higher
than that of III-1–18 (29.41 ± 2.41 min); however,
III-1–18A showed very short half-life (1.06 ±
0.02 min) compared with III-1–18 (Table 1). These
results suggested that M375T positively contributed to
the thermostability, whereas Q343H had little effect.
Interestingly, the specific activity of III-1–18 and III-1–
18B significantly decreased to about 13% of the wild-
type level, indicating that M375T negatively affected
the enzyme activity of BTMA, even though it stabil-
ized the enzyme at high temperature.
Four mutations (Y18C, V51I, F195L, and N263S)
added in III-1–19 were separated into two mutants,
III-1–19A (Y18C, V51I, N147D, F397S) and III-1–19B
(F195L, N263S, N147D, F397S). The thermostability
of III-1–19B was similar to III-1–19, while that of
III-1–19A was analogous to the parental enzyme II
(Table 1). To examine the contribution of F195L and
N263S further, III-1–19B1 (F195L, N147D, F397S)
and III-1–19B2 (N263S, N147D, F397S) were con-
structed. Although III-1–19B1 and III-1–19B2 had
shorter half-lives than III-1–19B, they were 3.8- and
3.0-fold longer than that of II, respectively, suggesting
that F195L and N263S both individually contributed
to the thermostability of BTMA (Table 1).

Two mutants were constructed from III-3–13: III-3–
13A (N508D, N147D, F397S) and III-3–13B (L549S,
N147D, F397S). The thermostability of III-3–13A was
comparable to that of III-3–13, while that of III-3–13B
Wild-type
N147D F397S
N147D, F397S
N147D, F397S
D311G* A344V*
N147D, F397S
Y18C* V51I*
F195L* N263S*
N147D, F397S
N508D* L549S*
N147D, F397S
Q343H* M375T*
N147D, F397S
D311G, A344V,
F195L , N263S
N508D
N147D, F397S
D311G, A344V,
F195L , N263S
N508D, M375T
III-4-85
III-1-19
III-3-13 III-1-18
II
I-1-79 I-5-49
III-1

III-2
N147D, F397S
I95M* S550P*
A569V*
II-1-56
1st round
2nd round
3rd round
Fig. 1. Lineage of BTMA shuffling mutants. Newly introduced
mutations in each generation are marked with asterisks. Positive
mutations are highlighted in bold.
Table 1. Thermostabilities of BTMA wild-type and mutants
obtained from DNA shuffling.
Enzymes Mutations t
1 ⁄ 2
(min)
a
Wild-type 0.63 ± 0.10
I-1–79 N147D 1.06 ± 0.14
I-5–49 F397S 1.86 ± 0.18
II-1–56 N147D, F397S, I95M, S550P, A569V 2.34 ± 0.08
II N147D, F397S 2.51 ± 0.20
II-1–56 A N147D, F397S, I95M 2.46 ± 0.16
II-1–56B N147D, F397S, S550P, A569V 2.43 ± 0.17
III-1–18 N147D, F397S, Q343H, M375T 29.41 ± 2.41
III-1–18 A N147D, F397S, Q343H 1.06 ± 0.02
III-1–18B N147D, F397S, M375T 56.02 ± 4.49
III-1–19 N147D, F195L, Y18C, V51I,
N263S, F397S
11.43 ± 0.40

III-1–19 A N147D, F397S, Y18C, V51I 1.76 ± 0.15
III-1–19B N147D, F397S, F195L, N263S 11.09 ± 0.69
III-1–19B1 N147D, F397S, F195L 9.47 ± 0.37
III-1–19B2 N147D, F397S, N263S 7.54 ± 0.36
III-3–13 N147D, F397S, N508D, L549S 4.63 ± 0.38
III-3–13 A N147D, F397S, N508D 4.43 ± 0.52
III-3–13B N147D, F397S, L549S 2.01 ± 0.14
III-4–85 N147D, F397S, D311G, A344V 10.05 ± 0.45
III-4–85 A N147D, F397S, D311G 7.17 ± 0.41
III-4–85B N147D, F397S, A344V 4.96 ± 0.41
a
t
1 ⁄ 2
, half-life, determined with 0.04 mgÆmL
)1
protein concentration
at 75 °C.
S Y. Tang et al. Enhancing MAase thermostability by DNA shuffling
FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS 3337
was analogous to the parental enzyme II. Conse-
quently, we suspected that N508D was responsible for
the increase of thermostability in III-3–13 (Table 1).
Finally, two mutants with all positively affecting
mutations were made by general recombination using
restriction enzymes and site-directed mutation. Mutant
III-2 carried all positive mutations found in each shuf-
fling mutant, whereas mutant III-1 included all muta-
tions except M375T, which negatively affects the
specific activity of the enzyme. As expected, the specific
activity of III-1 was similar to that of the wild-type

and other shuffling mutants, except III-1–18; however,
the specific activity of III-2 decreased to 20% of the
specific activity of the wild-type (Table 2). The optimal
reaction temperature of both mutants was 80 °C. The
half-lives of III-1 and III-2 observed at 78, 80, and
85 °C were all longer than those of other shuffling
mutants and wild-type BTMA. The half-life of III-1
was 20 times greater than that of the wild-type at
78 °C. The thermostability of III-2 was much higher
than that of III-1. The half-life of III-2 was 568 min
while that of wild-type was less than 1 min at 80 °C.
When the thermostability of the shuffling mutants of
BTMA was compared with that of the ThMA mutant,
the half-life of III-2 was about three times longer than
that of the ThMA shuffling mutant at 85 °C (Table 2).
T
m
of BTMA mutants
The melting temperatures (T
m
) of BTMA wild-type
enzyme and various mutants, II, III-1, and III-2,
obtained from shuffling reactions were determined by
differential scanning calorimetry (DSC). A variety of
shuffling mutants exhibited higher T
m
than the wild-
type. The T
m
of II was 3.5 °C higher than that of the

wild-type, while the T
m
values of III-1 and III-2 were
6.1 and 11.4 °C higher than that of wild-type, respect-
ively. The enthalpy change at T
m
also increased when
the T
m
increased (Table 3, Fig. 2), which showed that
higher temperatures and more energy were needed for
the denaturation of the mutants.
Predicting thermostabilization mechanisms
The homology-based models of the BTMA wild-type
and shuffling mutants were constructed by the swiss-
model program [31] with reference to the known
Table 2. Specific activities and thermostabilities of BTMA wild-type and mutants.
Enzymes
Specific activity
(UÆmg
)1
)
t
1 ⁄ 2
(min)
a
78 80 85
Wild-type 452.9 1.94 ± 0.07 ND ND
II 484.0 14.31 ± 0.69 3.57 ± 0.25 ND
III-3–13 A 481.3 22.17 ± 1.36 3.90 ± 0.13 ND

III-4–85 474.1 30.39 ± 2.98 6.62 ± 0.36 ND
III-1–19B 474.3 31.84 ± 2.29 7.41 ± 0.16 ND
III-1–18B 64.6 ND 157.02 ± 4.11 6.50 ± 0.17
III-1 484.3 37.29 ± 2.90 8.35 ± 0.35 ND
III-2 84.4 ND 567.78 ± 67.94 12.45 ± 0.47
ThMA-DM (6) 20.0 ND ND 4.47 ± 0.46
a
t
1 ⁄ 2
, half-life, determined with 0.12 mgÆmL
)1
protein concentration; ND, not determined.
Table 3. Differential scanning calorimetry analysis of BTMA wild-
type and mutants.
Enzymes T
m
(°C) Enthalpy (10
4
kJÆmol
)1
)
Wild-type 76.68 132.9
II 80.22 148.6
III-1 82.76 167.3
III-2 88.10 171.6
Temperature (
o
C)
60 70 80 90
/

elo
m/lack(p
C
o
)C
-2.0e+4
0.0
2.0e+4
4.0e+4
6.0e+4
8.0e+4
1.0e+5
1.2e+5
1.4e+5
1.6e+5
wild-type
II
III-1
III-2
Fig. 2. Differential scanning calorimetry analysis of BTMA wild-type
and mutants. Wild-type and mutant BTMAs were concentrated to
1mgÆmL
)1
in 50 mM glycine–NaOH buffer (pH 8.0) using a Microcon
filter (Millipore Co., Billerica, MA, USA). The samples were scanned
at temperatures from 40 to 110 °C with a scan rate of 1°CÆmin
)1
.
Enhancing MAase thermostability by DNA shuffling S Y. Tang et al.
3338 FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS

structure of wild-type ThMA [32] and neopullulanases
from Bacillus stearothermophilus [33], whose amino
acid sequences were 64 and 70% identical to BTMA,
respectively.
In hyperthermophilic proteins, there is a reduction
in the number of glutamines and asparagines, as these
two amino acids are easily deamidated at elevated
temperatures [34–36]. Examination of amino acid
compositions of the total protein content of eight
mesophilic and seven hyperthermophilic microorgan-
isms, for which genome sequences data are available,
revealed a 55% reduction in the number of glutam-
ines and a 28% reduction in asparagines in hyper-
thermophilic proteins [37]. N263S might contribute to
thermostability for this reason. Deamidation of N263
and N261, N264 surrounding it could lead to a pack-
ing of negative charges and a destabilization of the
native structure (Fig. 3A). When the asparagine was
mutated to serine, however, the hydroxyl group of
serine could form a hydrogen bond with the lysine at
position 268, while in the wild-type, this hydrogen
bond could not form between K268 and N263
(Fig. 3A).
It has been shown that improved calcium binding
could help the thermostability of the enzyme in many
cases, for example, a-amylase from B. licheniformis
[38] and Bacillus amyloliquefaciens [39], and influenza
virus neuraminidase [40]. N147 was one of the residues
in the calcium-binding site in BTMA (Fig. 3B). After
N147D substitution, the negatively charged aspartate

could interact better with a calcium ion than aspara-
gine and therefore enhance the thermostability of the
enzyme.
The peptide chain between V490 and I509 was flex-
ible, so when N508 was mutated to aspartate, the neg-
atively charged aspartate might adjust its direction and
interact with the positively charged arginine at position
26 (Fig. 3C). The newly formed electrostatic interac-
tion might improve the interaction between the
N-domain and C-domain and subsequently contribute
to the thermostability of the enzyme.
D311 was located at a-helix 3 in the (a ⁄ b)
8
domain.
The surrounding amino acids, Y308, L309, L310,
V312, A313, and Y315, were all hydrophobic amino
acids, whereas aspartate is hydrophilic. The mutation
of aspartate to glycine enhanced the hydrophobicity of
this area, and subsequently enhanced the thermostabil-
ity of the enzyme (Fig. 4A). The same mechanism was
predicted for A344V, on which many hydrophobic
amino acids (L310, I317, F337, F341, and I348) were
surrounded, forming a strong hydrophobic area. A
substitution of alanine with valine made this area more
hydrophobic (Fig. 4B).
N261
N264
N263S
K268
2.75Å

A
N153
G172
D174
N147D
N149
B
C
R26
N508D
N-domain
C-domain
Fig. 3. (A) Predicted thermostabilization mechanism with N263S
mutation. Deamidation of N261, N263 and N264 could lead to a
crowding of negative charges. S263 could form a hydrogen bond
with K268. (B) Calcium-binding site in BTMA involving N147D muta-
tion. The negatively charged aspartate could improve the calcium
binding. (C) Predicted electrostatic interaction between the N- and
C-domains of BTMA involving N508D mutation.
S Y. Tang et al. Enhancing MAase thermostability by DNA shuffling
FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS 3339
F195 is a residue buried inside the enzyme, and a
hydrophobic core was formed, with the hydrophobic
amino acids I178, I198, L241, V239, and I193 sur-
rounding it. The size of phenylalanine is much larger
than those of the other amino acids (isoleucine, leu-
cine, and valine), possibly decreasing the ease with
which they could be packed. The mutation of phenyl-
alanine to leucine, an amino acid smaller in size, might
therefore help in the packing of the hydrophobic core

and subsequently enhance the thermostability of the
enzyme (Fig. 5A). By computing the total occurrences
of all possible mutations in 16 families, Gromiha et al.
[41] gave the differences in preference of residue substi-
tutions from mesophilic to thermophilic proteins. The
sixth strongest preference of replacement was observed
for the mutation F fi L, as the steric hindrance of
phenylalanine might make it less preferable than leu-
cine in thermostable proteins.
The M375T mutation was located in the middle of
the barrel at the C-terminal of b-strand 6, close to the
catalytic center of the enzyme (Fig. 5B). This mutation
not only enhanced the thermostability dramatically but
also caused a decrease in specific activity of the enzyme
(Tables 1 and 2). D328, E357, and D424 were
A
Y315
A313
V312
Y308
L309
L310
D311G
B
I317
I348
A344V
F341
L310
F337

Fig. 4. (A) Predicted hydrophobic interactions with D311G mutation.
D311G enhanced the hydrophobicity of the area formed by Y308,
L309, L310, V312, A313, and Y315. (B) Predicted hydrophobic inter-
actions with A344V mutation. A344V enhanced the hydrophobicity
of the area surrounded by L310, I317, F337, F341, and I348.
I178
I193
F195L
V239
I198
L241
A
B
D424
E357
D328
M375T
I358
3.5Å
Fig. 5. (A) Predicted packing with F195L mutation. F195L might
help the packing of the hydrophobic core formed by I178, I198,
L241, V239, and I193. (B) Predicted thermostabilization mechanism
with M375T mutation. The small size of threonine and the newly
formed hydrogen bond between it and I358 could make the pack-
ing of this area tighter, which could be responsible for the
enhanced thermostability and decreased catalytic activity. The cata-
lytic residues at the active site are shown.
Enhancing MAase thermostability by DNA shuffling S Y. Tang et al.
3340 FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS
identified as the catalytic residues of the enzyme [32].

As the size of sulfur atom is larger than other atoms,
the bulky side chain of methionine at position 375
might assist the overall folding of the active site and
provide enough space and flexibility for the catalytic
activity. When methionine was mutated to the relat-
ively smaller residue, threonine, the conformation
around the active site might have changed so that the
hydroxyl of threonine might form a hydrogen bond
with the nitrogen atom of I358. The small size of thre-
onine and the newly formed hydrogen bond could
make the packing of this area tighter, which could be
responsible for the enhanced thermostability and
decreased catalytic activity.
Discussion
With three rounds of DNA shuffling method and
recombination of mutations, two types of thermostable
mutants of maltogenic amylase from B. thermoalkalo-
philus ET2 were obtained. While the specific activity of
one type of mutant remained unchanged relative to the
wild-type enzyme, the other type of mutant decreased
its catalytic capacity. In total, seven amino acid residue
substitutions were found to enhance the thermostabili-
ty of the enzyme without affecting the activity while
one mutation caused the specific activity to decrease,
but at the same time improved thermostability.
Hydrogen bonding is one of the major mechanisms
for protein thermostability. Substantial free energies of
interaction have been observed in enzyme–ligand inter-
actions, even in the presence of water. In Fersht’s
experiments involving mutations in tyrosyl-tRNA syn-

thetase, which compared ligand affinities as reflected
by k
cat
⁄ K
m
, hydrogen bonds between neutral partners
typically seemed to have free energies of formation
of 0.5 to 1.8 kcal, and hydrogen bonds involving
one charged partner attain effective values of 3.5–
4.5 kcalÆmol
)1
[42]. Even at the solvent-accessible
molecular surface, intramolecular hydrogen bonding is
favorable. In a mutational analysis of the protein bar-
nase, it was found that a side chain to a backbone
hydrogen bond at the N-terminus of either helix stabil-
ized the protein by up to 2.5 kcalÆmol
)1
relative to a
nonhydrogen-bonding residue in the corresponding
position [43]. In this study, M375T and N263S substi-
tutions generated new hydrogen bonds that did not
exist in the wild-type. In M375T, the hydrogen bond
was created with neutral partners while in N263S, it
was established with one charged partner.
Hydrophobic interaction is another powerful strat-
egy for protein thermostability. Removal of nonpolar
molecules or groups from water is accompanied by an
increase in entropy and a compensating uptake of heat
from the surroundings. Hydrophobic bonds are thus

distinguished by a tendency to become stronger with
increasing temperature [44]. In free energy contribu-
tions due to hydrophobic, electrostatic, hydrogen
bonding, disulfide bonding, and van der Waals interac-
tions, hydrophobic free energy due to carbon and
nitrogen atoms and such combinations of free energy
components play a vital role in the thermostability of
proteins [45]. In many cases, the correlation between
hydrophobic interactions and protein stability has been
demonstrated [5,46–49]. In the cases of D311G and
A344V, it was thought that the increased hydrophobic-
ity of the environment after the substitutions contribu-
ted to the enhanced thermostability of the enzyme.
Electrostatic interactions, such as salt bridges and
their networks, have important roles in protein folding
[50,51]. A correlation between salt bridge networks
and melting temperature has been observed for the
d-glyceraldehyde-3-phosphate dehydrogenase family
[52]. A similar correlation was also observed between
salt bridge networks and thermostabilities of glutamate
dehydrogenases from several hyperthermophilic, ther-
mophilic, and mesophilic sources [53]. Grimsley et al.
[54] have shown that the stability of RNase T1 can be
increased by improving long-range electrostatic interac-
tions among charged groups on the protein surface. In
this study, N508D might belong to this category, in
which the electrostatic interaction between aspartate
and arginine could tighten the interaction between the
N- and C-domains of the enzyme.
Packing is also an undeniable factor in protein sta-

bilization. The molecular interior of a folded protein
should be well packed; otherwise dispersion forces
would favor denaturation, since efficient packing can
presumably be achieved between protein and solvent in
the denatured state [55]. We postulate that the
improved thermostability by F195L was due to the
better packing of the protein after the residue substitu-
tion.
Experimental procedures
Bacterial strains and plasmid
The vector p6xHTKXb119, which carried the B. lichenifor-
mis maltogenic amylase promoter and N-terminal hexahisti-
dine tag, was used for over-expression of the six-His tagged
wild-type and mutant of MAase gene from B. thermoalkalo-
philus ET2 [6,56]. Escherichia coli MC1061 [F
–
, araD139,
recA13, D(araABC-leu)7696, galU, galK, DlacX74, rpsL,
thi, hsdR2, mcrB] was used as the host. The E. coli trans-
formants were cultured at 37 °C in Luria–Bertani medium
S Y. Tang et al. Enhancing MAase thermostability by DNA shuffling
FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS 3341
[LB; 1% (w ⁄ v) Bacto–tryptone, 0.5% (w ⁄ v) yeast extract,
and 0.5% (w ⁄ v) NaCl].
DNA shuffling for random mutagenesis of BTMA
DNA shuffling was carried out according to the method of
Zhao and Arnold [18]. The wild-type or mutant BTMA gene
on p6xHTKXb119 vector was isolated by digestion with
XbaI and HindIII and degraded randomly by DNaseI inreac-
tion buffer at 15 °C for 7 min. The fragments smaller than

300 bp were isolated from 2% (w ⁄ v) agarose gel and used
for self-assembly PCR with Taq polymerase (Takara Bio
Inc., Shiga, Japan). The thermocycling reaction was carried
out as follows: one initial denaturation step at 94 °C for
3 min and 40 cycles of denaturation at 94 °C for 1 min,
annealing at 55 °C for 1 min, and extension at 72 ° C for
1 min, with one final extension step at 72 °C for 7 min. To
isolate the products of full-length self-assembly, 3 lL of the
self-assembled DNA mixture was amplified using BTMA1,
the 5¢-flanking primer (5¢-caccattctagaatgttgaaagaagccatt
tatca-3¢), and BTMA2, the 3¢-flanking primer (5¢-gcgagag
gaagcttccttctcgcctttta-3¢). The PCR was conducted using the
conditions mentioned above. PCR products of full-length
assembly were purified, digested with XbaI and HindIII, and
ligated into p6xHTKXb119.
Screening of mutants with enhanced
thermostability
The transformants were picked onto LB agar plates con-
taining kanamycin (20 lgÆmL
)1
) and incubated at 37 °C for
12 h. The cells were then transferred onto nylon membranes
(Hybond-N; Amersham Pharmacia Biotech, Uppsala, Swe-
den), lysed, and fixed according to the method described by
Song and Rhee [57]. The membrane was incubated under
the appropriate conditions for each DNA shuffling round
(specific screening conditions for each round are described
in Results). The heat-treated membranes were placed on
fresh LB plates that contained 1% (w ⁄ v) soluble starch
(Showa Manufacturing Co., Fukuoka, Japan) and incuba-

ted at 60 °C for 12 h. Finally, mutants showing larger clear
zones than the parental enzyme when stained with an iod-
ine solution (0.203 g I
2
and 5.2 g KI in 100 mL H
2
O) were
selected. The mutations were verified by DNA sequencing
using the BigDye terminator cycle sequencing kit for the
ABI 377 Prism (Applied Biosystems, Foster City, CA,
USA).
Enzyme purification and activity assays
Wild-type BTMA and various mutant proteins tagged with
six-histidine residues were purified from E. coli harboring
the corresponding genes on p6xHTKXb119 using a nickel–
nitrilotriacetic acid (Ni–NTA) column (Qiagen Inc., Valen-
cia, CA, USA) as described previously [2]. The purity of
the enzyme was determined by SDS ⁄ PAGE [10%] as
described by Laemmli [58]. The hydrolytic activities of
wild-type and mutant enzymes were assayed in 0.5% (w ⁄ v)
b-cyclodextrin (b-CD) (Sigma Chemical Co., St Louis, MO,
USA) in 50 mm glycine–NaOH buffer (pH 8.0) according
to the 3,5-dinitrosalicylic acid method using maltose as the
standard [59]. One unit of enzyme activity was defined as
the amount of enzyme that produced 1 lmol of maltose per
minute. The protein concentration was measured using the
Bradford method [60] with bovine serum albumin (Sigma
Chemical Co.) as the standard.
Analysis of the thermostability of enzymes
The thermostability of enzymes was measured by determin-

ing the half-life of thermal inactivation and the melting
temperature (T
m
). Enzyme samples at certain protein con-
centrations were incubated in water baths at different tem-
peratures (75, 78, 80, and 85 °C). Aliquots were taken at
various time points and placed immediately in an ice bath.
The residual b-CD hydrolyzing activities of the aliquots
were measured at optimal reaction temperatures. The first-
order rate constant, k
d
, of irreversible thermal denaturation
was obtained from the slope of the plots of ln(residual
activity) vs. time, and the half-life was calculated as ln2 ⁄ k
d
.
To determine the T
m
of the mutant enzymes, DSC was
performed with a VP-DSC MicroCalorimeter (MicroCal,
Northampton, MA, USA). Wild-type and mutant BTMAs
(1 mgÆmL
)1
)in50mm glycine–NaOH buffer (pH 8.0) were
heated from 40 to 110 °C at a heating rate of 1 °C min
)1
using 50 mm glycine–NaOH buffer (pH 8.0) as a reference.
T
m
and enthalpy were calculated with microcal origin 5.0

software.
Acknowledgements
This work was supported by a grant from the Korea
Health 21 R & D Project, Ministry of Health & Wel-
fare, Korea (A050376) and partly supported by the
Korea Research Foundation Grant funded by the
Korean Government (KRF-2004–005-F00054). We
acknowledge the financial support of the Korea
Research Foundation Grant (320095) in the form of a
scholarship to S Y. Tang.
References
1 Jorgensen S, Vorgias CE & Antranikian G (1997) Clon-
ing, sequencing, characterization and expression of an
extracellular a-amylase from the hyperthermophilic
archaeon Pyrococcus furiosus. Escherichia coli and Bacil-
lus subtilis. J Biol Chem 272, 16335–16342.
Enhancing MAase thermostability by DNA shuffling S Y. Tang et al.
3342 FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS
2 Kim TJ, Kim MJ, Kim BC, Kim JC, Cheong TK, Kim
JW & Park KH (1999) Modes of action of acarbose
hydrolysis and transglycosylation catalyzed by a ther-
mostable maltogenic amylase, the gene for which was
cloned from a Thermus strain. Appl Environ Microbiol
65, 1644–1651.
3 Rose NN, Li D, Park JT, Cha HJ, Kim MS, Kim JW
& Park KH (2005) Characterization and application of
a novel thermostable glucoamylase cloned from a
hyperthermophilic archaeon Sulfolobus tokodaii. Food
Sci Biotechnol 14, 860–865.
4 Yang SJ, Lee HS, Park CS, Kim YR, Moon TW &

Park KH (2004) Enzymatic analysis of an amylolytic
enzyme from the hyperthermophilic archaeon Pyrococ-
cus furiosus reveals its novel catalytic properties as both
an a-amylase and a cyclodextrin-hydrolyzing enzyme.
Appl Environ Microbiol 70, 5988–5995.
5 Declerck N, Machius M, Chambert R, Wiegand G,
Huber R & Gaillardin C (1997) Hyperthermostable
mutants of Bacillus licheniformis a-amylase: thermody-
namic studies and structural interpretation. Protein Eng
10, 541–549.
6 Kim YW, Choi JH, Kim JW, Park C, Kim JW, Cha
HJ, Oh BH, Moon TW & Park KH (2003) Directed
evolution of Thermus maltogenic amylase toward
enhanced thermal resistance. Appl Environ Microbiol 69,
4866–4874.
7 Davies GJ, Gamblin SJ, Littlechild JA & Watson HC
(1993) The structure of a thermally stable 3-phosphogly-
cerate kinase and a comparison with its mesophilic
equivalent. Proteins 15, 283–289.
8 Wallon G, Kryger G, Lovett ST, Oshima T, Ringe D &
Petsko GA (1997) Crystal structures of Escherichia coli
and Salmonella typhimurium 3-isopropylmalate dehydro-
genase and comparison with their thermophilic counter-
part from Thermus thermophilus. J Mol Biol 266, 1016–
1031.
9 Haney PJ, Stees M & Konisky J (1999) Analysis of
thermal stabilizing interactions in mesophilic and ther-
mophilic adenylate kinases from the genus Methanococ-
cus. J Biol Chem 274, 28453–28458.
10 Thompson MJ & Eisenberg D (1999) Transpro-

teomic evidence of a loop-deletion mechanism for
enhancing protein thermostability. J Mol Biol 290,
595–604.
11 Chakravarty S & Varadarajan R (2000) Elucidation of
determinants of protein stability through genome
sequence analysis. FEBS Lett 470, 65–69.
12 Cambillau C & Claverie JM (2000) Structural and geno-
mic correlates of hyperthermostability. J Biol Chem 275,
32383–32386.
13 Fukuchi S & Nishikawa K (2001) Protein surface
amino acid compositions distinctively differ between
thermophilic and mesophilic bacteria. J Mol Biol 309,
835–843.
14 Petsko GA (2001) Structural basis of thermostability in
hyperthermophilic proteins, or ‘there’s more than one
way to skin a cat’. Methods Enzymol 334, 469–478.
15 Declerck N, Machius M, Joyet P, Wiegand G, Huber R
& Gaillardin C (2002) Engineering the thermostability
of Bacillus licheniformis a-amylase. Biologia 57 (Suppl.
11), 203–211.
16 Declerck N, Machius M, Joyet P, Wiegand G, Huber R
& Gaillardin C (2003) Hyperthermostabilization of
Bacillus licheniformis a-amylase and modulation of its
stability over a 50 degrees C temperature range. Protein
Eng 16, 287–293.
17 Gromiha MM, Oobatake M, Kono H, Uedaira H &
Sarai A (1999) Role of structural and sequence informa-
tion in the prediction of protein stability changes: com-
parison between buried and partially buried mutations.
Protein Eng 12, 549–555.

18 Zhao H & Arnold FH (1997) Optimization of DNA
shuffling for high fidelity recombination. Nucleic Acids
Res 25, 1307–1308.
19 Stemmer WPC (1994) Rapid evolution of a protein
In vitro by DNA shuffling. Nature 370, 389–391.
20 Zhang J, Dawes G & Stemmer WPC (1997) Evolution
of a fucosidase from a galactosidase by DNA shuf-
fling and screening. Proc Natl Acad Sci USA 94,
4504–4509.
21 Crameri A, Dawes G, Rodriguez E, Silver S & Stemmer
WPC (1997) Molecular evolution of an arsenate detoxi-
fication pathway by DNA shuffling. Nature Biotech 15,
436–438.
22 Hsu JS, Yang YB, Deng CH, Wei CL, Liaw SH & Tsai
YC (2004) Family shuffling of expandase genes to
enhance substrate specificity for penicillin G. Appl
Environ Microbiol 70, 6257–6263.
23 Suen WC, Zhang N, Xiao L, Madison V & Zaks A
(2004) Improved activity and thermostability of Candida
antarctica lipase B by DNA family shuffling. Protein
Eng Des Sel 17, 133–140.
24 Oslancova A & Janecek S (2002) Oligo-1,6-glucosidase
and neopullulanase enzyme subfamilies from the a-amy-
lase family defined by the fifth conserved sequence
region. Cell Mol Life Sci 59, 1945–1959.
25 Lee HS, Kim MS, Cho HS, Kim JI, Kim TJ, Choi JH,
Park C, Lee HS, Oh BH & Park KH (2002) Cyclomal-
todextrinase, neopullulanase, and maltogenic amylase
are nearly indistinguishable from each other. J Biol
Chem 277, 21891–21897.

26 Cha HJ, Yoon HG, Kim YW, Lee HS, Kim JW,
Kweon KS, Oh BH & Park KH (1998) Molecular and
enzymatic characterization of novel maltogenic amylase
that hydrolyzes and transglycosylates acarbose. Eur J
Biochem 253, 251–262.
27 Crab W & Mitchinson C (1997) Enzymes involved in
the processing of starch to sugars. Trends Biotechnol 15,
349–352.
S Y. Tang et al. Enhancing MAase thermostability by DNA shuffling
FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS 3343
28 Poonam N & Dalel S (1995) Enzyme and microbial sys-
tems involved in starch processing. Enzyme Microb
Technol 17, 770–778.
29 Pandey A, Nigam P, Soccol CR, Soccol VT, Singh D
& Mohan R (2000) Advances in microbial amylases.
Biotechnol Appl Biochem 31, 135–152.
30 Li D, Park SH, Shim JH, Lee HS, Tang SY, Park CS &
Park KH (2004) In vitro enzymatic modification of
puerarin to puerarin glycosides by maltogenic amylase.
Carbohydrate Res 339, 2789–2797.
31 Peitsch MC & Guex N (1997) Swiss-Model and the
Swiss-PDBViewer: an environment for comparative pro-
tein modeling. Electrophoresis 18, 2714–2723.
32 Kim JS, Cha SS, Kim HJ, Kim TJ, Ha NC, Oh ST,
Cho HS, Cho MJ, Kim MJ, Lee HS et al. (1999) Crys-
tal structure of a maltogenic amylase provides insights
into a catalytic versatility. J Biol Chem 274, 26279–
26286.
33 Hondoh H, Kuriki T & Matsuura Y (2003) Three-
dimensional structure and substrate binding of Bacillus

stearothermophilus neopullulanase. J Mol Biol 326, 177–
188.
34 Ahern TJ & Klibanov AM (1985) The mechanisms of
irreversible enzyme inactivation at 100°C. Science 228,
1280–1284.
35 Tomazic SJ & Klibanov AM (1988) Mechanisms of irre-
versible thermal inactivation of Bacillus a-amylases.
J Biol Chem 263, 3086–3091.
36 Tomazic SJ & Klibanov AM (1988) Why is one Bacillus
a-amylase more resistant against irreversible thermoin-
activation than another? J Biol Chem 263, 3092–3096.
37 Vieille C, Epting KL, Kelly RM & Zeikus JG (2001)
Bivalent cations and amino-acid composition contribute
to the thermostability of Bacillus licheniformis xylose
isomerase. Eur J Biochem 268, 6291–6301.
38 Declerck N, Machius M, Wiegand G, Huber R &
Gaillardin C (2000) Probing structural determinants
pecifying high thermostability in Bacillus licheniformis
a-amylase. J Mol Biol 301, 1041–1057.
39 Tanaka A & Hoshino E (2002) Calcium-binding para-
meter of Bacillus amyloliquefaciens a-amylase deter-
mined by inactivation kinetics. Biochem J 364, 635–639.
40 Burmeister WP, Cusack S & Ruigrok RW (1994) Cal-
cium is needed for the thermostability of influenza B
virus neuraminidase. J General Virol 75, 381–388.
41 Gromiha MM, Oobatake M & Sarai A (1999) Impor-
tant amino acid properties for enhanced thermostability
from mesophilic to thermophilic proteins. Biophys Chem
82, 51–67.
42 Fersht AR (1987) The hydrogen bond in molecular

recognition. Trends Biochem Sci 12, 301–304.
43 Serrano L & Fersht AR (1989) Capping and a-helix sta-
bility. Nature 342, 296–299.
44 Nemethy G & Scherega HA (1962) Structure of water
and hydrophobic bonding in proteins. I. A model for
the thermodynamic properties of liquid water. J Chem
Phys 36, 3382–3400.
45 Saraboji K, Gromiha MM & Ponnuswamy MN (2005)
Importance of main-chain hydrophobic free energy to
the stability of thermophilic proteins. Int J Biol Macro-
mol 35, 211–220.
46 Arrizubieta MJ & Polaina J (2000) Increased thermal
resistance and modification of the catalytic properties of
a b-glucosidase by random mutagenesis and in vitro
recombination. J Biol Chem 275, 28843–28848.
47 Macedo-Ribeiro S, Martins BM, Pereira PJ, Buse G,
Huber R & Soulimane T (2001) New insights into the
thermostability of bacterial ferredoxins: high-resolution
crystal structure of the seven-iron ferredoxin from
Thermus thermophilus . J Biol Inorg Chem 6, 663–674.
48 Numata K, Hayashi-Iwasaki Y, Kawaguchi J, Sakurai
M, Moriyama H, Tanaka N & Oshima T (2001) Ther-
mostabilization of a chimeric enzyme by residue substi-
tutions: four amino acid residues in loop regions are
responsible for the thermostability of Thermus thermo-
philus isopropylmalate dehydrogenase. Biochim Biophys
Acta 1545, 174–183.
49 Tanaka S, Igarashi S, Ferri S & Sode K (2005) Increas-
ing stability of water-soluble PQQ glucose dehydrogen-
ase by increasing hydrophobic interaction at dimeric

interface. BMC Biochem 6, 1–6.
50 Kumar S, Ma B, Tsai CJ & Nussinov R (2000) Electro-
static strengths of salt bridges in thermophilic and
mesophilic glutamate dehydrogenase monomers. Pro-
teins 38, 368–383.
51 Kumar S & Nussinov R (1999) Salt bridge stability in
monomeric proteins. J Mol Biol 293, 1241–1255.
52 Tanner JJ, Hecht RM & Krause KL (1996) Determinants
of enzyme thermostability observed in the molecular
structure of Thermus aquaticus d-glyceraldehyde-3-phos-
phate dehydrogenase at 2.5 A
˚
resolution. Biochemistry
35, 2597–2609.
53 Yip KSP, Britton KL, Stillman TJ, Lebbink J, De Vos
WM, Robb FT, Vetriani C, Maeder D & Rice DW
(1998) Insights into the molecular basis of thermal stabi-
lity from the analysis of ion pair networks in the gluta-
mate dehydrogenase family. Eur J Biochem 255, 336–
346.
54 Grimsley GR, Shaw KL, Fee LR, Alston RW, Huyg-
hues-Despointes BM, Thurlkill RL, Scholtz JM &
Pace CN (1999) Increasing protein stability by altering
long-range coulombic interactions. Protein Sci 8,
1843–1849.
55 Rose GD & Wolfenden R (1993) Hydrogen bonding,
hydrophobicity, packing, and protein folding. Ann Rev
Biophys Biomol Str 22, 381–415.
56 Cheong KA, Tang SY, Cheong TK, Cha H, Kim JW &
Park KH (2005) Thermostable and alkalophilic malto-

genic amylases of Bacillus thermoalkalophilus ET2.
Biocatal Biotransform 23, 79–87.
Enhancing MAase thermostability by DNA shuffling S Y. Tang et al.
3344 FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS
57 Song JK & Rhee JS (2000) Simultaneous enhancement
of thermostability and catalytic activity of phospholi-
pase A
1
by evolutionary molecular engineering. Appl
Environ Microbiol 66, 890–894.
58 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.
59 Miller GL (1959) Use of dinitrosalicylic acid reagent for
determination of reducing sugar. Anal Chem 31, 426–
428.
60 Bradford MA (1976) Rapid and sensitive method for
the quantitation of microgram quantities of protein util-
izing the principle of protein-dye binding. Anal Biochem
72, 248–254.
S Y. Tang et al. Enhancing MAase thermostability by DNA shuffling
FEBS Journal 273 (2006) 3335–3345 ª 2006 The Authors Journal compilation ª 2006 FEBS 3345