Báo cáo khoa học: Implication for buried polar contacts and ion pairs in hyperthermostable enzymes pot

MINIREVIEW
Implication for buried polar contacts and ion pairs in
hyperthermostable enzymes
Ikuo Matsui and Kazuaki Harata
Biological Information Research Center, National Institute of Advanced Industrial Science and Technology (AIST), Ibaraki, Japan
Introduction
Hyperthermophiles grow optimally at temperatures of
80–110 °C [1]. Only represented by bacterial and
archaeal species, these organisms have been isolated
from all types of terristerial and marine hot environ-
ments. Some of the enzymes from hyperthermophiles
are active at temperatures as high as 110 °C and above
[2]. Recently, much effort has been directed towards
the isolation and characterization of enzymes from
hyperthermophilic archaea. Interest in these enzymes
has increased because of their potential biotechnolo-
gical applications [3,4] and because of the need for a
better understanding of their intrinstic heating
and denaturing resistance. Elucidating the stabilizing
mechanisms has been one of the greatest challenges in
biochemistry and biotechnology [3–5].
This minireview encompasses the molecular determi-
nants of protein stability, and compares the various
molecular structures of amino acid aminotransferase
(AT), b-glycosidases (BG), and a-amylases (AM), from
different origins, including hyperthermophiles, thermo-
philes living at around 65–75 °C, and mesophiles living
at room temperature. The three representative
enzymes, AT, BG, and AM were selected due to the
numerous structural information and stability data
Keywords

accessible surface area; buried polar
contact; hyperthermophilic archaea;
hyperthermostable enzyme; ion pair;
molecular structure; Pyrococcus; subunit
interaction; thermostability
Correspondence
I. Matsui, Biological Information Research
Center, National Institute of Advanced
Industrial Science and Technology (AIST),
Tsukuba, Ibaraki 305, Japan
Fax: +81 29 8616151
Tel: +81 29 8616142
E-mail:
(Received 28 February 2007, accepted
17 May 2007)
doi:10.1111/j.1742-4658.2007.05956.x
Understanding the structural basis of thermostability and thermoactivity,
and their interdependence, is central to the successful future exploitation of
extremophilic enzymes in biotechnology. However, the structural basis of
thermostability is still not fully characterized. Ionizable residues play essen-
tial roles in proteins, modulating protein stability, folding and function.
The dominant roles of the buried polar contacts and ion pairs have been
reviewed by distinguishing between the inside polar contacts and the total
intramolecular polar contacts, and by evaluating their contribution as
molecular determinants for protein stability using various protein structures
from hyperthermophiles, thermophiles and mesophilic organisms. The anal-
ysis revealed that the remarkably increased number of internal polar con-
tacts in a monomeric structure probably play a central role in enhancing
the melting temperature value up to 120 °C for hyperthermophilic enzymes
from the genus Pyrococcus. These results provide a promising contribution

for improving the thermostability of enzymes by modulating buried polar
contacts and ion pairs.
Abbreviations
AM, a-amylase; AT, aminotransferase; BG, b-glycosidase; DERA, 2-deoxy-
D-ribose-5-phosphate aldolase; DSC, differential scanning
calorimetry; GluDH, glutamate dehydrogenase; T
m
, melting temperature.
4012 FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS
available from widely distributing origins, including
hyperthermophiles, thermophiles and mesophiles.
Background: there is no single
dominating factor for the
thermostability of proteins
Many crystal structures of hyperthermophilic enzymes
have been reported, and several factors responsible for
their extreme thermostability have been suggested. It
was proposed that the stability of a protein could be
increased by selected amino acid substitutions that
decrease the configurational entropy of unfolding [6].
Proline reduces the flexibility of the polypeptide chain.
The mutations, Gly fi Xaa or Xaa fi Pro should
decrease the entropy of a protein’s unfolded state and
stabilize the protein. A number of the thermophilic
and hyperthermophilic proteins also use this stabiliza-
tion mechanism [1]. The stabilizing role of a large and
more hydrophobic core was proposed based on experi-
mental evidence obtained using chimeric methanocco-
cal adenylate kinases [7]. The stability properties of the
chimera constructed from the Methanococcus voltae

and Methanococcus jannaschii adenylate kinases indi-
cated that cooperative interaction within the hydro-
phobic protein core plays an integral role in increasing
the thermalstability. Another potential stabilizing fac-
tor is the high packing density of the molecule. A com-
parison of the hyperthermophilic aldehyde ferredoxin
oxidoreductase with the mesophilic aldehyde ferre-
doxin oxidoreductase revealed that the former has a
small solvent-exposed surface area [8]. A comparison
of citrate synthases from a hyperthermophile
(Pyrococcus furiosus), thermophile (Thermoplasma
acidophilum) and mesophilic organism (pig) indicated
that increased compactness of the enzyme might be one
of the major factors required for its thermostability [9].
An increase in the number of ion pairs and hydro-
gen bonds is also important for thermostability. The
crystal structure of glyceroaldehyde-3-phosphate dehy-
drogenase from a thermophile, Thermus aquaticus, has
been compared with that of three glyceroaldehyde-
3-phosphate dehydrogenases of different origins [10].
From the results, a strong correlation between thermo-
stability and the number of hydrogen bonds between
charged side chains and neutral partners was found.
There are two reasons why proteins may use charged-
neutral hydrogen bonds rather than salt links or neu-
tral-neutral hydrogen bonds to stabilize protein [10].
First, the desolvation penalty associated with burying
charged-neutral hydrogen bonds would be less than
that of burying ion pairs because of only one charged
residue being involved. Second, the enthalpic reward

of charged-neutral hydrogen bonds is greater than that
of neutral-neutral hydrogen bonds because of the
charge–dipole interaction. By a structure comparison
of [Fe
3
S
4
]-feredoxin from the hyperthermophilic archa-
eon P. furiosus with those from the thermophile and
mesophile, further significant roles of the hydrogen
bonds on the hyperthermostablity have been reported
[11–13]. The P. furiosus feredoxin structure shows a
higher degree of hydrogen-bond network than other
homologus ferredoxins, and this is believed to be the
main reason for the observed increased thermostability
with a denaturation temperature of over 100 °C.
An increase in the number of ion pairs, especially
in networks, is observed nearly in every hyperthermo-
stable protein [1,14]. The structral comparison of an
O
6
-methylguanine-DNA methyltransferase from a
hyperthermophilic archaeon, Thermococcus kodakara-
ensis, with the mesophilic counterpart from Escherichi-
a coli suggested that four additional buried ion pairs
between a-helices might play a key role in its stabiliz-
ing mechanism [15]. The amino acid residues forming
interhelix ion pairs are buried relatively more often
than those of intrahelix ion pairs because the aver-
aged solvent-accesible surface areas of amino acid res-

idues forming inter- and intrahelix ion pairs are
39.5 A
˚
2
and 98.9 A
˚
2
per residue, respectively. This
suggests the internal location of interhelix ion pairs in
the molecule. The interhelix ion pairs in the interior
of the protein presumably enhance the stability of the
internal packing (tertiary structure). Furthermore, a
stabilizing function has also been proposed for buried
ion pairs in Thermosphaera aggregans BG (BGTa)
[16]. In four different sequences of hyperthermophilic
BGs (BGTa from T. aggregans,BGSs from Sulfolobus
solfataricus,BGPf and b-mannosidase from P. furio-
sus), 28% of the residues are strictly conserved. In the
aligned sequences, a strict conservation is observed
among the residues participating in forming internal
ion pairs; however, only 26% of the surface ion pairs
are conserved, consistent with the average sequence
conservation among these sequences. The homohexa-
meric structure of hyperthermophilic glutamate dehy-
drogenase (GluDHPf)(t
1 ⁄ 2
, 12 h; 100 °C) from
P. furiosus was compared with the mesophilic
GluDHCs (t
1 ⁄ 2

, 30 min; 52 °C) from Clostridium sym-
biosum. The comparison revealed that the hyperther-
mostable enzyme contains a striking series of ion pair
networks on the surface of the protein subunits, and
partially buried at intersubunit and inter-domain
interfaces, not found in the mesophilic counterparts
[1,17–19]. The importance of intersubunit ion pairs to
the structural stability of GluDHTk from a hyperther-
mophile, Thermococcus kodakaraensis, was examined
I. Matsui and K. Harata Strucural elememts providing hyperthermostability
FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS 4013
by site-directed mutagenesis, involving systematic
addition or removal of ion pairs [20]. These results
proved the important role for the intersubunit ion
pairs in stabilizing the GluDHTk molecule. However,
completely different results were reported from a
structural analysis of GluDHPi from a hyperthermo-
phile, Pyrobaculum islandicum [21]. The number of in-
tersubunit ion pairs in the homohexameric GluDHPi
molecule was much smaller than that in GluDHPf or
the mesophilic GluDHCs. These findings suggest that
the major molecular strategy for thermostability may
differ for each hyperthermophilic enzyme [21]. The
significant role of the entropic effect, due to shorter
surface loops, on the thermostability of tryptophan
synthase a-subunit from P. furiosus (a-subunit-Pf) was
reported [22]. The thermostability of the a-subunit-Pf
molecule was examined by differential scanning calo-
rimetry (DSC) and by comparing the molecular struc-
ture with that of a mesophilic a-subunit-St molecule

from Salmonella typhimurium. The DSC data indicated
that the greater stability of the a-subunit-Pf molecule
was not caused by an enthalpic factor. From these
results, it was concluded that hydrophobic interactions
in the protein interior do not contribute to the higher
stability of the a-subunit-Pf molecule. The increased
number of ion pairs, smaller cavity volume, and entro-
pic effects due to a shorter polypeptide chains, are
important in the hyperthermostability of the a-subunit-
Pf molecule.
A comparative analysis of the proteins Thermococ-
cus kodakaraensis KOD ribulose-bisphosphate carboxy-
lase ⁄ oxygenase [23], Thermotoga maritima dihydrofolate
reductase [24] and phosphoribosylanthranilate isomer-
ase [25], and Aeropyrum pernix 2-deoxy-d-ribose-
5-phosphate aldolase (DERA) [26] suggested that
oligomerization of subunits appears to be the factor
responsible for the hyperthermostability. The area of
the subunit–subunit interface in the dimer of the
A. pernix DERA is much larger compared with that of
the E. coli enzyme. Furthermore, the A. pernix DERA
has an additional N-terminal helix that induces the
formation of a characteristic dimer–dimer interface.
These results suggest that the hyperthermostability of
the A. pernix DERA could be enhanced by the forma-
tion of a unique tetrameric structure unlike the dimeric
structure of the mesophilic counterparts (i.e. the E. coli
enzyme) [26].
Hence, protein stability appears to be attributable
to a combination of factors, which are related to

each other and their contribution to vary depending
on the proteins. It is proposed that there is no single
dominating factor for the thermostability of proteins
[14].
Evaluating buried polar contacts and
ion pairs as structural elements related
to the thermal stability
Because intermolecular and intramolecular polar inter-
actions such as hydrogen bonds [11–13] and salt link-
ages [1,14–22], appeared to be major factors that are
responsible for hyperthermostablility, interatomic con-
tacts involving main chain peptide groups and polar
side chain groups of Asp, Glu, Arg, Lys, Asn, Gln,
His, Thr and Ser were counted and classified on their
location inside or outside of the molecule. The number
of these contacts was divided by the total number of
polar contacts and the results was used to evaluate the
rigidity of the core region and the hydrophilic property
of the molecular surface. For oligomeric proteins, in-
termolecular polar contacts between subunits were also
calculated. The solvent-accessible surface area was cal-
culated by the Lee and Richards algorithm (probe
radius 1.6 A
˚
) [27]. Next, the accessible surface area of
a protein molecule was divided by the number of
amino acid residues and used as an indicator to com-
pare the compactness of the protein structure.
Structural elements responsible for
thermostability: increase in molecular

compactness, hydrophilicity of the
molecular surface, and buried polar
contacts
ATs have been widely applied to the large-scale bio-
synthesis of unnatural amino acids, which are in
increasing demand by the pharmaceutical industry for
peptidomimic and other single-enantiomer drugs [28].
An aspartate aminotransferase gene homolog (ORF:
PH1371) was identified by sequencing the genome of
a hyperthermophilic archaeon, Pyrococcus horikoshii
[29,30]. The gene (ArATPh) was expressed in E. coli,
and the product was purified to homogeneity. The
enzyme ArATPh was proven to be an aromatic amino-
transferase [31]. ArATPh is one of the most thermosta-
ble aminotransferases reported to date far, with a
melting temperature (T
m
) of 120 °C. The crystal struc-
ture of ArATPh was determined at a resolution of
2.1 A
˚
[31] and shown in Fig. 1 as protein databank
accession code (PDB ID): 1dju. ArATPh has a homo-
dimeric structure in which each subunit has two
domains similar to other aminotransferases. As shown
in Fig. 1, the ArATPh structure is more compact due to
shortened loops (colored in green) compared to those
observed in thermophilic (Thermus thermophilus,
PDB ID: 1bjw) and mesophilic (E. coli, PDB ID: 1ars).
According to the thermodynamic database of proteins

Strucural elememts providing hyperthermostability I. Matsui and K. Harata
4014 FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS
and mutants (ProTherm; tech.
ac.jp/jouhou/protherm/protherm.html) [32], the highest
T
m
of an enzyme measured directly by DSC was
121.6 °C for cytochrome c3 from Desulfovibrio vulgaris
[33], although the T
m
value of PhCutA1 from P. hori-
koshii was reported recently to be 150 °C [34].
From the ArATPh structure, the accumulated acces-
sible surface area and intermolecular polar contacts at
distances shorter than 3.3 A
˚
were calculated (Table 1).
Inside–inside contacts refer to amino acid residues that
are buried inside the molecule and they are not sol-
vent accessible, whereas surface–surface contacts refer
to residues that are exposed to the solvent, even par-
tially. Such parameters that measure structure features
related to thermal stability were calculated for eight
aminotransferase molecules derived from hyperthermo-
philes, thermophiles and mesophiles, including mam-
mals such as pig and human [31,35–41], and are
summarized in Table 2. The optimal growing tempera-
ture of each living organism, the enzyme name, the
PDB ID, and the T
m

measured by DSC are also
shown in Table 2. The accessible surface area divided
by the total residue number of the dimer was used as a
reference to evaluate molecular compactness. The
occupancy of charged residues in the solvent-accessible
area was useful to evaluate the hydrophilicity of the
molecular surface.
All aminotransferases listed in Table 2 are
homodimers. Their Z score and rmsd values range
between 14.8 and 7.0 A
˚
and between 1.07 and 2.38 A
˚
,
Fig. 1. The crystal structures of ATs and
BGs. The figures were produced using the
program
TURBO-FRODO. (Left) C
a
-tracing of
the hyperthermophilic ArATPh dimer
(PDB ID: 1dju), thermophilic AT dimer (from
Thermus thermophilus, PDB ID: 1bjw), and
mesophilic AT dimer (from Escherichia coli,
PDB ID: 1ars). a-Helices, b-sheets and loops
are colored in pink, blue and green, respec-
tively. The cofactors, pyridoxal 5¢-phosphate
(PLP) molecules covalently binding to the
essential Lys residue, are shown with a
space filling model. (Right) C

a
-tracing of
the hyperthermophilic BGPh molecule
(PDB ID: 1vff) and mesophilic BG (from
Paenibacillus polymyxa, PDB ID: 1bga). The
model is viewed along the axis of the barrel.
a-Helices, b-sheets and loops are colored in
pink, blue and green, respectively.
Table 1. The solvent-accessible surface area and intramolecular
polar contacts less than 3.3 A
˚
of an aromatic amino acid amino-
transferase (ArATPh) as a dimer from an hyperthermophilic archa-
eon, Pyrococcus horikoshii.
Total residues
a
Total accessible
surface area (A
˚
2
)
b
Total buried
residues
748 (100%) 20924.37 304 (40.6%)
Total
The number and ratio of intramolecular polar
contacts less than 3.3 A
˚
Inside–

inside
Inside–
surface
Surface–
surface
Subunit–
subunit
1030 (100%) 188 (18.3%) 471 (45.7%) 371 (36.0%) 46 (4.5%)
a
Total residue number consisting of dimer excluding the disordered
regions of the molecule.
b
The 86.4% of the total surface area is
made up of side chains, which corresponds to 18088.29 A
˚
2
.
I. Matsui and K. Harata Strucural elememts providing hyperthermostability
FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS 4015
Table 2. Comparison of structural similarity, melting temperature, molecular compactness, hydrophilicity of the molecular surface and intermolecular polar contact as the rigidity of the
core region for various ATs of different origins, including hyperthermophiles, thermophile and mesophilic organisms. ND, not determined.
Biological diversity
Optimally
growing
temperature
(°C) Enzyme name
PDB
ID
Structure similarity
a

Melting
temperature
by DSC (°C)
Accessible
surface
area ⁄ residue
number (A
˚
2
)
Occupancy
of charged
residues in
the accesible
surface
area (%)
Intermolecular polar contact less
than 3.3 A
˚
Reference
Homology
(%)
Z
score
rmsd
(A
˚
)
Inside–
inside

(%)
Surface–
surface
(%)
Subunit–
subunit
(%)
Nonsurface
ionic
(%)
Hyperthermophile
Pyrococcus horikoshii 98 Aromatic
amino acid
aminotransferase
1dju 100 20.5 0 120 28.0 73.3 18.3 36.0 4.5 4.3 [1,31]
Pyrococcus horikoshii Human kynurenine
aminitransferase II
homolog
1x0m 27 10.3 1.93 ND 33.0 65.4 13.8 49.0 3.6 2.5 [35]
Thermophile 70–75
Thermus thermophilus Aspartate
aminotransferase
1bjw 43 14.8 1.07 ND 32.5 56.0 10.2 57.7 5.2 0.7 [1,36]
Mesophilic organisms Room
Escherichia coli temperature Aspartate
aminotransferase
1ars 19 7.8 2.11 63 33.7 45.8 11.1 52.7 4.0 1.8 [37,42]
Paracoccus denitrificans Aromatic
amino acid
aminotransferase

1ay4 17 7.0 2.38 ND 33.6 48.1 10.3 51.3 3.2 1.9 [38]
Trypanosoma cruzi Tyrosine
aminotransferase
1bw0 25 12.1 1.63 ND 31.6 52.4 9.0 48.0 3.3 2.1 [39]
Pig Cytosolic aspartate
aminotransferase
1ajs 18 7.1 2.29 ND 32.8 43.9 11.8 56.8 3.8 0.0 [40]
Human Kynurenine
aminotransferase
1w7n 30 11.4 1.77 ND 32.8 45.6 11.7 46.3 4.3 0.6 [41]
a
The homology (%), Z score and rmsd values against the ArATPh molecule were retrieved by protein structure matching in a macromolecular structure database (EMBL-EBI) (http://
www.ebi.ac.uk/msd-srv/ssm/cgi-bin/ssmserver).
Strucural elememts providing hyperthermostability I. Matsui and K. Harata
4016 FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS
respectively, reflecting a high structural similarity to
ArATPh, although the sequence similarity varied
between 43% and 17%. The T
m
of the mesophilic
ATEc from E. coli measured directly by DSC is
approximately 63 °C [42], whereas that of the hyper-
thermophilic ArATPh is 120 °C [31]. A comparison of
the accessible surface area per amino acid residue for
the enzymes (Table 2) shows that the value of ArATPh
is the lowest (28.0 A
˚
2
), suggesting tightest molecular
packing. Another prominent feature of ArATPh is the

largest occupancy of charged residues (Asp, Glu, Lys,
and Arg) at its surface (up to 73.3%), indicating a
hydrophilic molecular surface. Moreover, the fre-
quency of buried polar contacts among all polar con-
tacts in distances less than 3.3 A
˚
in the ArATPh
molecule is highest (18.3%) as shown in Tables 1 and
2. By contrast, the frequency of polar contact on the
surface of ArATPh is the lowest (36.0%) relative to
that of the other enzymes listed in Tables 1 and 2.
Interestingly, the presence of polar contact at the inter-
face between the monomers is essentially the same for
all enzymes tested. With an increase in the optimum
growing temperature of each organism, the molecular
compactness, surface hydrophilicity and ratio of buried
polar contacts of each aminotransferase appears to be
increased.
BGs are a group of enzymes that hydrolyze the
b-glycosidic linkage between carbohydrates or between
a carbohydrate and a noncarbohydrate moiety. The
BGs from the hyperthermophilic archaeon P. horiko-
shii (BGPh) were crystallized in the presence of a neu-
tral surfactant, and the crystal structure was solved at
2.5 A
˚
resolution [43] (Fig. 1). Notably, the major dif-
ference of the amino acid sequence among BGPh,
BGTa from T. aggregans [16], and BGSs from
S. solfataricus [44], is the deletion of more than 50

residues from the BGPh sequence that are assigned to
loops. As shown in Fig. 1, the overall structure of the
hyperthermophilic BGPh (PDB ID: 1vff) looks very
similar to that of mesophilic BGs from Paenibacillus
polymyxa (BGPp) (PDB ID: 1bga). BGPh is a mem-
brane-bound enzyme that is extremely thermostable
and it has been shown to have high affinity for
alkyl-b-glycosides. Hence, this enzyme may be used in
industrial applications for the degradation of sugar
derivatives and in the synthesis of various alkyl-glyco-
sides via transglycosidation or ‘reverse hydrolysis’ [45].
Parameters relevant to thermal stability were calcu-
lated for six BG molecules derived from hyperthermo-
philes, mesophiles and plants, and are summarized in
Table 3 [43–49]. The oligomeric state of the BG mole-
cules listed in Table 3 varies from monomer (BGPh)to
homooctamer (BGPp); however, their Z score and
rmsd values range between 15.3 and 12.4 A
˚
and
between 1.47 and 1.66 A
˚
, respectively. The sequence
similarity varied between 35% and 30%. These data
show that there is a high structural similarity between
BG molecules; the main difference among them being
their oligomeric state. BGPh is more thermostable
than the mesophilic BGPp molecule; t
1 ⁄ 2
of BGPh at

90 °C is 15 h [45], whereas t
1 ⁄ 2
of BGPp at 35 °C has
been reported to be 15 min [46]. The prevalence of
charged residues at the surface of hyperthermophilic
BG is higher than that of the homologus proteins from
mesophilic organisms (BGPh; 56.2%, BGTa; 52.1%);
this suggests that the hyperthermophilic enzyme has a
more hydrophilic surface. The frequency of buried
polar contacts in the BGPh molecule is the highest
(14.4%) as shown in Table 3, whereas the accessible
surface area per amino acid residue showed no promi-
nent difference between hyperthermophilic and meso-
philic BGs. The molecular compactness of BGs was
calculated for the monomeric form regardless of the
oligomeric state, which varied from monomer to
homooctamer: with increasing thermostability, the sur-
face hydrophilicity and the percentage of buried polar
contacts of the BG enzymes was also increased.
AMs catalyze the hydrolysis of a-(1,4)glycosidic
linkages of starch components, glycogen and various
oligosaccharides, and is an important industrial
enzyme. The a-amylase (AMPw) from the hyper-
thermophilic archaeon Pyrococcus woesei, which grows
optimally at 100–103 °C, was crystallized, and the
molecular structure was solved at 2.0 A
˚
resolution
[50]. Many a-amylases have been isolated and charac-
terized from hyperthermophiles to mesophilic organ-

isms [50–58]. The AMs listed in Table 4 are
monomers with known crystal structures. Except for a
distant relative of a-amylase, the aminomaltase from
T. aquaticus, the Z score and rmsd values of the other
AMs range between 13.3 and 10.0 A
˚
and between
1.41 and 2.52 A
˚
, respectively, representing fairly good
structural similarity to AMPw. The sequence similar-
ity of these AMs varied between 33% and 18%
(aminomaltase was excluded from the comparison).
The T
m
value of AMPw measured by DSC is 112 °C
[32], whereas the T
m
values of mesophilic a-amylases
from a fungi, Aspergillus oryzae, and from human
were reported to be 62 °C and 67.4 °C, respectively
[32]. A significant difference was observed in the fre-
quency of internal polar contacts of the AM mole-
cules; the value of the AMPw molecule is the
highest (17.1%) among the AMs listed in Table 4.
Furthermore, with an increase in thermostability, the
ratio of buried polar contacts also increased. How-
ever, the molecular compactness, surface hydrophilicity
I. Matsui and K. Harata Strucural elememts providing hyperthermostability
FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS 4017

Table 3. Comparison of structural similarity, melting temperature, molecular compactness, hydrophilicity of the molecular surface, and intermolecular polar contact as the rigidity of the
core region for various BGs of different origins including hyperthermophiles and mesophilic organisms. ND, not determined.
Biological diversity
Optimally
growing
temperature
(°C)
Enzyme name
(oligomeric
state)
PDB
ID
Structure similarity
a
Half-life
t
1 ⁄ 2
(°C)
Accessible
surface
area ⁄ residue
number (A
˚
2
)
Occupancy
of charged
residues in
the accesible
surface

area (%)
Intermolecular polar contact
less than 3.3 A
˚
Reference
Homology
(%)
Z
score
rmsd
(A
˚
)
Inside–
inside
(%)
Surface–
surface
(%)
Nonsurface
ionic
(%)
Hyperthermophile
Pyrococcus horikoshii 98 b-Glycosidase
(monomer)
1vff 100 23.5 0 15 h (90 °C) 36.0 56.2 14.4 52.9 1.0 [1,43,45]
Sulfolbus solfataricus 87 b-Galactosidase
(homotetramer)
1gow 34 12.7 1.66 ND 34.6 48.3 12.1 55.4 1.3 [1,44]
Thermosphera

aggregans
85 b-Glycosidase
(homotetramer)
1gvb 35 12.4 1.55 ND 35.6 52.1 12.7 47.2 1.5 [1,16]
Mesophilic organisms Room
Paenibacillus polymixa temperature b-Glucosidase
(homooctamer)
1bga 34 15.3 1.49 15 min (35 °C) 33.4 43.2 11.8 49.4 1.8 [46,47]
Plant
(Sinapis alba)
Myrosinase
(homodimer)
1dwa 31 13.2 1.54 ND 33.0 48.8 11.9 51.7 1.7 [48]
Plant
(Trifolium repens)
Cyanigenic
b-Glucosidase
(homodimer)
1cbg 30 13.6 1.65 ND 31.6 48.3 12.9 52.7 0.9 [49]
a
The homology (%), Z score and rmsd values against the BGPh molecule were retrieved by protein structure matching in a macromolecular structure database (EMBL-EBI) (http://www.
ebi.ac.uk/msd-srv/ssm/cgi-bin/ssmserver).
Strucural elememts providing hyperthermostability I. Matsui and K. Harata
4018 FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS
Table 4. Comparison of structural similarity, melting temperature, molecular compactness, hydrophilicity of the molecular surface, and intermolecular polar contact as the rigidity of the
core region for various AMs of different origins including hyperthermophile, thermophiles, and mesophilic organisms. ND, not determined.
Biological diversity
Optimally
growing
temperature

(°C) Enzyme name
PDB
ID
Structure similarity
a
Melting
temperature
by DSC (°C)
Accessible
surface
area ⁄ residue
number (A
˚
2
)
Occupancy
of charged
residues in
the accesible
surface
area (%)
Intermolecular polar contact less
than 3.3 A
˚
Reference
Homology
(%)
Z
score
rmsd

(A
˚
)
Inside–
inside
(%)
Surface–
surface
(%)
Nonsurface
ionic
(%)
Hyperthermophile
Pyrococcus woesei 100–103 a-Amylase 1mxd 100 24.2 0 112 34.2 30.3 17.1 46.4 0.0 [1,32,50,51]
Thermophile
Thermus aquaticus 70–75 Amylomaltase 1esw 16 6.6 2.60 ND 38.0 51.8 13.7 50.8 0.0 [1,52]
Bacillus
stearothermophilus
65–69 a-Amylase 1hvx 33 12.7 1.92 t
1 ⁄ 2
,50 min
(90 °C)
33.0 25.8 9.9 53.1 0.0 [53,54]
Mesophilic organisms Room
Alkalophilic
Bacullus sp.707
temperature G
6
-producing
amylase

1wp6 30 12.3 1.97 ND 32.6 28.9 12.6 41.7 0.0 [55]
Aspergillus oryzae a-Amylase 6taa 21 10.2 2.14 62 32.3 32.5 9.1 48.6 0.4 [32,56]
Plant (Barley) a-Amylase 1amy 32 13.3 1.41 ND 34.5 42.4 10.5 47.7 0.0 [57]
Human a-Amylase 1b2y 18 10.0 2.52 67.4 33.3 30.6 10.5 47.4 0.0 [32,58]
a
The homology (%), Z score and rmsd values against the AMPw molecule were retrieved by protein structure matching in a macromolecular structure database (EMBL-EBI) (http://
www.ebi.ac.uk/msd-srv/ssm/cgi-bin/ssmserver).
I. Matsui and K. Harata Strucural elememts providing hyperthermostability
FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS 4019
and percentage of polar contact on the surface did
not show a significant trend from the mesophilic to
the hyperthermophilic AM (Table 4).
Conclusion
As shown in Tables 2–4, the most thermostable
enzymes from Pyrococcus species (e.g. ArATPh,
BGPh, and AMPw) belonging to three different pro-
tein families, have the highest rate of buried polar con-
tacts compared to that of their mosophilic and
thermophilic counterparts. In addition, ArATPh has a
much higher rate of buried ion pairs than ATs from
other species. Recent surveys on the exposure of ioniz-
able groups to solvent [59], ion pairs [60], and the des-
olvation energy of these residues [61] using the protein
structure database, show that more than 30% of the
ionizable residues are fully or partially buried and ion-
ized in the internal of the molecule [62]. Buried ionized
residues appear to be more conserved than those on
the surface [62]. Here, the dominant roles of the buried
polar contacts and ion pairs were reviewed by distin-
guishing between the inside polar contacts and the

total intramolecular polar contacts, and by evaluating
their contribution as molecular determinants for pro-
tein stability using various protein structures from
hyperthermophiles, thermophiles and mesophilic
organisms. Although more abundant data for the
structure ⁄ stability relationships of various proteins
other than the representatives, AT, BG and AM, if
available, should be considered, the results reported
suggest strategies for improving the thermostability of
enzymes by modulating the internal polar contacts and
ion pairs.
Acknowledgements
We thank Hideshi Yokoyama at University of Shi-
zuoka, School of Pharmaceutical Science, and Eriko
Matsui for their valuable advice and discussion.
References
1 Vieille C & Zeikus JG (2001) Hyperthermophilic
enzymes: sources, uses, and molecular mechanisms for
thermostability. Microbiol Mol Biol Rev 65, 1–43.
2 Vieille C, Burdette DS & Zeikus JG (1996) Thermo-
zymes. Biotechnol Annu Rev 2, 1–83.
3 Cowan DA, Daniel RM & Morgan HW (1985) Thermo-
philic proteases. Trends Biotechnol 3, 68–72.
4 Cowan DA (1992) Biotechnology of the archaea. Trends
Biotechnol 10, 315–332.
5 Adams MW, Perler FB & Kelly RM (1995) Extremo-
zymes: expanding the limits of biocatalysis. Biotechnol-
ogy 13, 662–668.
6 Matthews BW, Nicholson H & Becktel WJ (1987)
Enhanced protein thermostability from site-directed

mutations that decrease the entropy of unfolding. Proc
Natl Acad Sci USA 84, 6663–6667.
7 Haney PJ, Stees M & Konisky J (1999) Analysis of
thermal stabilizing interactions in mesophilic and ther-
mophilic adenylate kinases from the genus
Methanococcus. J Biol Chem 274, 28453–28458.
8 Chan MK, Mukund S, Kletzin A, Adams MW & Rees
DC (1995) Structure of a hyperthermophilic tungstop-
terin enzyme, aldehyde ferredoxin oxidoreductase.
Science 267, 1463–1469.
9 Russell RJ, Ferguson JM, Hough DW, Danson MJ &
Taylor GL (1997) The crystal structure of citrate syn-
thase from the hyperthermophilic archaeon Pyrococcus
furiosus at 1.9 A
˚
resolution. Biochemistry 36, 9983–9994.
10 Tanner JJ, Hecht RM & Krause KL (1996) Determi-
nants of enzyme thermostability observed in the
molecular structure of Thermus aquaticus d-glyceral-
dehyde-3-phosphate dehydrogenase at 25 Angstroms
resolution. Biochemistry 35, 2597–2609.
11 Nielsen MS, Harris P, Ooi BL & Christensen HEM
(2004) The 1.5 A
˚
resolution crystal structure of [Fe
3
S
4
]-
feredoxin from the hyperthermophilic archaeon

Pyrococcus furiosus. Biochemistry 43, 5188–5194.
12 Macedo-Ribeiro S, Darimont B, Sterner R & Huber R
(1996) Small structure changes account for the high
thermostability of 1[4Fe-4S] ferredoxin from the hyper-
thermophilic bacterium Thermotoga maritima. Structure
4, 1291–1301.
13 Pfeil W, Gesierich U, Kleemann GR & Sterner R (1997)
Ferredoxin from the hyperthermophile Thermotoga
maritima is stable beyond the boiling point of water.
J Mol Biol 272, 591–596.
14 Petsko GA (2001) Structural bases of thermostability in
hyperthermophilic proteins, or ‘there’s more than one
way to skin a cat’. Methods Enzymol 334, 469–478.
15 Hashimoto H, Inoue T, Nishioka M, Fujiwara S, Tak-
agi M, Imanaka T & Kai Y (1999) Hyperthermostable
protein structure maintained by intra and inter-helix
ion-pairs in archaeal O
6
-methylguanine-DNA methyl-
transferase. J Mol Biol 292, 707–716.
16 Chi YI, Martinez-Cruz LA, Jancarik J, Swanson RV,
Robertson DE & Kim SH (1999) Crystal structure
of the b-glycosidase from the hyperthermophile
Thermosphaera aggregans: insights into its activity and
thermostability. FEBS Lett 445, 375–383.
17 Yip KSP, Stillman TJ, Britton KL, Artymiuk PJ,
Baker PJ, Sedelnikova SE, Engel PC, Pasquo A, Chi-
araluce R, Consalvi V et al. (1995) The structure of
Pyrococcus furiosus glutamate dehydrogenase reveals a
Strucural elememts providing hyperthermostability I. Matsui and K. Harata

4020 FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS
key role for ion-pair networks in maintaining enzyme
stability at extreme temperatures. Structure 3, 1147–
1158.
18 Rice DW, Yip KSP, Stillman TJ, Britton KL, Fuentes
A, Connerton I, Pasquo A, Scandurra R & Engel PC
(1996) Insights into the molecular basis of thermal sta-
bility from the structure determination of Pyrococcus
furiosus glutamate dehydrogenase. FEMS Microbiol Rev
18, 105–117.
19 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
stability from the analysis of ion-pair networks in the
glutamate dehydrogenase family. Eur J Biochem 255,
336–346.
20 Rahman RNZA, Fujiwara S, Nakamura H, Takagi M
& Imanaka T (1998) Ion pairs involved in maintaining
a thermostable structure of glutamate dehydrogenase
from a hyperthermophilic archaeon. Biochem Biophys
Res Commun 248, 920–926.
21 Bhuiya MW, Sakuraba H, Ohshima T, Imagawa T,
Katsunuma N & Tsuge H (2005) The first crystal struc-
ture of hyperthermostable NAD-dependent glutamate
dehydrogenase from Pyrobaculum islandicum. J Mol Biol
345, 325–337.
22 Yamagata Y, Ogasahara K, Hioki Y, Lee SJ, Nakaga-
wa A, Nakamura H, Ishida M, Kuramitsu S & Yutani
K (2001) Entropic stabilization of the tryptophan syn-
thase alpha-subunit from a hyperthermophile, Pyro-

coccus furiosus. X-ray analysis and calorimetry. J Biol
Chem 276, 11062–11071.
23 Maeda N, Kanai T, Atomi H & Imanaka T (2002) The
unique pentagonal structure of an archaeal Rubisco is
essential for its high thermostability. J Biol Chem 277,
31656–31662.
24 Dams T, Auerbach G, Bader G, Jacob U, Ploom T,
Huber R & Jaenicke R (2000) The crystal structure of
dihydrofolate reductase from Thermotoga maritima:
molecular features of thermostability. J Mol Biol 297,
659–672.
25 Thoma R, Hennig M, Sterner R & Kirschner K
(2000) Structure and function of mutationally gener-
ated monomers of dimeric phosphoribosylanthranilate
isomerase from Thermotoga maritima. Structure 8,
265–276.
26 Sakuraba H, Tsuge H, Shimoya I, Kawakami R, Goda
S, Kawarabayasi Y, Katsunuma N, Ago H, Miyano M
& Ohshima T (2003) The first crystal structure of archaeal
aldolase. Unique tetrameric structure of 2-deoxy-
D-ribose-5-phosphate aldolase from the hyperthermo-
philic archaea Aeropyrum pernix. J Biol Chem 278,
10799–10806.
27 Lee B & Richards FM (1971) The interpretation of pro-
tein structures: estimation of static accessibility. J Mol
Biol 55, 379–400.
28 Taylor PP, Pantaleone DP, Senkpeil RF & Fothering-
ham lan G (1998) Novel biosynthetic approaches to the
production of unnatural amino acids using transaminas-
es. Trends Biotechnol 16, 412–418.

29 Kawarabayasi Y, Sawada M, Horikoshi H, Haikawa Y,
Hino Y, Yamamoto S, Sekine M, Baba S, Kosugi H,
Hosoyama A et al. (1998) Complete sequence and gene
organization of the genome of a hyper-thermophilic
archaebacterium, Pyrococcus horikoshii OT3. DNA Res
5, 55–76.
30 Kawarabayasi Y, Sawada M, Horikoshi H, Haikawa Y,
Hino Y, Yamamoto S, Sekine M, Baba S, Kosugi H,
Hosoyama A et al. (1998) Complete sequence and gene
organization of the genome of a hyper-thermophilic
archaebacterium, Pyrococcus horikoshii OT3 (supple-
ment). DNA Res 5, 147–155.
31 Matsui I, Matsui E, Sakai Y, Kikuchi H, Kawarabayashi
Y, Ura H, Kawaguchi S, Kuramitsu S & Harata K (2000)
The Molecular structure of hyperthermostable aromatic
aminotransferase with novel substrate specificity from
Pyrococcus horikoshii. J Biol Chem 275, 4871–4879.
32 Abdulla Bava K, Gromiha MM, Uedaira H, Kitajima
K & Sarai A (2004) Protherm, version 4.0: thermody-
natic database for proteins and mutants. Nucleic Acids
Res 32, D120–D121.
33 Dolla A, Arnoux P, Protasevich I, Lobachov V, Brugna
M, Giudici-Orticoni MT, Haser R, Czjzek M, Makarov
A & Bruschi M (1999) Key role of phenylalanine 20 in
cytochrome C3: structure, stability, and functionstudies.
Biochemistry 38, 33–41.
34 Tanaka T, Sawano M, Ogasahara K, Sakaguchi Y,
Bagautdinov B, Katoh E, Kuroishi C, Shinkai A,
Yokoyama S & Yutani K (2006) Hyper-thermostability
of CutA1 protein, with a denaturation temperature of

nearly 150 °C. FEBS Lett 580, 4224–4230.
35 Chon H, Matsumura H, Koga Y, Takano K & Kanaya
S (2005) Crystal structure of a human kynurenine ami-
notransferase II homologue from Pyrococcus horikoshii
OT3 at 2.20 A
˚
resolution. Proteins 61, 685–688.
36 Nakai T, Okada K, Akutsu S, Miyahara I, Kawaguchi
S, Kato R, Kuramitsu S & Hirotsu K (1999) Structure
of Thermus thermophilus HB8 aspartate aminotransfer-
ase and its complex with maleate. Biochemistry 38,
2413–2424.
37 Okamoto A, Higuchi T, Hirotsu K, Kuramitsu S &
Kagamiyama H (1994) X-ray crystallographic study of
pyridoxal 5¢-phosphate-type aspartate aminotransferases
from Escherichia coli in open and closed form. J Bio-
chem 116, 95–107.
38 Okamoto A, Nakai Y, Hayashi H, Hirotsu K &
Kagamiyama H (1998) Crystal structures of Para-
coccus denitrificans aromatic amino acid aminotrans-
ferase: a substrate recognition site constructed by
rearrangement of hydrogen bond network. J Mol Biol
280, 443–461.
I. Matsui and K. Harata Strucural elememts providing hyperthermostability
FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS 4021
39 Blankenfeldt W, Nowicki C, Montemartini-Kalisz M,
Kalisz HM & Hecht HJ (1999) Crystal structure of
Trypanosoma cruzi tyrosine aminotransferase: substrate
specificity is influenced by cofactor binding mode.
Protein Sci 8, 2406–2417.

40 Rhee S, Silva MM, Hyde CC, Rogers PH, Metzler CM,
Metzler DE & Arnone A (1997) Refinement and
comparisons of the crystal structures of pig cytosolic
aspartate aminotransferase and its complex with
2-methylaspartate. J Biol Chem 272, 17293–17302.
41 Rossi F, Han Q, Li J, Li J & Rizzi M (2004) Crystal
structure of human kynurenine aminotransferase I.
J Biol Chem 279, 50214–50220.
42 Gloss LM, Planas A & Kirsch JF (1992) Contribution
to catalysis and stability of the five cysteines in Escheri-
chia coli aspartate aminotransferase. Preparation and
properties of a cysteine-free enzyme. Biochemistry 31,
32–39.
43 Akiba T, Nishio M, Matsui I & Harata K (2004) X-ray
structure of a membrane-bound b-glycosidase from the
hyperthermophilic archaeon Pyrococcus horikoshii.
Proteins 57, 422–431.
44 Aguilar CF, Sanderson I, Moracci M, Ciaramella M,
Nucci R, Rossi M & Pearl LH (1997) Crystal structure
of the b-glycosidase from the hyperthermophilic archeon
Sulfolobus solfataricus: resilience as a key factor in ther-
mostability. J Mol Biol 271, 789–802.
45 Matsui I, Sakai Y, Matsui E, Kikuchi H, Kawarabayasi
Y & Honda K (2000) Novel substrate specificity of a
membrane-bound b-glycosidase from the hyperthermo-
philic archaeon Pyrococcus horikoshii . FEBS Lett 467,
195–200.
46 Gonzalez-Blasco G, Sanz-Aparicio J, Gonzalez B,
Hermoso JA & Polaina J (2000) Directed evolution of
b-glucosidase A from Paenibacillus polymyxa to thermal

resistance. J Biol Chem 275, 13708–13712.
47 Sanz-Aparicio J, Hermoso JA, Martinez-Ripoll M,
Lequerica JL & Polaina J (1998) Crystal structure of
b-glucosidase A from Bacillus polymyxa: insights into
the catalytic activity in family 1 glycosyl hydrolases.
J Mol Biol 275, 491–502.
48 Burmeister WP (2000) Structural changes in a cryo-
cooled protein crystal owing to radiation damage. Acta
Crystallogr Sect D 56, 328–341.
49 Barrett T, Suresh CG, Tolley SP, Dodson EJ & Hughes
MA (1995) The crystal structure of a cyanogenic b-glu-
cosidase from white clover, a family 1 glycosyl hydro-
lase. Structure 3, 951–960.
50 Linden A, Mayans O, Meyer-Klaucke W, Antranikian
G & Wilmanns M (2003) Differential regulation of a
hyperthermophilic a-amylase with a novel (Ca,Zn) two-
metal center by zinc. J Biol Chem 278, 9875–9884.
51 Dong G, Vieille C, Savchenko A & Zeikus JG (1997)
Cloning, sequencing, and expression of the gene encod-
ing extracellular a
-amylase from Pyrococcus furiosus
and biochemical characterization of the recombinant
enzyme. Appl Environ Microbiol 63, 3569–3576.
52 Przylas I, Terada Y, Fujii K, Takaha T, Saenger W &
Strater N (2000) X-ray structure of acarbose bound to
amylomaltase from Thermus aquaticus. Implications for
the synthesis of large cyclic glucans. Eur J Biochem 267,
6903–6913.
53 Suvd D, Fujimoto Z, Takase K, Matsumura M &
Mizuno H (2001) Crystal structure of Bacillus

stearothermophilus a-amylase: possible factors deter-
mining the thermostability. J Biol Chem 129, 461–468.
54 Tomazic SJ & Klibanov AM (1988) Mechanisms of irre-
versible thermal inactivation of Bacillus a-amylases.
J Biol Chem 263, 3086–3091.
55 Kanai R, Haga K, Akiba T, Yamane K & Harata K
(2004) Biochemical and crystallographic analyses of
maltohexaose-producing amylase from alkalophilic
Bacillus sp. 707. Biochemistry 43, 14047–14056.
56 Swift HJ, Brady L, Derewenda ZS, Dodson EJ, Dodson
GG, Turkenburg JP & Wilkinson AJ (1991) Structure
and molecular model refinement of Aspergillus oryzae
(TAKA) a-amylase: an application of the simulated-
annealing method. Acta Crystallogr Sect B 47, 535–544.
57 Kadziola A, Abe J, Svensson B & Haser R (1994) Crys-
tal and molecular structure of barley a-amylase. J Mol
Biol 239, 104–121.
58 Qian M, Haser R, Buisson G, Duee E & Payan F
(1994) The active center of a mammalian a-amylase.
Structure of the complex of a pancreatic a-amylase with
a carbohydrate inhibitor refined to 2.2-A
˚
resolution.
Biochemistry 33, 6284–6294.
59 Kajander T, Kahn PC, Passila SH, Cohen DC, Lehtio
L, Adolfsen W, Warwicker J, Schell U & Goldman A
(2000) Buried charged surface in proteins. Structure 8,
1203–1214.
60 Kumar S & Nussinov R (1999) Salt bridge stability in
monomeric proteins. J Mol Biol 293, 1241–1255.

61 Gunner MR, Saleh MA, Cross E, ud-Doula A & Wise
M (2000) Backbone dipoles generate positive potentials
in all proteins: origins and implications of the effect.
Biophys J 78, 1126–1144.
62 Kim J, Mao J & Gunner MR (2005) Are acidic and
basic groups in buried proteins predicted to be ionized?
J Mol Biol 348, 1283–1298.
Strucural elememts providing hyperthermostability I. Matsui and K. Harata
4022 FEBS Journal 274 (2007) 4012–4022 ª 2007 The Authors Journal compilation ª 2007 FEBS