Kinetic and crystallographic analyses of the catalytic domain of chitinase from Pyrococcus furiosus- the role of conserved residues in the active site

Article abstract of DOI:10.1111/j.1742-4658.2010.07685.x

The hyperthermostable chitinase from the hyperthermophilic archaeon Pyrococcus furiosus has a unique multidomain structure containing two chitin-binding domains and two catalytic domains, and exhibits strong crystalline chitin hydrolyzing activity at high

Full text of DOI:10.1111/j.1742-4658.2010.07685.x

Kinetic and crystallographic analyses of the catalytic
domain of chitinase from Pyrococcus furiosus – the role of
conserved residues in the active site
Hiroaki Tsuji1, Shigenori Nishimura1, Takashi Inui1, Yuji Kado2, Kazuhiko Ishikawa2, Tsutomu
Nakamura2 and Koichi Uegaki2
1 Laboratory of Protein Sciences, Graduate School of Life and Environmental Sciences, Osaka Prefecture University, Japan
2 National Institute of Advanced Industrial Science and Technology, Osaka, Japan
Keywords
The hyperthermostable chitinase from the hyperthermophilic archaeon
Pyrococcus furiosus has a unique multidomain structure containing two chi-
chitinase; crystal structure; DXDXE motif;
glycoside hydrolase family;
tin-binding domains and two catalytic domains, and exhibits strong crystal-
line chitin hydrolyzing activity at high temperature. In order to investigate
the structure–function relationship of this chitinase, we analyzed one of the
Pyrococcus furiosus
Correspondence
catalytic domains (AD2) using mutational and kinetic approaches, and
determined the crystal structure of AD2 complexed with chito-oligosaccha-
ride substrate. Kinetic studies showed that, among the acidic residues in
the signature sequence of family 18 chitinases (DXDXE motif), the second
Asp (D2) and Glu (E) residues play critical roles in the catalysis of archaeal
chitinase. Crystallographic analyses showed that the side-chain of the cata-
lytic proton-donating E residue is restrained into the favorable conformer
for proton donation by a hydrogen bond interaction with the adjacent D2
residue. The comparison of active site conformations of family 18 chitinas-
es provides a new criterion for the subclassification of family 18 chitinase
based on the conformational change of the D2 residue.
S. Nishimura, Laboratory of Protein
Sciences, Graduate School of Life and
Environmental Sciences, Osaka Prefecture
University, 1-1 Gakuencho, Sakai, Osaka
599-8531, Japan
Fax: +81 72 254 9462
Tel: +81 72 254 9462
E-mail: tigers@bioinfo.osakafu-u.ac.jp
K. Uegaki, National Institute of Advanced
Industrial Science and Technology, 1-8-31
Midorigaoka, Ikeda, Osaka 563-8577, Japan
Fax: +81 72 751 8370
Tel: +81 72 751 9526
E-mail: k-uegaki@aist.go.jp
Database
Structural data are available at the Protein
Data Bank under the accession numbers
3A4W (E526A–substrate complex), 3A4X
(D524A–substrate complex) and 3AFB
(D524A apo-form)
(Received 25 February 2010, revised 10
April 2010, accepted 13 April 2010)
doi:10.1111/j.1742-4658.2010.07685.x
Introduction
Chitin, a highly stable homopolysaccharide of b-(1,4)-
linked N-acetyl-d-glucosamine (NAG), is an important
structural component of the shells of insects and
crustaceans, fungal cell walls and the exoskeletons of
Abbreviations
AD, active (catalytic) domain; BcChiA1, chitinase A1 from Bacillus circulans; CcCTS1, chitinase 1 from Coccidioides immitis; ChBD, chitin-
binding domain; GH, glycoside hydrolase; NAG, N-acetyl-b-D-glucosamine; (NAG)n, b-(1,4)-linked oligomers of NAG residue where n = 1–6;
Pf-ChiA, chitinase from Pyrococcus furiosus; PNP-(NAG)2, p-nitrophenyl-chitobiose; ScCTS1, chitinase 1 from Saccharomyces cerevisiae;
SmChiB, chitinase B from Serratia marcescens; TK-ChiA, chitinases A from Thermococcus kodakaraensis.
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS
2683

Archaeal chitinase complexed with substrate
H. Tsuji et al.
arthropods. Chitinases (EC 3.2.1.14) are important
enzymes that hydrolyze chitin into smaller chito-oligo-
saccharide fragments. They are found in a wide range
of organisms, including bacteria, fungi, plants and ani-
mals. The presence of chitinases in such organisms is
closely associated with the physiological roles of their
substrates. For instance, bacteria produce chitinases so
that they can use chitin as a source of carbon and
nitrogen for growth [1–3], whereas chitinases in yeasts
and other fungi are important for autolysis, nutritional
and morphogenetic functions [4,5]. Plant chitinases
play a role as defensive agents against pathogenic fungi
and some parasites by disrupting their cell walls [6–8],
whereas viral chitinases are involved in the pathogene-
sis of host cells. Animal chitinases are involved in die-
tary uptake processes [9]. Human chitinases are
particularly associated with anti-inflammatory effects
against T-helper-2-driven diseases, such as allergic
asthma [10–12].
(CAZy) database, http://www.cazy.org/]. These two
families show no homology in either primary or ter-
tiary structures. Family 19 chitinases are almost
exclusively derived from plants, and have a high
degree of sequence similarity. The catalytic domain of
family 19 chitinases comprises two lobes, each of
which is rich in a-helical structure [16,17]. In con-
trast, family 18 includes chitinases from microbes,
plants and animals, and has a substantial sequence
divergence. In spite of their diverse primary struc-
tures, all the catalytic domains of family 18 chitinases
have a common TIM-barrel (b⁄ a)8-fold [18–23] and
are characterized by a highly conserved signature
sequence (DXDXE motif) on the b4-strand (Fig. 1).
The Glu (E) in this motif acts as the catalytic proton
donor, and the second Asp (D2) is supposed to con-
tribute to the stabilization of the essential distortion
of the substrate [24].
We have reported previously that PF1234 and
PF1233, which are adjacent open reading frames of
the hyperthermophilic archaeon Pyrococcus furiosus
with an interval of 37 bp [25], are homologous to the
first and second halves, respectively, of a chitinase
from Thermococcus kodakaraensis (TK-ChiA) [26]. We
In a classification of glycoside hydrolases (GHs)
based on amino acid sequence similarity, established
by Henrissat and coworkers [13–15], chitinases are
classified into two different families: GH families 18
and 19 [described in the carbohydrate active enzyme
′
′
Fig. 1. Sequence alignment of three family 18 chitinases based on secondary structure similarity. AD2, hevamine from Hevea brasiliensis
and chitinase 1 from Saccharomyces cerevisiae (ScCTS1) are shown. The overall conserved amino acid residues are highlighted in black
boxes. Conserved secondary structure elements are indicated above the sequence alignment. The open diamonds represent the highly con-
served (among family 18 chitinases) DXDXE motif, and the filled circle represents the solvent-exposed tryptophan residue. The alignment
was performed using the MATRAS server (http://biunit.naist.jp/matras/).
2684
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS

H. Tsuji et al.
Archaeal chitinase complexed with substrate
combined them into one gene by a frame shift muta-
tion, and the gene product yielded a recombinant
chitinase (Pf-ChiA) homologous to TK-ChiA. Interest-
ingly, Pf-ChiA effectively hydrolyzed not only colloidal
chitin, but also crystalline chitin [25]. The optimum
temperature of Pf-ChiA for the hydrolysis of crystal-
line chitin was extremely high, measured to be over
90 ꢀC. Recently, the enzymatic degradation of chitin
waste using chitinases has attracted much attention as
an environmentally friendly alternative to conventional
chemical degradation methods, because chitin deriva-
tives provide a diverse range of applications in areas
such as biomedicines, food additives and cosmetics
[27]. Hence Pf-ChiA, which exhibits hyperthermostabil-
ity and high hydrolyzing activity towards crystalline
chitin, is useful as an efficient catalyst for the biocon-
version of chitin into valuable oligosaccharide deriva-
tives for various industrial applications.
respect to the conformational change of the D2 residue
on substrate binding, as well as the overall folding.
Results
Site-directed mutagenesis and enzyme
purification
First, we constructed a number of single point mutants
of AD2 by site-directed mutagenesis. Figure 1 shows
the sequence alignment of three family 18 chitinases.
The side-chains of three residues (Asp522, Asp524 and
Glu526) in the DXDXE motif were mutated into the
corresponding amide (Asn or Gln) and Ala. All the
AD2 mutants (D522N, D522A, D524N, D524A,
E526Q and E526A) were overexpressed in E. coli and
purified by the same procedures as the wild-type
enzyme described previously [29]. Far-UV CD spectra
(200–255 nm) of AD2 wild-type and all mutants at 25,
50 and 85 ꢀC were almost identical (data not shown),
indicating that all the mutant enzymes retained thermo-
stability and similar secondary structures.
Pf-ChiA has a unique multidomain structure con-
taining two chitin-binding domains (ChBD1 and
ChBD2) and two catalytic (active) domains (AD1 and
AD2) [25]. Both catalytic domains belong to GH fam-
ily 18. We have not performed any kinetic or struc-
tural studies of the complete Pf-ChiA because of its
low expression level in Escherichia coli, but have
focused instead on the properties of the individual
domains. We have already determined the structures of
ChBD2 and AD2 by means of NMR spectroscopy and
X-ray crystallography, respectively [23,28]. We found
that the overall structure of AD2 is a TIM-barrel
(b⁄ a)8-fold with a groove-like active site architecture,
which is a typical feature of endo-chitinases.
Kinetic properties of AD2 mutants
Table 1 shows the apparent kinetic constants kcat and
Km for the hydrolysis of p-nitrophenyl-chitobiose
[PNP-(NAG)2] catalyzed by seven enzymes (wild-type,
D522N, D522A, D524N, D524A, E526Q and E526A).
The D522N and D522A mutants retained about 40%
and 20%, respectively, of the wild-type kcat values. The
D524N mutation increased the Km value slightly, and
decreased the kcat value by about 2.7-fold. These kcat
and Km values were comparable with those of Asp522
mutants (D522N and D522A). In contrast, the D524A
mutation affected both kcat and Km values signifi-
cantly, which were 1⁄ 340 and 1 ⁄5 of the wild-type
values, respectively. This mutational change caused
a decrease of about 60-fold in enzymatic efficiency
(kcat ⁄ Km). Replacing Glu526 with Gln and Ala
As with other family 18 chitinases, AD2 contains a
highly conserved DXDXE motif [corresponding to
Asp522(D1)-Ile523-Asp524(D2)-Phe525-Glu526(E)] on
the b4-strand (Fig. 1). In this study, we focused on
these three conserved acidic residues in AD2 and car-
ried out mutational and crystallographic analyses in
order to clarify their catalytic role. Our kinetic study
indicated that D2 and E residues play particularly
important roles in catalysis. By using AD2 D524A and
E526A mutants, whose enzymatic activities have been
greatly depressed, we determined the crystal structures
of these mutants complexed with chito-oligosaccharide
substrate. The results of the kinetic analyses confirmed
that the Glu526 residue has a proton-donating func-
tion like other family 18 chitinases. Asp524 was con-
sidered to act to restrain the side-chain of catalytic
Glu526 into the favorable conformer for proton dona-
tion by hydrogen bond interaction. In addition, by
comparing the structures of AD2 with those of other
family 18 chitinases, we proposed a new criterion for
the subclassification of family 18 chitinases with
Table 1. Kinetic constants of AD2 wild-type and mutants for the
hydrolysis of PNP-(NAG)2. ND, not detected.
Enzyme
kcat (s)1
)
Km (mM)
kcat ⁄ Km (mM)1Æs)1
)
Wild-type
D522N
D522A
D524N
D524A
E526Q
E526A
W664A
6.7 0.4
0.46 0.06
0.54 0.07
0.74 0.06
0.61 0.04
0.09 0.01
0.12 0.01
14.6 2.1
4.3 0.6
2.36 0.09
1.49 0.04
2.47 0.05
0.022 0.001
0.045 0.002
ND
2.0 0.2
4.1 0.3
0.25 0.04
0.38 0.04
0.022 0.002
12.7 2.1
0.0017 0.0003
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS
2685

Archaeal chitinase complexed with substrate
H. Tsuji et al.
Table 2. Data collection and refinement statistics for AD2 E526A
influenced the catalytic activity drastically. The E526Q
mutation caused a reduction of about 130-fold in the
wild-type kcat value, and the E526A mutant abolished
the enzymatic activity. Our kinetic results clearly dem-
onstrate that Asp524 and Glu526 play important roles
in the catalytic mechanism of AD2, whereas Asp522
has only a minor role.
and D524A complexed with substrate.
Protein Data Bank code
Protein
3A4W
3A4X
E526A mutant
D524A
mutant
Derivatization methoda
Diffraction data
Cocrystallization Soaking
Space group
P212121
P212121
Unit cell parameters
˚
a (A)
˚
b (A)
Structural determination of AD2 mutants bound
to chito-oligosaccharide substrate
90.0
92.0
107.5
89.8
91.9
˚
c (A)
107.1
The molecular activity (kcat) of the AD2 E526A and
D524A mutants was much lower than that of the wild-
type (Table 1), and so we expected that these two
mutants would be more suitable for observing the
enzyme–substrate complex without any degradation of
the substrate. We obtained crystals of these mutant
enzymes complexed with chito-oligosaccharide sub-
strate by means of cocrystallization and soaking meth-
ods, respectively, and determined their tertiary
structures. We collected X-ray diffraction data for the
AD2 E526A and D524A mutants and refined them to
Number of observed reflections 596 577
641 671
88 323
Number of unique reflections
b
83 345
˚
Resolution range (A)
30.0–1.80
(1.86–1.80)
100 (100)
9.0 (36.4)
20.2 (5.7)
7.2 (6.8)
10.5
50.0–1.76
(1.79–1.76)
99.4 (90.4)
9.1 (37.4)
13.0 (2.2)
7.3 (4.9)
11.3
Completeness (%)b
Rmerge (%)b,c
I ⁄ r (I)b
Redundancyb
B-factors of data from Wilson
2
plot (A )
˚
Refinement
˚
Resolution range (A)
29.5–1.80
37.8–1.76
˚
resolutions of 1.80 and 1.76 A, respectively. A sum-
d
e
Rcryst (%) ⁄ Rfree (%)
R.m.s. deviations from ideality
15.4 ⁄ 17.4
15.3 ⁄ 17.5
mary of crystallographic data collection and refinement
statistics is given in Table 2.
˚
Bond length (A)
0.011
1.30
0.011
1.43
Bond angle (deg)
Superimposition of the overall (b ⁄a)8-barrel struc-
tures of AD2 wild-type (Protein Data Bank code
2DSK [23]), E526A and D524A mutants gave 300
equivalent Ca coordinates with r.m.s. deviations of
Average of B-factor values
2
All atoms (A )
2
Main-chain (A )
2
Side-chain (A )
2
Substrate (A )
˚
9.3
9.0
˚
8.3
8.1
˚
10.3
8.5
9.6
˚
approximately 0.3 A (Fig. S1). Some small conforma-
˚
12.5
24.6
2
Water (A )
˚
tional differences were observed in the surface loop
region comprising Gly488–Gly492 (a maximum Ca–Ca
˚
distance from the wild-type of 0.81 A). However, these
20.5
R.m.s. DB values
2
Main-chain (A )
2
Side-chain (A )
˚
0.6
2.0
0.6
1.8
˚
minor changes did not affect the overall structural
integrity of these mutants compared with the wild-type
(Fig. S1). Therefore, the significant depression of enzy-
matic activity by the introduction of E526A and
D524A substitutions (Table 1) is not a result of con-
formational changes, but of the removal of negative
charge at these residues.
Ramachandran plot statisticsf
Favored (%)
99.2
0.4
99.2
0.4
Allowed (%)
R.m.s. deviations of the two
monomers in the asymmetric
0.20
0.21
g
˚
unit (A)
a
b
See Experimental procedures. Values in parentheses are for
c
the highest resolution shells. Rmerge = R|I ) <I>| ⁄ RI, where I is
the intensity of observation I and <I> is the mean intensity of the
Conformation of chito-oligosaccharides bound to
the active site cleft
d
reflection. Rcryst = R||Fobs| ) |Fcalc|| ⁄ R|Fobs|, where Fobs and Fcalc
are the observed and calculated structure factor amplitudes,
e
respectively. Rfree was calculated using a randomly selected 5%
For the structural determination of the AD2–substrate
complex, we used a NAG pentamer [(NAG)5] as sub-
strate. In the AD2 E526A mutant, on the surface of
the active site cleft, a clear, connected electron density
corresponding to (NAG)5 was observed into which
each NAG residue could fit. The NAG units in
(NAG)5 are numbered 1–5 from the nonreducing end
towards the reducing end (i.e. NAG1–NAG5). We
observed an electron density corresponding to (NAG)4
of the dataset that was omitted through all stages of refinement.
f
Ramachandran plots were created for all residues other than Gly
g
and Pro. R.m.s. deviations were calculated for 300 Ca atoms of
the two molecules in the asymmetric unit.
in the AD2 D524A mutant. Presumably, this might be
caused by a partial disorder of terminal NAG residues
at the nonreducing end. We fitted the (NAG)4 molecu-
lar model corresponding to NAG2–NAG5 of (NAG)5
2686
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS

H. Tsuji et al.
Archaeal chitinase complexed with substrate
in the E526A mutant into the electron density and
carried out further refinements.
Enzyme–substrate interactions
The final refined 2Fobs ) Fcalc maps of the substrates
bound to AD2 E526A and D524A mutants are illus-
trated in Fig. 2. The conformations of (NAG)5 in the
E526A mutant and (NAG)4 in the D524A mutant were
almost identical, giving all matching atoms with an
˚
r.m.s. deviation of 0.10 A, and they made a sharp turn
at NAG3. Although NAG1, NAG2, NAG4 and
4
NAG5 residues adopted standard C1 chair conforma-
The crystal structures of the AD2 E526A and D524A
mutants complexed with substrate show that a number
of amino acid residues contribute to the binding of the
substrate by hydrogen bonding and ⁄ or hydrophobic
interactions. Using the ligplot program [31], we
investigated the specific interactions between enzymes
and each NAG residue in detail (Table 4). The most
significant enzyme–substrate interactions were localized
in NAG3, whose pyranose ring was distorted into the
‘boat’ conformation. Three residues (Ala490, Asp524
and Asp636) formed hydrogen bond interactions and
six residues (Tyr421, Phe448, Met585, Met587,
Met631 and Trp664) participated in hydrophobic
interactions. We believe that these residues stabilize
the distortion of the NAG3 residue. In these interac-
tions, we particularly focused on the Trp664 residue,
which is located at the bottom of the active site cleft.
The indole ring of Trp664 is hydrophobically stacked
with the pyranose ring of NAG3 (Fig. 2). We con-
ducted kinetic analysis to discover the effect of the
tions, the central NAG3 residue was distorted into the
1,4B boat conformation. In addition, the dihedral
angles of the third glycosidic bond (NAG3–NAG4)
were very different from those of the other glycosidic
bonds (NAG1–NAG2, NAG2–NAG3 and NAG4–
NAG5) (Table 3). This similar distortion and twist of
the bound substrate has been observed previously in
the crystal structure of the bacterial chitinase ChiB
from Serratia marcescens (SmChiB) complexed with
(NAG)5 [24,30]. AD2 causes the distortion and twist-
ing of the substrate, so that the glycosidic oxygen faces
towards the bottom of the deep cleft.
Tyr590
Tyr590
A
Asp636
Asp636
NAG2
NAG1
NAG2
NAG1
NAG3
NAG3
NAG4
NAG5
NAG5
NAG4
Ser425
Asp423
Ser425
Asp423
Trp664
Trp664
Ala526
Ala526
Asp524
Asp524
Asp522
Asp522
Tyr590
Tyr590
B
Asp636
NAG4
Asp636
NAG4
NAG2
NAG3
NAG2
NAG3
(NAG1)
(NAG1)
NAG5
NAG5
Trp664
Trp664
Glu526
Glu526
Ala524
Ala524
Asp522
Asp522
Fig. 2. Stereo figures of the model of the bound substrate in the AD2 E526A mutant (A) and D524A mutant (B). The structures of bound
sugars and the side-chains of three acidic residues in the conserved DXDXE motif are indicated in a stick representation. The mesh repre-
sents 2Fobs ) Fcalc electron density maps contoured at the 1.5r level. Residues involved in hydrogen bond interactions are also shown as
sticks. The broken lines represent hydrogen bond interactions.
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS
2687

Archaeal chitinase complexed with substrate
H. Tsuji et al.
Table 3. Dihedral angles around the glycosidic bonds in the bound
substrates. u is the O5–C1–O4¢–C4¢ angle and w is the C1–O4¢–
C4¢–C5¢ angle, where O4 represents the oxygen of the glycosidic
bond and atoms of the adjacent NAG unit are primed.
remarkable change in the conformation of the Glu526
side-chain. The 2Fobs ) Fcalc electron density of the
Glu526 side-chain in the D524A–substrate complex is
not clear compared with that of the wild-type sub-
strate-free form (Fig. 3A, C). However, the Fobs ) Fcalc
omit map of Glu526 clearly shows two conformers of
this side-chain: the A- and B-form (Fig. 3C). We esti-
mated the occupancy of the side-chain in these two
conformers to be 0.5 : 0.5 using the cns program [32].
In the wild-type substrate-free and D524A–substrate
complex structures (Fig. 3A, C), the positions and ori-
entations of the Glu526 side-chain in the A-form were
almost identical, and the maximum coordinate shift
E526A–substrate
complex
D524A–substrate
complex
Glycosidic bond
u
w
u
w
NAG1–NAG2
NAG2–NAG3
NAG3–NAG4
NAG4–NAG5
)120.5
)78.2
)57.5
)89.7
)162.5
)151.0
)98.0
–
–
)64.1
)54.7
)79.2
)150.8
)87.3
)163.7
)161.4
˚
after superimposition of the two structures was 0.78 A.
W664A mutant (Table 1). This mutation decreased the
wild-type kcat value by 300-fold and increased the
wild-type Km value by 30-fold, reducing kcat ⁄ Km by
about 9000-fold. We confirmed that hydrophobic
stacking by Trp664 is crucial for both catalysis and
substrate binding.
In the B-form, in contrast, the Glu526 side-chain
rotated 55ꢀ around v1 relative to the A-form and was
exposed to the solvent. The two oxygen atoms of the
Glu526 side-chain in the A-form were positioned close
to the proximal glycosidic oxygen atom (O1) at dis-
˚
tances of 3.0 and 3.1 A (Fig. 3C). This indicates that
the hydrolytic reaction occurs at the third b-(1,4)-gly-
cosidic bond between NAG3 and NAG4, and Glu526
acts as a catalytic proton donor. Therefore, AD2 pos-
sesses at least five sugar-binding subsites, )3, )2, )1,
+1, +2, as shown in Fig. 3B.
Characterization of acidic residues in the
conserved DXDXE motif
Figure 3 shows a comparison of the crystal structures
of the AD2 E526A–substrate and D524A–substrate
complexes with that of the wild-type (substrate-free
form) [23], focusing on the highly conserved DXDXE
motif close to the bound substrate. The conformations
of the bound substrate in E526A and D524A are
almost identical, but the D524A mutation resulted in a
Discussion
We performed mutational analyses of three conserved
acidic residues (Asp522, Asp524 and Glu526) in AD2
Table 4. Hydrophobic and hydrogen bond interactions in the AD2 E526A and D524A mutants complexed with substrate.
Hydrophobic interaction
Sugar atom
Hydrogen bond interaction
˚
Distance (A)
Sugar no.
Protein residue
Ala461
Sugar atom
Protein atom
NAG1 ()3)a
C7, C8
N2
O3
O6
O7
2.88
2.72
3.18
2.84
Od2 of Asp423
OH of Ser425
NH of Ala490
NE1 of Trp664
NAG2 ()2)
NAG3 ()1)
C6
Val491
Val677
Ser678
Tyr421
Phe448
Glu526b
Met585
Met587
Met631
Trp664
Ala490
Glu526b
Met585
Pro555
Ser556
Tyr590
C8
C8
C8
O3
O6
O7
O7
2.89
2.64
2.94
2.40
NH of Ala490
C7
Od2 of Asp636
Od1 of Asp524a
Od2 of Asp524a
C1
C7, C8
C1
C6
C5, C6, C7, C8
NAG4 (+1)
NAG5 (+2)
C3
O7
2.63
OH of Tyr590
C4, C5, C6
C6
C7,C8
C8
C5, C6
a
b
In the E526A–substrate complex only. In the D524A–substrate complex only.
2688
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS

H. Tsuji et al.
Archaeal chitinase complexed with substrate
A
B
Fig. 3. Close-up views of the active site in
the AD2 wild-type (A), E526A mutant (B)
and D524A mutant (C). All structures are
drawn from the same direction after super-
imposition. The side-chain structures are
imposed onto a 2Fobs ) Fcalc electron den-
sity map (orange mesh), contoured at 1.2r.
In (C), the blue mesh represents an
C
Fobs ) Fcalc electron density map contoured
at 2.4r in which the Glu526 side-chain has
been excluded from the calculation. The
broken lines represent hydrogen bond inter-
actions. Five subsites ()3, )2, )1, +1, +2)
deduced from the solved structures are also
shown, following the nomenclature system
for sugar-binding subsites in GH [53].
and determined the structure of AD2 catalytic site
mutants, E526A and D524A, complexed with (NAG)5.
To the best of our knowledge, these structures repre-
sent the first examples of an archaeal chitinase com-
plexed with natural chito-oligosaccharide substrate.
So far, the three-dimensional structures of family 18
chitinases have been determined for hevamine from
Hevea brasiliensis [19], chitinase 1 from Saccharo-
myces cerevisiae (ScCTS1) [33], chitinase B from
S. marcescens (SmChiB) [22], chitinase 1 from Coccidi-
oides immitis (CcCTS1) [21] and chitinase A1 from
Bacillus circulans (BcChiA1) [20] (Fig. 4B–F). On the
basis of their structures, family 18 chitinases are sub-
classified into ‘plant-type’ and ‘bacterial-type’ [33,34].
‘Plant-type’ family 18 chitinases (hevamine and
ScCTS1) contain a simple (b⁄ a)8-barrel structure with
a shallow substrate-binding groove (Fig. 4B, C), with
one solvent-exposed tryptophan residue at the –1 sub-
site. As AD2 contains a simple (b ⁄a)8-barrel fold with
an open active site architecture, and has one trypto-
phan residue, Trp664, in the active site groove
(Fig. 4A), archaeal chitinase AD2 belongs to the
‘plant-type’ family 18 chitinases. In contrast, ‘bacterial-
type’ chitinases (SmChiB, CcCTS1 and BcChiA1)
consist of the (b ⁄a)8-barrel embellished with a tightly
associated a⁄ b-insertion domain and several long loops
(Fig. 4D–F), resulting in a deep substrate-binding
groove (cleft). This groove contains a large number of
aromatic residues (Fig. 4D–F) which are thought to
participate in substrate binding [35,36].
the carboxyl group of the Glu526 side-chain is essen-
tial for the enzymatic activity of AD2, and this group
cannot be replaced by a neutral amide group (Table 1).
In addition, the side-chain of Glu526 is located close
to the scissile glycosidic bond (Fig. 3C). These results
confirm that the acidic character of the carboxyl group
of Glu526 has a catalytic proton-donating function as
in other family 18 chitinases. The D524N mutant
retained approximately 40% of the wild-type kcat
value, whereas the D524A mutant retained only 0.3%
(Table 1). Thus, the carboxyl group of Asp524 is not
necessarily indispensable and can be replaced by a neu-
tral amide group for the catalytic activity, implying
that the Asp524 side-chain participates in a hydrogen
bond interaction with the bound substrate or proximal
residues. Indeed, in the substrate-free wild-type struc-
ture, the Asp524 side-chain faces towards catalytic
Glu526, forming a hydrogen bond between the Oe
˚
atom of Glu526 and the Od atom of Asp524 in 2.5 A
(Fig. 3A). Interestingly, in the D524A–substrate com-
plex, an altered conformation of the Glu526 side-chain
(B-form) was observed in addition to the favorable
conformer for proton transfer (A-form) (Fig. 3C). In
the B-form, the shortest distance between the carboxyl
oxygen atoms (Oe) of the Glu526 side-chain and the
˚
scissile glycosidic oxygen atom (O1) is 5.0 A. Accord-
ingly, the B-form structure of the Glu526 side-chain is
believed to be unable to donate a proton. In the sub-
strate-free D524A mutant structure, on the other hand,
only the A-form of catalytic Glu526 was observed
(Table S1 and Fig. S2). The relative position of its
side-chain was almost identical to that of the wild-type
substrate-free form (Fig. 3A and Fig. S2B), despite the
We used a combination of kinetic and crystallo-
graphic approaches to characterize the function of the
DXDXE motif in AD2. Kinetic results showed that
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS
2689

Archaeal chitinase complexed with substrate
H. Tsuji et al.
A
B
C
D
E
F
Fig. 4. Structural comparison of overall (b ⁄ a)8-folds for six family 18 chitinases. The overall structure of the AD2 E526A–substrate complex
(A) is compared with the hevamine–allosamidin complex (Protein Data Bank code 1LLO) (B), ScCTS1–acetazolamide complex (Protein Data
Bank code 2UY4) (C), SmChiB E144Q–(NAG)5 complex (Protein Data Bank code 1E6N) (D), CcCTS1–allosamidin complex (Protein Data Bank
code 1LL4) (E) and BcChiA1 substrate-free form (Protein Data Bank code 1ITX) (F). The b-strands and a-helices are denoted in blue and red,
respectively. Catalytic Glu corresponding to Glu526 (replaced by Ala) in AD2 is shown as cyan carbon atoms. In the SmChiB–(NAG)5 struc-
ture, catalytic Glu144 is replaced by Gln. Solvent-exposed aromatic residues lining the active site groove are shown as yellow carbon atoms.
Substrate (inhibitor) structures are shown as green carbon atoms.
lack of hydrogen bond interaction between mutated
Ala524 and Glu526. This suggests that the A-form is
much more energetically favorable than the B-form,
implying that an alternative B-form structure of cata-
lytic Glu526 is induced by substrate binding onto the
active site. From these results, taken together, we con-
clude that the Glu526 side-chain can adopt two con-
formers (A-form and B-form) in the substrate-bound
form, and Asp524 acts to restrain the Glu526 side-
chain into the A-form by hydrogen bond interaction,
promoting Glu526 to donate a proton to a proximal
glycosidic oxygen atom. The D524A–substrate com-
plex is unique in that the substrate was detected in the
active site of the D524A mutant, which does not com-
pletely lose catalytic activity (Table 1). It is possible
that the conformational diversity of the Glu526 side-
chain observed in this complex reflects movement dur-
ing the catalytic cycle.
For BcChiA1, only the substrate-free form structure is
displayed because no substrate-bound structure is
available (Fig. 5F). The crystallographic studies of
SmChiB and CcCTS1 have demonstrated that catalysis
by these ‘bacterial-type’ chitinases involves a confor-
mational change of the second Asp (D2) in the
DXDXE motif on substrate binding (Fig. 5D, E)
[24,37]. Thus, the D2 residue interacts with catalytic
Glu (E) and the first Asp (D1) in the presence and
absence of the bound substrate, respectively (Fig. 5D,
E). This ‘flip’-like conformational change may also
play an important role in ‘cycling’ the pKa of catalytic
Glu during catalysis [24,38]. The mutation of the D1
residue to Asn (D140N) in SmChiB caused a 500-fold
decrease in activity [39]. On the other hand, in ‘plant-
type’ chitinases (AD2, hevamine and ScCTS1), the
side-chain of the D2 residue always faces towards the
catalytic Glu whether the substrate binds or not
(Fig. 5A–C), and so does not interact directly with the
adjacent D1 residue. For AD2, the shortest distance
between the side-chain atoms of Asp522 (D1) and
Through these structural analyses, we have found a
remarkable difference between ‘plant-type’ and ‘bacte-
rial-type’ family 18 chitinases in the conformational
change of the second Asp (D2) of the conserved
DXDXE motif. Figure 5 focuses on the DXDXE
motifs, comparing each structure in substrate-free
and substrate-bound forms for ‘plant-type’ (AD2,
hevamine, ScCTS1; Fig. 5A–C) and ‘bacterial-type’
(SmChiB, CcCTS1, BcChiA1; Fig. 5D–F) chitinases.
˚
Asp524 (D2) is actually 4.4 A in the wild-type sub-
strate-free structure (Fig. 3A). Kinetic results showed
that the D522A mutant retained approximately 20%
of its wild-type kcat value. These crystallographic and
kinetic results clearly demonstrate that, for ‘plant-type’
chitinase, the carboxyl group of the side-chain of the
2690
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS

H. Tsuji et al.
Archaeal chitinase complexed with substrate
A
B
C
D
E
F
Fig. 5. Structural comparison of active sites for six family 18 chitinases, focusing on the conserved DXDXE motif. The close-up view of the
active site in AD2 (A) is compared with hevamine (Protein Data Bank code 2HVM and 1LLO) (B), ScCTS1 (Protein Data Bank code 2UY2 and
2UY4) (C), SmChiB (Protein Data Bank code 1E15 and 1E6N) (D), CcCTS1 (Protein Data Bank code 1D2K and 1LL4) (E) and BcChiA1 (Protein
Data Bank code 1ITX) (F). Each diagram is the superimposition of ligand (substrate or inhibitor)-free and ligand-bound structures. In (F), only
the ligand-free structure is shown because no ligand-bound structure is available. The side-chains of three DXDXE acidic residues in ligand-
free and ligand-bound forms are shown as yellow and cyan carbon atoms, respectively. Ligand structures are shown as green carbon atoms.
Hydrogen bond interactions are indicated by broken lines, which are the same color as protein side-chain structures.
D1 residue in the DXDXE motif is not involved
A
B
directly in the catalytic mechanism, but participates in
the hydrogen bond network which stabilizes the core
of the (b ⁄a)8-barrel. Thus, we may propose a new
criterion for the classification of ‘plant-type’ and
‘bacterial-type’ family 18 chitinases based on the con-
formational change of the second Asp residue in the
DXDXE motif on substrate binding.
As suggested by X-ray crystallographic analyses of
SmChiB, catalysis in family 18 chitinases involves the
N-acetyl group of the sugar bound at the –1 subsite of
the enzyme (substrate-assisted catalysis) [24,40–42].
Protonation of the glycosidic bond by catalytic Glu
leads to a distortion of the sugar residue at the )1 sub-
site into a ‘boat’ conformation, and the departure of
the group is accompanied by a nucleophilic attack by
the N-acetyl oxygen (O7) on the anomeric carbon
(C1), thus yielding a positively charged, transient,
oxazolinium ion intermediate. In the AD2 E526A–sub-
strate structure, the N-acetyl oxygen of the )1 sugar
faces towards Asp524, which is opposite to the
direction in which it points in the SmChiB–(NAG)5
structure (Fig. 6) [24]. The N-acetyl oxygen (O7) is
located far from the anomeric carbon (C1) in an
unfavorable position for a direct nucleophilic attack
on the C1 carbon by the O7 oxygen (Fig. 6A). There-
fore, a drastic flip-like conformational change of the
Fig. 6. Comparison of the active sites in the AD2 E526A–(NAG)5
(A) and SmChiB E144Q–(NAG)5 (B) complexes (Protein Data Bank
code 1E6N), focusing on the conformation of bound substrates. For
clarity, only the sugar residues at subsites )1 and +1 are shown. In
the structures of AD2 and SmChiB, catalytic Glu (Glu526 and
Glu144) is replaced by Ala and Gln, respectively. The Tyr residue,
which is highly conserved among family 18 chitinases, is replaced
by Met in AD2. The anomeric carbons (C1), which are subjected to
nucleophilic attack by the carbonyl oxygen (O7) of the N-acetyl
group, are represented by asterisks. Hydrogen bond interactions
are shown as broken lines.
N-acetyl group should occur during the catalytic cycle
of AD2. In the current proposed catalytic models, con-
served Tyr residues (Tyr214 in SmChiB, Tyr183 in
hevamine) cooperate with the DXDXE motif to help
the catalytic reactions by stabilizing substrate distor-
tion (Fig. 6B) [22,24,40]. In AD2, however, this residue
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS
2691

Archaeal chitinase complexed with substrate
H. Tsuji et al.
substrate PNP-(NAG)2 (Seikagaku Co., Tokyo, Japan) [45].
Standard reaction mixtures contained purified enzyme and
0.01–5 mm of PNP-(NAG)2 in 0.2 m sodium acetate buffer
(pH 4.8) to a final volume of 400 lL. Enzyme concentra-
tions were adapted to the varying activities of the AD2
mutants. Reaction mixtures were incubated for 10 min at
50 ꢀC, after which the reaction was terminated with the
addition of 400 lL of 2 m sodium carbonate buffer
(pH 10.1). The amount of released p-nitrophenol was quan-
tified spectrophotometrically by the absorbance at 405 nm.
The standard employed p-nitrophenol at a concentration
range covering those of the substrates used in the kinetic
experiments. The production of p-nitrophenol was linear
with time for the incubation period, and < 5% of the
available substrate was hydrolyzed. The initial velocity was
saturable with increasing substrate concentration, and the
best-fit values of the apparent kinetic constants, kcat and
Km, in the Michaelis-Menten equation were obtained using
nonlinear regression analysis with origin software (Origin-
Lab Co., Northampton, MA, USA).
is replaced by Met, which does not seem to interact
with N-acetyl groups by forming a hydrogen bond in a
similar manner to Tyr (Fig. 6A) [23]. In the catalytic
mechanism of AD2, an oxazolinium ion intermediate
could be formed with the assistance of an amino acid
residue other than the DXDXE motif, as originally
proposed by Tews et al. [43] This is simpler than the
mechanism of the other family 18 chitinases.
Experimental procedures
Site-directed mutagenesis and enzyme
purification
Site-directed mutagenesis was introduced into a plasmid vec-
tor pET32_AD2Pf-ChiA [29] with the ‘QuikChange Site-direc-
ted Mutagenesis Kit’ (Stratagene, La Jolla, CA, USA)
according to the manufacturer’s protocol, with a minor mod-
ification: instead of Pfu DNA polymerase, we used KOD
plus polymerase (TOYOBO, Osaka, Japan). Target primers
for the generation of D522N, D522A, D524N, D524A,
E526Q, E526A and W664A mutations were 5¢-GCCACT
TACTTGAACTTTGACATAGAAGCCGG-3¢, 5¢-GCCAC
TTACTTGGCATTTGACATAGAAGCC-3¢, 5¢-GCCACT
TACTTGGACTTTAACATAGAAGCCGG-3¢, 5¢-GCCAC
TTACTTGGACTTTGCGATAGAAGCCGG-3¢, 5¢-GGAC
TTTGACATACAAGCCGGTATCGATGC-3¢, 5¢-GGACT
TTGACATAGCGGCCGGTATCGATGC-3¢ and 5¢-GGA
TCACTAGCCTTCGCGAGTGTAGACAGAG-3¢, respec-
tively, in which the mutated codons are in bold. The resulting
recombinant plasmids were verified by DNA sequencing with
an ABI Prismꢁ 310 Genetic Analyzer (Applied Biosystems,
Foster City, CA, USA) and transformed into expression host
E. coli Rosetta (DE3) cells.
Crystallization
We used AD2 E526A and D524A mutants to determine the
AD2–substrate complex structure. An E526A mutant com-
plexed with chito-oligosaccharides was cocrystallized by the
hanging drop vapor diffusion method. A portion (1 lL) of
E526A enzyme solution [20 mgÆmL)1 in 20 mm Tris ⁄ HCl
(pH 8.0), 50 mm NaCl] was mixed with 1 lL of reservoir
solution [0.1 m Mes (pH 6.5), 1.6 m MgSO4] containing
5 mm (NAG)5 (chitopentaose; Seikagaku Co., Tokyo,
Japan), and equilibrated against 0.35 mL of reservoir solu-
tion at 25 ꢀC. Crystals suitable for X-ray diffraction mea-
surement appeared within 1 week in the drops at
a
maximum size of 0.1 mm · 0.1 mm · 0.5 mm. We obtained
crystals of the D524A mutant complexed with chito-oligo-
saccharides by soaking experiments. Substrate-free D524A
crystals were prepared using procedures similar to those
employed previously for the wild-type [29]. A single D524A
crystal was soaked for 30 min at room temperature in a
reservoir solution [0.1 m Mes (pH 6.5), 1.6 m MgSO4] con-
taining 5 mm (NAG)5.
Overexpression and purification of all recombinant AD2
mutants were carried out using the same procedure as
described for the wild-type enzyme [29]. Briefly, cultures
were produced in Luria–Bertani (LB) broth containing
50 lgÆmL)1 of ampicillin at 37 ꢀC; enzyme expression was
induced with 0.5 mm isopropyl-1-thio-b-d-galactopyrano-
side and purification was conducted by a combination of
immobilized metal affinity chromatography using a HiTrap
Chelating HP column (GE Healthcare, Little Chalfont,
Buckinghamshire, UK) and anion-exchange chromatogra-
phy using a HiTrap Q HP column (GE Healthcare).
Enzyme purity was assessed by SDS ⁄ PAGE [44], followed
by Coomassie brilliant blue staining. The enzyme concen-
tration was determined using UV absorbance at 280 nm
and calculated extinction coefficients [29].
X-Ray crystal structure determination
˚
X-Ray diffraction data were collected using 0.90 A syn-
chrotron radiation at the undulator beamline BL44XU at
SPring-8 (Harima, Japan). For data collection, the crystals
were cryoprotected in the reservoir solution [0.1 m Mes
(pH 6.5), 1.6 m MgSO4] supplemented with 20% glycerol
(v ⁄ v), followed by flash cooling at 100 K by a nitrogen gas
stream. Diffraction data were integrated and scaled using
the programs denzo and scalepack from the hkl2000
package [46]. Cross-validation was applied by excluding 5%
of the reflections throughout the refinement procedure (free
Enzymology
The kinetic constants kcat and Km of the AD2 wild-type
and mutants were determined using the chromogenic
2692
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS

H. Tsuji et al.
Archaeal chitinase complexed with substrate
R-factor) [47]. All refinements were initiated by rigid body
refinement with refmac in the ccp4 suite [48,49], followed
by substrate model building. Substrate starting structures
and topologies were generated with the prodrg server [50].
The molecular models were built using coot [51] based on
2Fobs ) Fcalc and Fobs ) Fcalc electron density maps, and
were improved through many steps by alternate cycles of
manual model modification and crystallographic restrained
refinement with refmac until the R-factor and free R-fac-
tor converged. Occupancy refinement for two conformers
of the Glu526 side-chain observed in the D524A–substrate
complex (Fig. 3C, see Results for details) was carried out
using the cns program [32]. The stereochemistry of the
model (structural validation) was analyzed using procheck
[52]. The data collection and refinement statistics are sum-
marized in Table 2. For the final model, the structures of
the two independent monomers in the asymmetric unit were
almost identical (Table 2). In the interest of simplicity, the
structures are discussed consistently using the first mono-
mer of the coordinate files, unless otherwise stated. The
structures and electron density maps were created and dis-
played using the pymol program (http://www.pymol.org).
7 Collinge DB, Kragh KM, Mikkelsen JD, Nielsen KK,
Rasmussen U & Vad K (1993) Plant chitinases. Plant J
3, 31–40.
8 Melchers LS, Apotheker-de Groot M, van der Knaap
JA, Ponstein AS, Sela-Buurlage MB, Bol JF, Cornelissen
BJ, van den Elzen PJ & Linthorst HJ (1994) A new class
of tobacco chitinases homologous to bacterial exo-chitin-
ases displays antifungal activity. Plant J 5, 469–480.
9 Jeuniaux C (1961) Chitinase: an addition to the list of
hydrolases in the digestive tract of vertebrates. Nature
192, 135–136.
10 Zhu Z, Zheng T, Homer RJ, Kim YK, Chen NY, Cohn
L, Hamid Q & Elias JA (2004) Acidic mammalian
chitinase in asthmatic Th2 inflammation and IL-13
pathway activation. Science 304, 1678–1682.
11 Donnelly LE & Barnes PJ (2004) Acidic mammalian
chitinase: a potential target for asthma therapy. Trends
Pharmacol Sci 25, 509–511.
12 Kawada M, Hachiya Y, Arihiro A & Mizoguchi E
(2007) Role of mammalian chitinases in inflammatory
conditions. Keio J Med 56, 21–27.
13 Henrissat B (1991) A classification of glycosyl hydrolas-
es based on amino acid sequence similarities. Biochem J
280, 309–316.
Acknowledgements
14 Henrissat B & Bairoch A (1993) New families in the
classification of glycosyl hydrolases based on amino
acid sequence similarities. Biochem J 293, 781–788.
15 Henrissat B & Davies G (1997) Structural and
sequence-based classification of glycoside hydrolases.
Curr Opin Struct Biol 7, 637–644.
The authors thank Ms C. Kageyama for technical
assistance with protein purification and crystallization
and Dr Joseph Rodrigue for critical reading of the
manuscript. The X-ray diffraction studies were carried
out at the Japan Synchrotron Radiation Research
Institute.
16 Hart PJ, Pfluger HD, Monzingo AF, Hollis T &
Robertus JD (1995) The refined crystal structure of an
˚
endochitinase from Hordeum vulgare L. seeds at 1.8 A
References
resolution. J Mol Biol 248, 402–413.
17 Hoell IA, Dalhus B, Heggset EB, Aspmo SI & Eijsink
VG (2006) Crystal structure and enzymatic properties
of a bacterial family 19 chitinase reveal differences from
plant enzymes. FEBS J 273, 4889–4900.
1 Cottrell MT, Moore JA & Kirchman DL (1999) Chitin-
ases from uncultured marine microorganisms. Appl
Environ Microbiol 65, 2553–2557.
2 Bhattacharya D, Nagpure A & Gupta RK (2007) Bacte-
rial chitinases: properties and potential. Crit Rev
Biotechnol 27, 21–28.
18 Perrakis A, Tews I, Dauter Z, Oppenheim AB, Chet I,
Wilson KS & Vorgias CE (1994) Crystal structure of a
˚
bacterial chitinase at 2.3 A resolution. Structure 2,
3 Keyhani NO & Roseman S (1999) Physiological aspects
of chitin catabolism in marine bacteria. Biochim Biophys
Acta 1473, 108–122.
1169–1180.
19 Terwisschavan Scheltinga AC, Hennig M & Dijkstra
˚
BW (1996) The 1.8 A resolution structure of hevamine,
4 Kuranda MJ & Robbins PW (1991) Chitinase is
required for cell separation during growth of Saccharo-
myces cerevisiae. J Biol Chem 266, 19758–19767.
5 Sahai AS & Manocha MS (1993) Chitinases of fungi
and plants: their involvement in morphogenesis and
host–parasite interaction. FEMS Microbiol Rev 11, 317–
338.
a plant chitinase ⁄ lysozyme, and analysis of the con-
served sequence and structure motifs of glycosyl hydro-
lase family 18. J Mol Biol 262, 243–257.
20 Matsumoto T, Nonaka T, Hashimoto M, Watanabe T &
Mitsui Y (1999) Three-dimensional structure of the
catalytic domain of chitinase A1 from Bacillus circulans
WL-12 at a very high resolution. Proc Jpn Acad Ser B 75,
269–274.
6 Leah R, Tommerup H, Svendsen I & Mundy J (1991)
Biochemical and molecular characterization of three
barley seed proteins with antifungal properties. J Biol
Chem 266, 1564–1573.
21 Hollis T, Monzingo AF, Bortone K, Ernst S, Cox R &
Robertus JD (2000) The X-ray structure of a chitinase
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS
2693

Archaeal chitinase complexed with substrate
H. Tsuji et al.
from the pathogenic fungus Coccidioides immitis.
Protein Sci 9, 544–551.
ing-based discovery of potent inhibitors. Chem Biol 14,
589–599.
22 van Aalten DM, Synstad B, Brurberg MB, Hough E,
Riise BW, Eijsink VG & Wierenga RK (2000) Structure
of a two-domain chitotriosidase from Serratia marces-
34 Rao FV, Houston DR, Boot RG, Aerts JM, Hodkinson
M, Adams DJ, Shiomi K, Omura S & van Aalten DM
(2005) Specificity and affinity of natural product cyclo-
pentapeptide inhibitors against A. fumigatus, human,
and bacterial chitinases. Chem Biol 12, 65–76.
˚
cens at 1.9-A resolution. Proc Natl Acad Sci USA 97,
5842–5847.
23 Nakamura T, Mine S, Hagihara Y, Ishikawa K &
Uegaki K (2007) Structure of the catalytic domain of the
hyperthermophilic chitinase from Pyrococcus furiosus.
Acta Crystallogr Sect F: Struct Biol Cryst Commun 63,
7–11.
35 Watanabe T, Ariga Y, Sato U, Toratani T, Hashimoto
M, Nikaidou N, Kezuka Y, Nonaka T & Sugiyama J
(2003) Aromatic residues within the substrate-binding
cleft of Bacillus circulans chitinase A1 are essential for
hydrolysis of crystalline chitin. Biochem J 376, 237–244.
36 Katouno F, Taguchi M, Sakurai K, Uchiyama T,
Nikaidou N, Nonaka T, Sugiyama J & Watanabe T
(2004) Importance of exposed aromatic residues in
chitinase B from Serratia marcescens 2170 for crystal-
line chitin hydrolysis. J Biochem 136, 163–168.
37 Bortone K, Monzingo AF, Ernst S & Robertus JD
(2002) The structure of an allosamidin complex with the
Coccidioides immitis chitinase defines a role for a second
acid residue in substrate-assisted mechanism. J Mol Biol
320, 293–302.
24 van Aalten DM, Komander D, Synstad B, Gaseidnes S,
˚
Peter MG & Eijsink VG (2001) Structural insights into
the catalytic mechanism of a family 18 exo-chitinase.
Proc Natl Acad Sci USA 98, 8979–8984.
25 Oku T & Ishikawa K (2006) Analysis of the hyper-
thermophilic chitinase from Pyrococcus furiosus: activity
toward crystalline chitin. Biosci Biotechnol Biochem 70,
1696–1701.
26 Tanaka T, Fujiwara S, Nishikori S, Fukui T, Takagi M
& Imanaka T (1999) A unique chitinase with dual
active sites and triple substrate binding sites from the
hyperthermophilic archaeon Pyrococcus kodakaraensis
KOD1. Appl Environ Microbiol 65, 5338–5344.
27 Dahiya N, Tewari R & Hoondal GS (2006) Biotechno-
logical aspects of chitinolytic enzymes: a review. Appl
Microbiol Biotechnol 71, 773–782.
38 Kolstad G, Synstad B, Eijsink VG & van Aalten DM
(2002) Structure of the D140N mutant of chitinase B
˚
from Serratia marcescens at 1.45 A resolution. Acta
Crystallogr D: Biol Crystallogr 58, 377–379.
39 Synstad B, Gaseidnes S, Vriend G, Nielsen JE &
˚
Eijsink VGH (2000) On the contribution of conserved
acidic residues to the catalytic activity of chitinase B
from Serratia marcescens. In Advances in Chitin Science,
Vol. 4 (Peter MG, Muzzarelli RAA & Domard A, eds),
pp. 524–529. European Chitin Society.
28 Nakamura T, Mine S, Hagihara Y, Ishikawa K,
Ikegami T & Uegaki K (2008) Tertiary structure and
carbohydrate recognition by the chitin-binding domain
of a hyperthermophilic chitinase from Pyrococcus furio-
sus. J Mol Biol 381, 670–680.
40 Bokma E, Rozeboom HJ, Sibbald M, Dijkstra BW &
Beintema JJ (2002) Expression and characterization of
active site mutants of hevamine, a chitinase from the
rubber tree Hevea brasiliensis. Eur J Biochem 269,
893–901.
29 Mine S, Nakamura T, Hirata K, Ishikawa K,
Hagihara Y & Uegaki K (2006) Crystallization and
X-ray diffraction analysis of a catalytic domain of
hyperthermophilic chitinase from Pyrococcus furiosus.
Acta Crystallogr Sect F: Struct Biol Cryst Commun
62, 791–793.
41 Synstad B, Gaseidnes S, van Aalten DM, Vriend G,
˚
Nielsen JE & Eijsink VG (2004) Mutational and com-
putational analysis of the role of conserved residues in
the active site of a family 18 chitinase. Eur J Biochem
271, 253–262.
30 Papanikolau Y, Prag G, Tavlas G, Vorgias CE, Oppen-
heim AB & Petratos K (2001) High resolution struc-
tural analyses of mutant chitinase A complexes with
substrates provide new insight into the mechanism of
catalysis. Biochemistry 40, 11338–11343.
42 Vaaje-Kolstad G, Houston DR, Rao FV, Peter MG,
Synstad B, van Aalten DM & Eijsink VG (2004)
Structure of the D142N mutant of the family 18
chitinase ChiB from Serratia marcescens and its com-
plex with allosamidin. Biochim Biophys Acta 1696,
103–111.
31 Wallace AC, Laskowski RA & Thornton JM (1995)
LIGPLOT: a program to generate schematic diagrams
of protein–ligand interactions. Protein Eng 8, 127–134.
32 Brunger AT, Adams PD, Clore GM, DeLano WL,
¨
Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J,
Nilges M, Pannu NS et al. (1998) Crystallography &
NMR system: a new software suite for macromolecular
structure determination. Acta Crystallogr D: Biol
Crystallogr 54, 905–921.
43 Tews I, Terwisscha van Scheltinga AC, Perrakis A,
Wilson KS & Dijkstra BW (1997) Substrate-assisted
catalysis unifies two families of chitinolytic enzymes.
J Am Chem Soc 119, 7954–7959.
44 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.
33 Hurtado-Guerrero R & van Aalten DM (2007) Struc-
ture of Saccharomyces cerevisiae chitinase 1 and screen-
2694
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS

H. Tsuji et al.
Archaeal chitinase complexed with substrate
45 Suginta W, Songsiriritthigul C, Kobdaj A, Opassiri R
& Svasti J (2007) Mutations of Trp275 and Trp397
altered the binding selectivity of Vibrio carchariae
chitinase A. Biochim Biophys Acta 1770, 1151–
1160.
52 Laskowski RA, Moss DS & Thornton JM (1993) Main-
chain bond lengths and bond angles in protein struc-
tures. J Mol Biol 231, 1049–1067.
53 Davies GJ, Wilson KS & Henrissat B (1997) Nomencla-
ture for sugar-binding subsites in glycosyl hydrolases.
Biochem J 321, 557–559.
46 Otwinowski Z & Minor W (1997) Processing of X-ray
diffraction data collected in oscillation mode. Methods
Enzymol 276, 307–326.
Supporting information
47 Brunger AT (1992) Free R value: a novel statistical
¨
quantity for assessing the accuracy of crystal structures.
Nature 355, 472–475.
The following supplementary material is available:
Fig. S1. Overall structures of AD2 catalytic site
mutants complexed with chito-oligosaccharides.
Fig. S2. Structure of AD2 D524A substrate-free form.
Table S1. Data collection and refinement statistics for
the AD2 D524A substrate-free form.
48 Collaborative Computational Project, Number 4 (1994)
The CCP4 suite: programs for protein crystallography.
Acta Crystallogr D: Biol Crystallogr 50, 760–763.
49 Murshudov GN, Vagin AA & Dodson EJ (1997)
Refinement of macromolecular structures by the maxi-
mum-likelihood method. Acta Crystallogr D: Biol Crys-
tallogr 53, 240–255.
This supplementary material can be found in the
online version of this article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
50 Schuttelkopf AW & van Aalten DM (2004) PRODRG:
¨
a tool for high-throughput crystallography of protein–
ligand complexes. Acta Crystallogr D: Biol Crystallogr
60, 1355–1363.
51 Emsley P & Cowtan K (2004) Coot: model-building
tools for molecular graphics. Acta Crystallogr D: Biol
Crystallogr 60, 2126–2132.
FEBS Journal 277 (2010) 2683–2695 ª 2010 The Authors Journal compilation ª 2010 FEBS
2695