Báo cáo khoa học: Mitochondria regulate platelet metamorphosis induced by opsonized zymosan A – activation and long-term commitment to cell death potx

Accessory active site residues of Streptomyces sp. N174
chitosanase
Variations on a common theme in the lysozyme superfamily
Marie-E
`
ve Lacombe-Harvey
1
, Tamo Fukamizo
2
, Julie Gagnon
1
, Mariana G. Ghinet
1
,
Nicole Dennhart
3
, Thomas Letzel
3
and Ryszard Brzezinski
1
1De
´
partement de Biologie, Centre d’E
´
tude et de Valorisation de la Diversite
´
Microbienne, Universite
´
de Sherbrooke, Canada
2 Department of Advanced Bioscience, Kinki University, Nara, Japan
3 Department for Basic Life Sciences, Technische Universita

¨
tMu
¨
nchen, Freising-Weihenstephan, Germany
The chitosanase from Streptomyces sp. N174
(CsnN174) catalyzes the hydrolysis of b-1,4-glycosidic
links in chitosan, a water-soluble derivative of chitin
composed of d-glucosamine (GlcN) with a variable but
minor proportion of N-acetyl-d-glucosamine (GlcNAc)
[1]. Research on the enzymatic hydrolysis of chitosan
is driven by the fact that this polymer has numerous
potential applications and that its properties often
depend on its molecular mass [2]. CsnN174 belongs to
family 46 of the glycoside hydrolases (GH46), endohy-
drolase-type enzymes acting via an inverting mecha-
nism [3,4]. GH46 enzymes belong to the GH-I clan [5]
together with lysozymes from family GH24 (the most
studied being the lysozyme from T4 phage). Enzymes
from these two families share the same catalytic mech-
anism and are folded similarly, with two globular
Keywords
chitinase; chitosanase; glycoside hydrolase;
inverting mechanism; lysozyme
Correspondence
R. Brzezinski, De
´
partement de Biologie,
Universite
´
de Sherbrooke, 2500 boul. de

l’Universite
´
, Sherbrooke, QC J1K 2R1,
Canada
Fax: +1 819 821 8049
Tel: +1 819 821 8000; ext 61077
E-mail:
(Received 22 September 2008, revised 26
November 2008, accepted 3 December
2008)
doi:10.1111/j.1742-4658.2008.06830.x
The chitosanase from Streptomyces sp. N174 (CsnN174) is an inverting
glycoside hydrolase belonging to family 46. Previous studies identified
Asp40 as the general base residue. Mutation of Asp40 into glycine revealed
an unexpectedly high residual activity. D40G mutation did not affect the
stereochemical mechanism of catalysis or the mode of interaction with sub-
strate. To explain the D40G residual activity, putative accessory catalytic
residues were examined. Mutation of Glu36 was highly deleterious in a
D40G background. Possibly, the D40G mutation reconfigured the catalytic
center in a way that allowed Glu36 to be positioned favorably to perform
catalysis. Thr45 was also found to be essential. Thr45 is thought to orientate
the nucleophilic water molecule in a position to attack the glycosidic link.
The finding that expression of heterologous CsnN174 in Escherichia coli
protects cells against the antimicrobial effect of chitosan, allowed the selec-
tion of active chitosanase variants after saturation mutagenesis. Thr45 could
be replaced only by serine, indicating the importance of the hydroxyl group.
The newly identified accessory catalytic residues, Glu36 and Thr45 are
located on a three-strand b sheet highly conserved in GH19, 22, 23, 24 and
46, all members of the ‘lysozyme superfamily’. Structural comparisons
reveal that each family has its catalytic residues located among a small num-

ber of critical positions in this b sheet. The position of Glu36 in CsnN174 is
equivalent to general base residue in GH19 chitinases, whereas Thr45 is
located similarly to the catalytic residue Asp52 of GH22 lysozyme. These
examples reinforce the evolutionary link among these five GH families.
Abbreviations
(GlcN)
n
, b-D-glucosamine oligosaccharide with n monomer units; CsnN174, chitosanase from Streptomyces sp. N174; GH, glycoside
hydrolase family; GlcN,
D-glucosamine; GlcNAc, N-acetylglucosamine.
FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS 857
(mostly a-helical) domains separated by a substrate
binding cleft [6–9].
Site-directed mutagenesis studies of CsnN174, com-
bined with crystallography data, catalytic functions to
be assigned to residues Glu22 (the general acid) and
Asp40 (the general base) [10]. These residues are
strictly conserved among all GH46 proteins for which
chitosanase activity has been confirmed by biochemical
studies. Glu22 is close to the C-terminus of an a helix,
belonging to a central structural core consisting of two
a helices and a b sheet [8]. This structure is shared
with GH24 enzymes, and is also found in GH19
chitinases and GH22 or GH23 lysozymes, represented
respectively by the extensively studied chitinase from
barley seeds and lysozymes from hen and goose egg-
white. This group of enzymes is sometimes designated
as the lysozyme superfamily [6,8]. Despite their struc-
tural similarity and a highly equivalent positioning of
the general acid residue, GH22 enzymes differ from

the others in this group in that they act by a mecha-
nism with anomeric retention [11]. Asp40 of CsnN174
is found inside another element of the central con-
served core: a sheet formed by three antiparallel
b strands separated by loops of varying lengths [7].
In contrast to the general acid residue, the general
base residues are not localized in equivalent positions
in these five families [8]. In the extreme case, no resi-
due with general base function has so far been pro-
posed for goose egg-white lysozyme [12]. Details of the
catalytic mechanism should vary among these structur-
ally related enzyme families.
This study was initiated by the observation that a
CsnN174 mutant in which the general base residue has
been substituted by a glycine (D40G) retained a signifi-
cant proportion of the wild-type activity. Studies of
residues that caused complete loss of activity in this
mutant led to the identification of two residues with
accessory catalytic functions. We developed a method
for revertant chitosanase identification among a popu-
lation of inactive enzyme-encoding genes based on the
discovery that the heterologous CsnN174 expression
protects Escherichia coli against the antimicrobial
effect of chitosan. Finally, we discuss the evolutionary
implications of the presence of such accessory catalytic
residues in GH46 chitosanases.
Results
N174 chitosanase devoid of Asp40 retains
significant enzymatic activity
In a previous study, we identified Asp40 as the best

candidate for the general base in the inverting mecha-
nism. This was supported by its positioning in the 3D
structure [7] and the substantial loss of activity
observed for enzymes mutated in this position, because
the replacement of Asp40 by conservative Glu or Asn
residues resulted in a decrease of k
cat
to 1 ⁄ 125 and
1 ⁄ 485 of wild-type, respectively [10]. These values were
typical of similar mutations in other inverting glyco-
side hydrolases [13].
A mutation path was suggested by Brameld and
Goddard to rebuild the GH19 barley inverting chitin-
ase into a retaining enzyme [14]. This set of mutations,
resulting from molecular dynamics simulations,
included mutation of the general base residue Glu89
into glycine followed by mutation of Gly113 into glu-
tamate. Although (to our best knowledge) a GH19
enzyme with a retaining reaction mechanism has not
been disclosed in the literature, it was worth trying to
introduce analogous mutations in CsnN174, consider-
ing the similarity of the structural cores among GH46
and GH19 enzymes [8]. We thus mutated Asp40 of
chitosanase into Gly [15]. In preliminary studies, the
D40G mutant revealed significant activity, unexpected
for an enzyme devoid of its general base catalytic resi-
due. We further proceeded with kinetic analysis which
revealed that K
m
remained similar to wild-type

(Table 1), whereas k
cat
was  28 times lower than
wild-type but, respectively, 4.5 and 17.5 times higher
than that of mutants D40E and D40N studied previ-
ously [10]. The data suggested that D40G mutant
chitosanase was impaired in its catalytic activity
although its substrate-binding mode remained essen-
tially unchanged.
Table 1. Specific activities and kinetic parameters of purified wild-
type and mutant CsnN174. All specific activities were determined
at a single chitosan concentration (800 lgÆmL
)1
). Kinetic parame-
ters were calculated using the non linear least-square fitting proce-
dure for Michaelis–Menten equation in
PRISM software v. 5.0. ND,
not determined.
Enzyme
Specific activity
(unitsÆmg
)1
protein)
K
m
(lgÆmL
)1
)
k
cat

(min
)1
)
Wild-type 52.9 24.8 743.9
D40G 1.6 22.9 26.4
E36A 29.2 40.2 688.0
E36Q 44.7 34.8 699.8
E36D 29.1 25.8 500.2
E36N 21.9 36.1 409.6
E36Q + D40G 0.3 41.6 4.1
E36A + D40G 0.07 52.4 1.44
T45H < 0.01 ND ND
T45E 0.011 23.5 2.21
T45S 37.6 25.8 628.9
D40G + T45D 0.06 31.2 1.0
D40G + T45E 0.02 ND ND
Active site residues of family 46 chitosanase M E
`
. Lacombe-Harvey et al.
858 FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS
Interaction of the D40G mutant with the substrate
could be investigated in several ways. First, the sub-
strate-binding ability was assessed by thermal unfolding
experiments (Fig. 1). (GlcN)
3
binding to the wild-type
enzyme increased the transition temperature (T
m
)by
5.7 °C, and its binding to D40G increased T

m
by
2.8 °C. Thus, (GlcN)
3
binding to D40G stabilizes the
protein structure to a similar extent as in the case of
the wild-type enzyme.
The mode of hydrolysis of glucosamine oligosaccha-
rides [5,15,16], however, provides insight into the inter-
action of the enzyme with the substrate during the
reaction, because a mutation of a residue involved in
substrate binding is expected to result in an altered
time course of hydrolysis. The reaction time course of
D40G mutant enzyme was thus investigated with
(GlcN)
5
and (GlcN)
4
substrates and monitored by
real-time MS [15]. As shown in Fig. 2, the specific
activity of D40G chitosanase, determined from the
degradation rate of these substrates was found to be
7.4Æmin
)1
(wild-type = 385Æmin
)1
) for (GlcN)
5
and
3.1Æmin

)1
(wild-type = 140 Æmin
)1
) for (GlcN)
4
. In both
cases, the D40G chitosanase degradation rate is  2%
that of wild-type, which is in the range obtained with
high-molecular mass chitosan substrate (Table 1). The
time course profile of hydrolysis by D40G mutant is
similar to that of the wild-type (Fig. 2), indicating that
this mutant is not impaired in substrate binding. A
control experiment with (GlcNAc)
6
indicated that nei-
ther wild-type nor D40G chitosanase are able to cleave
GlcNAc–GlcNAc bonds (data not shown).
The stereochemistry of the D40G chitosanase reac-
tion was investigated by
1
H-NMR. As shown in
Fig. 3, D40G mutant is still an inverter, because the
time course of anomer formation is essentially the
same as for the wild-type. In D40G, the water mole-
cule was found to attack the C1 carbon of the transi-
tion state sugar residue from the side identical to that
in the wild-type [3].
It was shown recently that a mutation of the gen-
eral base residue can be rescued by sodium azide in
retaining glycoside hydrolases [17] and also in an

inverting a-glycosidase [18]. We thus investigated the
effect of azide ion on the activity of D40G mutant
chitosanase. The time courses of (GlcN)
6
degradation
in the absence or presence of sodium azide (0.65 and
2.6 m) are shown in Fig. 4A. The rate of (GlcN)
6
degradation was significantly enhanced by the addi-
tion of the azide ion. The effect of the azide concen-
tration on the reaction rate, shown in Fig. 4B, clearly
demonstrates that the rate enhancement depends upon
the azide concentration. The results indicate that
Asp40 acts as a catalytic base, which activates a
water molecule.
In summary, substitution of the general base Asp40
by glycine resulted in an enzyme that is distinguished
from wild-type only by a lower activity, without
changing the mechanism of hydrolysis or the mode of
interaction with substrate.
Glu36 as a possible alternative general
base residue
A possible explanation of the higher activity of D40G
chitosanase compared with mutants D40N or D40E
was that the mutant D40G reconfigured its three
b-strands motif such that another residue could
become localized in a favorable position to perform
catalysis. Glu36 was found to be the best candidate
Fig. 1. Thermal unfolding curves of wild-type (A) and D40G (B)
chitosanases in the presence or absence of (GlcN)

3
. The enzyme
and the trisaccharide were mixed in 50 m
M sodium acetate buffer
pH 5.5. The final concentrations are 2.3 l
M for the enzyme and
2.3 m
M for the saccharide. The unfolding process was monitored
by CD at 222 nm.
M E
`
. Lacombe-Harvey et al. Active site residues of family 46 chitosanase
FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS 859
because its side chain points towards the substrate-
binding cleft (Fig. 5A). Glu36 appears to be a minor
player in the wild-type configuration, because its sub-
stitution by Asp, Asn, Gln or even Ala had minor
effects on activity, decreasing the catalytic constant at
most by one third and slightly increasing the K
m
value
A B
C D
Fig. 2. Time courses of (GlcN)
5
and (GlcN)
4
hydrolysis catalyzed by wild-type and D40G endochitosanases monitored by real-time MS. The
enzymatic reactions were carried out in 10 m
M ammonium acetate-containing aqueous solutions pH 5.2 at 20 °C. (A) (GlcN)

n
hydrolysis time
courses obtained for wild-type endochitosanase (5.0 n
M) catalyzed reaction performed with 25.0 lM of the substrate (GlcN)
5
(100% = 2.3 · 10
6
counts); (B) (GlcN)
n
hydrolysis time courses obtained for D40G endochitosanase (200 nM) catalyzed reaction performed
with 25.0 l
M of the substrate (GlcN)
5
(100% = 2.2 · 10
6
counts); (C) (GlcN)
n
hydrolysis time courses obtained for wild-type endochitosanase
(5.0 n
M) catalyzed reaction performed with 25.0 lM of the substrate (GlcN)
4
(100% = 2.5 · 10
6
counts); (D) (GlcN)
n
hydrolysis time courses
obtained for D40G endochitosanase (200 n
M) catalyzed reaction performed with 25.0 lM of the substrate (GlcN)
4
(100% = 2.8 · 10

6
counts).
(A) and (C) were adapted from Dennhart et al. [15] with permission.
Fig. 3. Anomer production from the D40G
mutant chitosanase hydrolysis of (GlcN)
6
.
(A) Time-dependent
1
H-NMR spectra. (B)
Time course of anomer production. The
enzymatic reaction was conducted in 50 m
M
deuterated sodium acetate buffer pH 5.0 in
an NMR tube thermostated at 30 °C.
Active site residues of family 46 chitosanase M E
`
. Lacombe-Harvey et al.
860 FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS
Fig. 4. Chemical rescue experiments. (A) Time courses of the
enzymatic degradation of (GlcN)
6
by D40G in the absence or pres-
ence of sodium azide (0.65 and 2.6
M). The enzymatic reaction was
conducted in 50 m
M sodium acetate buffer pH 5.0 and at 40 °C.
The enzyme concentration was 8.5 l
M. Only some examples of
tested concentrations are shown. (B) Effect of sodium azide con-

centration on the reaction rate of D40G. (C) Time courses of the
enzymatic degradation of (GlcN)
6
by E36A + D40G in the absence
or presence of sodium azide (2.3
M).
A
B
Fig. 5. (A) Structural view of the active site cleft of chitosanase
Csn-N174. The image represents a portion of the chain A from
1CHK file in Protein Data Bank [7]. L(1–2); loop between sheets b-1
and b-2; L(2–3), loop between sheets b-2 and b-3. Asp57, Glu197
and Glu201 are residues involved in chitosan substrate binding at
)2, )1 and +2 subsite, respectively [37]. The model was drawn
using
PYMOL software (version 0.99; DeLano Scientific, San Fran-
cisco, CA, USA). (B) Alignment of portions of the primary structure
of GH46 chitosanases including active site residues. Numbering
refers to the distance of the first residue from the N-terminus of
the mature protein (Csn-N174; chitosanase from B. circulans
MH-K1) or of the precursor protein as stored in GenBank (other
chitosanases). Arrows indicate the residues discussed in this work.
Symbol explanation (bacterial names followed by accession num-
bers for GenBank database): BAC_CIRC, B. circulans MH-K1
(D10624); BAC-EHIM, Paenibacillus ehimensis EAG1 (AB008788);
BUR_GLAD, Burkholderia gladioli (AB029336); BAC_SUBT, Bacil-
lus subtilis (U93875); BAC_AMYL, Bacillus amyloliquefaciens
(ABS75305); BAC_KFB, Bacillus sp. KFB-CO4 (AF160195); PBCV-1,
Chlorella virus 1 of Paramecium bursaria (U42580); CVK2, Chlorella
virus CVK2 (D88191); CsnN174, Streptomyces sp. N174 (L07779);

NOC_N106, Nocardioides sp. N106 (L40408); STR_COEL, Strepto-
myces coelicolor A3(2) (AL109849.1 ORF SC3A3.02).
0.12
0.1
0.08
0.06
0.04
0.02
0
0.12
0.1
0.08
0.06
0.04
0.02
0
0.12
0.1
0.08
0.06
0.04
0.02
0
0 40 80 120
Reaction time (min)
Concentration (M)
160 200 240
0 40 80 120 160 200 240
0 40 80 120 160 200 240
(GIcN)

2
(GIcN)
6
(GIcN)
3
(GIcN)
3
(GIcN)
6
(GIcN)
2
(GIcN)
4
(GIcN)
2
(GIcN)
4
(GIcN)
3
(GIcN)
6
0 M NaN
3
A
0.65 M NaN
3
2.6 M NaN
3
(GIcN)
4

B
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5
Sodium azide (
M)
Specific acivity (×10
–4
M·min
–1
·µM
–1
)
2 2.5 3
C
0 M NaN
3
2.3 M NaN
3
Concentration (m
M
)
Reaction time (min)
(GIcN)
6

(GIcN)
3
(GIcN)
4
(GIcN)
6
(GIcN)
3
(GIcN)
4
70
60
50
40
30
20
10
0
70
60
50
40
30
20
10
0
0 100 200 300 400 500 600 700
0 100 200 300 400
500 600 700
M E

`
. Lacombe-Harvey et al. Active site residues of family 46 chitosanase
FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS 861
(Table 1). The effect of the E36 mutation was quite
different in the enzyme with glycine-substituted Asp40.
The k
cat
of the double mutant E36Q + D40G is more
than five times lower than that of the single mutant
D40G and  18 times lower for E36A + D40G, putt-
ing it into the very low range observed for D40N or
D40E, whereas K
m
in both these double mutants is
increased only by a factor of 2. The rate of (GlcN)
6
degradation was also enhanced by sodium azide in
E36A + D40G. The enhancement was less intensive
than that in D40G but significant as shown in Fig. 4C.
In combination with the D40G mutation, these
data suggest that the carboxylate group of Glu36 is in
position to act as a general base in the inverting
mechanism.
Thr45 is essential for catalytic activity both in
wild-type chitosanase and mutant D40G
Residues with hydroxyl groups were found in the
microenvironment of general base residues in some
inverting glycoside hydrolases. They are thought to
orientate the nucleophilic water molecule in a position
optimal for catalysis. Tyr203 of the inverting GH8

xylanase from Pseudoalteromonas haloplanktis [19] or
Ser190 of GH19 Streptomyces griseus chitinase ChiC
(equivalent to Ser120 in barley seed chitinase) [20,21]
are examples of such residues. However, no residue
with this function has been proposed in chitosanases.
From this point of view, we examined Thr45 as a pos-
sible candidate, a residue highly conserved in GH46
chitosanases (Fig. 5B). Thr45 was first mutated into
His or Glu. T45H mutation resulted in a complete loss
of activity, whereas T45E mutant had a very low resid-
ual activity but sufficient to perform kinetic analysis
allowing the conclusion that the loss of activity of
T45E can be explained by a severe decrease of k
cat
(Table 1). Interestingly, the activity of this mutant
could not be enhanced by sodium azide (data not
shown).
We then verified whether the Thr45 residue is also
essential in the chitosanase reconfigured by the D40G
mutation. Two double mutants were examined:
D40G + T45E and D40G + T45D. The D40G +
T45E mutant had only 0.03% of wild-type specific
activity when tested on chitosan substrate; a value sim-
ilar to the single T45E mutant. The mutant D40G +
T45D was slightly more active (0.1% of wild-type
activity). Again, kinetic analysis of this double mutant
has shown that the loss of activity was explained by a
dramatic decrease in k
cat
, although K

m
remained simi-
lar to wild-type. This could be confirmed by the
reaction time course of D40G + T45D mutant investi-
gated with (GlcN)
6
substrate and monitored by real-
time MS. As shown in Fig. 6, even if the enzyme con-
centration was 500-fold higher in the D40G + T45D
reaction (Fig. 6B), time course profiles were compara-
ble between the wild-type and double mutant. This
also indicates a dramatic decrease in enzymatic activity
in the double mutant.
A
B
Fig. 6. Time courses of (GlcN)
6
hydrolysis catalyzed by wild-type
and D40G + T45D chitosanases monitored by real-time mass spec-
trometry. The enzymatic reactions were carried out in 10 m
M
ammonium acetate-containing aqueous solutions pH 5.2 at 20 °C.
(A) (GlcN)
n
hydrolysis time courses obtained for wild-type endochi-
tosanase (5.0 n
M) catalyzed reaction performed with 25.0 lM of the
substrate (GlcN)
6
(100% = 3.1 · 10

6
counts); (B) (GlcN)
n
hydrolysis
time courses obtained for D40G + T45D endochitosanase (2.5 l
M)
catalyzed reaction performed with 25.0 l
M of the substrate (GlcN)
6
(100% = 3.0 · 10
6
counts).
Active site residues of family 46 chitosanase M E
`
. Lacombe-Harvey et al.
862 FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS
Saturation mutagenesis analysis of residue 45
including selection on chitosan medium
In order to identify any residues that could replace
Thr45 CsnN174, allowing the enzymatic activity to be
kept at levels close to that of the wild-type enzyme, we
studied the reversion of the inactive mutant T45H by
saturation mutagenesis. We constructed a library of
several hundred E. coli clones with randomly intro-
duced codons at position 45 of the csnN174 gene
harbored by the plasmid pAlter-csn [10].
Chitosan polymer solubilized in growth medium has
antibacterial activity and severely inhibits the growth of
E. coli [22,23]. The extent of growth inhibition is
dependent on chitosan concentration, average molecu-

lar mass, the pH of the medium and salt composition.
We noticed that growth inhibition could be suppressed
by the expression of CsnN174 in E. coli JM109 (data
not shown). This led to the development of a method
for chitosanase revertant selection. The composition of
the selective medium was optimized using the mutants
V148T and T45H encoding chitosanases having, res-
pectively,  10% and < 0.1% of wild-type activity
(I. Boucher & R. Brzezinski, unpublished data, and
Table 1). We used chitosan (0.3 gÆL
)1
) with an average
molecular mass (M
n
) reduced to  15 kDa by enzymatic
hydrolysis, which exhibited a severe antimicrobial effect
against E. coli [23] although much more soluble in aque-
ous solutions than native chitosan. A low concentration
(5 mm) citrate buffer (pH 6.0) was included in the med-
ium to keep the pH slightly acidic and avoid chitosan
precipitation (usually occurring at pH > 6.5). The
chitosan concentration was adjusted to allow growth of
E. coli strains expressing wild-type or V148T chitosan-
ase although inhibiting strains expressing the T45H
chitosanase or harboring the empty pAlter-1 vector.
During the optimization of the selection medium, we
observed that growth inhibition was highly dependent
on the bacterial density on the Petri plates. We further
adjusted the monovalent (Na
+

) and divalent (Mg
2+
)
ion concentrations for a density of 500 colony forming
units per plate. As the saturation mutagenesis library
contained presumably only a small minority of chito-
sanase-positive revertants, some chitosanase-negative
colonies could still grow on chitosan medium due to a
kind of ‘protective effect’ resulting from higher local
cell density. We thus estimated the number of false-
positive colonies recovered on this medium by mixing
various proportions of the V148T chitosanase-express-
ing cells and chitosanase-negative cells and plating
them on media with various salt compositions. Both
colonies could be distinguished thanks to supplemen-
tation with 5-bromo-4-chloro-3-indolyl-b-d-galactopyr-
anoside and isopropyl thio-b-d-galactoside, because
the empty pAlter-1 vector directs b-galactosidase syn-
thesis in E. coli and false positives appeared as blue
colonies. The lowest proportion of false positives
(1 ⁄ 250) was obtained after the addition of 200 mm
NaCl and 3 mm MgSO
4
. This salt composition was
adopted for the revertant selection experiment.
After plating the complete T45H-saturation muta-
genesis library ( 450 clones) on the optimized chito-
san medium, we obtained 55 colonies of putative
chitosanase-positive revertants. We sequenced csn
genes in 15 randomly chosen clones, revealing Thr resi-

dues in nine revertants (three ACC codons, three
ACT, two ACG and one ACA) and Ser residues in six
revertants (three TCT codons, one AGC, one AGT,
one ACT). Codon diversity indicated that the muta-
genesis has been performed without bias in nucleotide
substitution. We concluded that the Thr residue could
be replaced only by Ser, showing the importance of
the hydroxyl group in position 45.
We purified the T45S chitosanase mutant after its
introduction into Streptomyces lividans. This mutant
was quite active, keeping  71% of specific activity of
the wild-type enzyme. This confirmed the utility of the
chitosan medium for isolation of chitosanase rever-
tants. Kinetic analysis showed that the T45S mutation
results in almost unchanged K
m
and slightly decreased
k
cat
(Table 1).
Discussion
Proposed functions for E36 and T45 residues in
CsnN174
In this study, we confirmed that Asp40 functions as a
general base in CsnN174 catalysis, because the activity
of mutants devoid of this aspartate (D40G and
E36A + D40G) can be enhanced by sodium azide.
The lack of effect of sodium azide on double mutant
D40G + T45E activity indicates that the rate enhance-
ment observed in the D40G single mutant is derived

from complementing the D40 function. In this case,
the azide ion acts as a general base which enhances the
nucleophilicity of the water molecule, as proposed by
Miyake et al. [18]. Unexpectedly, substitution of Asp40
by Gly resulted in an enzyme with a residual activity
much higher than predicted for a mutation involving a
catalytic residue. Asp40 is localized on a loop between
the b-1 and b-2 strands of CsnN174 (Fig. 5A). Analo-
gous loops in related glycosyl hydrolases such as T4
lysozyme or barley chitinase show a conformational
diversity, indicating that the loop is potentially mobile
[7,8]; a tendency accentuated further by the Asp40 to
M E
`
. Lacombe-Harvey et al. Active site residues of family 46 chitosanase
FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS 863
Gly substitution, because this results in the addition of
a fourth glycine residue to this loop, in addition to
Gly39, Gly41 and Gly43 already present in wild-type.
Two of these glycines (Gly41 and Gly43) are present
in most GH46 chitosanases (Fig. 5B) suggesting that
the flexibility of this loop is important. Finally, exami-
nation of the structure with the what if program [24]
suggested that the Asp40 to Gly mutation eliminates
the salt bridge-type of interaction with the Arg42 resi-
due observed in the wild-type enzyme, allowing for
further loop mobility. It is then likely that the Glu36
residue, localized in the neighboring b-1 strand
(Fig. 5A) could be placed in a position competent for
catalysis. In the native chitosanase crystal, the distance

between the catalytic carboxylates (13.8 A
˚
) is greater
than that usually observed in inverting glycoside
hydrolases, so the binding of substrate must induce a
substantial conformational change to allow catalysis.
Glu36 carboxylate is localized 14.9 A
˚
from the general
acid residue Glu22, but after the reconfiguration result-
ing from substrate binding it could be in position to
perform catalysis. To date, the exact conformational
changes occurring after substrate binding could not be
described in GH46 chitosanases, because co-crystals
with substrate could not be obtained; the same difficul-
ties being reported for the structurally related GH19
chitinases [4,7,20,21,25].
Whereas mutations of Asp40 allowed for substantial
residual activity, those of Thr45 had more severe con-
sequences. Mutations involving Thr45 reduced the
activity by at least three orders of magnitude and they
had equally severe consequences when introduced in
the enzyme reconfigured by the D40G mutation. The
T45 residue appears to be essential for catalysis. Satu-
ration mutagenesis revealed however that the residue
can be replaced by a serine with a very moderate loss
of activity, suggesting the importance of a hydroxyl
group but with some tolerance regarding its exact
position. Interestingly, although this threonine is
highly conserved among GH46 chitosanases, one

sequenced chitosanase (from Paenibacillus ehimensis;
Fig. 5B) has a serine in the corresponding position
[26], which indirectly confirms our revertant analysis.
Possibly, the sulfhydryl group in a T45C mutant
could also adequately orientate the water molecule
resulting in decent enzyme activity. The absence of
such a mutant among the revertants isolated after
saturation mutagenesis could simply result from statis-
tical probability, but we also remark that residue 45
in CsnN174 is in the immediate proximity of residue
Cys52. Mutation of Thr45 to cysteine could result in
the creation of a disulphide bond with Cys52, making
the sulfhydryl group unavailable for the orientation of
the water molecule and implying loss of enzymatic
activity.
Thr45 of CsnN174 lies in a position analogous to
Thr26 in T4 lysozyme [8], an extensively studied resi-
due. T26H mutation resulted in conversion of an
inverting enzyme into a retaining one [27], whereas
T26E mutation resulted in an inactive enzyme forming
a covalent bond with the substrate [28]. None of these
effects was observed in the corresponding CsnN174
mutants. Structural differences between CsnN174 and
T4 lysozyme could account for this different behavior,
because the mutual positions of the discussed threo-
nines and the general base residues (Asp40 in CsnN174
and Asp20 in T4 lysozyme) are not totally equivalent;
the hydroxyl group being closer to the general base
carboxylate in T4 lysozyme (3.6 A
˚

) than in CsnN174
(4.9 A
˚
).
Another explanation for this different behavior was
raised by Zechel and Withers [29]: the retaining mecha-
nism of the T26H mutant of T4 lysozyme could
involve the acetamide group of the substrate, as
observed in GH18 or GH20 enzymes hydrolyzing
chitin polymers. As the mutation effects are, in our
case, observed with GlcN oligomers lacking GlcNAc
residues, this is unlikely for CsnN174. Besides these
differences in mutant behavior, the requirement for a
hydroxyl in residue 45 in CsnN174 and the almost
complete loss of activity in mutants indicate that posi-
tioning of the attacking water is a plausible function
for this residue. In barley chitinase, Ser120 is thought
to play the same role [21] and the alignment of pri-
mary structures of GH19 chitinases (not shown)
reveals that this serine is replaced in many enzymes by
a threonine. Interaction of these hydroxyl amino acids
with water is observed in the crystal structures of the
GH46 chitosanase from Bacillus circulans MH-K1 (res-
idue Thr60) and the GH19 chitinase of S. griseus
(Ser190) [4,20].
Catalytic residues in the lysozyme superfamily:
variations on a common theme
It is now generally accepted that strict positioning of
the catalytic base is not required for inverting glycosid-
ases [30]. This flexibility results in a variety of confor-

mations for the residues supporting the ‘nucleophilic
side’ of the catalytic mechanism observed in the lyso-
zyme superfamily. A closer look at the three b-strands
segment of the conserved structural core in this super-
family [8] reveals a small number of key structural ele-
ments, the residues of which play various functions
depending on the enzyme family. For example, Glu36
of CsnN174 discussed here lies in a position equivalent
Active site residues of family 46 chitosanase M E
`
. Lacombe-Harvey et al.
864 FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS
to the general base residue of barley chitinase [7,21]
and to the nucleophile of the recently characterized
invertebrate-type lysozyme from Tapes japonica [31]
whereas Thr45 is localized similarly to the nucleophile
of hen egg-white lysozyme [8]. Further examples are
shown in Table 2. During their evolution from a hypo-
thetical common ancestor, each group of enzymes
selected the best positions for essential catalytic resi-
dues choosing among a small number of possibilities;
optimizing their configuration to perform hydrolysis
on a particular substrate in a given condition. This
‘mosaic’ of positions remains in sharp contrast with
the invariant position occupied by the general acid
residues in the entire superfamily (Table 2).
Chitosanase as a resistance determinant against
antimicrobial action of chitosan
The finding that a heterologous chitosanase can protect
E. coli against the antimicrobial activity of chitosan is

novel and raises the possibility of a new function for
chitosanases. As described by several authors [2,23,32]
chitosan shows its maximal antimicrobial effect against
E. coli at relatively high molecular mass, whereas chito-
san oligosaccharides or short-chain chitosan forms
(< 4 kDa) have no inhibitory effect. Such a pattern is
also observed for some other bacterial species. By
shortening the chain length of chitosan, chitosanase
could function as a resistance factor against the toxic
effect of chitosan. Besides a strictly metabolic function,
consisting of the endohydrolysis of high molecular
mass chitosan into oligosaccharides that can be trans-
ported inside the cell to be used as C and N source,
chitosanase could also play the role of a stress
enzyme, protecting the microbial cells against chito-
san. This possibility deserves further studies. Formal
genetic experiments with chitosanase-producing micro-
organisms are in progress in our group.
Experimental procedures
Bacterial strains and plasmids
E. coli strains JM109 (endA1, thi, gyrA96, hsdR17
(rk
)
,mk
)
), relA1, supE44, D(lac-proAB),[F¢, traD36,
proAB, lacl
q
zDM15]) and BMH 71-18 (thi, supE, D(lac-
proAB), [mutS::Tn10] [F¢, proAB, lacI

q
zDM15]) were used
for routine plasmid propagation and as hosts in site-direc-
ted mutagenesis procedures of D40G, T45E, T45H,
D40G + T45E, T45E + T45D mutants (Promega, Madi-
son, WI, USA). E. coli strain DH5a (F) u80lacZDM15
D(lacZYA-argF)U169 recA1, endA1, hsdR17(rk), mk+)
phoA, supE44, thi-1, gyrA96, rel A1 k)) was used for rou-
tine plasmid propagation and as host in site-directed muta-
genesis procedures of E36A, E36Q and D40G + E36A.
Recombinant strains of S. lividans TK24 were used for
chitosanase production [10]. The vector pAlter-1 (for site-
directed mutagenesis of D40 and T45 mutants), the vector
pUC19 (for site-directed mutagenesis of residue E36) and
the shuttle vector pFD666 have been described previously
[10,33,34]. In some experiments, pFD ES, a smaller deriva-
tive of pFD666, kindly provided by E. Sanssouci and
C. Beaulieu, was used as vector for expression of mutated
chitosanase genes. This derivative has been obtained by
pFD666 digestion with AclI and NruI followed by intramo-
lecular ligation.
Site-directed mutagenesis
The procedure used to generate mutants D40G, T45E,
T45H, D40G + T45E and D40G + T45D has been
described previously [10]. A variant of this procedure has
been used to perform saturation mutagenesis of the T45
Table 2. Key structural motifs for active site residues in the lysozyme superfamily. GH 19, 23, 24, 46 enzymes act by inverting mechanism;
GH22 enzymes act by retaining mechanism.
Structural motif
a

CsnN174 (GH46) T4 lysozyme (GH24)
Hen egg-white
lysozyme (GH22)
T. japonica lysozyme
(i-type) (GH22)
Goose egg-white
lysozyme (GH23)
Barley chitinase
(GH19)
C-end of a-1 helix E22 (general acid) E11 (general acid) E35 (acid–base
residue)
E18 (acid–base
residue)
E73 (general acid) E67 (general
acid)
b-1 strand E36 (alternative
general base)
D30 (nucleophile) E89 (general
base)
Loop between b-1
and b2 strands
D40
b
(general base)
D20
b
(general base)
b-2 strand T45 (water
positioning)
T26 (water

positioning)
D52 (nucleophile)
b-3 strand D97 (putative
general base)
S120 (water
positioning)
a
Nomenclature as in CsnN174 [7]. See also Fig. 5A.
b
Although localized in the same loop, these two residues are not in equivalent
positions when both structures are superimposed [8].
M E
`
. Lacombe-Harvey et al. Active site residues of family 46 chitosanase
FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS 865
position: the DNA of the chitosanase gene harboring a
T45H mutation was obtained in single-stranded form and
hybridized with the oligonucleotide 5¢-CAGAAGCCGATGA
TGGCCGCCNNNGTAGCCCCGGCCGTCACCGATGT-3¢
(N representing any of the four nucleotides inserted at the
positions encoding the 45th residue). This oligonucleotide
harbored also a silent mutation abolishing the sole SacII
restriction site present in the vector. After elongation of the
second DNA strand, the resulting double-stranded plasmids
were transformed into E. coli BMH 71-18. The transfor-
mants were cultivated overnight in Luria broth with tetracy-
cline. Plasmid DNA was isolated and digested with SacII (to
linearize plasmids that did not incorporate the mutagenic oli-
gonucleotide sequence). Plasmid DNA rescued from this
digestion (highly enriched in mutated forms) was trans-

formed into E. coli JM109. After selection on tetracycline, a
library of T45-mutated E. coli transformants was collected.
Mutants of the E36 residue have been produced by a
site-directed mutagenesis method involving PCR using
Easy-A
Ò
High-Fidelity PCR Cloning Enzyme (Stratagene,
La Jolla, CA, USA) [35]. The procedure was applied to the
csnN174 gene (wild-type or D40G mutant) cloned in the
pUC19 vector in which the E36 codon was localized
between unique restriction sites BamHI and BstXI. A first
series of amplifications was performed by using a common
forward primer adjacent to the BamHI (BamHI-F, 5¢-GCT
CACTCATTAGGCACC-3¢) site and the reverse primer for
each specific mutation (E36A-R, 5¢-CCGATGTCCGCGAT
GTACTTG-3¢; E36Q-R, 5¢-CCGATGTCCTGGATGTAC
TTG-3¢). A parallel series of amplifications was performed
by using a common forward primer adjacent to BstXI site
(BstXI-F, 5¢-CTCAGCTGTTGATGAGGT-3¢) and the for-
ward primer for each specific mutation (E36A-F, 5¢-AGTA
CATCGCGGACATCGGTG-3¢; E36Q-F, 5¢-AGTACATC
CAGGACATCGGTG-3¢). After purification of the PCR
products, a second series of PCR was performed with the
same external primers. The resulting 1215 bp mutated frag-
ments were cloned between the BamHI and BstXI sites of
pFD-ES vector for expression in S. lividans. The mutated
DNA sequences were confirmed by DNA sequencing.
Revertant selection
Chitosanase-positive revertants after saturation mutagenesis
were selected on toxic chitosan medium. Chitosan (N-acety-

lation degree of 21%) was dissolved at 15 gÆL
)1
in sodium
acetate buffer pH 5.3 and hydrolyzed with CsnN174
(0.025 UÆmL
)1
) for 10 min at 37 °C. The hydrolyzate was
boiled for 30 min to stop the reaction, chilled on ice and
lyophilized. Number average molecular mass (M
n
)of
hydrolyzed chitosan was determined using the reducing sug-
ars assay of Lever [36]. The toxic chitosan medium was pre-
pared as follows: to a sterile, melted base medium
(tryptone; 10 gÆL
)1
, yeast extract; 5 gÆL
)1
, agar, 15 gÆL
)1
in
distilled water), we added (in that order, with constant gen-
tle shaking) sodium citrate buffer (5 mm final concentra-
tion, pH 6.0), NaCl (200 mm); MgSO
4
(3 mm) and chitosan
(M
n
 15 kDa; 0.3 gÆL
)1

). Salt and chitosan concentrations
were optimized according to the required level of medium
toxicity [22] as described above. In some cases, 5-bromo-4-
chloro-3-indolyl-b-d-galactopyranoside (54 lgÆ mL
)1
), iso-
propyl thio-b-d-galactoside (54 lgÆmL
)1
) and tetracycline
(15 lgÆmL
)1
) were also included.
A mixture of E. coli cells, members of the saturation
mutagenesis library, was diluted to an approximate cell
density of 5 · 10
3
ÆmL
)1
and plated (100 lLÆplate
)1
)on
chitosan medium. Plates were incubated at 37 °C and
colonies were picked up after 48–72 h. Revertant charac-
terization was completed by sequencing their chitosanase
genes.
Chitosanase purification and assay
Chitosanase and protein assays were performed as
described previously [37]. All chitosanase forms were puri-
fied from recombinant S. lividans TK24 culture super-
natants as described previously [10] except that the

gel-filtration step was replaced by the more rapid hydroxy-
apatite chromatography [15]. The CD spectra of the chito-
sanase preparations thus obtained were identical to that
of the wild-type enzyme, indicating that the global con-
formation was not significantly affected by the individual
mutations.
Chitosanase assays were performed determined using
chitosan Sigma-Aldrich (St Louis, MO, USA) (characterized
by an N-acetylation degree of 18%) as substrate at 37 °Cin
sodium acetate buffer (pH 5.5). In standard assay, a concen-
tration of 0.8 mgÆmL
)1
was used. In kinetic assays, 0.4 mL
reaction mixtures were set up containing eight different con-
centrations (0.02–0.8 mgÆmL
)1
) of chitosan in eight replicas
using microtiter plates. Protein concentration and reaction
time was adjusted to obtain the same overall hydrolysis level
for all studied proteins. Reaction time was 10 min for wild-
type, E36A, E36D, E36N and E36Q chitosanases, 20 min for
the D40G and T45S chitosanases, 50 min for D40G + E36A
chitosanase, and for 100 min for D40G + T45D and
D40G + E36Q chitosanases. Liberation of reducing sugars
was measured as described previously [37]. K
m
and k
cat
values
were calculated using the non linear least-square fitting pro-

cedure for Michaelis–Menten equation in prism software
(version 5.0 for Windows, San Diego, CA, USA).
MS set-up and signal correction
Several experiments were performed in continuous-flow
mode directly coupled with ESI-MS using a time-of-flight
mass spectrometer (Agilent, Santa Clara, CA, USA). The
analytical set-up as well as the chosen signal corrections
were as published recently [15,38].
Active site residues of family 46 chitosanase M E
`
. Lacombe-Harvey et al.
866 FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS
Enzymatic assays obtained by real-time MS
Batches were prepared in 10 mm ammonium acetate
(pH 5.2) aqueous solution containing 5.0 nm wild-type
CsnN174 or 200 nm D40G mutant, 1.0 lm malantide
(internal standard) and 25.0 lm (GlcN)
5
or (GlcN)
4
, respec-
tively. Control experiments were performed using solutions
containing 5.0 nm wild-type CsnN174 or 1.0 lm D40G
mutant, 1.0 lm internal standard and 25.0 lm (GlcNAc)
6
.
Further batches were prepared in 10 mm ammonium ace-
tate (pH 5.2) aqueous solution containing 5.0 nm wild-type
CsnN174 or 2.5 lm D40G + T45D mutant, 1.0 lm inter-
nal standard and 25.0 lm (GlcN)

6
. Several experiments
were performed in duplicate. Enzyme specific activity was
estimated from the slope of the linear portion of degrada-
tion curve of the respective oligosaccharides (GlcN)
6
or
(GlcN)
5
substrate during time course analysis.
Stereochemical course of the enzymatic reaction
The substrate (GlcN)
6
was lyophilized three times from
D
2
O, and then dissolved in 0.5 mL of 10 mm deuterated
sodium acetate buffer, pH 5.0. The substrate solution was
placed in a 5 mm NMR tube, and the enzyme (1.5 nmol)
added. The NMR tube was immediately set into the NMR
probe which was thermostatically controled at 30 °C. After
an appropriate reaction time, accumulation of
1
H-NMR
spectra was started. Each accumulation required 3 min.
The substrate concentration was 8.0 mm.
Thermal unfolding experiments
Far-UV CD spectra of the chitosanases were obtained in
20 mm sodium phosphate buffer, pH 7.0, using a Jasco
J-720 spectropolarimeter (cell length, 0.1 cm). For obtain-

ing thermal unfolding curves of the enzymes, the CD
value at 222 nm was monitored while raising the solution
temperature at a rate of 1 °CÆmin
)1
. By setting the ther-
mocouple in the cell, solution temperature was directly
measured using a DP-500 thermometer (Rikagaku Kogyo).
The same experiments were conducted in the presence of
(GlcN)
3
in order to examine the stabilization effect caused
by the trisaccharide binding to the enzymes. To facilitate
comparison between the unfolding curves obtained, the
experimental data were normalized as follows; fractions of
unfolded protein at individual temperatures were calcu-
lated from the CD value by linearly extrapolating the pre-
and post-transition baselines into the transition zone, and
plotted against temperature. Assuming that the unfolding
transition of the chitosanase follows a two-state mecha-
nism [39], the unfolding curves obtained by CD were ana-
lyzed by least square curve fitting to obtain the midpoint
temperatures (T
m
). Reversibility values for the unfolding
transition were estimated from comparison of the CD
value obtained after annealing with that obtained before
raising the temperature, and were > 75% for both
proteins tested.
Chemical rescue experiments
The enzymatic reaction of D40G toward (GlcN)

6
was mon-
itored by HPLC and refractometric detection in the pres-
ence or absence of sodium azide. The substrate solutions
(72 mm) containing sodium azide were at first prepared
with 50 mm sodium acetate buffer, and then each solution
pH was adjusted to 4.5. A small amount of D40G enzyme
solution was added to each substrate solution, and the reac-
tion mixture was incubated at 40 °C. The final concentra-
tions of sodium azide were 0, 0.22, 0.65, 1.6 and 2.6 m, and
that of D40G was 8.5 lm. When D36A + D40G was used
instead of D40G, the final concentrations of the double
mutant and sodium azide were 26 lm and 2.3 m, respec-
tively. After an appropriate incubation period, a portion of
the reaction mixture was withdrawn, and the enzymatic
reaction was terminated by mixing with an equal volume of
0.1 m NaOH. The resulting solution was applied to
the HPLC column of TSK-GEL NH
2
-60 (Tosoh,
4.6 · 250 mm), eluting with 60% acetonitrile at a flow rate
of 0.8 mLÆmin
)1
. GlcN oligosaccharides were detected with
a refractive index monitor, and the individual peak areas in
the HPLC profile were converted into molar concentrations
using the standard curves obtained by authentic saccharide
solutions. The molar concentrations of the individual oligo-
saccharides were plotted against the reaction time to obtain
the reaction time course.

Reagents
Restriction enzymes were purchased from New England
Biolabs (Beverly, MA, USA). Chitosan was from Sigma-
Aldrich (St Louis, MO, USA). Chitosan oligosaccharides
(GlcN, GlcN
n
, n = 4–6) and chitin oligosaccharide (Glc-
NAc
6
) were purchased from Seikagaku Kogyo Co. (Tokyo,
Japan). Malantide (internal standard, ‡ 97%) was purchased
from Sigma-Aldrich (Steinheim, Germany). Ammonium
acetate (> 98%) was obtained from Merck (Darmstadt,
Germany), and high-purity water was from a Milli-Q system
(Millipore, Eschborn, Germany). All the other reagents and
enzyme substrates were of analytical grade and are commer-
cially available. Culture media components were obtained
from Difco (Mississauga, Canada).
Acknowledgements
Work at Universite
´
de Sherbrooke was supported by a
Discovery grant from the Natural Science and Engi-
neering Research Council of Canada to RB. M-E
`
L-H
is a recipient of a doctoral student fellowship from les
Fonds Que
´
becois de la Recherche sur la Nature et les

M E
`
. Lacombe-Harvey et al. Active site residues of family 46 chitosanase
FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS 867
Technologies. We thank Isabelle Boucher and Hugo
Tremblay for assistance in studies of chitosanase
mutants. Work at Kinki University was supported in
part by Academic Frontier Project for Private Univer-
sities, matching fund subsidy from MEXT, 2004–2008.
Work at the TU Mu
¨
nchen was supported in part by a
grant from the Vereinigung zur Fo
¨
rderung der Milch-
wissenschaftlichen Forschung an der TUM in Freising-
Weihenstephan e.V. and from the Bund der Freunde
der Technischen Universita
¨
tMu
¨
nchen e. V.
References
1 Boucher I, Dupuy A, Vidal P, Neugebauer WA & Brze-
zinski R (1992) Purification and characterization of a
chitosanase from Streptomyces N174. Appl Microbiol
Biotechnol 38, 188–193.
2 Blanchard J, Park JK, Boucher I & Brzezinski R (2003)
Industrial applications of chitosanases. In Recent
Advances in Marine Biotechnology, Biomaterials and Bio-

processing (Fingerman M & Nagabhushanam R, eds)
vol. 9, pp. 257–277. Science Publishers, Enfield, UK.
3 Fukamizo T, Honda Y, Goto S, Boucher I & Brzezinski
R (1995) Reaction mechanism of chitosanase from
Streptomyces sp. N174. Biochem J 311, 377–383.
4 Saito J, Kita A, Higuchi Y, Nagata Y, Ando A & Miki
K (1999) Crystal structure of chitosanase from Bacillus
circulans MH-K1 at 1.6 A
˚
resolution and its substrate
recognition mechanism. J Biol Chem 274, 30818–30825.
5 Coutinho PM & Henrissat B (1999) Carbohydrate-
active enzymes: an integrated database approach. In
Recent Advances in Carbohydrate Bioengineering (Gil-
bert HJ, Davies B, Henrissat B & Svensson B, eds), pp.
3–12. Royal Society of Chemistry, Cambridge.
6 Holm L & Sander C (1994) Structural similarity of
plant chitinase and lysozymes from animals and phage.
An evolutionary connection. FEBS Lett 340, 129–132.
7 Marcotte E, Monzingo AF, Ernst SR, Brzezinski R &
Robertus JD (1996) X-ray structure of an anti-fungal
chitosanase from Streptomyces N174. Nat Struct Biol 3,
155–162.
8 Monzingo AF, Marcotte EM, Hart PJ & Robertus JD
(1996) Chitinases, chitosanases, and lysozymes can be
divided into prokaryotic and eukaryotic families sharing
a conserved core. Nat Struct Biol 3, 133–140.
9 Fukamizo T (2000) Chitinolytic enzymes: catalysis,
substrate binding, and their application. Curr Protein
Peptide Sci 1, 105–124.

10 Boucher I, Fukamizo T, Honda Y, Willick GE, Neuge-
bauer WA & Brzezinski R (1995) Site-directed muta-
genesis of evolutionary conserved carboxylic amino
acids in the chitosanase from Streptomyces sp. N174
reveals two residues essential for catalysis. J Biol Chem
270, 31077–31082.
11 Vocadlo DJ, Davies GJ, Laine R & Withers SG (2001)
Catalysis by hen egg-white lysozyme proceeds via a
covalent intermediate. Nature 412, 835–838.
12 Weaver LH, Gru
¨
tter MG & Matthews BW (1995) The
refined structures of goose lysozyme and its complex
with a bound trisaccharide show that the ‘goose-type’
lysozymes lack a catalytic aspartate residue. J Mol Biol
245, 54–68.
13 Ly HD & Withers SG (1999) Mutagenesis of glycosid-
ases. Annu Rev Biochem 68, 487–522.
14 Brameld KA & Goddard WA III (1998) The role of
enzyme distortion in the single displacement mechanism
of family 19 chitinases. Proc Natl Acad Sci USA 95,
4276–4281.
15 Dennhart N, Fukamizo T, Brzezinski R, Lacombe-Har-
vey M-E
`
& Letzel T (2008) Oligosaccharide hydrolysis by
chitosanase enzymes monitored by real-time electrospray
ionization-mass spectrometry.
J Biotechnol 134, 253–260.
16 Letzel T (2008) Real-time mass spectrometry in enzy-

mology. Anal Bioanal Chem 390, 257–261.
17 Wang Q, Graham RW, Trimbur D, Warren RAJ &
Withers SG (1994) Changing enzymatic reaction mecha-
nisms by mutagenesis – conversion of a retaining gluco-
sidase into an inverting enzyme. J Am Chem Soc 116 ,
11594–11595.
18 Miyake H, Otsuka C, Nishimura S & Nitta Y (2002)
Catalytic mechanism of b -amylase from Bacillus cereus
var. mycoides: chemical rescue of hydrolytic activity for
a catalytic site mutant (Glu367 fi Ala) by azide. J Bio-
chem 131, 587–591.
19 Collins T, De Vos D, Hoyoux A, Savvides SN, Gerday
C, Van Beeumen J & Feller G (2005) Study of the
active site residues of a glycoside hydrolase family 8
xylanase. J Mol Biol 354, 425–435.
20 Kezuka Y, Ohishi M, Itoh Y, Watanabe J, Mitsutomi
M, Watanabe T & Nonaka T (2006) Structural studies
of a two-domain chitinase from Streptomyces griseus
HUT6037. J Mol Biol 358, 472–484.
21 Hart PJ, Pfluger HD, Monzingo AF, Hollis T & Rober-
tus JD (1995) The refined crystal structure of an endo-
chitinase from Hordeum vulgare L. seeds at 1.8 A
˚
resolution. J Mol Biol 248, 402–413.
22 Tsai GJ & Su WH (1999) Antibacterial activity of
shrimp chitosan against Escherichia coli. J Food Protect
62, 239–243.
23 Ueno K, Yamaguchi T, Sakairi N, Nishi N & Tokura S
(1997) Antimicrobial activity by fractionated chitosan
oligomers. In Advances in Chitin Science, vol. II (Dom-

ard A, Roberts GAF & Va
˚
rum KM, eds), pp. 156–161.
Jacques Andre
´
, Lyon, France.
24 Vriend G (1990) WHAT IF: a molecular modeling and
drug design program. J Mol Graph 8, 52–56.
25 Hoell IA, Dalhus B, Heggset EB, Aspmo SI & Eijsink
GH (2006) Crystal structure and enzymatic properties
Active site residues of family 46 chitosanase M E
`
. Lacombe-Harvey et al.
868 FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS
of a bacterial family 19 chitinase reveal differences from
plant enzymes. FEBS J 273, 4889–4900.
26 Akiyama K, Fujita T, Kuroshima K, Sakane T, Yokota
A & Takata R (1999) Purification and gene cloning of a
chitosanase from Bacillus ehimensis EAG1. J Biosci Bio-
eng 87, 383–385.
27 Kuroki R, Weaver LH & Matthews BW (1995) Struc-
ture-based design of a lysozyme with altered catalytic
activity. Nat Struct Biol 2, 1007–1011.
28 Kuroki R, Weaver LH & Matthews BW (1993) A cova-
lent enzyme–substrate intermediate with saccharide
distortion in a mutant T4 lysozyme. Science 262,
2030–2033.
29 Zechel DL & Withers SG (2001) Dissection of nucleo-
philic and acid–base catalysis in glycosidases. Curr Opin
Chem Biol 5, 643–649.

30 Sauer J, Sigurskjold BW, Christensen U, Frandsen TP,
Mirgorodskaya E, Harrison M, Roepstorff P & Svens-
son B (2000) Glucoamylase: structure ⁄ function relation-
ships, and protein engineering. Biochem Biophys Acta
1543, 275–293.
31 Goto T, Abe Y, Kakuta Y, Takeshita K, Imoto T &
Ueda T (2007) Crystal structure of Tapes japonica lyso-
zyme with substrate analogue. J Biol Chem 282, 27459–
27467.
32 Rabea EI, Badawy ME-T, Stevens CV, Smagghe G &
Steurbaut W (2003) Chitosan as antimicrobial agent:
applications and mode of action. Biomacromolecules 4,
1457–1465.
33 Denis F & Brzezinski R (1992) A versatile shuttle
cosmid vector for use in Escherichia coli and actino-
mycetes. Gene 111, 115–118.
34 Messing J & Vieira J (1982) A new pair of M13 vectors
for selecting either DNA strand of double-digest restric-
tion fragments. Gene 19, 269–276.
35 Ho SN, Hunt HD, Horton RM, Pullen JK & Pease LR
(1989) Site-directed mutagenesis by overlap extension
using the polymerase chain reaction. Gene 77, 51–59.
36 Lever M (1972) A new reaction for colorimetric deter-
mination of carbohydrates. Anal Biochem 47, 273–279.
37 Katsumi T, Lacombe-Harvey M-E
`
, Tremblay H, Brze-
zinski R & Fukamizo K (2005) Role of acidic amino
acid residues in chitooligosaccharide-binding to Strep-
tomyces sp. N174 chitosanase. Biochem Biophys Res

Commun 338, 1839–1844.
38 Dennhart N & Letzel T (2006) Mass spectrometric
real-time monitoring of enzymatic glycosidic hydrolysis,
enzymatic inhibition and enzyme complexes. Anal
Bioanal Chem 386, 689–698.
39 Honda Y, Fukamizo T, Boucher I & Brzezinski R
(1997) Substrate binding to inactive mutants of Strepto-
myces sp. N174 chitosanase: indirect evaluation from
the thermal unfolding experiments. FEBS Lett
411,
346–350.
M E
`
. Lacombe-Harvey et al. Active site residues of family 46 chitosanase
FEBS Journal 276 (2009) 857–869 ª 2009 The Authors Journal compilation ª 2009 FEBS 869