Báo cáo khoa học: Role of active-site residues of dispersin B, a biofilm-releasing b-hexosaminidase from a periodontal pathogen, in substrate hydrolysis pptx

Role of active-site residues of dispersin B, a
biofilm-releasing b-hexosaminidase from a periodontal
pathogen, in substrate hydrolysis
Suba G. A. Manuel, Chandran Ragunath, Hameetha B. R. Sait, Era A. Izano, Jeffrey B. Kaplan
and Narayanan Ramasubbu
Department of Oral Biology, University of Medicine and Dentistry of New Jersey, Newark, NJ, USA
The biofilm attachment and detachment properties of
the oral pathogen Aggregatibacter actinomycetemcomi-
tans (formerly Actinobacillus actinomycetemcomitans)
are mediated by a soluble b-N-acetylglucosaminidase
(dispersin B abbreviated to DspB; EC 3.2.1.52) [1].
Interestingly, DspB exhibits biofilm-detachment activ-
ity not only for A. actinomycetemcomitans, but also
for biofilms produced by several bacterial species
including Gram-positive species (Staphylococcus epide-
rmidis) [2] and Gram-negative species (Actinobacillus
Keywords
biofilm detachment; dispersin B; enzyme
mechanism; hydrolysis; site-directed
mutagenesis
Correspondence
N. Ramasubbu, Department of Oral Biology,
C-634, MSB, UMDNJ, 185 South Orange
Ave, Newark NJ 07103, USA
Fax: +1 973 972 0045
Tel: +1 973 972 0704
E-mail:
(Received 28 August 2007, revised 30
September 2007, accepted 1 October 2007)
doi:10.1111/j.1742-4658.2007.06121.x
Dispersin B (DspB), a family 20 b-hexosaminidase from the oral pathogen

Aggregatibacter actinomycetemcomitans, cleaves b(1,6)-linked N-acetylglu-
cosamine polymer. In order to understand the substrate specificity of
DspB, we have undertaken to characterize several conserved and noncon-
served residues in the vicinity of the active site. The active sites of DspB
and other family 20 hexosaminidases possess three highly conserved acidic
residues, several aromatic residues and an arginine at subsite )1. These res-
idues were mutated using site-directed mutagenesis and characterized for
their enzyme activity. Our results show that a highly conserved acid pair in
b-hexosaminidases D183 and E184, and E332 play a critical role in the
hydrolysis of the substrates. pH activity profile analysis showed a shift to a
higher pH (6.8) in the optimal activity for the E184Q mutant, suggesting
that this residue might act as the acid ⁄ base catalyst. The reduction in k
cat
observed for Y187A and Y278A mutants suggests that the Y187 residue
(unique to DspB) located on a loop might play a role in substrate specific-
ity and be a part of subsite +1, whereas the hydrogen-bond interaction
between Y278A and the N-acetyl group might help to stabilize the transi-
tion state. Mutation of W237 and W330 residues abolished hydrolytic
activity completely suggesting that alteration at these positions might col-
lapse the binding pocket for the N-acetyl group. Mutation of the conserved
R27 residue (to R27A or R27K) also caused significant reduction in k
cat
suggesting that R27 might be involved in stabilization of the transition
state. From these results, we conclude that in DspB, and possibly in other
structurally similar family 20 hydrolases, some residues at the active site
assist in orienting the N-acetyl group to participate in the substrate-assisted
mechanism, whereas other residues such as R27 and E332 assist in holding
the terminal N-acetylglucosamine during the hydrolysis.
Abbreviations
DspB, dispersin B; Hex A, human b-hexosaminidase; MuGlcNAc, 4-methylumbelliferyl-b-

D-N-Acetylglucosamine; PGA, homopolymer of
b(1,6)-linked GlcNAc; pNPGlcNAc, p-nitrophenyl-b-
D-N-acetylglucosamine; SpHex, S. plicatus hexosaminidase.
FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS 5987
pleuropneumoniae; Escherichia coli, Yersinia pestis and
Pseudomonas fluorescens) [3,4]. DspB belongs to fam-
ily 20 of the glycoside hydrolase classification scheme,
members of which are exo-acting hexosaminidases that
cleave terminal monosaccharide residues from the non-
reducing end. The substrate of DspB is a hexosamine-
containing matrix polysaccharide consisting of a linear
homoglycan of N-acetyl-d-glucosamine residues in
b(1,6)-linkages (PGA) [5–9]. DspB catalytic activity
has been investigated using PGA from E. coli, which
showed that the reaction products included GlcNAc
and other unassigned GlcNAc oligomers [4]. In an ear-
lier study, we reported the crystal structure of DspB in
complex with glycerol and an acetate at the active site,
which suggested that the active-site architecture of
DspB is similar to that of the family 20 b-hexosami-
nidases [10]. However, very little information is avail-
able on the mechanistic aspects of DspB.
It has been proposed that family 20 and the closely
related family 18 b-hexosaminidases follow a retaining
mechanism (Fig. 1) with anchimeric assistance pro-
vided by the C2-acetamido group of the substrate and
an acidic active-site residue acting as the catalytic acid
[11,12]. A unique feature of this mechanism is the acet-
amido group acting as the nucleophile in catalysis.
Interestingly, a substrate-assisted mechanism is opera-

tive in other sequentially dissimilar but functionally
similar enzymes such as family 56 hyaluronid-
ase (EC 3.2.1.35) and family 84 O-GlcNAcase
(EC 3.2.1.52) [13,14]. Using the crystal structure of
DspB enzyme [10], we identified residues at subsite )1
of the active site that are likely to be involved in catal-
ysis and ⁄ or substrate binding. We generated a struc-
ture-based sequence alignment through superposition
of the active-site regions of the reported crystal struc-
tures of b-hexosaminidases [15–18] and noted that, in
addition to the well-conserved pair of acidic residues
D–E at positions 183 and 184 (DspB numbering),
acidic residue E332 is also structurally conserved
(Fig. 2A,B). Juxtaposition of this residue appears to be
suitable for exerting a significant role in substrate
binding. In b-hexosaminidases, highly conserved tryp-
tophan residues at subsite )1 create a binding pocket
for the terminal GlcNAc, especially the N-acetyl
group. Because there is a clear difference in the sub-
strate specificity between DspB [acts on b(1,6)-linked
poly(N-acetylglucosamine)] and the other members of
the family 20 b-hexosaminidases [act on b(1,4)-linked
N-acetylglucosamine], we tested the roles of conserved
and nonconserved residues at the active site using site-
directed mutagenesis. Our results demonstrate that: (a)
DspB follows a substrate-assisted mechanism proposed
for the family 20 b-hexosaminidases with participation
of the conserved acidic pair D183 and E184; (b) the
aromatic residues W237 and Y278 along with D183,
play a role in hydrolysis and assist the N-acetyl group

to properly orient itself to take part in catalysis; (c)
the aromatic residue Y187, which is part of a loop at
subsite +1, might play a role in the specificity exhib-
ited by DspB; and (d) three residues, R27, E332 and
W330 interacting with the GlcNAc at the C4 side pro-
vide transition state stabilization during hydrolysis.
Results
Mutations were performed using the primers listed in
Table 1. Wild-type DspB and the mutants were gen-
erated with a His6 tag to facilitate purification using
metal-affinity chromatography as described previously
[10]. SDS ⁄ PAGE analysis of the cell lysates indicated
that all the mutant enzymes were present in a solu-
ble form and were expressed at levels comparable
with wild-type DspB. All proteins were purified to
homogeneity using Ni
2
+
-affinity chromatography
with yields ranging from 20 to 30 mgÆL
)1
of culture.
Our earlier studies had shown that the half-life of
DspB was  3–4 h at 37 °C [19]. However, DspB
was stable and retained its activity for several
months when stored in buffer comprising 50 mm
sodium phosphate, pH 5.8, 50 mm NaCl and 50%
glycerol at )20 °C. Purified enzymes were stored in
this buffer until further use.
Fig. 1. Hydrolysis mechanism proposed for

family 20 b-hexosaminidases. In this
substrate-assisted mechanism one acidic
residue, Glu184, acts as the acid ⁄ base. The
nucleophile is the N-acetyl group of the
substrate and is assisted by Asp183.
Active site of dispersin B S. G. A. Manuel et al.
5988 FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS
Role of active-site acidic residues
Comparison of the amino acid sequence and crystal
structure of DspB with other family 20 b-hexosaminid-
ases clearly suggests that three acidic residues, D183,
E184 and E332, are highly conserved (Fig. 2B).
Among these, the residues equivalent to D183 and
E184 in other family 20 b-hexosaminidases have been
shown to play a critical role in the catalytic reaction
[11,12,17]. Thus, the E184 equivalent has been
proposed to act as the acid ⁄ base catalyst, whereas the
D183 equivalent has been suggested to orient the
N-acetyl group [11]. In general, in family 20 b-hexos-
aminidases, the E184 equivalent is juxtaposed close to
the anomeric center, whereas the D183 equivalent is
well positioned to interact with the N- of the acetam-
ido group. Many crystal structures also show that a
glutamate residue (equivalent to E332) is located in
close proximity to the C3- and C4-hydroxyl groups on
the opposite side of the bound sugar at subsite )1. To
investigate the proposed role of the D183 and E184
acidic residues and the potential role of E332 in DspB,
these residues were replaced by either N or Q and their
biochemical properties analyzed using the substrate

p-nitrophenyl-b-d-N-acetylglucosamine (pNPGlcNAc).
All three acidic mutants display significantly reduced
specific activities as shown in Table 2. The specific
activity of the catalytic residue mutant E184Q was
122-fold less, and that of D183N was 11 000-fold
lower than wild-type. The observed results for the
mutant E184Q are entirely consistent with loss of the
acid–base residue, whereas the significant loss in activ-
ity exhibited by D183N is consistent with its proposed
role as a residue that activates ⁄ orients the N-acetyl
group to act as a nucleophile [11]. Inhibition studies
were carried out with the mechanism-based inhibitor
NAG-thiazoline [17]. IC
50
for inhibition was found to
be 147 ± 5 lm compared with 47 ± 4 lm for the jack
bean enzyme. Inhibition of DspB by NAG-thiazoline
is consistent with an anchimeric assistance mechanism
occurring during catalysis in DspB because NAG-
thiazoline is a mechanism-based inhibitor of b-hexos-
aminidases. The  2700-fold lower specific activity
Glu332
Glu184
Asp183
A
B
Fig. 2. Sequence and structure alignment of b-hexosaminidases
with DspB. (A) Multiple amino acid sequence alignment of DspB
and b-hexosaminidases whose crystal structures have been
reported. Subsite )1 residues are indicated by an asterisk (*). Note

the conserved residues WXE near the C-terminus corresponding to
W330 and E332. Although all four aromatic residues, W216, W237,
Y278 and W330 are conserved, the sequence homology in the
region around Y187 of DspB is weak. This region might correspond
to residues in subsite +1 and contribute to the substrate specificity
being b(1,6) in DspB. Sequences shown are: DspB, 1yht [10]; chito-
biase from Serratia marcescens 1qbb [20]; b-hexosaminidase from
Streptomyces plicatus, 1hp5 [17]; human lysosomal b-hexosamini-
dase isoform b, 1now [16]; human b-hexosaminidase from placenta,
b-chain, 2gjx [30]. The secondary structure as observed in DspB is
shown on top. Arrows represent b strands and cylinders represent
helices. (B) Superposition of the various crystal structures of b-
hexosaminidases using the suite
SWISS-PDBVIEWER [31]. The fit of
D183, E184 and E332 residues in the proximity of the subsite )1is
shown. Structures shown are: DspB; PDB Code: 1yht (gray) [10];
PDB Code: 2gjx (red) [30]; PDB Code: 1hp5 (green) [17]; PDB Code:
1now (blue) [16]; PDB Code: 1qbb (yellow) [20].
S. G. A. Manuel et al. Active site of dispersin B
FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS 5989
exhibited by E332Q suggests that this residue plays a
significant role in catalysis. The location of this resi-
due, although away from the anomeric carbon, might
be critical in providing necessary stabilization in the
transition state while the terminal GlcNAc is undergo-
ing conformational changes during catalysis (Fig. 2B).
To further substantiate our observations, kinetic
parameters were determined for the mutants using
pNPGlcNAc as substrate (Table 2). All three mutants
exhibited substantially lower K

m
and significantly
reduced k
cat
values resulting in lower catalytic effi-
ciency (k
cat
⁄ K
m
) for the mutant enzymes (71-fold lower
for E184Q, 13 333-fold lower for D183N and 2000-
fold lower for E332Q) compared with wild-type. All
three mutants required higher enzyme concentrations
to exhibit measurable enzyme activity compared with
Table 1. Primers used for mutational analysis of the active-site residues.
Mutation Primer sequence (5¢-to3¢) Type
R27A CTGGACATCGCC
GCGCATTTTTATTCAC Forward
GTGAATAAAAATG
CGCGGCGATGTCCAG Reverse
R27K GCTGGACATCGCC
AAACATTTTTATTCACCCG Forward
CGGGTGAATAAAAATG
TTTGGCGATGTCCAGC Reverse
D183N GGTGGC
AACGAATTTGGTTATTCTGTGG Forward
CCACAGAATAACCAAATTC
GTTGCCACC Reverse
E184Q GGTGGCGAT
CAATTTGGTTATTCTGTGG Forward

CCACAGAATAACCAAA
TTGATCGCCACC Reverse
Y187A GATGAATTTGGT
GCGTCTGTGGAAAG Forward
CTTTCCACAGA
CGCACCAAATTCATC Reverse
W237A CCGAATATTGAAATTACTTAT
GCGAGCTATGATGGCG Forward
CGCCATCATAGCT
CGCATAAGTAATTTCAATATTCGG Reverse
Y278A CCTATTATCTT
GCGATTGTTCCGAAAGC Forward
GCTTTCGGAACAAT
CGCAAGATAATAGG Reverse
W330Y GCAGCATTATCGATC
TACGGAGAAGATGC Forward
GCATCTTCTCC
GTAGATCGATAATGCTGC Reverse
E332Q CATTATCGATCTGGGGA
CAAGATGCAAAAGC Forward
GCTTTTGCATC
TTGTCCCCAGATCGATAATG Reverse
Table 2. Specific activities and kinetic parameters of DspB and its variants for hydrolysis with pNPGlcNAc. Kinetic assays were performed
with pNPGlcNAc as described in Experimental procedures. Reactions were carried out in NaCl ⁄ P
i
at 37 °C. The molar absorptivity of p-nitro-
phenol used 9900 (
M
)1
Æcm

)1
) at 405 nm. Results are the average of three independent experiments. Standard errors for the values of appar-
ent K
m
and k
cat
are given in parentheses. ND, not determined because the reaction was too slow to be detected.
Enzyme
Conc
(l
M) Specific activity
k
cat
(min
)1
)
K
M
(mM)
k
cat
⁄ K
M
(min
)1
ÆmM
)1
)
Wild-type
pNPGlcNAc 0.1 536 (34) 55.6 (2.7) 46.2 (2.7) 1.20

MuGlcNAc 4 – 1.41 (0.2) 0.39 (0.04) 3.61
D183N
a
25 0.05 (0.01) 0.002 (0.001) 19.4 (2.5) 0.00009
E184Q
a
1 4.4 (0.8) 0.073 (0.001) 3.7 (0.5) 0.017
E332Q
a
12 0.2 (0.2) 0.005 (0.001) 6.5 (2.3) 0.0006
W237A
a
16 ND ND ND ND
W330Y
a
9ND ND ND ND
Y187A
pNPGlcNAc 1 16.8 (4.4) 3.6 (0.2) 101.1 (6.7) 0.03
MuGlcNAc 4 – 0.43 0.59 0.73
Y278A
pNPGlcNAc 1 3.1 (1.3) 0.25 (0.007) 35.9 (1.4) 0.0013
MuGlcNAc 4 – 0.09 1.13 0.08
R27A
a
6.0 0.60 (0.2) 0.023 (0.06) 14.5 (1.7) 0.0005
R27K
a
5.5 0.31 (0.2) 0.05 (0.01) 76.2 (11) 0.0007
a
No measurable activity for the substrate MuGlcNAc.

Active site of dispersin B S. G. A. Manuel et al.
5990 FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS
wild-type (E184Q, 1 lm; E332Q, 12 lm; and D183N,
25 lm compared with 0.1 lm for wild-type; Table 2).
In addition, we also used another substrate, 4-methyl-
umbelliferyl-b-d-N-acetylglucosamine (MuGlcNAc), to
study the effect of leaving aglycone (Table 2). Interest-
ingly, this substrate was also poorly hydrolyzed by the
wild-type enzyme and required a higher enzyme con-
centration (4 l m), compared with 0.1 lm for pNPGlc-
NAc. There was no measurable activity for the mutant
enzymes, D183N, E184Q and E332Q.
The pH activity profiles of the wild-type and variant
enzymes were also determined using pNPGlcNAc and
are shown in Fig. 3. The wild-type enzyme exhibited
an approximately bell-shaped profile with a pH opti-
mum of 5.8, which might be interpreted as resulting
from the ionization of two residues with pK
a
values on
either side of the maximum. The apparent pK
a
value
for the basic limb corresponding to the acid ⁄ base of
the wild-type profile was estimated to be 6.8 ± 0.1.
Interestingly, the pH optimum for the E184Q variant
is shifted to 6.8 with an apparent pK
a
of 8.3 ± 0.1 for
the basic limb consistent with the substitution of E184

carboxyl group to an amide. The pH activity profile of
E184Q mutant supports its presumed role in catalysis
as the acid ⁄ base (Fig. 1). The extremely low activity of
D183N and E332Q variants for the pNPGlcNAc
substrate prevented us from obtaining a pH profile for
these mutants.
Role of aromatic residues at the active site
One of the characteristics of family 20 b-hexosaminid-
ases is the substrate-binding pocket provided by a
number of conserved tryptophan residues. As shown
in Fig. 2A, W216, W237 and W330 are well con-
served and together create a hydrophobic pocket into
which the N-acetyl group of the GlcNAc moiety at
subsite )1 binds [15]. Together with the acidic resi-
dues, such binding juxtaposes the acetamido group
for neighboring-group participation and enables intra-
molecular nucleophilic attack at the anomeric center
of the sugar molecule at subsite )1 [16]. Residue
Y278 occupies a space close to the bound acetyl
group and in several crystal structures, including
DspB, is known to interact with the O of the acetyl
group via a hydrogen bond. Residue Y187 is unique
to DspB in that it is located in a mobile loop that is
disordered in the crystal structure (PDB code 1YHT)
[10] and appears to be present at the entrance to the
catalytic site. To investigate the roles of these aro-
matic residues in DspB, we designed mutations at
these positions (Table 1) and the results of the bio-
chemical analysis are given in Table 2. Clearly, muta-
tion of the tryptophan residues at positions 237

(W237A) and 330 (W330Y), which create the N-ace-
tyl-binding pocket, had a drastic effect on enzyme
activity. We observed that even at concentrations of
9–16 lm, enzyme activity was undetectable using
pNPGlcNAc as the substrate (Table 2). Both
mutants, W237A and W330Y, were also ineffective
when MuGlcNAc was used as a substrate.
Mutation of the two tyrosines at the active site
(Y187 and Y278) showed lower specific activity than
the wild-type (Table 2) by as much as 33-fold (Y187A)
or 176-fold (Y278A). Mutant Y187A has a higher K
m
and a lower k
cat
in the hydrolysis of pNPGlcNAc
(Table 2). The lower k
cat
⁄ K
m
value (40-fold less for
pNPGlcNAc and 4.9-fold less MuGlcNAc) for Y187A
suggests that the mutation at Y187 affects substrate
binding. Involvement of this residue, which is farther
away from subsite )1(6A
˚
), suggests that DspB might
have multiple subsites for substrate binding. Mutant
Y278A, by contrast, showed a greater decrease in
k
cat

⁄ K
m
value (923-fold for pNPGlcNAc and 45-fold
for MuGlcNAc) compared with wild-type suggesting
that the hydrogen bond to the acetamido group is sig-
nificant for enzyme activity. Because this hydrogen
bond involves the hydroxyl group of Y278 and car-
bonyl oxygen of the acetamido group, the reduction in
the kinetic parameters suggests that Y278 might partic-
ipate in orienting the acetyl group in conjunction with
the D183 residue.
4 5 6 7 8 9
0
50
100
150
200
250
300
wild type
E184Q
R27K
pH
Activity
5 6 7 8 9
0
3
6
9
12

15
R27K
E184Q
pH
Activity
Fig. 3. pH dependence of wild-type DspB and mutants. The pH
profile was measured using pNPGlcNAc as the substrate at pH val-
ues of 3.5–10.0. The solid lines refer to the fitted curve through the
data points using a Gaussian form using GraphPad
PRISM3. (Inset)
Bell-shaped curve for E184Q showing the shift in optimum pH for
the mutant. Activity values were arbitrarily scaled by a factor of 10
to show the Gaussian fit of the profile. The pH profile for the
mutants of D183N and E332Q could not be measured due to their
very low activities. Note that no pH shift was observed for the
R27K mutant.
S. G. A. Manuel et al. Active site of dispersin B
FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS 5991
Role of R27 at the active site
b-Hexosaminidases have one arginine at the active site,
equivalent to R27 of DspB, that is structurally con-
served and involved in substrate binding [11,12].
Crystal structure analyses of hexosaminidases with
substrate ⁄ inhibitor complexes have shown that R27-
equivalent arginine enters into a hydrogen-bonding
interaction with C3 and C4 oxygen atoms. In the crys-
tal structure of DspB with glycerol at the active site, it
was observed that R27 hydrogen bonds with the
hydroxyl groups of glycerol [10]. As shown in Table 2,
a conservative mutation at position 27 of DspB

(R27K) affects the activity less than a nonconservative
mutation (R27A). Nevertheless, both substitutions ren-
dered the enzyme significantly less active with a 2400-
fold reduction in k
cat
⁄ K
m
for the R27K mutant
(5.5 lm) and a 1714-fold reduction for R27A (6.0 lm).
Both mutants were ineffective against MuGlcNAc.
Biofilm detachment and hydrolytic activities
of DspB and its mutants
We previously used the biofilm detachment assay in
which the substrate is the PGA polymer to study the
activity of DspB on biofilms of various bacteria
[1,2,19]. The PGA polymer from S. epidermidis is a
natural substrate with a chain length of  130 GlcNAc
residues in b(1,6)-linkages [5]. We used this biofilm-
detachment assay to validate the hydrolytic activities
of the active-site mutants using biofilm from S. epide-
rmidis. As shown in Fig. 4A, the E184Q mutant
showed very low activity that was not measurable even
at an enzyme concentration of 1.4 lm. Although both
D183N and E332Q variants are less effective than
E184Q in the hydrolysis of pNPGlcNAc, these variants
behaved differently towards the detachment of S. epi-
dermidis biofilm. Both D183N and E332Q mutants
exhibited biofilm-detachment activity [time required
for 50% detachment, T
50

, is 7.2 min for E332Q at
0.055 lm (163-fold lower) and 8.0 min for D183N at
0.275 lm (28-fold lower than wild-type; Table 3)].
Mutants W237A and W330Y exhibited biofilm-detach-
ment activity with a T
50
of  1 min for W330Y and
 9 min for W237A, each at an enzyme concentration
of 0.275 lm (Fig. 4B). In contrast to pNPGlcNAc
hydrolysis, the biofilm-detachment activity of the tyro-
sine mutants is not severely affected (Fig. 4C). Thus,
the T
50
for the mutant Y187A is 5.5 min at 0.022 lm
and for the Y278A mutant is  2.8 min at 0.055 lm.
0
1
2
3
4
5
ABE
CD
E332Q (0.055 µM)
E184Q (1.375 µM)
wild type (0.011 µM)
D183N (0.275 µM)
Absorbanceat 590 nmAbsorbanceat 590 nm
Absorbanceat 590 nm Absorbanceat 590 nm
0

1
2
3
4
5
W237A (0.275 µ
M)
Wild type (0.011 µM)
W330Y (0.275 µM)
0
1
2
3
4
5
R27A (0.056 µM)
R27K (0.055 µM)
Wild type (0.011 µM)
0 1 2 3 4 5 6 7 8 9 10 11
0
1
2
3
4
5
Y278A (0.055 µM)
Y187A (0.022 µM)
wild type (0.011 µM)
Time (minutes)
01234567891011

Time (minutes)
012345678910
Time (minutes)
012345678910
Time (minutes)
DspB
Y187A
W237A
Y278A
W330Y
D183N
E184Q
E332Q
R27A
R27K
0
10
20
30
Efficiency
0
25
50
75
100
GlcNAc
Enzyme
GlcNAc released (nmol)
% Relative Efficiency
Fig. 4. S. epidermidis biofilm detachment activity of DspB and its mutants. (A) Biofilm detachment activity of DspB and its acidic variants

D183N, E184Q and E332Q. Biofilms were treated with NaCl ⁄ P
i
or the enzyme variant for the indicated time intervals and assayed in 96-well
microtiter plates [1,2,19]. Biofilms were then rinsed and stained with crystal violet. The amount of bound crystal violet dye which is propor-
tional to the biofilm mass was quantitated by measuring its absorbance at 590 nm. (B) Biofilm detachment activity of the aromatic residue
mutants W237A and W330Y. (C) Biofilm detachment activity of the aromatic residue mutants Y187A and Y278A. (D) Biofilm detachment
activity of the mutants R27A and R27K. Values plotted correspond to the mean absorbance of triplicate wells. Error bars indicate standard
deviations calculated using the program
PRISM3 (GraphPad). Legends include the final concentrations of the enzymes used in the detachment
assay. (E) Plot of relative efficiency of biofilm detachment (shown on the right y-axis) and the amount of GlcNAc generated (left y-axis) by
the various mutants in comparison with wild-type DspB.
Active site of dispersin B S. G. A. Manuel et al.
5992 FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS
Interestingly, the two mutants R27A and R27K
showed biofilm-detachment activity at concentrations
0.055 lm with T
50
 5–7 min, which is 20–36-fold less
efficient than the wild-type (Fig. 4D). The relative bio-
film-detachment activities of the mutants were com-
pared with the amount of GlcNAc generated due to
hydrolysis of PGA in the biofilm (Fig. 4E). One nota-
ble outcome from the GlcNAc-releasing activities of
the R27 mutants is that R27K exhibited much higher
activity than R27A. Nevertheless, mutation of residues
that interact with the acetyl group (W237, W330,
D183, Y278) or the catalytic residue E184 showed low
activity in both biofilm detachment and pNPGlNAc
hydrolysis.
Discussion

It is generally accepted that family 20 hydrolases such
as chitobiase from Serratia marcescens and human
b-hexosaminidases operate via a retaining mechanism
[17]. In such a mechanism, a general acid ⁄ base catalyst
plays a dual role by first protonating the departing
aglycone and then deprotonating the incoming water
molecule (Fig. 1). Because of a lack of nucleophile
derived from protein, family 20 b-hexosaminidases and
other functionally similar enzymes such family 56
hyaluronidase and family 84 O-GlcNAcase utilize the
C2-acetamido group to help stabilize the incipient
oxazolinium ion at the anomeric carbon [13,14,17,20].
Formation of the cyclic oxazolinium ion intermediate
is assisted by another acidic residue, which interacts
with the nitrogen atom of the N-acetyl group and
helps stabilize the transition state. The cyclic interme-
diate formed is then hydrolyzed by the general base-
catalyzed attack of the water at the anomeric center.
The strongest candidate for the acidic residue that sta-
bilizes the transition state and the acid ⁄ base residue
has been suggested to be a highly conserved DE or
DD pair [21], equivalent to D183 and E184 in DspB.
Because of the involvement of this pair, family 20, 56
and 84 enzymes have been suggested to follow a sub-
strate-assisted mechanism [11,13–16,20]. Based on
superposition of the available crystal structures of vari-
ous b-hexosaminidases with that of DspB structure
(PDB code 1YHT), subsite )1 in DspB has been
deduced to be formed by residues R27, D183, E184,
W216, W237, Y278, W330 and E332 (Fig. 5A). Thus,

the active-site architecture in DspB, at least at
subsite )1, has several salient features that are charac-
teristic of b-hexosaminidases exhibiting the substrate-
assisted mechanism.
Role of acidic residues
Among the three acidic residues studied here, D183 is
the putative residue that assists the N-acetyl group in
neighboring-group participation. Replacement of Asp
with Asn at position 183 in DspB rendered this mutant
inactive. Mutations of D183-equivalent residues pro-
duced very low enzyme activity in human b-hexosamin-
idase (Hex A), Streptomyces plicatus b-hexosaminidase
and human O-GlcNAcase enzymes. Thus, the mutant
D354N (human enzyme) exhibited a k
cat
value that
was only 0.04% that of the wild-type [12], whereas
D313N (SpHex enzyme) produced a 560 000-fold
reduction in the k
cat
⁄ K
m
value [11] and that of the
O-GlcNAcase exhibited a 6727-fold reduction in
k
cat
⁄ K
m
[21]. The very low activity of the D313N
mutant in SpHex was rationalized by the increased

bulk (COO
–
to CONH
2
) due to the mutation and ⁄ or
by the adaptation of alternate conformations of the
N-acetyl group of the terminal GlcNAc residue [11]. In
addition to the assistance provided to activate the
N-acetyl group, it has been suggested that this Asp
might control the orientation of the C2-acetamido
group. The inactivity of the D183N mutant suggests
that residue D183 in DspB, like its counterparts in
other b-hexosaminidases, participates in cleavage of
the glycosidic bond and formation of the intermediate
by assisting the acetamido group during catalysis. The
conclusion that DspB utilizes a substrate-assisted catal-
ysis not only relies on comparison with the work of
others, but is also supported by the ability of
NAG-thiazoline to inhibit DspB.
Our results support the role of E184 in DspB as an
acid ⁄ base catalyst because the E184Q mutant shows
significant reduction in k
cat
⁄ K
m
(67-fold lower than the
Table 3. Relative biofilm detachment efficiency of DspB and its
mutants with respect to DspB. Values are given in time required
for removal of 50% of biofilm. Relative efficiency was calculated
as: 100 · (T

50
for DspB ⁄ T
50
for mutant) · (concentration of
DspB ⁄ concentration of mutant).
Enzyme
Concentration
used (l
M)
T
50
(min)
% Relative
efficiency
Wild-type 0.011 1.3 (0.7) 100
D183N 0.275 8.0 (1.4) 0.65
E184Q 1.375 –
a
–
a
E332Q 0.055 7.2 (0.3) 3.6
W237A 0.275 9.0 (1.4) 0.6
W330Y 0.275 1.0 (0.8) 5.2
Y187A 0.022 5.5 (0.2) 12.0
Y278A 0.055 2.8 (0.2) 9.3
R27K 0.055 5.2 (0.2) 5.0
R27A 0.056 6.9 (0.3) 3.8
a
Data could not be fitted because of very low activity.
S. G. A. Manuel et al. Active site of dispersin B

FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS 5993
wild-type) and a shift in the apparent pK
a
for the
acid ⁄ base. Mutational studies on E314 from S. plicatus
also showed a reduction in the kinetic parameters and
a shift in the pH optimum. The acid ⁄ base role for this
residue is identical to that of the equivalent gluta-
mate ⁄ aspartate residues in family 20 b-hexosaminidas-
es and family 56 and family 84 enzymes.
Our findings reveal an important role for E332 in
DspB and possibly other hexosaminidases as well. As
shown in Fig. 2B, E332 of DspB and its equivalent res-
idues (residue Glu739 in the human Hex enzyme and
residue E444 in the S. plicatus enzyme) in fam-
ily 20 b-hexosaminidases occupy the same space on the
opposite side of the bound ligand at subsite )1. These
residues are positioned in space to interact with the
hydroxyl group at C4. Kinetic parameters obtained for
the E332Q mutant, reflecting the loss of activity upon
mutation, unequivocally show that a loss of transition-
state stabilization has occurred. In this regard, E332
residue might be acting similar to D300, one of the
three catalytic residues in human salivary a-amylase
and other a-amylases that belong to the family 13 hy-
drolases [22,23]. Overall, the acidic residue mutations
in DspB tend to destabilize the transition state and the
oxazolinium intermediate (Fig. 1) that might be accu-
mulated (low K
m

values), and undergo a slower hydro-
lysis (reduced k
cat
) by a suitably juxtaposed water
molecule.
Role of aromatic residues
Although subsite )1 is highly conserved in family 20
b-hexosaminidases, significant differences in the active
site of DspB should be noted. For example, the bind-
ing pockets created by the aromatic residues for sub-
site )1 in DspB and S. plicatus enzyme differ in size
and shape (Fig. 5B). It is possible that the difference
in pocket shape might lead to subtle differences in the
binding of the terminal GlcNAc residue and enable
the active site of DspB to be flexible. Differences
R27
E332
W330
W237
W216
Y278
E184
A
B
D183
GlcNAc
Fig. 5. Active site of DspB. (A) Active-site
residues of DspB and their potential
interactions with a docked GlcNAc at
subsite )1. GlcNAc was docked onto the

DspB active site manually using SpHex
structure (PDB code: 1hp5) [17] as a
reference. Superposition of the crystal
structures was carried out using
SWISS-MODEL [31]. (B) Superposition of the
active-site cavity formed by subsite )1in
DspB (blue) and S. plicatus enzyme (yellow).
Two views 90° to each other are shown in
the left and right panels. Note the difference
in the cavity size and shape (207 A
˚
3
for
DspB and 222 A
˚
3
for S. plicatus enzyme)
suggesting that the binding of the terminal
GlcNAc residue might be different in the
two enzymes. Cavity was calculated using
the server at informatics.
leeds.ac.uk/cgi-bin/pocketfinder/.
Active site of dispersin B S. G. A. Manuel et al.
5994 FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS
in the substrate specificity in DspB versus other
b-hexosaminidases might arise because of differences
in the residues that create subsite +1. In particular,
tryptophan residues W685 (human Hex A) and W408
(S. plicatus), respectively, interact with the subsite +1
moiety via hydrophobic-stacking interactions [16,17].

In DspB, an equivalent tryptophan residue is absent.
In this regard, the juxtaposition of two loops segment
in DspB (residues 185–191 containing Y187) and a sec-
ond loop (residues 238–247 containing Q244) might be
important. As a result, there are two pockets in the
active site, formed by the intrusion of Y187, and
unnatural substrates such as pNPGlcNAc may be ori-
ented to bind in these pockets, albeit poorly (Fig. 6).
Another commonly used substrate, MuGlcNAc, is also
poorly cleaved by DspB (Table 2) suggesting that agly-
cone might be critical in DspB. The first-order rate
constant for this substrate is 39-fold lower than for
pNPGlcNAc. No measurable activity was observed for
the mutants D183N, E184Q, E332Q, W330Y, W237A,
R27A and R27K, whereas values for the other two
mutants, Y187A and Y278A are consistent with their
proposed roles. In the case of SpHex enzyme, however,
the kinetic parameters for the two substrates, pNP-
GlcNAc and MuGlcNAc, are comparable ( k
cat
:
193 ± 3 s
)1
versus 180 ± 7 s
)1
; K
m
: 0.049 ± 0.004
mm versus 0.054 ± 0.03 mm; and k
cat

⁄ K
m
:
3900 ± 400 s
)1
Æmm
)1
versus 3300 ± 300 s
)1
Æmm
)1
)
[24]. The importance of the aglycone in DspB is in
contrast to the existing body of work on other fam-
ily 20 enzymes that cleave b(1,4)-linked substrates,
wherein the aglycone does not appear to have a signifi-
cant impact. Analysis of the reaction mixtures by TLC
with chitin oligosaccharides using DspB and the
mutants did not generate any hydrolytic products
(data not shown) further substantiating the substrate
specificity of DspB. The active site of DspB is clearly
designed to bind 1,6-linked substrates whereas 1,4-
linked substrates and other substrate analogs mimick-
ing 1,4-linked substrates either do not bind at all, or
bind only poorly.
Note that the acetyl group and W237 are involved
in a stacking interaction as has been observed in many
b-hexosaminidase crystal structures bound to a ligand.
Even in DspB, a similar stacking is present between
W237 and the bound acetate mimicking the N-acetyl

group [10]. In the W237A mutant, binding of the ace-
tyl group might have been affected thus reducing the
enzyme activity against pNPGlcNAc substrate. By con-
trast, the observed activity of the tryptophan mutants
W237A and W330Y in biofilm detachment could be
explained because the active site is designed to bind
the 1,6-linked substrate as present in the biofilm
matrix. These results have been further confirmed by
total glucosamine assay measuring the hydrolysis
rather than the biofilm detachment (Fig. 4E). The rela-
tive hydrolytic activity, as measured by the release of
GlcNAc (Fig. 4E), might be considered to reflect true
activities of these mutations than the use of pNPGlc-
NAc as a substrate. It is likely that DspB possesses
multiple subsites for GlcNAc residues to bind and pro-
vides potential interactions between GlcNAc residues
at subsites +1 and ⁄ or +2 and protein atoms.
The role of arginine at the active site
The available crystal structure data show that R27 is
directly involved in substrate binding and helps to
dock a GlcNAc at subsite )1. This is facilitated by
bridging interactions with the C3- and C4-hydroxyl
groups, as observed in the crystal structures of several
b-hexosaminidases reported in the literature. The lower
k
cat
value for mutant R27K suggests that R27 might
be involved in stabilization of the transition state
[18,25]. Interestingly, distinct differences in GlcNAc-
releasing activity from the biofilm have been noted for

the two R27 mutants (Fig. 4E). The low activities of
the two mutants with respect to pNPGlcNAc hydroly-
sis, compared with the wild-type, may be because this
substrate does not truly mimic the natural substrate.
The difference between substitution with a neutral side
Fig. 6. Representation of the surface cavity in DspB. The large
substrate-binding pocket is partitioned by the residues Y187 and
Q244 positioned in the middle. The residue W237 provides a wall
for a stacking interaction with the bound acetate ⁄ N-acetyl group.
The pocket extends on the right and might be utilized to bind
b(1,6)-linked GlcNAc polymer to generate multiple subsites.
S. G. A. Manuel et al. Active site of dispersin B
FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS 5995
chain (R27A) and a charged side chain (R27K)
becomes apparent when one compares biofilm-detach-
ment activities. The small increase in efficiency for
mutant R27K (Fig. 4D and Table 3) is likely to arise
because of the residual potential stabilization of the
transition state due to retention of the positive charge
(R27 versus K27). However, the biofilm is a complex
environment and detachment activity requires action
on the surface-attached PGA or PGA very close to the
abiotic surface. By contrast, when one compares the
GlcNAc-releasing activity of the mutant and wild-type
enzymes, charge retention at the transition state in
R27K clearly reflects the role of R27 in DspB. Thus,
R27K exhibits a significantly higher activity (Fig. 4E)
than R27A. Unlike biofilm detachment, GlcNAc
release can occur from the exposed chains of PGA in
the biofilm each of which consists up to 130 b(1,6)-

linked GlcNAc moieties [5] further supporting the idea
that GlcNAc-releasing activity might truly reflect the
roles of these residues in DspB.
Active-site residues of DspB and their role
in the substrate-assisted mechanism
The hydrolysis of pNPGlcNAc and the biofilm-
detachment activity exhibited by the mutants studied
clearly suggest that several residues at subsite )1 play
a significant role in the hydrolytic reaction. It is well
established that the position of the N-acetyl group is
critical for b-hexosaminidase enzymes to exhibit a
substrate-assisted mechanism. Whereas earlier reports
focused on the individual role played by D183 in this
regard, our results show that, in addition to D183,
residues such as W237 and Y278 assist in orienting
the acetyl group by participating in a critical-stacking
or hydrogen-bond interaction with it. Mutation of
R27 or E332 has a drastic effect on enzyme activity.
Because these residues are on the same side of the
bound ligand and interact with C4 oxygen, the stabil-
ization provided by the residues in the transition state
appears to be critical as the saccharide is undergoing
a conformational change. A detailed study on the
probable conformational changes during the sub-
strate-assisted mechanism has suggested that the ter-
minal saccharide changes from a skew conformation
in the Michaelis complex, to an envelope (or perhaps
half chair) conformation in the transition state, and
then to a chair conformation in the intermediate [14].
In this regard, residue W330, also on the side of C4,

provides a wall for the GlcNAc-binding pocket and
thus might work in concert with R27 and E332 to
provide the necessary stabilization in the transition
state. Our results suggest that hydrolysis in DspB
occurs via a substrate-assisted mechanism. Further-
more, the notable differences in the activity of DspB
mutant Y187A towards pNPGlcNAc and PGA
biofilm from S. epidermidis, suggest that DspB might
utilize the loop containing Y187 for its substrate
specificity. In conclusion, our mutational analysis has
shown that every residue at the subsite )1 in DspB is
essential for optimal activity.
Experimental procedures
Expression and purification of DspB and mutants
DspB and its site-specific mutants were expressed and puri-
fied using plasmid pRC3 carrying the dspB gene (encoding
amino acids 21–381 fused directly to a hexahistidine metal-
binding C-terminal tail located downstream from an iso-
propyl thio-b-d-galactoside-inducible tac promoter) [10].
Mutants of DspB were generated using the primers listed in
Table 1. Native and the mutant enzymes were expressed as
previously described [10]. Briefly, a 2 L Erlenmeyer flask
containing 500 mL Luria–Bertani broth supplemented with
30 lg kanamycin per mL was inoculated with 5 mL of an
overnight culture of E. coli strain Rosettaä (DE3) (Nov-
agen, Madison, WI) [26] transformed with the plasmid
pRC3 containing the appropriate mutation. The flask
was incubated at 37 °C with agitation (200 r.p.m.) until
A
600

¼ 0.6 (c. 3 h). Isopropyl thio-b-d-galactoside was
added to a final concentration of 0.1 mm, and the flask was
incubated for 5 h with agitation. Cells were harvested by
centrifugation for 15 min at 6000 g.
The cell pellet was resuspended in 20 mL lysis buffer
(20 mm Tris ⁄ HCl, pH 8.0, 500 mm NaCl, 1 mm phenyl-
methylsulfonyl fluoride and 0.1% Nonidet P-40). The cell
suspension was then sonicated on ice for 30 s (· 5 with
2 min intervals) at 30% capacity with a 30% duty cycle
using a Branson model 450 sonicator equipped with a
microprobe. The cell debris was pelleted by centrifugation
(15 000 g for 20 min; 4 °C), and the supernatant was
loaded onto a 3 mL (bed volume) activated Ni–NTA aga-
rose affinity column (Qiagen, Valencia, CA) according to
the manufacturer’s instructions. The column was washed
with 50 mL wash buffer (20 mm Tris, pH 8.0, 500 mm
NaCl) containing 5 mm imidazole. Bound proteins were
eluted with 25 mL wash buffer containing 50 mm imidaz-
ole. Fractions of the eluate (5 mL) were collected and
assayed for the presence of DspB by SDS–PAGE and
Coomassie Brilliant Blue R250 staining [27]. Fractions
containing pure protein were pooled and dialyzed overnight
against water (2 · 4 L changes at 4 °C) by using a 12–14
kDa cut-off dialysis membrane (Spectra ⁄ Por2, Spectrum
Laboratories, Inc., Rancho Dominguez, CA), and lyophil-
lized. The purity and molecular mass of DspB proteins
were determined by MS analysis.
Active site of dispersin B S. G. A. Manuel et al.
5996 FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS
Enzyme assays

pNPGlcNAc (Sigma-Aldrich, St Louis, MO) or MuGlcNAc
was used to determine the enzyme activity. Typically,
enzyme reactions were carried out in 100-lL mixtures con-
taining 50 mm NaCl ⁄ P
i
(pH 5.8), 50 mm NaCl, and varying
concentrations of substrate (0.312–10 mm). The final
enzyme concentrations used are provided in Table 2.
Because of the solubility constraints of the substrate, the
maximum substrate concentration was restricted to 10 mm.
All assays were carried out in 96-well microtiter plates
placed in a 37 °C incubator. Reactions were terminated at
various times by adding 1 lLof5m NaOH. The increase
in absorption resulting from the release of p-nitrophenol in
each well was measured as nitrophenolate with a Bench-
mark microplate reader (Bio-Rad, Hercules, CA) set at
405 nm. The specific activities of DspB and its mutants
were determined using pNPGlcNAc where 1 unit of enzyme
activity was defined as the amount of enzyme needed to
hydrolyze 1 lmol of pNPGlcNAc to p-nitrophenol and
N-acetylglucosamine per min at pH 5.8 at 37 °Cin50mm
NaCl ⁄ P
i
containing 50 mm NaCl [11]. Michaelis–Menten
kinetic parameters were obtained from double-reciprocal
plots (1⁄ v versus 1 ⁄ S) using first-order kinetics because of
the low solubility of pNPGlcNAc. Inhibition studies were
carried out using the mechanism-based inhibitor NAG-
thiazoline [16] whose concentration was varied from 0 to
1000 lm. For comparison of the inhibitory efficiency

against DspB, jack bean b-hexosaminidase (Sigma-Aldrich)
was used. All assays were performed in triplicate on at least
three separate occasions, which exhibited similar results
with minimal variation among them.
Determination of pH optima
DspB activity was measured using pNPGlcNAc at various
pH values ranging from 3.5 to 10.0. The buffers used were
Mes (pH 3.5–6.5), Hepes (pH 7.0–8.0) and Tris ⁄ HCl
(pH 8.5–10.0). The pH dependence on activity was mea-
sured using a final substrate concentration of 5 mm. The
molar absorption coefficient of the liberated p-nitrophenol
varies with pH from 7280 to 19 000 depending upon the
degree of ionization [28]. The appropriate molar absorption
coefficient for the tested pH was used in the calculations.
p-Nitrophenol release was measured at 400 nm over a time
interval after the addition of enzyme to initiate the reaction.
The enzyme concentrations used were 100 nm DspB,
5.5 lm R27K, and 1 lm E184Q. The change in absorbance
was fitted to a first-order rate equation to yield the pseudo-
first order rate constant at each pH value tested. The exper-
imental points were fitted to a sigmoid curve using the
program prism 3.0 (GraphPad Software, San Diego, CA).
The apparent pK
a
values for the acidic and basic limbs
were estimated by fitting the data to the equation v ¼
V
max
⁄ (1 + [H] ⁄ K
a

+ K
b
⁄ [H]), where v is rate at a given
pH, V
max
is the maximum rate at optimum pH, [H] is the
hydrogen ion concentration and K
a
and K
b
represent the
acid dissociation constants. pH dependence was measured
on three separate occasions with similar results. Because of
the very low activity of D183N and E332Q, no pH optima
for these mutants were determined.
Bacterial strains, media, and growth conditions
S. epidermidis strain NJ9709 (isolated from the surface of
infected intravenous catheters removed from patients at
University Hospital, Newark, NJ) [2] was streaked onto
blood agar plates and incubated for 24 h at 37 °C. Plates
were stored at 4 °C, and bacteria were passaged weekly.
Biofilms were grown in Trypticase soy broth (Becton-Dick-
inson, Franklin Lakes, NJ) supplemented with 6 g yeast
extract and 8 g glucose per L. Isolation and identification
of the bacterial strains were part of routine diagnostic pro-
cedures at the hospital and were carried out with the under-
standing and consent of each subject.
Biofilm growth and detachment assay
The bacterial inoculum was prepared as follows. Two loops
of colonies scraped from the surface of an agar plate were

transferred to a microfuge tube containing 200 lL fresh
medium. Cells were homogenized with a disposable pellet
Kontes pestle (Fischer, Itasca, IL) and vortex agitated at
high speed for 30 s. Twenty-two microliters was transferred
to a 50 mL polystyrene tube with 22 mL fresh medium. The
resulting inoculum contained 10
9
–10
10
cfuÆmL
)1
. Biofilms
were grown in 96-well tissue-culture-treated polystyrene mi-
crotiter plates (Corning model 3595, Sigma-Aldrich). Wells
filled with 200 lL inoculum were incubated for 16 h at
37 °C. For detachment studies, wells were rinsed by sub-
merging the entire plate in a tub of cold, running tap water.
Varying concentrations of DspB or mutant enzymes in
NaCl ⁄ P
i
were added at specific time intervals ranging from 1
to 10 min. Biofilms were washed in tub of cold, running tap
water and were stained with crystal violet as previously
described [1]. The absorbance values of the well solutions
were determined by using a Bio-Rad Benchmark microplate
reader set at 590 nm. All assays were performed in triplicate
wells on at least three separate occasions, which exhibited
similar results with minimal variation among them.
Total hexosamine assay
Confluent biofilms were grown in 100-mm diameter tissue-

culture-treated polystyrene Petri dishes. The biofilm that
formed on the surface of the plate was rinsed with NaCl ⁄ P
i
(3·) and then scraped from the surface into 30 mL of
NaCl ⁄ P
i
by using a cell scraper. Cells were pelleted by cen-
trifugation for 10 min at 2684 g in a Sorvall model RT 7
S. G. A. Manuel et al. Active site of dispersin B
FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS 5997
centrifuge. The cell pellet (50 mg) was resuspended in
NaCl ⁄ P
i
containing 5.75 lm of DspB or mutant enzymes
and the reaction mixture (200 lL) was incubated at 37 °C
for 30 min with occasional stirring. After centrifugation,
total hexosamine was measured in the supernatant by using
the Morgan–Elson colorimetric assay [29]. All assays were
performed at least three times with similar results.
Acknowledgements
This project was supported by the USPHS Grant
DE16291 (NR) and DE15124 (JBK). We would like to
thank Ms. Samin Nawaz, Ms. Maryam Sheikh and Mr
James Rankin for technical assistance. We thank Dr
Spencer A. Knapp for generous gift of the NAG-thiaz-
oline inhibitor used in this study.
References
1 Kaplan JB, Meyenhofer MF & Fine DH (2003) Biofilm
growth and detachment of Actinobacillus actinomyce-
temcomitans. J Bacteriol 185, 1399–1404.

2 Kaplan JB, Ragunath C, Velliyagounder K, Fine DH &
Ramasubbu N (2004) Enzymatic detachment of Staphy-
lococcus epidermidis biofilms. Antimicrob Agents Chemo-
ther 48, 2633–2636.
3 Izano EA, Sadovskaya I, Vinogradov E, Mulks MH,
Velliyagounder K, Ragunath C, Kher WB, Ramasubbu
N, Jabbouri S, Perry MB et al. (2007) Poly-N-acetylglu-
cosamine mediates biofilm formation and antibiotic
resistance in Actinobacillus pleuropneumoniae. Microb
Pathol 43, 1–9.
4 Itoh Y, Wang X, Hinnebusch BJ, Preston JF 3rd &
Romeo T (2005) Depolymerization of beta-1,6-N-acetyl-
d-glucosamine disrupts the integrity of diverse bacterial
biofilms. J Bacteriol 187, 382–387.
5 Mack D, Fischer W, Krokotsch A, Leopold K, Hart-
mann R, Egge H & Laufs R (1996) The intercellular
adhesin involved in biofilm accumulation of Staphylo-
coccus epidermidis is a linear beta-1,6-linked glucosami-
noglycan: purification and structural analysis.
J Bacteriol 178, 175–183.
6 Gerke C, Kraft A, Sussmuth R, Schweitzer O & Gotz F
(1998) Characterization of the N-acetylglucosaminyl-
transferase activity involved in the biosynthesis of the
Staphylococcus epidermidis polysaccharide intercellular
adhesin. J Biol Chem 273, 18586–18593.
7 Maira-Litran T, Kropec A, Abeygunawardana C, Joyce
J, Mark G, Goldmann DA 3rd & Pier GB (2002)
Immunochemical properties of the staphylococcal
poly-N-acetylglucosamine surface polysaccharide. Infect
Immun 70, 4433–4440.

8 Joyce JG, Abeygunawardana C, Xu Q, Cook JC,
Hepler R, Przysiecki CT, Grimm KM, Roper K, Ip CC,
Cope L et al. (2003) Isolation, structural characteriza-
tion, and immunological evaluation of a high-molecu-
lar-weight exopolysaccharide from Staphylococcus
aureus. Carbohydr Res 338, 903–922.
9 Wang X, Preston JF 3rd & Romeo T (2004) The
pgaABCD locus of Escherichia coli promotes the synthe-
sis of a polysaccharide adhesin required for biofilm for-
mation. J Bacteriol 186, 2724–2734.
10 Ramasubbu N, Thomas LM, Ragunath C & Kaplan JB
(2005) Structural analysis of dispersin B, a biofilm-
releasing glycoside hydrolase from the periodontopatho-
gen Actinobacillus actinomycetemcomitans. J Mol Biol
349, 475–486.
11 Williams SJ, Mark BL, Vocadlo DJ, James MN &
Withers SG (2002) Aspartate 313 in the Streptomyces
plicatus hexosaminidase plays a critical role in substrate-
assisted catalysis by orienting the 2-acetamido group
and stabilizing the transition state. J Biol Chem 277,
40055–40065.
12 Hou Y, Vocadlo DJ, Leung A, Withers SG & Mahuran
D (2001) Characterization of the Glu and Asp residues
in the active site of human beta-hexosaminidase B.
Biochemistry 40, 2201–2209.
13 Sheldon WL, Macauley MS, Taylor EJ, Robinson CE,
Charnock SJ, Davies GJ, Vocadlo DJ & Black GW
(2006) Functional analysis of a group A streptococcal
glycoside hydrolase Spy1600 from family 84 reveals it is
a beta-N-acetylglucosaminidase and not a hyaluroni-

dase. Biochem J 399, 241–247.
14 Whitworth GE, Macauley MS, Stubbs KA, Dennis RJ,
Taylor EJ, Davies GJ, Greig IR & Vocadlo DJ (2007)
Analysis of PUGNAc and NAG-thiazoline as transition
state analogues for human O-GlcNAcase: mechanistic
and structural insights into inhibitor selectivity and
transition state poise. J Am Chem Soc 129, 635–644.
15 Maier T, Strater N, Schuette CG, Klingenstein R, Sand-
hoff K & Saenger W (2003) The X-ray crystal structure
of human beta-hexosaminidase B provides new insights
into Sandhoff disease. J Mol Biol 328, 669–681.
16 Mark BL, Mahuran DJ, Cherney MM, Zhao D, Knapp
S & James MN (2003) Crystal structure of human beta-
hexosaminidase B: understanding the molecular basis of
Sandhoff and Tay–Sachs disease. J Mol Biol 327, 1093–
1109.
17 Mark BL, Vocadlo DJ, Knapp S, Triggs-Raine BL,
Withers SG & James MN (2001) Crystallographic evi-
dence for substrate-assisted catalysis in a bacterial beta-
hexosaminidase. J Biol Chem 276, 10330–10337.
18 Mark BL, Wasney GA, Salo TJ, Khan AR, Cao Z,
Robbins PW, James MN & Triggs-Raine BL (1998)
Structural and functional characterization of Streptomy-
ces plicatus beta-N-acetylhexosaminidase by comparative
molecular modeling and site-directed mutagenesis. J Biol
Chem 273, 19618–19624.
Active site of dispersin B S. G. A. Manuel et al.
5998 FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS
19 Kaplan JB, Ragunath C, Ramasubbu N & Fine DH
(2003) Detachment of Actinobacillus actinomycetemcomi-

tans biofilm cells by an endogenous beta-hexosaminidase
activity. J Bacteriol 185, 4693–4698.
20 Tews I, Perrakis A, Oppenheim A, Dauter Z, Wilson
KS & Vorgias CE (1996) Bacterial chitobiase structure
provides insight into catalytic mechanism and the basis
of Tay–Sachs disease. Nat Struct Biol 3, 638–648.
21 Cetinbas N, Macauley MS, Stubbs KA, Drapala R &
Vocadlo DJ (2006) Identification of Asp174 and Asp175
as the key catalytic residues of human O-GlcNAcase by
functional analysis of site-directed mutants. Biochemis-
try 45, 3835–3844.
22 Ramasubbu N, Ragunath C & Mishra PJ (2003) Prob-
ing the role of a mobile loop in substrate binding and
enzyme activity of human salivary amylase. J Mol Biol
325, 1061–1076.
23 Ramasubbu N, Ragunath C, Mishra PJ, Thomas LM,
Gyemant G & Kandra L (2004) Human salivary
alpha-amylase Trp58 situated at subsite-2 is critical for
enzyme activity. Eur J Biochem 271, 2517–2529.
24 Vocadlo DJ & Withers SG (2005) Detailed comparative
analysis of the catalytic mechanisms of beta-N-acetyl-
glucosaminidases from families 3 and 20 of glycoside
hydrolases. Biochemistry 44, 12809–12818.
25 Hou Y, Vocadlo D, Withers S & Mahuran D
(2000) Role of beta Arg211 in the active site of human
beta-hexosaminidase B. Biochemistry 39, 6219–
6227.
26 Dubendorff JW & Studier FW (1991) Creation of a T7
autogene. Cloning and expression of the gene for bacte-
riophage T7 RNA polymerase under control of its

cognate promoter. J Mol Biol 219, 61–68.
27 Sambrook J, Fritsch E & Maniatis T (1989) Molecular
Cloning. A Laborartory Manual, 2nd edn. Cold Spring.
Harbor Laboratory Press, Cold Spring Harbor, NY.
28 Rauscher E, Neumann U, Schaich E, von Bulow S &
Wahlefeld AW (1985) Optimized conditions for deter-
mining activity concentration of alpha-amylase in
serum, with 1,4-alpha-d-4-nitrophenylmaltoheptaoside
as substrate. Clin Chem 31, 14–19.
29 Strominger JL, Park JT & Thompson RE (1959) Com-
position of the cell wall of Staphylococcus aureus: its
relation to the mechanism of action of penicillin. J Biol
Chem 234, 3263–3268.
30 Lemieux MJ, Mark BL, Cherney MM, Withers SG,
Mahuran DJ & James MN (2006) Crystallographic
structure of human beta-hexosaminidase A: interpreta-
tion of Tay–Sachs mutations and loss of GM2 ganglio-
side hydrolysis. J Mol Biol 359, 913–929.
31 Schwede T, Kopp J, Guex N & Peitsch MC (2003)
SWISS-MODEL: an automated protein homology-
modeling server. Nucleic Acids Res 31, 3381–3385.
S. G. A. Manuel et al. Active site of dispersin B
FEBS Journal 274 (2007) 5987–5999 ª 2007 The Authors Journal compilation ª 2007 FEBS 5999