The identification of the catalytic nucleophiles of two β-galactosidases from glycoside hydrolase family 35

Article abstract of DOI:10.1016/S0008-6215(01)00108-2

The β-galactosidases from Xanthomonas manihotis (β-Gal Xmn) and Bacillus circulans (β-Gal-3 Bcir) are retaining glycosidases that hydrolyze glycosidic bonds through a double displacement mechanism involving a covalent glycosyl-enzyme intermediate. The mec

Full text of DOI:10.1016/S0008-6215(01)00108-2

www.elsevier.nl/locate/carres
Carbohydrate Research 333 (2001) 7–17
The identification of the catalytic nucleophiles of two
b-galactosidases from glycoside hydrolase family 35
a
b
a
b
Jan E. Blanchard, Laurent Gal, Shouming He, Janine Foisy,
b
a,
R. Antony J. Warren, Stephen G. Withers *
a
Department of Chemistry, Protein Engineering Network of Centres of Excellence of Canada,
Uni6ersity of British Columbia, Vancou6er, BC, Canada V6T 1Z1
b
Department of Microbiology, Protein Engineering Network of Centres of Excellence of Canada,
Uni6ersity of British Columbia, Vancou6er, BC, Canada V6T 1Z1
Received 26 February 2001; accepted 9 April 2001
Abstract
The b-galactosidases from Xanthomonas manihotis (b-Gal Xmn) and Bacillus circulans (b-Gal-3 Bcir) are retaining
glycosidases that hydrolyze glycosidic bonds through a double displacement mechanism involving a covalent
glycosyl–enzyme intermediate. The mechanism-based inactivator 2,4-dinitrophenyl 2-deoxy-2-fluoro-b-D-galactopy-
ranoside was shown to inactivate b-Gal Xmn and b-Gal-3 Bcir through the accumulation of 2-deoxy-2-fluorogalact-
osyl enzyme intermediates with half lives of 40 and 625 h, respectively. Peptic digestion of these labeled enzymes and
260
233
analysis by LC–MS identified Glu
and Glu
as the catalytic nucleophiles involved in the formation of the
glycosyl–enzyme intermediate during catalysis by b-Gal Xmn and b-Gal-3 Bcir, respectively. These findings confirm
the previous prediction of the position of these residues based on primary sequence similarities to other members of
the glycoside hydrolase family 35. © 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Glycosidases; Labeling; Fluorosugars; Galactosidase; Mass spectrometry
1
. Introduction
catalysis. This step results in the formation of
a covalent glycosyl–enzyme intermediate that
subsequently undergoes deglycosylation by the
general base-catalyzed nucleophilic attack of
either water (hydrolysis) or another moiety
(transglycosylation).
Retaining glycosidases catalyze the cleavage
of glycosidic bonds with net retention of
1,2
anomeric configuration. Hydrolysis is cata-
lyzed through a double displacement mecha-
nism, aided by a pair of carboxylic acids
strategically placed in the enzyme active site
The natural ability of glycosidases to trans-
glycosylate can be exploited to synthesize gly-
cosidic bonds. Although this method of
transglycosylation has been used successfully
(
Scheme 1). In the first step of catalysis (glyco-
sylation), a carboxylate attacks the anomeric
center of the scissile bond, while another car-
boxylic acid side-chain provides general acid
3
–
to synthesize a number of different linkages,
6 there are still drawbacks to this methodol-
ogy, including the subsequent hydrolysis of
transglycosylation products by the glycosi-
dase. The recent development of glycosyn-
*
Corresponding author. Tel.: +1-604-8223266; fax: +1-
7
–10
thases
circumvents this problem; these are
6
04-8222847.
E-mail address: withers@chem.ubc.ca (S.G. Withers).
glycosidases in which the amino acid
0008-6215/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(01)00108-2

8
J.E. Blanchard et al. / Carbohydrate Research 333 (2001) 7–17
Scheme 1. The mechanism of catalysis of a retaining b-galactosidase.
residue responsible for nucleophilic catalysis
has been specifically substituted to a non-
nucleophilic residue, typically alanine or ser-
ine. If a glycosyl fluoride donor with the
anomeric configuration opposite to that of the
natural substrate is supplied, along with a
suitable acceptor, transglycosylation can occur
in high yields (90–100%).
analyzed by LC–MS. The peptide covalently
attached to the fluoro-glycosyl moiety is iden-
tified by comparison to an equivalent digest of
non-inactivated enzyme and sequenced to
identify the catalytic amino acid residue. To
date, this procedure has been used successfully
to assign the position of the catalytic nucle-
1
3–17
ophile in a number of glycosidases.
The first step in the construction of a gly-
cosynthase is the identification of the residue
responsible for nucleophilic catalysis for a
given glycosidase. This can be achieved
through a methodology that involves labeling
the residue of interest with a fluoro-glycoside.
Glycosides substituted with fluorine at the C-2
or C-5 position have been demonstrated to be
excellent mechanism-based inactivators of gly-
The b-galactosidase from the pathogenic
bacterium Xanthomonas manihotis (EC
3.2.1.23, AC P48982; b-Gal Xmn) is a 66 kDa
exoglycosidase that hydrolyzes (1“3) and, to
a lesser degree, (1“4) terminally linked galact-
1
8,19
ose residues from oligosaccharides.
b-Gal
Xmn has a strong sequence similarity (42%
identity, 57% similarity) to a 67 kDa (1“3)
b-galactosidase from Bacillus circulans (EC
3.2.1.23, AC O31341; b-Gal-3 Bcir), the third
b-galactosidase to be isolated from this bac-
1
1,12
cosidases.
The fluorine substituent destabi-
lizes the oxacarbenium ion-like transition
states associated with glycosylation and degly-
cosylation, thereby slowing both steps of
catalysis. A good leaving group at the
anomeric center, such as dinitrophenyl or
fluoride, accelerates glycosylation relative to
deglycosylation, resulting in the accumulation
of the glycosyl–enzyme intermediate, thereby
rendering the enzyme inactive. This trapped
glycosyl–enzyme is then digested by a
protease, and the resultant peptide fragments
2
0
terium. The recently developed classification
system that assigns glycosidases into groups
based on primary amino acid sequence simi-
2
1–23
larities,
designates b-Gal Xmn and b-Gal-
†
3 Bcir as family 35 enzymes. This is a
relatively diverse family, containing over 25
retaining b-galactosidases, the vast majority of
which are eukaryotic in origin, including the
†
http://afmb.cnrs-mrs.fr/ꢀpedro/CAZY/ghf 35.html.
–

J.E. Blanchard et al. / Carbohydrate Research 333 (2001) 7–17
9
human lysosomal enzyme. Mutations at sev-
eral sites in this glycosidase result in the neu-
rological disorders GM1-gangliosidosis and
Aldrich Co. or Aldrich Chemical Company
unless noted otherwise. Pwo DNA polymerase
and deoxynucleoside triphosphates were from
Boehringer Mannheim. Restriction endonucle-
ases, Vent polymerase, and T4 DNA ligase
were from New England Biolabs (NEB).
Galactosyl-imidazole was a gift from Dr An-
drea Vasella (Institut f u¨ r Organische Chemie,
ETH-Z u¨ rich). Protein sequence alignments
were done with the National Center for Bio-
technology Information (NCBI) BLAST
24
Morquio B. syndrome. To date, there are
only four bacterial members of family 35:
b-Gal Xmn, b-Gal-3 Bcir, and b-galactosi-
25
dases from Arthrobacter sp. and Strepto-
2
6
myces coelicolor.
Due to the similarity
between b-Gal Xmn, b-Gal-3 Bcir, and the
human enzyme, mechanistic or structural
studies of the more readily studied b-Gal Xmn
and b-Gal-3 Bcir are of particular interest as
they could contribute to an understanding of
the human b-galactosidase and its associated
disorders. In addition, since b-Gal Xmn and
b-Gal-3 Bcir, in common with the mammalian
enzyme, preferentially hydrolyze (1“3) link-
ages, glycosynthases engineered from these
glycosidases are likely to be able to synthesize
these same linkages. 2-Acetamido-2-deoxy-3-
3
0‡
server site.
Preparation of oligonucleotide
primers, and DNA and protein sequencing
was performed by the Nucleic Acid-Protein
Service Unit, University of British Columbia.
3
1
Curve fitting was done with GRAFIT
.
En-
zyme activity assays were performed on a
Unicam UV4 UV–Vis spectrometer; the spec-
trometer and cuvettes were pre-equilibrated to
3
7 °C by the use of a circulating water bath.
O - (b -
D
- galactopyranosyl) - - glucopyranose
D
Polystyrol or acryl 1-mL disposable cuvettes
were used for all spectrometer measurements.
plays an integral role in complex carbohy-
drates in vivo and is therefore a desirable
target for the synthesis of carbohydrate conju-
gates for study. Although this target has been
synthesized via transglycosylation by b-Gal
Construction of His -tagged i-Gal Xmn.—
6
The gene encoding b-Gal Xmn minus the
amino-terminal 21 amino acid signal sequence
coding region (pCTb-Gal-100) was obtained
as a construct in vector pAGR3 (NEB) from
Dr Christopher Taron (Department of Bio-
chemistry, University of Illinois at Urbana-
Champaign). The b-Gal Xmn gene was
subcloned into pTUG10N18 using the
HindIII and NcoI restriction sites. This new
pTUG10N18/b-Gal construct (52 ng) was
used as the template in PCR with 21 pmol
oligonucleotide primers (see below), 200 mmol
of each of the four deoxynucleoside triphos-
2
7
28
Xmn and b-Gal-3 Bcir, as well as bovine
29
testes b-galactosidase, yields were relatively
low (12–22%). The use of b-Gal Xmn and
b-Gal-3 Bcir as glycosynthases would hope-
fully greatly improve these yields and thereby
expand the range of application. This ap-
proach therefore requires prior knowledge of
the identity of the nucleophile within each
enzyme sequence. This key catalytic residue
has previously been identified within the hu-
13
man b-galactosidase; however, given the di-
versity of this family, it seemed wise to
confirm the identity of the nucleophile in en-
zymes from a prokaryotic source.
phates and 5% Me SO in 50 mL of Vent
2
polymerase buffer. The mixture was heated at
9
6 °C for 60 s and 5 U of Vent polymerase
added to start the reaction. Twenty-five PCR
cycles (95 °C for 30 s, 55 °C for 30 s, 72 °C for
This paper describes the inactivation of b-
Gal Xmn and b-Gal-3 Bcir by 2,4-dinitro-
7
0 s) were performed in a thermal cycler
phenyl 2-deoxy-2-fluoro-b- -galactopyrano-
D
(Perkin–Elmer, GeneAmp PCR System 2400).
side (2FDNPGal) and subsequent experimen-
tal identification of the catalytic nucleophiles
of these glycosidases.
Primers were as follows: 5%-GTG GAA AAC
AGC GGC CGC ATC AAT-3% (forward
primer) and 5%-AG TCT AAG CTT GCT
AGC TTA GTG ATG GTG ATG GTG
ATG GCC GCG CAC GCT GGG GTG-3%
2
. Experimental
(reverse primer). The PCR product and the
General.—All buffers, substrates, and
reagents were purchased from either Sigma–
‡
http://www.ncbi.nlm.nih.gov/BLAST.

10
J.E. Blanchard et al. / Carbohydrate Research 333 (2001) 7–17
pTUG10N18/b-Gal construct were each di-
gested with HindIII and NotI and ligated.
mM imidazole in start buffer. Fractions that
showed b-galactosidase activity (determined as
Construction of His -tagged i-Gal-3
described below) were pooled (V =130 mL)
6
T
Bcir.—The b-Gal-3 Bcir gene was amplified
by PCR with genomic B. circulans
ATCC31382 DNA (prepared by the method
of Ref. 32) as the template (250 mg) and
oligonucleotide primers complementary to the
XhoI and NdeI restriction sites (200 nM) with
and dialyzed against 2×2.5 L of 20 mM
sodium phosphate buffer, 50 mM NaCl, pH
6.0 (SP buffer), then concentrated in 30 kDa
molecular weight cut-off Amicon Centriprep
concentrators and loaded onto two pre-
packed 5 mL Pharmacia Biotech Hi-Trap SP
columns connected in tandem. b-Gal Xmn
was eluted with a linear gradient of 50–350
mM NaCl in SP buffer. The purified b-Gal
Xmn appeared as a single band on silver-
stained SDS–PAGE. Amino-terminal se-
quencing of the purified enzyme (ATPESW)
2
50 mM of each of the four deoxynucleoside
triphosphates in 100 mL of Pwo polymerase
buffer. The mixture was heated at 95 °C for
1
20 s and 5 U of Pwo polymerase added to
start the reaction. Twenty-five PCR cycles
(
94 °C for 30 s, 59 °C for 45 s, 72 °C for 90 s)
1
9
was consistent with the published sequence.
were performed in a thermal cycler (Perkin–
Purification of i-Gal-3 Bcir.—The purifica-
tion of b-Gal-3 Bcir from the cell paste was
Elmer, GeneAmp PCR System 2400). The
1
760 bp PCR product was ligated into the
2
+
accomplished in a single step with a Ni
-
expression vector pET29b(+) (Novagen) us-
ing the XhoI and NdeI cloning sites.
chelating column, as described for b-Gal
Xmn, with the following changes: the start
and elution buffers were as stated, but in-
cluded 1 M sodium phosphate, pH 7.0; pooled
fractions that contained b-Gal-3 Bcir were
dialyzed against 1 M sodium phosphate
buffer, pH 7.0. Purified b-Gal-3 Bcir appeared
as a single band on Coomassie stained SDS–
PAGE. Amino-terminal sequencing of the
purified enzyme (SQLTYDDSF) was consis-
Expression of i-Gal Xmn and i-Gal-3
Bcir.—The constructs encoding the b-Gal
Xmn and b-Gal-3 Bcir genes were each trans-
33
formed into Escherichia coli JM101 via elec-
troporation. A single colony grown on
Luria-Bertani/ampicillin media was used to
inoculate a 30 mL Typ/Amp starter culture
(
16 g/L tryptone, 16 g/L yeast extract, 5 g/L
NaCl, 100 mg ampicillin, pH 7.0). After incu-
bation for 5.5 h at 37 °C, the starter culture
was used to inoculate a 2 L Typ/Amp culture.
After incubation at 30 °C for 3 h, isopropyl
20
tent with the published sequence.
Measurement of enzyme acti6ity of i-Gal
Xmn and i-Gal-3 Bcir.—Enzyme activity us-
ing pNPGal was determined by monitoring
the release of pNP at 400 nm (DA /min).
b- -thiogalactoside was added to a final con-
D
4
00
centration of 0.1 mM, and the culture was
Spectrometer cuvettes contained 0.1% BSA
incubated at 30 °C until OD =0.5. The cells
600
(
w/v) in 50 mM sodium phosphate buffer, pH
were harvested and spun at 4 °C for 15 min at
6
.0 (buffer A) for b-Gal Xmn or pH 7.0
1
0,000g to produce the cell paste.
Purification of i-Gal Xmn.—The cell paste
(
buffer B) for b-Gal-3 Bcir. Unless noted oth-
erwise, the activity of b-Gal Xmn and b-Gal-3
Bcir was determined with 1.2 mM and 0.94
mM pNPGal, respectively, at 37 °C.
was suspended in 50 mM sodium phosphate
buffer, 0.5 M NaCl, pH 6.0 (start buffer) (3.8
mL/g), PMSF (0.25 mM) and b-mercap-
toethanol (6.7 mM) and the suspension was
French pressed three times. The lysed cells
were centrifuged at 10,000g for 1 h and the
supernatant filtered through a 45 mm mem-
brane (final volume ca. 45 mL). The clarified
supernatant was loaded onto a Pharmacia
Biotech Chelating Sepharose Fast Flow
Hydrolysis kinetics.—Reaction rates were
determined as described above at a minimum
of nine different concentrations of pNPGal
ranging from 7 to 350 mM. Resultant rates
and substrate concentrations were fit to the
Michaelis–Menten expression to determine
values of K and V . Extinction coefficients
m
max
−
1
−1
for p-nitrophenol, 3.32 mM cm (pH 6.0)
2
+
−1
−1
column (18 mL) charged with Ni according
to the manufacturer’s instructions. b-Gal Xmn
was eluted using a linear gradient of 0–100
and 9.05 mM cm (pH 7.0) (as determined
by the method of Ref. 34), were used to
quantify the maximal reaction rate of each

J.E. Blanchard et al. / Carbohydrate Research 333 (2001) 7–17
11
enzyme. Enzyme concentrations were deter-
mined by absorbance at 280 nm, and the
extinction coefficients of b-Gal Xmn (1.97 mL
−1
respect to inhibitor was not achieved with
obs
b-Gal Xmn; k
and [I] values were therefore
inact
fit to Eq. (4) to derive the ratio K /kinact.
i
−
1
−1
mg cm ) and b-Gal-3 Bcir (1.46 mL mg
−
1
kinact[I]
obs
cm ) (as determined by the method of Ref.
k
=
(3)
(4)
inact
3
5), and molecular weights of 64,849 Da (b-
K +[I]
i
Gal Xmn) and 67,859 Da (b-Gal-3 Bcir) (as
determined by mass spectrometry) to calculate
pseudo first order rate constants, kcat.
1
1
ꢀ
Ki
ꢁ
1
kinact
=
+
obs
k
k
[I] k
inact
inact
The K value for inhibition of b-Gal Xmn
i
by galactosyl-imidazole was determined by
measuring rates at five concentrations of pNP-
Gal (within the range of 7.0–91.0 mM), both
in the presence (7.2 nM) and absence of galac-
tosyl-imidazole (P). Total reaction volume was
Protection against inacti6ation.—Values of
obs
inact were determined as described for b-Gal
Xmn (20 nM) with 2FDNPGal (46 mM) in
BSA-free buffer A (final vol=100 mL) with
obs
obs
zero (k
) and 40 nM (kip ) of galactosyl-imi-
i
9
00 mL, using 9.7 mg of b-Gal Xmn. The data
dazole. The same experiment was performed
for b-Gal-3 Bcir (100 nM) with 2FDNPGal
(9.6 mM) in BSA-free buffer B (final vol=100
mL) with zero and 500 nM galactosyl-
imidazole.
were plotted in a Lineweaver–Burk format
and the slopes fit to the expression K =[P]/
p
(
slope /slope −1) to derive Kp (where
i
o
slope =slope of the Lineweaver–Burk plot
o
when [P]=zero, and slope =slope of the
Catalytic competence of inacti6ated species
Generation of the inacti6ated enzyme. b-Gal
Xmn or b-Gal-3 Bcir (2.4–5.5 mg) was incu-
bated with 2FDNPGal (0.20–0.40 mM) in
BSA-free buffer A or B, respectively (1 mL),
until B5% of the original activity remained.
Each inactivated enzyme mixture was concen-
trated using a 30 kDa molecular weight cut-
off Millipore Ultrafree 0.5 centrifugal filter to
a volume of 10–20 mL, and fresh BSA-free
buffer A or B added to a final volume of 200
mL. This process of concentration and dilution
was repeated until all excess inactivator was
removed, as determined by the maintenance of
full activity of an aliquot (0.01 mg) of non-in-
activated enzyme diluted one half with the
inactivated species. The concentration of inac-
tivated enzyme was determined by absorption
at 280 nm and used to calculate the maximal
i
Lineweaver–Burk plot when [P]"zero). The
same experiment was performed for b-Gal-3
Bcir (5.0 ng) at zero and 290 nM galactosyl-
imidazole in 800 mL buffer B.
Inacti6ation kinetics.—2FDNPGal (1.15–
9
2.0 mM) in BSA-free buffer A (for b-Gal
Xmn) or B (for b-Gal-3 Bcir) (100–200 mL)
was incubated in a 20 °C (for b-Gal Xmn) or
3
7 °C (for b-Gal-3 Bcir) water bath for 5 min.
b-Gal Xmn or b-Gal-3 Bcir (0.1–9.7 mg) was
added to initiate the reaction. Aliquots (5 mL)
of this inactivator mix were removed at vari-
ous times and residual enzyme activity was
assayed as described. The fractional loss in
activity at each 2FDNPGal concentration was
plotted against time and fit to Eq. (1) (for
b-Gal Xmn) or Eq. (2) (for b-Gal-3 Bcir),
where w/V is the fractional activity at time t,
o
obs
inact
activity possible (V ) if the enzyme was com-
A and C are constants, and k
served rate constant.
is the ob-
o
pletely active.
Demonstration of catalytic competence. In-
activated enzyme freed from excess inactivator
(50–175 mg in 75–100 mL) was incubated at
6
obs
−
k
inact
t
=
=
A(e
A(e
)+C
(1)
(2)
Vo
2
0 °C (b-Gal Xmn) or 37 °C (b-Gal-3 Bcir) in
6
obs
−
k
inact
t
)
BSA-free buffer A or B, respectively, and
aliquots were removed at various times and
assayed as described. The gain in fractional
activity was fit to a first order equation
Vo
obs
inact
Values of k
so derived for b-Gal Bcir
and the corresponding 2FDNPGal concentra-
tions ([I]) were fit to Eq. (3) to determine the
ꢀ
6
ꢁ
hydrol
−k t
inhibitor dissociation constant (K ) and inacti-
i
=A(1−e
)
vation rate constant (kinact). Saturation with
Vo

12
J.E. Blanchard et al. / Carbohydrate Research 333 (2001) 7–17
(
for b-Gal Xmn) or a linear expression (for
parent peptides of m/z 869.0, 951.0, 861.5 or
1025.5 from the first quadrupole (Q1) into the
collision cell (Q2) and observing the daughter
ions in the third quadrupole (Q3). The scan
range of Q3 was 50–1920 Da; step size, dwell
time, and OR as above; ISV was 4.8 kV.
b-Gal-3 Bcir). This procedure was repeated
for b-Gal Xmn in the presence of 31 mM
phenyl b- -glucopyranoside, and for b-Gal-3
D
Bcir in the presence of 34 mM 2-deoxy- -ara-
D
bino-hexopyranose to determine the rate con-
stant for reactivation by transglycosylation
(
ktrans).
3
. Results and discussion
Identification of the catalytic nucleophiles of
i-Gal Xmn and i-Gal-3 Bcir
b-Gal Xmn and b-Gal-3 Bcir were shown to
Digestion of inacti6ated enzyme. b-Gal Xmn
or b-Gal-3 Bcir (14–100 ng) were incubated
with 2FDNPGal (0.2–25.8 mmol) in BSA-free
buffer A or B, respectively (vol=65 mL), until
B5% of the original activity remained. An
aliquot (10 mL) of the inactivated enzyme was
frozen and kept at −20 °C until analyzed by
LC–MS. Each inactivated species (50 mL) was
added to pepsin in a ratio of ca. 12:1 (en-
zyme:pepsin) in 200 mM sodium phosphate,
pH 2.0 (vol=78 mL). Each digestion was per-
formed at rt for either 2 h (for b-Gal Xmn) or
hydrolyze p-nitrophenyl b- -galactopyran-
D
oside (pNPGal) with associated parameters of
K =0.05090.002 and 0.02190.001 mM,
m
−
1
and k =35.890.3 and 18.390.3 s , re-
cat
spectively. Incubation of b-Gal Xmn and b-
Gal-3 Bcir individually in the presence of
2
FDNPGal resulted in a time-dependent de-
crease of enzyme activity according to pseudo-
first order kinetics (Fig. 1(a, b)), behavior that
was modeled according to Scheme 2. Inactiva-
tion of b-Gal-3 Bcir proceeded to completion,
thus data were fit to the expression
0
.5 h (for b-Gal-3 Bcir); the samples were then
frozen and kept at −20 °C until analyzed by
mass spectrometry.
6
obs
inact
t
−
k
=
A(e
)
Vo
Electrospray mass spectrometry. Mass spec-
tra were recorded on a Perkin–Elmer Sciex
API 300 triple quadrupole mass spectrometer
with an ion spray ion source. The spectrome-
ter was interfaced with a Michrom BioRe-
sources Inc. Ultrafast Microprotein Analyzer
that separated peptides via reverse phase
HPLC. Digested proteins were loaded onto a
Reliasil C18 column (1×150 mm) equili-
brated with solvent A (0.05% trifluoroacetic
acid (TFA), 2% MeCN in water). Peptides
were eluted using a 0–60% gradient of solvent
B over 60 min, followed by 100% solvent B
over 2 min (solvent B: 0.045% TFA, 80%
MeCN in water). The flow rate was 50 mL/
min. Spectra were obtained in the single
quadrupole scan mode (LC–MS) or the
tandem MS daughter ion scan mode (MS–
MS).
to yield pseudo first order rate constants at
each inactivator concentration. As can be seen
in Fig. 1(a), inactivation of b-Gal Xmn did
not proceed to completion, thus data were fit
to the expression
6
obs
−kinact
t
=
A(e
)+C
V
o
Such ‘steady state’ inactivation has been seen
3
6
previously and shown to be the result of an
inhibited enzyme being reactivated in the pres-
ence of excess 2FDNPsugar through transgly-
cosylation over the course of the experiment.
Saturation of b-Gal Xmn with respect to
inactivator was not achieved since, at high
concentrations of 2FDNPGal, inactivation
was too rapid to monitor reliably. Therefore,
individual values for K and k
could not be
i
inact
accurately determined; however, the ratio Ki/
kinact was calculated to be 1791 mM min from
the double reciprocal plot of Fig. 1(c).
In LC–MS, the quadrupole mass analyzer
was scanned over a mass to charge ratio (m/z)
of 300–2400 Da with a step size of 0.5 Da and
a dwell time of 1 ms per step. The ion-source
voltage (ISV) was 5 kV and the orifice energy
Saturation with respect to inactivator was
demonstrated with b-Gal Bcir. Derived values
obs
of k
were thus fit with 2FDNPGal concen-
inact
(
OR) was 50 V. In MS–MS, the spectra were
tration to Eq. (3) to determine both k
(0.1090.01 s ) and K (0.03790.005 mM)
inact
−1
obtained by selectively introducing one of the
i

J.E. Blanchard et al. / Carbohydrate Research 333 (2001) 7–17
13
Fig. 1. Inactivation of b-Gal Xmn and b-Gal-3 Bcir by 2FDNPGal. (a) The time-dependent first order decrease in activity of
b-Gal Xmn upon incubation with ꢀ 1.15, ꢁ 2.30, ꢂ 11.5, ꢃ 23.0, and ꢄ 46.0 mM 2FDNPGal. (b) The time-dependent first
order decrease in activity of b-Gal-3 Bcir upon incubation with ꢄ 9.64, ꢁ 14.5, ꢂ 24.1, ꢅ 69.0, and ꢀ 92.0 mM 2FDNPGal.
(
k
c, d) Double reciprocal plots of observed rate constants versus inhibitor concentration. Note that for b-Gal-3 Bcir, values for
and K were calculated from a non-linear fit of Eq. (3); panel (d) is for illustrative purposes only.
inact
i
obs obs
(
Fig. 1(d)). The similarity of the values of Km
constant ratio (k /k ) of 0.42, which is in
ip
i
for pNPGal and K for 2FDNPGal suggests
reasonable agreement with the predicted value
of 0.35. The rate of inactivation of both gly-
cosidases by 2FDNPGal decreased by the ex-
i
that the substitution of a hydroxyl for fluorine
at the C-2 position does not significantly af-
fect ground state binding between the enzyme
and substrate.
pected amount in the presence of
a
competitive inhibitor, data that is consistent
with the 2-fluoro-glycoside inactivator being
active-site directed.
The potent, reversible, competitive inhibitor
galacto-configured pyridoimidazole (galacto-
Once b-Gal Xmn and b-Gal-3 Bcir were
inactivated by 2FDNPGal and the excess in-
hibitor removed, both enzymes spontaneously
regained hydrolytic activity over time (Fig. 3).
The reactivation data for b-Gal Xmn were fit
to a first order equation
syl-imidazole) (K =15 nM with b-Gal Xmn
i
and K =220 nM with b-Gal-3 Bcir) protects
i
b-Gal Xmn and b-Gal-3 Bcir against inactiva-
tion by 2FDNPGal. The fractional loss in
activity of b-Gal Xmn by 2FDNPGal in the
presence and absence of galactosyl-imidazole
obs
(
Fig. 2(a)) yielded a rate constant ratio (k /
ip
obs
i
k ) of 0.43, which is in close agreement with
the value of 0.48 predicted by Eq. (5) which
describes inactivation according to Scheme 2.
obs
k
K +[IX]
ip
i
=
(5)
obs
ki
ꢀ
[P]
Kp
ꢁ
K 1+
+[IX]
i
Similarly, the same experiment performed
for b-Gal-3 Bcir (Fig. 2(b)) yielded a rate
Scheme 2. Enzyme inactivation and reactivation, where E,
enzyme; IX, inactivator; and P, competitive inhibitor.

14
J.E. Blanchard et al. / Carbohydrate Research 333 (2001) 7–17
Fig. 2. Protection against inactivation with galactosyl-imidazole. (a) Fractional loss of activity of b-Gal Xmn over time by 46.0
mM 2FDNPGal in the presence ꢅ and absence ꢀ of 40 nM galactosyl-imidazole (inset). (b) Fractional loss of activity of b-Gal-3
Bcir by 9.6 mM 2FDNPGal in the presence ꢅ and absence ꢀ of 0.50 mM galactosyl-imidazole.
Fig. 3. Reactivation of the trapped intermediates of b-Gal Xmn and b-Gal-3 Bcir. (a) The spontaneous reactivation of
2
-deoxy-2-fluorogalactosyl-b-Gal Xmn in the presence ꢀ and absence ꢅ of 31 mM phenyl b-D-glucopyranoside. (b) The
spontaneous reactivation of 2-deoxy-2-fluorogalactosyl-b-Gal-3 Bcir in the presence ꢀ and absence ꢅ of 34 mM 2-deoxy-
bino-hexopyranose.
D-ara-
ꢀ
6
ꢁ
−
k
hydrol
t
oside, the rate of reactivation of b-Gal Xmn
increased fourfold (Fig. 3(a)). Similarly, the
reactivation of b-Gal-3 Bcir increased approx-
imately threefold in the presence of 34 mM
=
Ae
Vo
to derive a rate constant (khydrol) of 0.0189
.001 h , which corresponds to a half life of
0 h for the glycosyl enzyme intermediate.
−
1
0
2
-deoxy-
D
-arabino-hexopyranose. As demon-
4
strated in some detail for a family 1 b-glucosi-
Data for the reactivation of b-Gal-3 Bcir how-
ever, were limited to only the approximately
linear portion of the first-order expression. A
reliable value for khydrol could not therefore be
extrapolated, but a half life of :625 h for the
trapped intermediate was estimated from Fig.
3
6
dase, such behavior is entirely consistent
with the trapping of a catalytically competent
intermediate and then reactivation through
transglycosylation.
The stoichiometric reaction of b-Gal Xmn
with 2FDNPGal was illustrated by mass spec-
tral analysis. The inactivated enzyme is heav-
ier than the unlabelled enzyme by an amount
(171 Da) equivalent, within error, to the ex-
pected mass of a 2-fluorogalactosyl moiety
(165 Da). The numerous peptide fragments
from the digestion of inactivated b-Gal Xmn
were separated and detected by HPLC–ESMS
as shown in the total ion chromatogram (TIC)
3
(b).
The spontaneous reactivation of each en-
zyme observed is consistent with the trapping
of a catalytically relevant species. Further
demonstration of the catalytic competence of
the trapped species was provided by the in-
crease in the rate of enzyme reactivation
(
ktrans) in the presence of an acceptor sugar. In
the presence of 31 mM phenyl b- -glucopyran-
D

J.E. Blanchard et al. / Carbohydrate Research 333 (2001) 7–17
15
of Fig. 4(a). Comparison of the LC–MS
profiles of labeled and unlabelled digests re-
vealed one major difference: a peptide of m/z
error, one half of the expected mass increase
(165 Da), suggesting that these peptides are
+
doubly charged. Hence, (951.0×2)−2H =
+
1
900.0 Da and (869.0×2)−2H =1736.0 Da
8
69.0 in the control digest, eluting at ca. 28
were the masses of the labeled and control
peptides originating from the inactivated and
control enzymes, respectively. A search of the
primary amino acid sequence of b-Gal Xmn
for a peptide of mass 173692 Da revealed
several candidate sequences that contain an
Asp or Glu residue, either of which could
potentially act as the nucleophile, based on
the conservation of such catalytic residues in
other glycoside hydrolases. The number of
peptides that could possibly contain the cata-
lytic nucleophile was narrowed to two over-
min, was absent in the labeled digest, but
replaced by one of m/z 951.0 that was not
present in the control digest (Fig. 4(b, c)).
The difference between the m/z values for
these two peptides is 82.0 which is, within
2
48
lapping candidate fragments ( IKFRPDQ-
2
61
258
PRMVGEY
and
VGEYWAGWFDH-
2
71
WGK ) by the comparison of the b-Gal
Xmn peptides of mass 1736 Da to those frag-
ments identified through the same analysis of
b-Gal-3 Bcir. Upon digestion of inactivated
b-Gal-3 Bcir, comparison of the m/z profiles
of the separated peptides originating from the
control and inactivated enzyme samples re-
vealed labeled and control peptides having
m/z values of 1025.5 and 861.5 Da, respec-
tively. The difference between these m/z values
is 164 Da, which is within error of the ex-
pected mass difference showing that these pep-
tides are singly charged. A search of the
primary sequence of b-Gal-3 Bcir for peptides
with a mass of 86192 Da indicated, again,
several peptides that could potentially contain
the catalytic nucleophile. Comparison of the
1
8
73692 Da peptides of b-Gal Xmn and the
6192 Da peptides of b-Gal-3 Bcir indicated
that there was only one residue with the requi-
site carboxylic acid side chain that appeared in
at least one of the potential fragments for
each enzyme and was conserved between the
two glycosidases (with respect to the full se-
2
60
233
quences): Glu for b-Gal Xmn, and Glu
for b-Gal-3 Bcir.
Confirmation of the identity of the position
of the nucleophile of b-Gal Xmn was achieved
through further analysis of the m/z 951.0 and
8
69.0 ions of the labeled and control peptides,
Fig. 4. Detection of the 2-deoxy-2-fluorogalactosyl-b-Gal
Xmn species by ESMS. (a) TIC of the control b-Gal Xmn
after peptic digestion. Mass spectrum of peptides eluting at
ca. 28 min in (b) the control and (c) the labeled digest,
showing the m/z values of the major species.
respectively, by tandem mass spectrometry in
the daughter ion scan mode.
The MS–MS profile of the labeled peptide
shown in Fig. 5 is consistent with the expected

16
J.E. Blanchard et al. / Carbohydrate Research 333 (2001) 7–17
Fig. 5. Tandem MS analysis in the daughter ion scan mode of the parent m/z 951.0 peptide of labeled b-Gal Xmn. P corresponds
to the full sequence depicted above and L represents the covalently bound 2-deoxy-2-fluorogalactosyl moiety. Inset shows a 4×
expansion of the region from m/z 975–1350.
Fig. 6. Daughter ion scan of the labeled peptide of m/z 1025.5 Da originating from b-Gal-3 Bcir. Insets show 4× expansion of
the regions from m/z 250–300 to 580–825.
fragmentation pattern for the sequence
side chain within this sequence, it was con-
cluded that this amino acid functioned as the
catalytic nucleophile of b-Gal Xmn.
Similarly, the MS–MS fragmentation pat-
tern of the 1025.5 Da labeled peptide of b-
248
261
IKFRPDQPRMVGEY
with the fluoro-
galactosyl moiety bound to a residue within
260
the fragment PRMVGEY. As Glu
is the
only residue with the requisite carboxylic acid

J.E. Blanchard et al. / Carbohydrate Research 333 (2001) 7–17
17
9
. Mayer, C.; Zechel, D. L.; Reid, S. P.; Warren, R. A. J.;
Withers, S. G. FEBS Lett. 2000, 466, 40–44.
Gal-3 Bcir (Fig. 6) is consistent with that
233
238
expected for the sequence
EFWHGW
1
0. Fort, S.; Boyer, V.; Greffe, L.; Davies, G. J.; Moroz, O.;
Chistiansen, L.; Sch u¨ lein, M.; Cottaz, S.; Driguez, H. J.
Am. Chem. Soc. 2000, 122, 5429–5437.
with the fluoro-galactosyl moiety bound to
2
33
Glu .
1
1
1
1. Withers, S. G.; Rupitz, K.; Street, I. P. J. Biol. Chem.
1
988, 263, 7929–7932.
2. McCarter, J. D.; Withers, S. G. J. Am. Chem. Soc. 1996,
18, 241–242.
1
4
. Conclusions
3. McCarter, J. D.; Burgoyne, D. L.; Miao, S.; Zhang, S.;
Callahan, J. W.; Withers, S. G. J. Biol. Chem. 1997, 272,
2
60
The experimental identification of Glu
396–400.
2
33
and Glu
as the catalytic nucleophiles of
14. Tull, D.; Withers, S. G.; Gilkes, N. R.; Kilburn, D. G.;
Warren, R. A. J.; Aebersold, R. J. Biol. Chem. 1991, 266,
b-Gal Xmn and b-Gal-3 Bcir, respectively,
confirms the previous prediction of the loca-
tion of these residues that was based on pri-
mary sequence similarities. These findings are
the second and third such assignments of
mechanistically important residues from the
enzymes of glycoside hydrolase family 35 and
essentially confirm the alignments for this di-
verse set of enzymes. Work is currently under-
way to investigate the use of b-Gal Xmn and
b-Gal-3 Bcir as glycosynthases.
15621–15625.
1
1
5. McCarter, J. D.; Withers, S. G. J. Biol. Chem. 1996, 271,
6889–6894.
6. Mackenzie, L. F.; Brooke, G. S.; Cutfield, J. F.; Sullivan,
P. A.; Withers, S. G. J. Biol. Chem. 1997, 272, 3161–
3167.
17. Zechel, D. L.; He, S.; Dupont, C.; Withers, S. G.
Biochem. J. 1998, 336, 139–145.
1
1
2
8. Wong-Madden, S. T.; Landry, D. Glycobiology 1995, 5,
19–28.
9. Taron, C. H.; Benner, J. S.; Hornstra, L. J.; Guthrie, E.
P. Glycobiology 1995, 5, 603–610.
0. Ito, Y.; Sasaki, T. Biosci. Biotech. Biochem. 1997, 61,
1270–1276.
2
2
1. Henrissat, B. Biochem. J. 1991, 280, 309–316.
2. Henrissat, B.; Bairoch, A. Biochem. J. 1993, 293, 781–
788.
Acknowledgements
2
3. Henrissat, B.; Callebaut, I.; Fabrega, S.; Lehn, P.;
Mornon, J. P.; Davies, G. Proc. Natl. Acad. Sci. USA
1995, 92, 7090–7094.
We thank Dr Christopher Taron of the
Department of Biochemistry, University of Il-
linois at Urbana-Champaign, for the generous
gift of the b-Gal Xmn gene. We also thank
Karen Rupitz for technical assistance and the
Protein Engineering Network of Centers of
Excellence of Canada for financial support.
2
4. Nishimoto, J.; Nanba, E.; Inui, K.; Okada, S.; Suzuki, K.
Am. J. Hum. Genet. 1991, 49, 566–574.
2
5. Gutshall, K.; Wang, K.; Brenchley, J. E. J. Bacteriol.
1997, 179, 3064–3067.
26. Redenbach, M.; Kieser, H. M.; Denapaite, D.; Eichner,
A.; Cullum, J.; Kinashi, H.; Hopwood, D. A. Mol.
Microbiol. 1996, 21, 77–96.
2
2
7. Fujimoto, H. J. Carbohydr. Chem. 1997, 16, 967–970.
8. Fujimoto, H.; Miyasato, M.; Ito, Y.; Sasaki, T.; Ajisaka,
K. Glycoconjugate J. 1998, 15, 155–160.
2
9. Hedbys, L.; Johansson, E.; Mosbach, K.; Larsson, P. O.;
Gunnarsson, A.; Svensson, S.; L o¨ nn, H. Glycoconjugate
J. 1989, 6, 161–168.
References
1
2
3
. Koshland, D. E. Biol. Re6. 1953, 28, 416–436.
. Sinnott, M. L. Chem. Re6. 1990, 90, 1171–1202.
. Sakai, K.; Katsumi, R.; Ohi, H.; Usui, T.; Ishido, Y. J.
Carbohydr. Chem. 1992, 11, 553–565.
30. Altschul, S. F.; Madden, T. L.; Sch a¨ ffer, A. A.; Zhang,
J.; Zhang, Z.; Miller, W.; Lipman, D. J. Nucleic Acids
Res. 1997, 25, 3389–3402.
31. Leatherbarrow, R. J. GRAFIT, version 3.03; Erithacus
Software Ltd., 1994.
32. Pospiech, A.; Neumann, B. Trends Genet. 1995, 11, 217–
218.
33. Yanish-Perron, C.; Vierira, J.; Messing, J. Gene 1985, 33,
103–119.
34. Kempton, J. B.; Withers, S. G. Biochemistry 1992, 31,
9961–9969.
35. Pace, N. C.; Vajdos, F.; Fee, L.; Grimsely, G.; Gray, T.
Protein Sci. 1995, 4, 2411–2423.
36. Street, I. P.; Kempton, J. B.; Withers, S. G. Biochemistry
1992, 31, 9970–9978.
4
5
6
. Usi, T.; Kubota, S.; Ohi, H. Carbohydr. Res. 1993, 244,
315–323.
. Prade, H.; Mackenzie, L. F.; Withers, S. G. Carbohydr.
Res. 1998, 305, 371–381.
. Vocadlo, D. J.; Withers, S. G. In Carbohydrates in Chem-
istry and Biology; Ernst, B.; Hart, G. W.; Sinay, P., Eds.
Glycoside-catalyzed oligosaccharide synthesis.; Wiley-
VCH: Weinheim, Germany, 2000; Vol. 2, pp. 723–844.
. Mackenzie, L. F.; Wang, Q.; Warren, R. A. J.; Withers,
S. G. J. Am. Chem. Soc. 1998, 120, 5583–5584.
. Malet, C.; Planas, A. FEBS Lett. 1998, 440, 208–212.
7
8
.