Identification of the acid/base catalyst of a glycoside hydrolase family 3 (GH3) β-glucosidase from Aspergillus niger ASKU28

Article abstract of DOI:10.1016/j.bbagen.2012.11.014

Background The commercially important glycoside hydrolase family 3 (GH3) β-glucosidases from Aspergillus niger are anomeric-configuration-retaining enzymes that operate through the canonical double-displacement glycosidase mechanism. Whereas the catalytic nucleophile is readily identified across all GH3 members by sequence alignments, the acid/base catalyst in this family is phylogenetically variable and less readily divined. Methods In this report, we employed three-dimensional structure homology modeling and detailed kinetic analysis of site-directed mutants to identify the catalytic acid/base of a GH3 β-glucosidase from A. niger ASKU28. Results In comparison to the wild-type enzyme and other mutants, the E490A variant exhibited greatly reduced k _cat and k_cat/K_m values toward the natural substrate cellobiose (67,000- and 61,000-fold, respectively). Correspondingly smaller kinetic effects were observed for artificial chromogenic substrates p-nitrophenyl β-d-glucoside and 2,4-dinitrophenyl β-d-glucoside, the aglycone leaving groups of which are less dependent on acid catalysis, although changes in the rate-determining catalytic step were revealed for both. pH-rate profile analyses also implicated E490 as the general acid/base catalyst. Addition of azide as an exogenous nucleophile partially rescued the activity of the E490A variant with the aryl β-glucosides and yielded β-glucosyl azide as a product. Conclusions and general significance These results strongly support the assignment of E490 as the acid/base catalyst in a β-glucosidase from A. niger ASKU28, and provide crucial experimental support for the bioinformatic identification of the homologous residue in a range of related GH3 subfamily members.

Full text of DOI:10.1016/j.bbagen.2012.11.014

Biochimica et Biophysica Acta 1830 (2013) 2739–2749
Contents lists available at SciVerse ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbagen
Identification of the acid/base catalyst of a glycoside hydrolase family 3 (GH3)
β-glucosidase from Aspergillus niger ASKU28
Preeyanuch Thongpoo a,b, Lauren S. McKee b,c,1, Ana Catarina Araújo b,1
,
Prachumporn T. Kongsaeree a,d, Harry Brumer b,c,e,
⁎
a
Interdisciplinary Graduate Program in Genetic Engineering, Faculty of Graduate School, Kasetsart University, Bangkok 10900, Thailand
Division of Glycoscience, School of Biotechnology, Royal Institute of Technology (KTH), AlbaNova University Centre, 106 91 Stockholm, Sweden
Wallenberg Wood Science Center, Royal Institute of Technology (KTH), Teknikringen 56-58, 100 44 Stockholm, Sweden
Department of Biochemistry, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand
b
c
d
e
Michael Smith Laboratories and Department of Chemistry, University of British Columbia, 2185 East Mall, Vancouver, BC, Canada V6T 1Z4
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 July 2012
Received in revised form 7 November 2012
Accepted 19 November 2012
Available online 29 November 2012
Background: The commercially important glycoside hydrolase family 3 (GH3) β-glucosidases from Aspergillus
niger are anomeric-configuration-retaining enzymes that operate through the canonical double-displacement
glycosidase mechanism. Whereas the catalytic nucleophile is readily identified across all GH3 members by
sequence alignments, the acid/base catalyst in this family is phylogenetically variable and less readily divined.
Methods: In this report, we employed three-dimensional structure homology modeling and detailed kinetic
analysis of site-directed mutants to identify the catalytic acid/base of a GH3 β-glucosidase from A. niger ASKU28.
Results: In comparison to the wild-type enzyme and other mutants, the E490A variant exhibited greatly reduced
kcat and kcat/Km values toward the natural substrate cellobiose (67,000- and 61,000-fold, respectively). Corre-
spondingly smaller kinetic effects were observed for artificial chromogenic substrates p-nitrophenyl β-D-
glucoside and 2,4-dinitrophenyl β-D-glucoside, the aglycone leaving groups of which are less dependent on
acid catalysis, although changes in the rate-determining catalytic step were revealed for both. pH-rate profile
analyses also implicated E490 as the general acid/base catalyst. Addition of azide as an exogenous nucleophile
partially rescued the activity of the E490A variant with the aryl β-glucosides and yielded β-glucosyl azide as a
product.
Keywords:
β-Glucosidase
Acid/base catalyst
Aspergillus niger
Glycoside hydrolase family 3
Azide rescue
Conclusions and general significance: These results strongly support the assignment of E490 as the acid/base
catalyst in a β-glucosidase from A. niger ASKU28, and provide crucial experimental support for the bioinformatic
identification of the homologous residue in a range of related GH3 subfamily members.
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
anomeric configuration-inverting mechanism, whereas the remainder
operate through a double-displacement, configuration-retaining mech-
β-Glucosidases (EC 3.2.1.21) catalyze the hydrolysis of the glucosidic
linkage of aryl- and alkyl-β-glucosides, and liberate terminal non-
reducing β-glucosyl residues from oligosaccharides. Based on amino
acid sequence and structural similarity, known β-glucosidases have
been placed in Glycoside Hydrolase (GH) Families GH1, GH3, GH5,
GH9, GH30, and GH116 [1–3]. Members of GH9 utilize a one-step,
anism [1].
In the canonical “retaining” mechanism employed by GH3 β-
glucosidases [4–6], the carboxylic acid side chains of two active-site
residues function individually as a catalytic nucleophile and a general
acid/base residue [7]. The first step of the mechanism involves the dis-
placement of the leaving group (aglycone or saccharide) by the
active-site nucleophile, assisted by proton donation to the nucleofuge
from the conjugate acid form of the general acid/base residue. The
resulting covalent glycosyl-enzyme intermediate is decomposed in
the second step by an incoming water molecule, which is activated
through proton abstraction by the basic form of the general acid/base
residue. Alternatively, interception of the glycosyl-enzyme by an in-
coming saccharide can lead to transglycosylation in some enzymes [8].
The catalytic residues are, in general, highly conserved both within a
particular GH Family and, more broadly, among a group of families com-
prising a GH Clan. As such, identification of a particular catalytic residue in
one representative of a family allows facile prediction of the homologous
Abbreviations: pNP-Glc, p-nitrophenyl β-D-glucoside; 2,4-DNP-Glc, 2,4-dinitrophenyl
β-D-glucoside; beta-Glc-N3, beta-D-glucosyl azide; GH, Glycoside Hydrolase; CAZy,
carbohydrate-active enzymes from Bacillus subtilis; NagZ, β-N-acetylglucosaminidase;
ExoP, exo-1,3/1,4-β-glucanase from Pseudoalteromonas sp.; Bgl3B, β-glucosidase B from
Thermotoga neapolitana; ExoI, β-D-glucan exohydrolase isoenzyme from Hordeum
vulgare; KmBglI, β-glucosidase I from Kluyveromyces marxianus
⁎
Corresponding author at: Michael Smith Laboratories and Department of Chemistry,
University of British Columbia, 2185 East Mall, Vancouver, BC, Canada V6T 1Z4.
E-mail addresses: g5089003@ku.ac.th (P. Thongpoo), mckee@kth.se (L.S. McKee),
acaraujo@kth.se (A.C. Araújo), fscippt@ku.ac.th (P.T. Kongsaeree),
harry@biotech.kth.se, brumer@msl.ubc.ca (H. Brumer).
1
The equal contribution of A.C.A. and L.S.M. is highlighted.
0304-4165/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.bbagen.2012.11.014

2740
P. Thongpoo et al. / Biochimica et Biophysica Acta 1830 (2013) 2739–2749
residue in (nearly) all members [9]. Indeed, this is the case for the cata-
lytic nucleophile of GH3, which has been conclusively identified as a
conserved Asp residue via mechanism-based active-site labeling of a
number of members [6,10–12]. The catalytic acid–base residue, on the
other hand, has escaped general definition, as it appears to be presented
by different protein motifs in various GH3 members from bacteria, fungi,
and plants [12–18]. In its most divergent incarnation, a catalytic His–Asp
dyad, in which the His side chain directly mediates catalysis, replaces the
canonical carboxylate in some GH3 β-N-acetylglucosaminidases [13,19].
Aspergillus species, and A. niger in particular, are commercially im-
portant fungal sources of a wide range of carbohydrate-active enzymes,
including β-glucosidases of relevance to biomass saccharification and
other applications [20–22]. Homologous GH3 β-glycosidases from
Aspergillus spp. have been biochemically scrutinized to various degrees,
including the detailed analysis of transglycosylation potential [23]. In
the current absence of a three-dimensional structural representative,
some catalytically important active site residues in Aspergillus GH3
β-glucosidase have been pinpointed by seminal studies on the A. niger
homologue: Residues W49 and W262 modulate the ratio of substrate
hydrolysis to transglycosylation [24,25], while the neighboring D261
has been identified as the catalytic nucleophile by mechanism-based la-
beling with 2-deoxy-2-fluoro β-glucosyl fluoride and peptide mass
spectrometry [11]. However, definitive biochemical and structural
information on the nature of the general acid/base catalyst in GH3
β-glucosidases from Aspergillus spp. and other related fungi are current-
ly lacking.
2.3. Protein expression and purification of wild-type β-glucosidase from
A. niger ASKU28
The coding sequence of mature β-glucosidase from A. niger ASKU28
was subcloned into a pPICZαBNH8 expression vector [27] at the
EcoRI-XbaI sites, and transformed into E. coli strain DH5α by electropo-
ration. The positive transformants were selected by plating onto LB
medium containing 25 μg/ml zeocin at 37 °C. After that, pPICZαBNH8-
ASKU28 plasmid was extracted and linearized by PmeI before trans-
forming into P. pastoris strain Y11430 by electroporation. The trans-
formed clones were grown on YPDS agar containing 100 μg/ml zeocin
for 3–5 days at 30 °C.
Fifty positive clones of P. pastoris harboring pPICZαBNH8-ASKU28
were selected for small scale expression. The transformants were
grown in 5 ml BMGH medium overnight at 30 °C, 200 rpm, and
expressed in 5 ml BMMH medium. Methanol was added every day
to the final concentration of 0.5% (v/v) for 7 days. The level of
β-glucosidase expression was checked by activity assay in the reaction
containing 1 mM pNP-Glc in 0.1 M sodium acetate, pH 5.0, at 50 °C
for 30 min. The p-nitrophenol released was determined by reading ab-
sorbance at 400 nm after stopping the assays by addition of 0.2 M
Na2CO3. The P. pastoris clone that showed the highest activity was se-
lected for large-scale expression in 1 L medium under the same culture
conditions as above. Methanol was added every day to the final concen-
tration of 0.5–0.8% (v/v) until the activity did not increase any further.
The culture supernatant from large-scale expression was added with
2 M ammonium sulfate before loading onto a hydrophobic interaction
chromatography column containing phenyl sepharose beads (GE
Healthcare, Sweden), that was previously equilibrated with 0.1 M sodi-
um acetate, pH 5.0, containing 2 M ammonium sulfate. The recombi-
nant β-glucosidase was eluted from the column with a linear gradient
of 2–0 M ammonium sulfate in 0.1 M sodium acetate, pH 5.0. Fractions
containing β-glucosidase activities were pooled and concentrated, and
ammonium sulfate was removed by centrifugal ultrafiltration. The
protein content of recombinant β-glucosidase was determined at absor-
bance 595 nm by using the Bio-Rad protein assay reagent kit (Bio-Rad,
Hercules, CA, USA) and bovine serum albumin as the standard. 10%
SDS–PAGE was performed to check the purity of the purified enzyme.
To address this need, we utilized site-directed mutagenesis guided
by protein structure homology modeling, in harness with biochemical
analysis, to identify glutamic acid 490 as the general acid/base residue
in the GH3 β-glucosidase from A. niger strain ASKU28.
2. Materials and methods
2.1. Materials
Escherichia coli strains Top10 and DH5α were used for plasmid prop-
agation. Pichia pastoris strain Y11430 (Invitrogen, Carlsbad, CA, USA)
was used for protein expression. A. niger ASKU28 was from the existing
stock collection at the Department of Microbiology, Faculty of Science,
Kasetsart University, Thailand. The substrates used in this study were
cellobiose, which was purchased from Fluka (St. Louis, MO, USA),
p-nitrophenyl-β-D-glucopyranoside (pNP-Glc), which was purchased
from Sigma-Aldrich (St. Louis, MO, USA), and 2,4-dinitrophenyl-
β-D-glucopyranoside (2,4-DNP-Glc), which was kindly provided by Pro-
fessor Stephen G. Withers, University of British Columbia, Vancouver,
Canada. All other chemicals used in this study were of analytical grade.
2.4. Sequence alignment and molecular modeling
To select candidates for site-directed mutagenesis, the amino acid
sequence of the β-glucosidase from A. niger ASKU28 was aligned with
those of the 2 reported A. niger β-glucosidases (GenBank ID: CAB75696
and GenBank ID: CAK48740), 14 enzymes from subfamily 4 of GH3
[28], β-glucosidase B (Bgl3B) from Thermotoga neapolitana [14], β-D-
glucan exohydrolase isoenzyme (ExoI) from Hordeum vulgare (barley)
[15], β-glucosidase I (KmBglI) from Kluyveromyces marxianus [16],
2.2. Cloning of β-glucosidase from A. niger ASKU28
A. niger ASKU28 was grown in minimal medium [26] supplemented
with 2% beechwood xylan for 2 days at 30 °C, 100 rpm. Total RNA was
extracted from the fungal mycelia with an RNeasy Mini Kit (QIAGEN,
Hilden, Germany). cDNA was produced according to the protocol for
first-strand cDNA synthesis (Fermentas, Glen Burnie, MD, USA) by
using oligo (dT)18 as a primer. The coding sequence of mature
β-glucosidase from A. niger ASKU28 (from position 19 of the amino
acid sequence to the stop codon) was amplified from the first strand
cDNA with 3 units of Pfu Ultra HF DNA polymerase (Promega, Madison,
WI, USA) and specific primers, ASKU28_for and ASKU28_rev (Table 1),
that were designed based on the nucleotide sequence of exons 2–7 of
A. niger bgl1 gene (GenBank ID: AJ132386.1) [11]. The PCR products
were ligated into pZErO™-2 cloning vector (Invitrogen, Carlsbad, CA,
USA) at the EcoRV site, and transformed into E. coli strain Top10 by elec-
troporation. The positive transformants were selected by plating onto
LB medium containing 25 μg/ml kanamycin at 37 °C. The coding
sequence of A. niger β-glucosidase was checked by DNA sequencing.
Table 1
Oligonucleotide primers used in this study. The mismatch sequences in the mutagenic
primers are underlined.
Oligonucleotide
primer
Nucleotide sequence (5′ to 3′)
ASKU28_for
ASKU28_rev
D463A_f
D463A_r
D487A_f
D487A_r
E490A_f
E490A_r
D496A_f
D496A_r
D501A_f
D501A_r
TAG AAT TCT TGA TGA ATT GGC CTA CTC CCC
CGT ATC TAG ATT AGT GAA CAG TAG GCA GAG A
GGT GTA TTC ACC GCC ACC GCT AAC TGG GCT ATC GAT C
GAT CGA TAG CCC AGT TAG CGG TGG CGG TGA ATA CAC C
CTT TGT CAA CGC CGC TTC TGG TGA GGG TTA C
GTA ACC CTC ACC AGA AGC GGC GTT GAC AAA G
GTC AAC GCC GAC TCT GGT GCT GGT TAC ATC AAT GTG GAC
GTC CAC ATT GAT GTA ACC AGC ACC AGA GTC GGC GTT GAC
GAG GGT TAC ATC AAT GTG GCT GGA AAC CTG GGT GAC C
GGT CAC CCA GGT TTC CAG CCA CAT TGA TGT AAC CCT C
GTG GAC GGA AAC CTG GGT GCT CGC AGG AAC CTG ACC CTG
CAG GGT CAG GTT CCT GCG AGC ACC CAG GTT TCC GTC CAC

P. Thongpoo et al. / Biochimica et Biophysica Acta 1830 (2013) 2739–2749
2741
β-glucosidase from Flavobacterium meningosepticum [17] and
glucosylceramidase from Paenibacillus sp. TS12 [12] by using Clustal
Omega (1.1.0) [29].
standard samples of p-nitrophenol in all buffer systems after addition
of the carbonate stop-solution were within experimental error of this
published value.
In parallel, the amino acid sequence of β-glucosidase from ASKU28
was submitted to the Phyre2 server [30] to predict the fold of the enzyme.
The final model from the server was then subjected to the Chiron server
[31] to evaluate and resolve steric clashes in the model before further
analysis. The stereochemical qualities of the energy-minimized model
were assessed by the Procheck v3.5 [32] and Verify-3D [33] by submitting
the atomic coordinates to the Swiss Model server (swissmodel.expasy.
org) and the NIH MBI laboratory server (nihserver.mbi.ucla.edu), respec-
tively. This three-dimensional model was aligned with the crystal struc-
tures of ExoI from barley in complex with glucose (PDB ID: 1ex1) [15]
by using the “align” command in the PyMOL software program
(Schrödinger LLC, Portland, OR, USA).
The kinetic parameters of the recombinant wild-type β-glucosidase
and the E490A variant toward 2,4-DNP-Glc were determined in 0.1 M
sodium citrate, pH 3–6.4, and 0.1 M potassium phosphate, pH 6.0–8.0,
at 40 °C by a continuous assay wherein the change in UV absorbance
at 400 nm due to the release of 2,4-DNP was monitored spectrophoto-
metrically (Cary 300 spectrophotometer, Varian, Australia). The rates of
2,4-DNP-Glc hydrolysis were calculated from the initial slopes of the
change in A400 versus time (obtained by linear least-squares fitting)
by using extinction coefficients determined for 2,4-dinitrophenol stan-
dards at each pH value.
2.7. Azide rescue
2.5. Site-directed mutagenesis and production of enzyme variants
The kinetic parameters of the recombinant wild-type β-glucosidase
and the E490A variant toward pNP-Glc in the presence of sodium azide
were determined in 0.1 M sodium citrate, pH 6.0, containing 0–1 M so-
dium azide, for 10 min at 40 °C. Reactions were stopped by adding 2 M
sodium carbonate. The concentration of pNP released from the reactions
was measured as described in Section 2.6.
The kinetic parameters of the E490A variant toward 2,4-DNP-Glc
in the presence of sodium azide were determined in 0.1 M sodium
citrate, pH 6.0, containing 0–2 M sodium azide, at 40 °C. The rates
of 2,4-DNP-Glc hydrolysis were determined by a continuous assay
as described earlier. For this substrate and protein, a linear depen-
dence of the initial rate of reaction (v0) on substrate concentration
[S] was observed, indicative of [S]≪Km conditions; kcat/Km values
were therefore determined by fitting the following equation to the
data [36]: v0 =(kcat/Km) [E]t [S].
The selected acidic residues, namely D463, D487, E490, D496, and
D501, were replaced individually with alanine by site-directed muta-
genesis by using 2 units of Phusion™ high-fidelity DNA polymerase
(Finnzymes, Finland) according to the method published previously
[34]. The residue numbers indicate their positions in the mature se-
quence of β-glucosidase from A. niger as reported previously [11].
The sequences of sense/antisense mutagenic primer pairs are listed
in Table 1. The PCR products were treated with 10 units DpnI at
37 °C overnight and transformed into the competent E. coli Top10F′
by heat shock method. The positive transformants were selected by
plating onto LB agar plates containing 25 μg/mL zeocin at 37 °C over-
night. Plasmids containing the correct mutation were identified by re-
striction digestion and DNA sequencing. Then, the mutant plasmids
were linearized with PmeI, and transformed into P. pastoris Y11430
by electroporation. The positive transformed clones were selected
on YPDS agar containing 100 μg/mL zeocin after incubation for 3–
5 days at 30 °C. The positive clones from each mutation were selected
for small scale expression as described in Section 2.3. The P. pastoris
clones that showed the highest activities or the highest protein pro-
duction from each mutation were selected for large-scale expression,
and the enzyme variants were purified from the culture media of P.
pastoris by the same method as described in Section 2.3.
1H-NMR spectroscopy was utilized to determine the reaction
products of the E490A variant with pNP-Glc and 2,4-DNP-Glc in
the presence of sodium azide. 1H-NMR data were recorded on a
Bruker DMX 500 instrument at 500 MHz and 25 °C. Chemical shifts
(δ) were quoted in parts per million (ppm) referenced to the
solvent peak (HOD at 4.79 ppm). Determination of the anomeric con-
figurations was based on the observed chemical shifts and coupling
constants (J values, reported in Hz).
Table 2
2.6. Kinetic analysis
Kinetic parameters of the recombinant wild-type β-glucosidase and mutant forms of A.
niger ASKU28 β-glucosidase against cellobiose, pNP-Glc and 2,4-DNP-Glc in 0.1 M sodi-
um citrate, pH 4.0, at 40 °C.
Kinetic parameters of the mutant forms of β-glucosidase from
A. niger ASKU28 were determined toward cellobiose, pNP-Glc and
2,4-DNP-Glc, and compared to those of wild-type β-glucosidase. All kinet-
ic measurements were performed in duplicate. The kinetic parameters
were calculated by non-linear curve-fitting of the Michaelis–Menten
equation with Origin, version 8 (OriginLab, Northampton, MA, USA).
The kinetic parameters of the purified enzymes toward cellobiose
were determined in 0.1 M sodium citrate, pH 4.0, at 40 °C for 10 min
(30 min for the E490A variant). The reactions were stopped by
heating to 95 °C for 5 min, and the released glucose from the reaction
was reacted with a glucose oxidase/peroxidase reagent (Sigma, St.
Louis, MO, USA) for 15 min at 37 °C. The amount of glucose was de-
termined by reading the absorbance at 400 nm, and comparing it to
a standard curve of glucose.
The kinetic parameters of the purified enzymes toward pNP-Glc
were determined in 0.1 M sodium citrate, pH 4.0, at 40 °C for 10 min
to generate the data in Table 2. The pH dependence of the rate of
enzyme-catalyzed pNP-Glc hydrolysis was determined in either
50 mM sodium citrate (pH 2.0–6.4) or 50 mM potassium phosphate
(pH 6.0–8.0), at 40 °C for 10 min. In all cases, reactions were stopped
by the addition of 0.2 M sodium carbonate, and the amount of pNP re-
leased was calculated from A400 values using an extinction coefficient
of 18,300 M−1 cm−1 [35]. Measured extinction coefficients from
Substrate
Enzyme
Km
(mM)
kcat
kcat/Km
(s−1
)
(s−1 mM−1
)
Cellobiose
Wild-type
D463A
D487A
E490A
D496A
D501A
Wild-type
D463A
D487A
E490A
D496A
D501A
Wild-type
E490A
Wild-type
E490A
1.11 0.09
1.00 0.05
1.03 0.01
1.02 0.05
0.57 0.02
1.36 0.04
0.54 0.04
0.64 0.03
0.55 0.02
6.09 0.15
0.44 0.03
0.97 0.02
0.49 0.06
0.10 0.01
0.26 0.04
0.015 0.003
67.4 1.9
78.4 0.9
87.7 0.5
0.001 0.00003
78.3 1.5
47.3 0.5
60.6 5.4
78.7 4.4
85.2 0.8
0.001 0.0001
138
6
34.7 1.0
263 18
249 12
297 14
0.05 0.002
321 21
pNP-Glca
141
160
163
3
3
3
0.32 0.01
141
104
4
1
107
2
2,4-DNP-Glc
422 25
2.12 0.13
71 4.0
863 123
21.2 2.5
273 44
2,4-DNP-Glc
(pH 6.0)
0.34 0.02
22.6 4.7
a
Non-Michealian kinetics were observed for pNP-Glc for the wild-type and most
variants, which were indicative of substrate inhibition or transglycosylation at high
[S] (Supplementary Fig. 7A). To simplify comparative kinetic analysis, apparent kcat
and Km values are reported for pNP-Glc, which were obtained by fitting the standard
Michaelis–Menten equation to data at [S]b2.5 mM (e.g. Supplementary Fig. 7B), except
in the case of the E490A variant, where data in the range [S]=0–20 mM were used
(Supplementary Fig. 8).

2742
P. Thongpoo et al. / Biochimica et Biophysica Acta 1830 (2013) 2739–2749
Two reaction conditions were analyzed: (1) 36 μM of the E490A
1), which decreased to 105 kDa after deglycosylation by Endo Hf
enzyme (Supplementary Fig. 2), thus suggesting glycosylation by the
yeast host, P. pastoris [37]. In keeping with this observation, the se-
quence of the mature β-glucosidase is predicted to contain 11 con-
served N-glycosylation sites.
variant with 10 mM pNP-Glc in the presence of 50 mM sodium
azide in 0.1 M sodium citrate, pD 6.0, at 37 °C, and (2) 31 nM of the
E490A variant with 0.5 mM 2,4-DNP-Glc in the presence of 1 M sodi-
um azide in 0.1 M sodium citrate, pD 6.1, at 37 °C. The E490A variant
enzyme, sodium citrate (0.25 M, pH 6.0), sodium azide and substrates
were each freeze-dried and subsequently dissolved in D2O three
times to reduce the proton concentration in the samples.
3.2. Sequence alignment, molecular modeling and production of mutant
enzymes
An initial 1H-NMR spectrum of the mixture of sodium citrate buffer,
sodium azide and the substrates was performed for comparison, prior
to addition of the enzyme solution and subsequent NMR analysis over
various time intervals. When pNP-Glc was used as substrate, each ex-
periment was performed with 16 scans, whereas 128 scans acquired
when 2,4-DNP-Glc was used as substrate. The following experimentally
determined data for the substrates were used as spectral references:
The conserved acidic residues that could serve as the acid/base
catalyst of the β-glucosidase were predicted by primary and tertiary
structural alignments. Ninety-nine enzymes from GH3 were previ-
ously used to establish six subfamilies by phylogenetic analysis [28];
although the family now contains over 4000 members (http://www.
cazy.org/GH3.html), these subfamilies are still broadly useful for de-
scriptive purposes. In particular, the acid/base catalyst has been pro-
posed to be conserved within the same subfamily but not conserved
across subfamilies [17]. Sequence alignment of the A. niger ASKU28
β-glucosidase and β-glucosidases from A. aculeatus and A. kawachii
in subfamily 4 showed 83% and 98% sequence identity, respectively,
which confirmed assignment to this subfamily. The sequence of
A. niger ASKU28 β-glucosidase was then aligned with those of the
two previously reported β-glucosidases from A. niger [11,22], 14
enzymes from subfamily 4 of GH3 glycoside hydrolases [28], and 5
GH3 enzymes in which the identities of the acid/base catalysts are
known (Supplementary Fig. 3).
pNP-Glc: 1H NMR (500 MHz, D2O): δ 8.24–8.21 (m, 2H, Ar–H), 7.23–
7.20 (m, 2H, Ar–H), 5.27–5.22 (d, 1H, H1), 3.93 (dd, J=12.4 and
2.1 Hz, 1H, H6), 3.75 (dd, J=12.4 and 5.7 Hz, 1H, H6′), 3.69 (ddd,
J=9.7, 5.7 and 2.2 Hz, 1H, H5), 3.65–3.60 (m, 2H, H3, H4), 3.54–
3.48 (m, 1H, H2).
2,4-DNP-Glc: 1H NMR (500 MHz, D2O): δ 8.90 (d, J=2.8 Hz, 1H, Ar–
H), 8.54 (dd, J=9.4 and 2.8 Hz, 1H, Ar–H), 7.61 (d, J=9.3 Hz, 1H,
Ar–H), 5.43 (d, J=7.6 Hz, 1H, H1), 3.94 (dd, J=12.4 and 2.1 Hz,
1H, H6), 3.77 (dd, J=12.3 and 5.6 Hz, 1H, H6′), 3.74–3.68 (m, 2H,
H5, H3), 3.64 (at, 1H, J=9.3 Hz, H4), 3.54 (at, 1H, J=9.6 Hz, H2).
From this alignment, the conserved catalytic nucleophile was iden-
tified as D261 within the VMSDW motif previously demonstrated for
the A. niger B1 homologue [11]. However, potential candidate acid/
base residues were not clearly identifiable, and moreover appeared
to occur in a highly variable region of the sequence when considering
other GH3 enzymes whose acid/base catalyst had been indicated by
structural or kinetic analysis (Supplementary Fig. 3). In particular, it
was only possible to achieve tenuous alignment of E458 of Bgl3B
from T. neapolitana [14], E590 of KmBglI from K. marxianus [16],
E473 of β-glucosidase from F. meningosepticum [17], E411 of
glucosylceramidase from Paenibacillus sp. TS12 [12], and E491 of
ExoI from barley [15] with E490 of A. niger ASKU28 β-glucosidase,
which included significant gaps and mismatches in the neighboring
residues. The acid/base catalysts of β-N-acetylglucosaminidase
(NagZ) from Bacillus subtilis and β-N-acetylglucosaminidase 3A from
Clostridium paraputrificum are located much further away from this re-
gion [13,18]. There are seven conserved acidic residues in A. niger
ASKU28 β-glucosidase in the region flanking the reported acid/base cat-
alysts, namely D463, E471, D487, E490, D496, D501 and D512. Of these,
the residues E490 and D501 are fully conserved within subfamily GH3 4
enzymes.
In light of the ambiguity presented by sequence alignment, we then
employed three-dimensional homology modeling to further inform our
choice of target residues for site-directed mutagenesis. As of June 2012,
the crystal structures of 6 bacterial and 2 eukaryotic GH3 members have
been solved, as collected in the Carbohydrate Active EnZymes database
(CAZy) [2]. These include the bacterial representatives SYNPCC7002_
A0075 from Synechococcus sp. (unpublished, PDB ID: 3sql), DR1333
from Deinococcus radiodurans R1 (unpublished, PDB ID: 3tev), exo-1,3/
1,4-β-glucanase (ExoP) from Pseudoalteromonas sp. [38], Bgl3B from
T. neapolitana [14], β-N-acetylhexosaminidase from Vibrio cholerae
[39,40], and NagZ from B. subtilis [13]. The two eukaryotic structures
were those of the ExoI from barley [15] and KmBglI from K. marxianus
[16]. The common structure of GH3 members includes an N-terminal
(α/β)8 barrel which bears the catalytic nucleophilic residue. With the
exception of certain bacterial β-N-acetylglucosaminidases, this domain
is followed by a C-terminal (α/β)6 sandwich domain [1,13,28]. Rooted
in seminal structural studies of the barley ExoI β-glucosidase [10,15]
the catalytic acid/base residue appears to be generally borne on this
second domain, although the His–Asp dyad implicated in one- and
To detect the potential of the wild-type enzyme to utilize azide as
a transglycosylation acceptor, analogous reactions were performed
in 0.1 M sodium citrate, pH 6.0, at 37 °C and analyzed by thin-layer
chromatography (TLC). 150 nM of the wild-type enzyme or 36 μM
of the E490A variant were incubated with 10 mM pNP-Glc in the
presence or absence of 50 mM sodium azide. Likewise, 6 nM of the
wild-type enzyme or 31 μM of the E490A variant were incubated
with 5 mM 2,4-DNP-Glc in the presence or absence of 1 M sodium
azide. TLC was performed on silica gel 60 F254 aluminum plates
(Merck, Darmstadt, Germany), followed by development in 7:2:1
(v/v/v) ethyl acetate/methanol/water and visualization with UV illu-
mination and dipping into 10% sulfuric acid in methanol followed by
charring.
3. Results and discussion
3.1. Cloning and expression of β-glucosidase from A. niger ASKU28
The coding sequence of mature β-glucosidase from A. niger ASKU28
(from position 19 of the amino acid sequence to the stop codon) was
amplified with specific primers that were designed based on the nucle-
otide sequence of A. niger β-glucosidase (GenBank ID: CAB75696.1).
The resulting RT-PCR product was 2546 bp long (GenBank ID:
JX127252) and was found to be similar to the sequence of exons 2–7
of A. niger bgl1 gene (GenBank ID: AJ132386.1) [11] with 32 positional
differences. After translation to protein, the sequence of the recombi-
nant β-glucosidase from A. niger ASKU28 was found to be nearly identi-
cal to that of β-glucosidase from A. niger B1 (GenBank ID: CAB75696.1)
[11] with 1 amino acid difference of tyrosine at position 189 instead of
phenylanine. Interestingly, the amino acid sequence of a homologous
β-glucosidase from the genome sequence of A. niger CBS 513.88
(GenBank ID: CAK48740.1) has tyrosine at position 189, but differs
from the A. niger ASKU28 sequence at 21 positions [22].
The coding sequence of the mature β-glucosidase from A. niger
ASKU28 was then expressed as an N-terminal fusion protein with the
α-mating factor and a polyhistidine tag. The expected size of the mature
recombinant enzyme after the removal of the α-mating factor is
93 kDa. However, the recombinant enzyme showed an apparent mo-
lecular weight of 130 kDa by SDS–PAGE analysis (Supplementary Fig.

P. Thongpoo et al. / Biochimica et Biophysica Acta 1830 (2013) 2739–2749
2743
two-domain bacterial β-GlcNAc'ases is contained within the first
domain [13,19]. The structures of some GH3 enzymes are elaborated
with additional domains. The extra C-terminal domain of ExoP from
Pseudoalteromonas sp. appears to play a role in protein stability and
may assist in substrate orientation [38]. KmBglI from K. marxianus has
a PA14 domain inserted into the (α/β)6 sandwich domain that influ-
ences carbohydrate-binding, as well as a C-terminal fibronectin type III
domain [16]. A similar additional C-terminal fibronectin type III domain
is also found in Bgl3B from T. neapolitana and was suggested to be in-
volved in substrate recognition and affect enzyme thermostability [14].
In this context, the three-dimensional structure of A. niger ASKU28
β-glucosidase was predicted by Phyre2 to contain (α/β)8 barrel and
(α/β)6 sandwich domains similar to the structure of BglI from the
yeast K. marxianus (PDB ID: 3ac0, [16]); these enzymes share 33%
amino acid identity. The quality of the model was further improved
by an automated minimization by using the Chiron server. The
Ramachandran plot (Supplementary Table 1), ProSa Z-scores, and
Verify3D scores (Supplementary Fig. 4) indicated an acceptable
model quality. Alignment of this model with the crystal structure of
ExoI from barley in complex with glucose (PDB ID: 1ex1) revealed
that residue E490 of the A. niger ASKU28 β-glucosidase was well-
aligned with the predicted acid/base residue, E491, of barley ExoI
(Fig. 1). To ensure completeness, five residues located in the non-
conserved, flexible region of the A. niger β-glucosidase, namely
D463, D487, E490, D496 and D501 (Fig. 1), were systematically
subjected to site-directed mutagenesis to generate alanine variants
for kinetic analysis. Two other residues, E471 and D512, were located
in α-helices distal to the active site, and were not studied further.
Fig. 1. Structural superimposition of the three-dimensional model of A. niger ASKU28 β-glucosidase (cyan) with the crystal structure of barley ExoI in complex with glucose (PDB ID:
1ex1, brown). (A) Overview (B) close-up of the active site region. Side chains of key residues are represented as stick models. The nucleophile and acid/base catalyst of ExoI (D285
and E491, respectively) are shown in magenta and the glucose molecule in the binding pocket of ExoI is shown in green. The side chain of the nucleophile (D261) in A. niger ASKU28
β-glucosidase is shown in purple. The side chains of the seven conserved acidic residues in A. niger ASKU28 β-glucosidase (D463, E471, D487, E490, D496, D501, D512) are shown in
yellow. The picture was generated in PyMOL (Schrödinger LLC, Portland, OR, USA).

2744
P. Thongpoo et al. / Biochimica et Biophysica Acta 1830 (2013) 2739–2749
A
A
3.0
2.5
2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
100000
10000
1000
100
10
2
3
4
5
6
7
8
9
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0
pH
pH
B
3.5
3.0
2.5
2.0
1.5
1.0
B
100000
10000
1000
100
2
3
4
5
6
7
8
9
pH
10
Fig. 2. pH dependence of kinetic parameters of the wild-type A. niger ASKU28
β-glucosidase (circles) and the E490A variant (squares) for the hydrolysis of 2,4-DNP-Glc
in 0.1 M sodium citrate, pH 2.0–6.4 (filled symbols) and 0.1 M potassium phosphate, pH
6.0–8.0 (open symbols). (A) plot of log (kcat) vs pH; and (B) plot of log (kcat/Km) vs pH.
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0
pH
Fig. 3. pH dependence of the specific activities of the wild-type and all mutant forms of
β-glucosidase from A. niger ASKU28 in the hydrolysis of pNP-Glc. The reactions were
performed with pNP-Glc at 2 mM for the wild-type and three mutant enzymes,
4 mM for the D501A mutant, and 24 mM for the E490A mutant, in 50 mM sodium
citrate, pH 2.0–6.4 (solid lines) and 50 mM potassium phosphate, pH 6.0–8.0 (dash
lines). (A) pH profile of all enzymes; and (B) pH profile of the wild-type enzyme and
the E490A mutant in the presence (open symbols) and absence (filled symbols) of
50 mM sodium azide. Wild-type (●), D463A (▲), D487A (♦), E490A (■), D496A (*)
and D501A (+).
3.3. Kinetic analysis
The steady-state kinetic parameters of the recombinant wild-type
and mutant forms of A. niger ASKU28 β-glucosidase against cellobiose,
pNP-Glc and 2,4-DNP-Glc, were determined at the optimal pH, 4.0,
and 40 °C (Table 2 and Supplementary Figs. 5–10). For cellobiose and
pNP-Glc, the D463A, D487A, D496A, and D501A variants all exhibited
kinetic constants that were similar to the wild-type enzyme, although
small variations in kcat and Km values were observed, indicating subtle
effects on catalysis (Table 2).
In stark contrast, the E490A variant exhibited a large depression of
the kcat and kcat/Km values (at least 67,000- and 61,000-fold, respec-
tively) in comparison to the wild-type enzyme when cellobiose was
used as a substrate, although the Km value was within error of that
of the wild-type enzyme (Table 2, Supplementary Figs. 5 & 6). With
the aryl β-glucoside pNP-Glc, the E490A variant also exhibited signif-
icant decreases in the apparent kcat and kcat/Km values (440- and
5300-fold, respectively) when compared to the wild-type enzyme
(Table 2). The accompanying 11-fold increase in the apparent Km
value of the E490A variant for this substrate may be explained by
elimination of substrate transglycosylation [23] and a resulting shift
from non-Michaelian to classic Michaelis–Menten saturation kinetics
[41], as reflected in vo/[E]t vs. [S] plots (Supplementary Figs. 7 & 8).
When the activity of the E490A variant was tested on 2,4-DNP-Glc,
which bears the good leaving group 2,4-dinitrophenol (pKa 4.1), the
kcat and Km values were ca. 200- and 5-fold lower relative to the
wild-type enzyme, resulting in a ca. 40-fold decrease of the kcat/Km
value at the pH optimum of the wild-type enzyme, 4.0 (Table 2). As a
prelude to azide rescue studies and pH-rate profile measurements
(vide infra), we also measured activity toward 2,4-DNP-Glc at pH 6.0.
In this case, the kcat and Km values of the E490A variant were reduced
ca. 210- and 17-fold relative to the wild-type enzyme, resulting in a
ca. 12-fold decrease of the kcat/Km value (Table 2). At all pH values ex-
amined, plots of vo/[E]t vs. [2,4-DNP-Glc] for both the wild-type and
E490A variant were classically Michaelian over the substrate ranges
Table 3
The effect of sodium azide on kinetic parameters of the recombinant wild-type
β-glucosidase toward pNP-Glc at pH 6.0, 40 °C for 10 min.
[NaN3] (mM)
Km (mM)
kcat (s−1
)
kcat/Km (s−1 mM−1
)
0
5
0.35 0.01
0.38 0.02
0.39 0.02
0.37 0.003
0.38 0.004
0.32 0.01
0.31 0.008
37.2 0.2
37.7 1.1
37.7 1.3
36.1 0.2
34.6 0.1
24.2 0.5
15.4 0.2
105
2
98.2 6.5
97.6 6.7
96.7 1.1
91.4 1.0
74.6 3.0
48.7 1.3
10
30
50
200
500

P. Thongpoo et al. / Biochimica et Biophysica Acta 1830 (2013) 2739–2749
2745
Table 4
E490A variant indicated the removal of catalytic assistance from general
acid reside in the enzyme glycosylation step. A similar effect on kcat/Km
values has been observed for the acid/base catalyst variant of the
Cellulomonas fimi glycanase, Cex [48].
The effect of sodium azide on kinetic parameters of the E490A mutant toward pNP-Glc
at pH 6.0, 40 °C for 10 min.
[NaN3] (mM)
Km (mM)
kcat (s−1
)
kcat/Km (s−1 mM−1
)
For pNP-Glc, the precise effects of E490A mutation on the pH depen-
dence of the kcat and kcat/Km values are difficult to divine, because of the
complex kinetic profile exhibited by the wild-type enzyme (Supple-
mentary Fig. 7A), which is indicative of substrate transglycosylation
[41]. Contrasted with the simple Michaelian kinetics exhibited by the
E490A variant, it is therefore difficult to draw meaningful conclusions
regarding individual kinetic steps. Nonetheless, we measured specific
activity values for the wild-type and all variants at either the substrate
concentration yielding the maximum rate, or in the case of the E490A
variant, at a concentration four-fold greater than the Km value
(Fig. 3A). Whereas the wild-type enzyme and all of the other variants
exhibited curved pH-rate profiles as expected, that of the E490A was es-
sentially linear, mirroring the results for 2,4-DNP-Glc (Fig. 2).
0
1
2
5
10
30
50
200
500
800
1000
1.49 0.06
9.80 0.64
8.40 0.55
5.72 0.88
8.61 0.67
9.26 0.39
8.33 0.51
7.92 0.62
4.85 0.23
3.52 0.08
7.01 1.00
0.15 0.003
0.89 0.03
0.79 0.03
0.58 0.07
0.84 0.05
0.88 0.03
0.85 0.04
0.87 0.05
0.56 0.02
0.42 0.004
0.72 0.06
0.10 0.005
0.09 0.01
0.09 0.01
0.10 0.02
0.10 0.01
0.09 0.01
0.10 0.01
0.11 0.01
0.12 0.01
0.12 0.003
0.10 0.02
examined (≤1 mM for wild-type and≤0.1 mM for the E490A variant;
Supplementary Figs. 9–10 and data not shown).
Taken together, these results above are consistent with the assign-
ment of the side chain of E490 as acid/base catalyst in the A. niger
ASKU28 β-glucosidase. By analogy with the homologous enzyme from
A. niger strain B1 [11], the GH3 β-glucosidase can be reliably predicted
to use the canonical retaining mechanism of glycoside hydrolysis. As
such, the hydrolytic breakdown of the common α-glucosyl-enzyme in-
termediate can be expected to be slowed equally for all three substrates
examined here, due to removal of the acid/base catalyst in the mutant
enzyme. For substrates with poor leaving groups, effects on the first
chemical step, as revealed by kcat/Km values, are expected to be especial-
ly significant, as observed for the likely natural substrate, cellobiose
(leaving-group pKa ca. 16 [42]). The similar magnitude of reduction of
both the kcat and kcat/Km values for cellobiose in the E490A variant, be-
cause the Km value is unaffected relative to the wild-type, implies that
enzyme glycosylation (k2) is rate-determining for this substrate.
In contrast, significant reductions of the Km value are observed in
the E490A variant at both optimal (4.0) and non-optimal (6.0) pH
values on the artificial substrate 2,4-DNP-Glc (Table 2), the aglycone
of which is a comparatively good leaving group (pKa 4.1) that gener-
ally does not benefit greatly from proton assistance in glycosidases
[43]. Given the complex dependence of Km on both the rate of enzyme
glycosylation (k2) and deglycosylation (k3) in two-step enzymes such
as retaining glycosidases (Km =[k3/(k2 +k3)]·[(k2+k−1)/k1]; [44]),
these observations suggest that k3 is now significantly rate-limiting
due to removal of the acid/base catalyst. Consequently, kcat will reflect
an increased dependence on, and may become equivalent to, k3
(kcat =k2 ·k3/(k2 +k3); [44]). This effect is most pronounced at
pH 6.0 and above (Table 2 and Fig. 2).
3.4. Chemical rescue by azide
Azide rescue experiments were undertaken to further confirm the
identity of E490 as the catalytic acid/base of β-glucosidase from A. niger
ASKU28. The kinetic parameters for pNP-Glc and 2,4-DNP-Glc were de-
termined for the wild-type and E490A variant in the presence of in-
creasing amounts of sodium azide at 40 °C and pH 6.0 (Tables 3–5).
This non-optimal pH was chosen to minimize the amount of volatile,
toxic hydrazoic acid (HN3, pKa 4.6) in the assay. High concentrations
of azide had a slightly inhibitory effect on the wild-type enzyme (50%
inhibition at 500 mM) due exclusively to the reduction of the kcat
value (Table 3); a similar inhibitory effect of azide was previously
observed, for example, in a wild-type GH16 enzyme [44]. The addition
of azide did not significantly affect the pH-rate profile of the wild-type
A. niger ASKU28 β-glucosidase (Fig. 3B).
In contrast, the addition of sodium azide into assay mixtures of the
E490A variant and pNP-Glc resulted in the direct activation of the en-
zyme reaction by increasing kcat values, even at low concentrations
(1 mM, Table 4). Km values were simultaneously increased, possibly
reflecting decreased accumulation of the glycosyl-enzyme due to a
shift in the relative rates of formation and breakdown. Moreover,
the effect of azide activation on specific activity values was pH inde-
pendent over the range 3–8 (Fig. 3B).
Importantly, reaction product monitoring by NMR in a parallel reac-
tion containing 50 mM sodium azide and 10 mM pNP-Glc (Fig. 4) di-
rectly indicated the production of β-glucosyl azide, the expected
product from decomposition of the α-glucosyl-enzyme intermediate
by nucleophilic attack (Scheme 1). After 1.5 h of incubation, 12% of
the β-glucosyl azide product was produced, as indicated by the pres-
ence of a characteristic doublet at 4.73 ppm (J=8.8 Hz), corresponding
to the H1 peak [49,50]. Concomitantly, peaks due to the p-nitrophenol
by-product were also detected at 8.18–8.14 ppm (m, Ar–H) and 6.94–
6.90 ppm (m, Ar–H) as the intensity of the H1 peak of the starting
material (5.27 ppm) decreased. Ultimately, characteristic doublets
corresponding to the H1 peak of the hydrolysis products, α- and
β-D-glucose (5.22 ppm (J=3.8 Hz) and 4.63 ppm (J=8.0 Hz), respec-
tively [51]), were detected after 24 h and 48 h of incubation, at which
time 75% of the substrate had been consumed. A control experiment
containing 10 mM pNP-Glc, 50 mM sodium azide, and 0.1 M of sodium
citrate, pD 6.0, but lacking the enzyme, indicated no background de-
composition of the substrate after 24 h at 37 °C (data not shown).
Similar results were observed upon addition of sodium azide to the re-
action of 2,4-DNP-Glc catalyzed by the E490A variant. Although an essen-
tially linear dependence of vo/[E]t on donor substrate concentration
precluded extraction of individual kinetic parameters for all assays
containing sodium azide (50 mM–2 M), increased kcat/Km values were
indicative of activation (Table 5 and Supplementary Fig. 11). In parallel,
reaction monitoring by 1H NMR again revealed the generation of the
Indeed, we measured kinetic parameters for hydrolysis of 2,4-
DNP-Glc by the wild-type enzyme and the E490A variant over the pH
range 3–8 (Fig. 2). Plots of log(kcat) and log(kcat/Km) versus pH for the
wild-type enzyme were curved as expected for the canonical retaining
glycosidase mechanism involving two ionizable active-site residues.
The observation that the plot of log(kcat) versus pH for the E490A vari-
ant was not strictly linear, as previously observed for other acid/base
residue variants [12,45–47], may reflect a limited degree of buffer or
specific acid catalysis in the deglycosylation step in the low pH regime.
In contrast, the pH-independence of log(kcat/Km) observed for the
Table 5
The kinetic parameters of azide reactivation assays with 2,4-DNP-Glc at pH 6.0, 40 °C.
Enzyme
[NaN3] (mM)
kcat/Km (s−1 mM−1
)
Wild-type
E490A
0
0
50
500
1000
2000
197 29
18.4 5.5
66.1 7.4
174.3 5.7
217.1 9.2
249.2 7.2

2746
P. Thongpoo et al. / Biochimica et Biophysica Acta 1830 (2013) 2739–2749
Fig. 4. 1H-NMR spectra of the reaction between the E490A mutant with 10 mM pNP-Glc in the presence of 50 mM sodium azide in 0.1 M sodium citrate, pD 6.0, at 37 °C.
expected β-glucosyl azide from 0.5 mM 2,4-DNP-Glc and 1 M sodium
azide in the presence of the E490A variant (Fig. 5). After 1.5 h of incuba-
tion, the characteristic doublet of the H1 of β-glucosyl azide was observed
at 4.83 ppm (J=8.8 Hz), and the signal increased over time. The doublet
at 5.53 ppm (J=7.6 Hz), which corresponds to H1 of 2,4-DNP-Glc, con-
tinually decreased and was completely absent after 24 h of incubation.
No hydrolysis products (α- and β-D-glucose) were detected. The slight
shift of spectral peaks compared to reported values [49–51] is ascribed
to solute concentration effects due to complex sample preparation.
Thin-layer chromatographic analysis of analogous reaction mix-
tures containing either phenyl β-glucoside revealed that wild-type
A. niger ASKU28 β-glucosidase produced only the expected phenol
plus glucose, both in the presence and absence of the competitive
nucleophile azide (data not shown, see Section 2.7 for experimental
Scheme 1. Azide rescue mechanism of an alanine mutant of the acid/base catalyst β-glucosidase toward pNP-Glc or 2,4-DNP-Glc. In the absence of sodium azide, the catalytic acid/
base mutant enzyme was trapped in the glucosyl-enzyme intermediate form. The additional sodium azide then acted as an exogenous base to attack the glucosyl-enzyme interme-
diate and reactivated the mutant enzyme yielding β-glucosyl azide product.

P. Thongpoo et al. / Biochimica et Biophysica Acta 1830 (2013) 2739–2749
2747
Fig. 5. 1H-NMR spectra of the reaction between the E490A mutant with 0.5 mM 2,4-DNP-Glc in the presence of 1 M sodium azide in 0.1 M sodium citrate, pD 6.1, at 37 °C.
details). This indicated that transglycosylation to azide, which is
mechanistically possible, is kinetically insignificant in the wild-type
enzyme and a unique property of the E490A variant.
4. Conclusions
The data presented here provide essential experimental support for
the identification of glutamic acid 490 as the acid/base catalyst in the
β-glucosidase of A. niger ASKU28, and, by analogy, the assignment of
this catalytically important residue in a range of fungal GH3 members.
Furthermore, this analysis highlighted a clear link regarding catalytic
sequence conservation between these fungal members of subfamily 4
and bacterial/yeast members of subfamily 5. Given the vast expansion
in the number of members of GH3 since the turn of the millenium, a re-
vised, robust subfamily classification [52–54] is sorely needed to allow
extension of this analysis to the family at-large. We anticipate that our
present work will inform future bioinformatic and structural enzymol-
ogy studies concerning GH3 enzymes in the contexts of rapidly emerg-
ing (meta)genomic sequence data and the industrial utility of these
enzymes in biomass modification.
3.5. Updated comparison of acid/base catalysts among GH3 enzymes
As discussed above, global multiple sequence alignments of GH3
enzyme often fail to successfully identify the homologous catalytic
acid/base residue across diverse family members, due to low conser-
vation of primary protein structure in the flanking regions. The likely
acid/base of A. niger ASKU28 β-glucosidase, E490, occurs within the
sequence motif SGEGY (Fig. 6). E490 is highly conserved within the
consensus sequence (S/A)(G/S/T/R)E(G/N)(Y/V/F/E/T) of GH3 sub-
family 4 enzymes (see also Supplementary Fig. 3). Conservation of
the proposed acid/base also extends to the (S/E)(G/T/S)E(G/S)(Y/S)
motif of GH3 subfamily 5 enzymes whose acid/base catalysts have
been reported [12,14,16,17]. The sequence surrounding the acid/
base catalyst in the distantly related barley ExoI of subfamily 1 [28]
is often not readily aligned with the corresponding region in these
microbial GH3 members using automated alignment algorithms
(especially ClustalW, data not shown; also ClustalΩ, Supplementary
Fig. 3). This specifically highlights the challenge in broadly predicting
this residue across those GH3 subfamilies whose members are com-
prised of two domains [28] based on sequence data alone. In such
cases, three-dimensional structure homology modeling (Fig. 1) can
specifically inform manual sequence alignment (Fig. 6).
Acknowledgments
The authors especially thank Assoc. Prof. Vichien Kitpreechavanich
(Kasetsart University, Thailand) for providing Aspergillus niger ASKU28
and Prof. Stephen G. Withers (University of British Columbia, Canada)
for providing 2,4-DNP-Glc. H.B. also thanks Prof. Withers and Prof.
Gideon J. Davies (University of York, UK) for enlightening discussions.
P.T. is a recipient of the Ph.D. scholarship from the Office of the Higher
Education Commission, Thailand. A.C.A. was supported by the Swedish
Research Council Formas (via “CarboMat — The KTH Advanced

2748
P. Thongpoo et al. / Biochimica et Biophysica Acta 1830 (2013) 2739–2749
[6] D.J. Vocadlo, C. Mayer, S. He, S.G. Withers, Mechanism of action and identification of
Asp242 as the catalytic nucleophile of Vibrio furnisii N-acetyl-β-D-glucosaminidase
using 2-acetamido-2-deoxy-5-fluoro-alpha-L-idopyranosyl fluoride, Biochemistry
39 (2000) 117–126.
[7] D.L. Zechel, S.G. Withers, Glycosidase mechanisms: anatomy of a finely tuned catalyst,
Acc. Chem. Res. 33 (1999) 11–18.
[8] M.L. Sinnott, Catalytic mechanism of enzymic glycosyl transfer, Chem. Rev. 90
(1990) 1171–1202.
[9] G.J. Davies, M.L. Sinnott, Sorting the diverse: the sequence-based classifications of
carbohydrate-active enzymes, Biochem. J. (2008), http://dx.doi.org/10.1042/
bj20080382.
[10] M. Hrmova, J.N. Varghese, R. De Gori, B.J. Smith, H. Driguez, G.B. Fincher, Catalytic
mechanisms and reaction intermediates along the hydrolytic pathway of a plant
β-D-glucan glucohydrolase, Structure 9 (2001) 1005–1016.
[11] S. Dan, I. Marton, M. Dekel, B.A. Bravdo, S. He, S.G. Withers, O. Shoseyov, Cloning
expression characterization and nucleophile identification of family 3, Aspergillus
niger β-glucosidase, J. Biol. Chem. 275 (2000) 4973–4980.
[12] K. Paal, M. Ito, S.G. Withers, Paenibacillus sp. TS12 glucosylceramidase: kinetic
studies of a novel sub-family of family 3 glycosidases and identification of the
catalytic residues, Biochem. J. 378 (2004) 141–149.
[13] S. Litzinger, S. Fischer, P. Polzer, K. Diederichs, W. Welte, C. Mayer, Structural and
kinetic analysis of Bacillus subtilis N-acetylglucosaminidase reveals a unique Asp–His
dyad mechanism, J. Biol. Chem. 285 (2010) 35675–35684.
[14] T. Pozzo, J.L. Pasten, E.N. Karlsson, D.T. Logan, Structural and functional analyses
of β-glucosidase 3B from Thermotoga neapolitana: a thermostable three-domain
representative of glycoside hydrolase 3, J. Mol. Biol. 397 (2010) 724–739.
[15] J.N. Varghese, M. Hrmova, G.B. Fincher, Three-dimensional structure of a barley
β-D-glucan exohydrolase, a family 3 glycosyl hydrolase, Structure 7 (1999) 179–190.
[16] E. Yoshida, M. Hidaka, S. Fushinobu, T. Koyanagi, H. Minami, H. Tamaki, M.
Kitaoka, T. Katayama, H. Kumagai, Role of
substrate specificity of glycoside hydrolase family
Kluyveromyces marxianus, Biochem. J. 431 (2010) 39–49.
a
PA14 domain in determining
a
3
β-glucosidase from
[17] J. Chir, S. Withers, C.F. Wan, Y.K. Li, Identification of the two essential groups in
the family 3 β-glucosidase from Flavobacterium meningosepticum by labelling
and tandem mass spectrometric analysis, Biochem. J. 365 (2002) 857–863.
[18] H. Li, G. Zhao, H. Miyake, H. Umekawa, T. Kimura, K. Ohmiya, K. Sakka, Identification
of a catalytic residue of Clostridium paraputrificum N-acetyl-β-D-glucosaminidase
Nag3A by site-directed mutagenesis, Biosci. Biotechnol. Biochem. 70 (2006)
1127–1133.
[19] J.-P. Bacik, G.E. Whitworth, K.A. Stubbs, D.J. Vocadlo, B.L. Mark, Active Site Plastic-
ity within the Glycoside Hydrolase NagZ Underlies a Dynamic Mechanism of Sub-
strate Distortion, Chem. Biol. 19 (2012) 1471–1482.
[20] J. van den Brink, R.P. de Vries, Fungal enzyme sets for plant polysaccharide degra-
dation, Appl. Microbiol. Biotechnol. 91 (2011) 1477–1492.
[21] M.H. Jaime Eyzaguirre, A. Leschot, β-Glucosidases from filamentous fungi: properties
structure and applications, in: K.J. Yarema (Ed.), Handbook of Carbohydrate Engi-
neering, CRC Press, Florida, 2005, pp. 645–685.
[22] H.J. Pel, J.H. de Winde, D.B. Archer, P.S. Dyer, G. Hofmann, P.J. Schaap, G. Turner, R.P.
de Vries, R. Albang, K. Albermann, M.R. Andersen, J.D. Bendtsen, J.A.E. Benen, M. van
den Berg, S. Breestraat, M.X. Caddick, R. Contreras, M. Cornell, P.M. Coutinho, E.G.J.
Danchin, A.J.M. Debets, P. Dekker, P.W.M. van Dijck, A. van Dijk, L. Dijkhuizen,
A.J.M. Driessen, C. d'Enfert, S. Geysens, C. Goosen, G.S.P. Groot, P.W.J. de Groot, T.
Guillemette, B. Henrissat, M. Herweijer, J.P.T.W. van den Hombergh, C.A.M.J.J. van
den Hondel, R.T.J.M. van der Heijden, R.M. van der Kaaij, F.M. Klis, H.J. Kools, C.P.
Kubicek, P.A. van Kuyk, J. Lauber, X. Lu, M.J.E.C. van der Maarel, R. Meulenberg, H.
Menke, M.A. Mortimer, J. Nielsen, S.G. Oliver, M. Olsthoorn, K. Pal, N.N.M.E. van
Peij, A.F.J. Ram, U. Rinas, J.A. Roubos, C.M.J. Sagt, M. Schmoll, J. Sun, D. Ussery, J.
Varga, W. Vervecken, P.J.J. van de Vondervoort, H. Wedler, H.A.B. Wosten, A.-P.
Zeng, A.J.J. van Ooyen, J. Visser, H. Stam, Genome sequencing and analysis of the ver-
satile cell factory Aspergillus niger CBS 513.88, Nat. Biotechnol. 25 (2007) 221–231.
[23] C. Bohlin, E. Praestgaard,M. Baumann, K. Borch, J. Praestgaard, R.Monrad, P.Westh,
A comparative study of hydrolysis and transglycosylation activities of fungal
β-glucosidases, Appl. Microbiol. Biotechnol. (in press), http://dx.doi.org/10.
1007/s00253-012-3875-9.
[24] H.F. Seidle, K. McKenzie, I. Marten, O. Shoseyov, R.E. Huber, Trp-262 is a key
residue for the hydrolytic and transglucosidic reactivity of the Aspergillus niger family
3 β-glucosidase: Substitution results in enzymes with mainly transglucosidic activity,
Arch. Biochem. Biophys. 444 (2005) 66–75.
[25] H.F. Seidle, S.J. Allison, E. George, R.E. Huber, Trp-49 of the family 3 β-glucosidase
from Aspergillus niger is important for its transglucosidic activity: creation of
novel β-glucosidases with low transglucosidic efficiencies, Arch. Biochem.
Biophys. 455 (2006) 110–118.
Fig. 6. Manual alignment of partial sequences of GH3 subfamily 4 members including
subfamily 1 and 5 members with identified catalytic acid/base residues. The conserved
glutamic residues are shown as white letters against black background, except for those
that have been experimentally determined previously, which are highlighted in gray.
Full species names and Genbank IDs of the enzymes are as follows, Aspergillus niger
ASKU28 β-glucosidase, JX127252 (this study); A. niger_1 β-glucosidase, CAB75696.1;
A. niger_2 β-glucosidase bglI, CAK48740.1; A. aculeatus β-glucosidase precursor,
BAA10968.1; A. kawachii β-glucosidase, BAA19913.1; Coccidioides posadasii_1 β-
glucosidase, AAB67972.1; C. Posadasii_2 β-glucosidase precursor, AAF21242.1; Cochliobolus
heterostrophus β glucosidase homolog, AAB84005.1; Phaeosphaeria avenaria β-glucosidase,
CAB82861.1; Phanerochaete chrysosporium β-glucosidase, BAB85988.1; Saccharomycopsis
fibuligera_1 β-glucosidase 1 precursor, AAA34314.1; S. fibuligera_2 β-glucosidase 2 precursor,
AAA34315.1; Septoria lycopersici β-1,2-D-glucosidase, AAB08445.1; Trichoderma reesei
β-D-glucoside glucohydrolase, AAA18473.1; Flavobacterium meningosepticum β-glucosidase,
AAB66561.1; Kluyveromyces marxianus KmBglI, ACY95404.1; Paenibacillus sp. TS12
glucosylceramidase, BAC16750.1; Thermotoga neapolitana Bgl3B, ABI29899.1; Hordeum
vulgare ExoI, AAD23382.1.
Carbohydrate Materials Consortium”) and L.S.M. was supported by the
Knut & Alice Wallenberg Foundation (via the Wallenberg Wood Science
Centre). Additional funding from the Michael Smith Laboratories, the
National Research Council of Thailand and the Office of the Higher Edu-
cation Commission, Thailand, is gratefully acknowledged.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://
dx.doi.org/10.1016/j.bbagen.2012.11.014.
References
[26] C. Riou, J.-M. Salmon, M.-J. Vallier, Z. Günata, P. Barre, Purification, characterization,
and substrate specificity of a novel highly glucose-tolerant β-glucosidase from
Aspergillus oryzae, Appl. Environ. Microbiol. 64 (1998) 3607–3614.
[27] P. Toonkool, P. Metheenukul, P. Sujiwattanarat, P. Paiboon, N. Tongtubtim, M.
Ketudat-Cairns, J.R. Ketudat-Cairns, J. Svasti, Expression and purification of
dalcochinase, a β-glucosidase from Dalbergia cochinchinensis Pierre, in yeast and
bacterial hosts, Protein Expr. Purif. 48 (2006) 195–204.
[28] A.J. Harvey, M. Hrmova, R. De Gori, J.N. Varghese, G.B. Fincher, Comparative
modeling of the three-dimensional structures of family 3 glycoside hydrolases,
Proteins 41 (2000) 257–269.
[29] F. Sievers, A. Wilm, D. Dineen, T.J. Gibson, K. Karplus, W. Li, R. Lopez, H.
McWilliam, M. Remmert, J. Soding, J.D. Thompson, D.G. Higgins, Fast, scalable
generation of high-quality protein multiple sequence alignments using Clustal
Omega, Mol. Syst. Biol. 7 (2011).
[1] J.R. Ketudat Cairns, A. Esen, β-Glucosidases, Cell. Mol. Life Sci. 67 (2010) 3389–3405.
[2] B.L. Cantarel, P.M. Coutinho, C. Rancurel, T. Bernard, V. Lombard, B. Henrissat,
The Carbohydrate-Active EnZymes database (CAZy): an expert resource for
Glycogenomics, Nucleic Acids Res. 37 (2009) D233–D238.
[3] S. Sansenya, R. Opassiri, B. Kuaprasert, C.J. Chen, J.R. Ketudat Cairns, The crystal
structure of rice (Oryza sativa L.) Os4BGlu12, an oligosaccharide and tuberonic
acid glucoside-hydrolyzing β-glucosidase with significant thioglucohydrolase
activity, Arch. Biochem. Biophys. 510 (2011) 62–72.
[4] M. Hrmova, A.J. Harvey, J. Wang, N.J. Shirley, G.P. Jones, B.A. Stone, P.B. Høj, G.B.
Fincher, Barley β-D-glucan exohydrolases with β-D-glucosidase activity, J. Biol.
Chem. 271 (1996) 5277–5286.
[5] G. Legler, K.-R. Roeser, H.-K. Illig, Reaction of β-D-glucosidase A3 from Aspergillus
wentii with D-Glucal, Eur. J. Biochem. 101 (1979) 85–92.

P. Thongpoo et al. / Biochimica et Biophysica Acta 1830 (2013) 2739–2749
2749
[30] L.A. Kelley, M.J.E. Sternberg, Protein structure prediction on the Web: a case study
using the Phyre server, Nat. Protoc. 4 (2009) 363–371.
[43] H.D. Ly, S.G. Withers, Mutagenesis of glycosidases, Annu. Rev. Biochem. 68 (1999)
487–522.
[31] S. Ramachandran, P. Kota, F. Ding, N.V. Dokholyan, Automated minimization of
steric clashes in protein structures, Proteins 79 (2011) 261–270.
[32] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton, PROCHECK: a program
to check the stereochemical quality of protein structures, J. Appl. Crystallogr. 26
(1993) 283–291.
[33] D. Eisenberg, R. Lüthy, J.U. Bowie, [20] VERIFY3D: assessment of protein models
with three-dimensional profiles, R.M.S. Charles, W. Carter Jr. (Eds.), Methods
Enzymol. 277 (1997) 396–404, (Academic Press).
[34] J. Braman, C. Papworth, A. Greener, Site-directed mutagenesis using double-
stranded plasmid DNA templates, Methods Mol. Biol. 57 (1996) 31–44.
[35] R.J. Bryant, D.R. Rao, S. Ogutu, α and β-galactosidase activities and oligosaccharide
content in peanuts, Plant Foods Hum. Nutr. 58 (2003) 213–223.
[36] J. Wicki, J. Schloegl, C.A. Tarling, S.G. Withers, Recruitment of both uniform and
differential binding energy in enzymatic catalysis: Xylanases from families 10
and 11, Biochemistry 46 (2007) 6996–7005.
[37] J. Cregg, J. Cereghino, J. Shi, D. Higgins, Recombinant protein expression in Pichia
pastoris, Mol. Biotechnol. 16 (2000) 23–52.
[38] Y. Nakatani, S.M. Cutfield, N.P. Cowieson, J.F. Cutfield, Structure and activity of
exo-1,3/1,4-β-glucanase from marine bacterium Pseudoalteromonas sp. BB1
showing a novel C-terminal domain, FEBS J. 279 (2012) 464–478.
[39] M.D. Balcewich, K.A. Stubbs, Y. He, T.W. James, G.J. Davies, D.J. Vocadlo, B.L. Mark,
[44] J.L. Viladot, E. de Ramon, O. Durany, A. Planas, Probing the mechanism of Bacillus
1,3-1,4-β-D-glucan 4-glucanohydrolases by chemical rescue of inactive mutants
at catalytically essential residues, Biochemistry 37 (1998) 11332–11342.
[45] T. Bravman, V. Belakhov, D. Solomon, G. Shoham, B. Henrissat, T. Baasov, Y.
Shoham, Identification of the catalytic residues in family 52 glycoside hydrolase,
a beta-xylosidase from Geobacillus stearothermophilus T-6, J. Biol. Chem. 278
(2003) 26742–26749.
[46] Y.-K. Li, J. Chir, S. Tanaka, F.-Y. Chen, Identification of the general acid/base catalyst of
a family 3 β-glucosidase from Flavobacterium meningosepticum, Biochemistry 41
(2002) 2751–2759.
[47] Q. Wang, D. Trimbur, R. Graham, R.A.J. Warren, S.G. Withers, Identification of the
acid/base catalyst in Agrobacterium faecalis beta-glucosidase by kinetic analysis of
mutants, Biochemistry 34 (1995) 14554–14562.
[48] A.M. MacLeod, D. Tull, K. Rupitz, R.A.J. Warren, S.G. Withers, Mechanistic conse-
quences of mutation of active site carboxylates in a retaining β-1,4-glycanase
from Cellulomonas fimi, Biochemistry 35 (1996) 13165–13172.
[49] M. Moracci, A. Trincone, G. Perugino, M. Ciaramella, M. Rossi, Restoration of the
activity of active-site mutants of the hyperthermophilic β-glycosidase from
Sulfolobus solfataricus: dependence of the mechanism on the action of external
nucleophiles, Biochemistry 37 (1998) 17262–17270.
[50] U. Sirion, H.J. Kim, J.H. Lee, J.W. Seo, B.S. Lee, S.J. Lee, S.J. Oh, D.Y. Chi, An efficient F-18
labeling method for PET study: Huisgen 1,3-dipolar cycloaddition of bioactive
substances and F-18-labeled compounds, Tetrahedron Lett. 48 (2007) 3953–3957.
[51] J.E. Gurst, NMR and the structure of D-glucose, J. Chem. Educ. 68 (1991) 1003.
[52] H. Aspeborg, P.M. Coutinho, Y. Wang, H. Brumer, B. Henrissat, Evolution, substrate
specificity and subfamily classification of glycoside hydrolase family 5 (GH5),
BMC Evol. Biol. 12 (2012) 186.
Insight into
a strategy for attenuating AmpC-mediated β-lactam resistance:
Structural basis for selective inhibition of the glycoside hydrolase NagZ, Protein
Sci. 18 (2009) 1541–1551.
[40] K.A. Stubbs, M. Balcewich, B.L. Mark, D.J. Vocadlo, Small molecule inhibitors of a
glycoside hydrolase attenuate inducible AmpC-mediated β-lactam resistance,
J. Biol. Chem. 282 (2007) 21382–21391.
[41] H. Brumer, P.F.G. Sims, M.L. Sinnott, Lignocellulose degradation by Phanerochaete
chrysosporium: purification and characterization of the main alpha-galactosidase,
Biochem. J. 339 (1999) 43–53.
[53] V. Lombard, T. Bernard, C. Rancurel, H. Brumer, P.M. Coutinho, B. Henrissat, A
hierarchical classification of polysaccharide lyases for glycogenomics, Biochem. J.
432 (2010) 437–444.
[42] H.G. Damude, V. Ferro, S.G. Withers, R.A.J. Warren, Substrate specificity of
[54] M.R. Stam, E.G.J. Danchin, C. Rancurel, P.M. Coutinho, B. Henrissat, Dividing the large
glycoside hydrolase family 13 into subfamilies: towards improved functional anno-
tations of alpha-amylase-related proteins, Protein Eng. Des. Sel. 19 (2006) 555–562.
endoglucanase
A from Cellulomonas fimi: fundamental differences between
endoglucanases and exoglucanases from family 6, Biochem. J. 315 (1996) 467–472.